View
217
Download
4
Category
Preview:
Citation preview
Brunel University London
Department of Mechanical, Aerospace and Civil Engineering
College of Engineering, Design and Physical Sciences
Development of a Real-Time Discrete Event Monitoring and
Control Application for Manufacturing Plants with Novel
Key Performance Indicator Measurement
By
Washington Gabriel Barriga Baldeón
Supervisor: Dr Ali Mousavi
September 2016
A dissertation submitted in partial fulfilment of the award of the degree of
Master of Science in Advanced Manufacturing Systems
2
ABSTRACT
The primary objective of this research was the integration of real-time data acquisition (DAQ)
systems for the measurement of industrial Key Performance Indicators (KPIs) to
acknowledge the state of a system and respond to constant change and unpredictability
transforming traditional mechanic systems to viable systems which influence and modify the
environment to their advantage. Modern industrial systems are capable of capture data in real-
time and have the necessity to adjust to changing system requirements. KPIs metrics seek to
find optimal performance scenarios by assessing the impact of input activities to the output
yield. To implement and test this model, selected operational conditions were chosen for a
production line from which the real-time DAQ architecture is made. By monitoring
supporting elements (time and quantity measures) at discrete intervals in a production line,
several Basic KPIs are obtained which reflect a single feature of the state of the system. The
overall purpose of these metrics is to improve the operations of a company from the lean
approach, i.e., reducing non-value activities in the value chain and from the corporate
perspective of increasing business profitability reaching the strategic business goals. In this
application, the obtained Basic KPIs include Parts in Process, Number Waiting, Available
Production Time, Production Rate, Available Resource Time, Busy Resource Time, Idle
Resource Time, and Electricity Consumption. These metrics reflect relevant information
about the conditions of operation of resources, energy, queues in the system, and
throughput of products. From these single indicators, further computing allows to obtain
Global KPIs which gather these singles features of performance and indicate the state of
the production system using a single comprehensive KPI. The global KPIs derived are
Unit Consumption, Work-in-Process, and Greenhouse Gas Emissions (GHG) emissions.
The real-time model was then validated against a Discrete Event Model Simulation. The
successful validation of comprehensive KPI such as WIP gives a good reliability to the
entire model because it means that the measures of time and quantity in which the
computing of these metrics are based, and the single features KPIs are correctly modelled
to reflect the performance of the system in real-time.
3
ACKNOWLEDGEMENTS
Foremost, I would like to express my gratitude to my postgraduate sponsor the Secretariat
for Higher Education, Science, Technology and Innovation (Senescyt) in representation
of the national government of the Republic of Ecuador to whom I dedicate this research.
I am thankful to my supervisor Dr Ali Mousavi for his support and insightful advice
throughout this project from its original concept to the final stages. He also facilitated my
access to key resources of the University at the Systems Engineering Research Group
Laboratory
I am grateful to Eng. Daniel Vaca with whom I worked collaboratively in the development
of this research
I want to express my gratitude to my girlfriend Tatiana Litardo for her continuous
understanding and support. I am also thankful to my parents, brothers and friends for their
encouragement in the fulfilment of all my goals.
Author’s Declaration
I declare that the work in this dissertation was carried out in accordance with the
requirements of the University’s Regulations, including the section on plagiarism, and
I certify that the work presented is my own unless referenced.
Signature ………………………………………..
Date: September, 10th 2016
4
LIST OF ABBREVIATIONS
Abbreviation Stands for
ABC Activity-Based Cost
ACC Accumulated Value
CONWIP Constant Work-in-Process
CPS Cyber-Physical Systems
CTU Count Up Counter
DAQ Data Acquisition System
DBR Drum-Buffer-Rope
DECC Department of Energy and Climate Change
DEFRA Department for Environment, Food and Rural Affairs
DESM Discrete Event Simulation Modelling
DSTP Data Socket Transfer Protocol
EBIT Earnings Before Interest and Taxes
ERP Enterprise Resource Planning
GHG Greenhouse Gas
GWP Global Warming Potential
HMI Human Machine Interface
IDEF Integration Definition for Function Modeling
IEA International Energy Agency
IPCC
Intergovernmental Panel on Climate Change
JIT Just in time
KgCO2e Equivalent CO2 kilograms
KPIs Key Performance Indicators
MES Manufacturing Execution System
MLT Manufacturing Lead Time
MRP Material Requirement Planning
MTBF Mean Time between Failures
MTTR Mean Time to Repair
OEE Overall Equipment Effectiveness
OPC Open Platform Communications
5
OTE Output Energize instruction
OTIF On-Time In-Full delivery
PERA Purdue Enterprise Reference Architecture
PLC Programmable Logic Controller
PPCS Production planning and control system
RAM Random Access Memory
ROM Read-Only Memory
RTO Retentive Timer
SCADA Supervisory Control and Data Acquisition
SQL Structured Query Language
T&D
Transmission and Distribution
UC Unit Consumption
UI User-Interface
WIP Work-in-Process
XIC Evaluate if Closed Ladder Instruction
XIO Exaluate if Open Ladder Instruction
6
TABLE OF CONTENTS
ABSTRACT ...................................................................................................................... 2
ACKNOWLEDGEMENTS .............................................................................................. 3
LIST OF ABBREVIATIONS ........................................................................................... 4
TABLE OF CONTENTS .................................................................................................. 6
LIST OF FIGURES .......................................................................................................... 9
LIST OF TABLES .......................................................................................................... 11
Chapter 1: Introduction ............................................................................................... 12
1.1. Background ...................................................................................................... 12
1.2. Aim and Objectives .......................................................................................... 13
1.2.1. Objectives .................................................................................................. 13
1.3. Methodology .................................................................................................... 13
1.4. Significance ...................................................................................................... 14
Chapter 2: Literature Review ...................................................................................... 15
2.1. Introduction ...................................................................................................... 15
2.2. Manufacturing Execution Systems (MES)....................................................... 17
2.3. Real-Time Data in Operations Management.................................................... 18
2.4. Production Performance Metrics...................................................................... 20
2.5. Manufacturing Lead Time................................................................................ 21
2.6. Work – In-Process ............................................................................................ 22
2.6.1. Definition .................................................................................................. 22
2.6.2. WIP control models .................................................................................. 23
2.7. Energy Consumption in manufacturing industries ........................................... 24
2.7.1. Power Consumption per Resource states .................................................. 25
2.8. Environmental Impact of Manufacturing Operations ...................................... 26
2.8.1. Greenhouse-Gas Emissions Factors .......................................................... 28
Chapter 3: Methodology ............................................................................................. 31
3.1. Introduction ...................................................................................................... 31
7
3.2. Supervisory Control and Real-Time Data Acquisition structure ..................... 31
3.2.1. Programmable Logic Controller Architecture .......................................... 32
3.2.2. PLC programming and communication software ..................................... 33
3.2.3. Databridge ................................................................................................. 33
3.2.4. Human Machine Interface (HMI) ............................................................. 34
3.3. Classification of systems parameters ............................................................... 35
3.4. Work-In-Process .............................................................................................. 36
3.5. Electricity Consumption .................................................................................. 39
3.6. Methodology for Calculating the 2016 GHG Emission Factor ....................... 40
Chapter 4: Implementation .......................................................................................... 46
4.1. Introduction ...................................................................................................... 46
4.2. Programmable Logic Controller ladder code ................................................... 48
4.3. Key Performance Indicators User Interface ..................................................... 54
4.3.1. Production Rate per Hour ......................................................................... 55
4.3.2. Number Waiting ........................................................................................ 56
4.3.3. Work-In-Process Calculation .................................................................... 57
4.3.4. Electricity Consumption............................................................................ 59
4.3.5. Unit Consumption ..................................................................................... 60
4.3.6. Green Houses Gas Emissions.................................................................... 61
Chapter 5: Testing and Validation .............................................................................. 64
5.1. Introduction ...................................................................................................... 64
5.2. Testing of the Real-Time Model ...................................................................... 64
5.3. Discrete Event System Simulation ................................................................... 65
5.4. Verification and Validation of the system ....................................................... 69
5.4.1. Increasing the Confidence Interval for Terminating Conditions .............. 69
5.4.2. Validation of the collected data in Real-Time and Simulated Data using T-
Test ...................................................................................................................... 70
5.4.3. Validation of Collected data in Real-time and Simulated Data using F-Test
............................................................................................................................. 72
8
5.4.4. Graphical Validation of the Performance Indicators ................................ 74
Chapter 6: Conclusions ............................................................................................... 76
6.1. Meeting the Research Objectives ..................................................................... 76
6.2. Results and Findings ........................................................................................ 77
6.3. Future Work ..................................................................................................... 80
REFERENCES ................................................................................................................ 82
Appendix A: RsLogix 5000, RsLogix Emulate 5000, RsLinks Configuration .............. 87
Appendix B: PLC Ladder Code Command Lines.......................................................... 93
Appendix C: ARENA Simulation Results .................................................................... 104
9
LIST OF FIGURES
Figure 2-1 Example of an ABC cost method for a manufacturing system ..................... 16
Figure 2-2 Functional Hierarchy PERA model ............................................................... 18
Figure 2-3 Example of the DuPont Model ...................................................................... 20
Figure 2-4 Manufacturing Lead Time components ........................................................ 22
Figure 2-5 Flow of Energy and Material Inputs and Outputs. ........................................ 24
Figure 2-6 World production of electricity by source type ............................................. 27
Figure 2-7 Summary of Defra GHG emissions classification ........................................ 30
Figure 2-8 Basic structure of the Electric System ........................................................... 30
Figure 3-1 OPC process control ...................................................................................... 32
Figure 3-2 PLC architecture ............................................................................................ 33
Figure 3-3 Methodology for KPI categorization ............................................................. 36
Figure 4-1 Flow Line Processing Sequence .................................................................... 47
Figure 4-2 Line available operating time code ............................................................... 49
Figure 4-3 Start/Stop logic code for Machine 1 .............................................................. 50
Figure 4-4 Available time and Input sensor CTU for Machine 1 ................................... 50
Figure 4-5 Busy Time and CTU output for Machine 1 ................................................... 51
Figure 4-6 Parts in Process Machine 1 ............................................................................ 52
Figure 4-7 Inactive State for Machine 1 and Queue 1 OTE ........................................... 53
Figure 4-8 Number Waiting Queue 1 and Parts in Process Machine 2 .......................... 53
Figure 4-9 Reset Timers and Counter ............................................................................. 54
Figure 4-10 User Interface OPC server connection ........................................................ 55
Figure 4-11 Production Rate per Hour Formula Node and Front Panel Indicator Display
......................................................................................................................................... 56
Figure 4-12 Number Waiting Table Array...................................................................... 57
Figure 4-13 Work-in-Process Formula Node .................................................................. 57
Figure 4-14 Parts being Processed in the System Table Array ....................................... 58
Figure 4-15 Electricity Consumption (Total, Idle and Busy State) Formula Node ........ 60
10
Figure 4-16 Unit Consumption Formula Node ............................................................... 61
Figure 4-17 Electricity Consumed Factor in the years 2006 to 2015 ............................. 61
Figure 4-18 Electricity Consumed Factor 2016 .............................................................. 62
Figure 4-19 Average, Maximum, Minimum and 2016 projection Energy Consumed
Factor............................................................................................................................... 62
Figure 4-20 GHG emissions Formula Node ................................................................... 63
Figure 5-1 Discrete Event Modelling Logic ................................................................... 65
Figure 5-2 Create Parts Module ...................................................................................... 66
Figure 5-3 Entry of Parts Record Module ....................................................................... 66
Figure 5-4 Route and Transfer time between Stations .................................................... 67
Figure 5-5 Station-Process-Route (Production Line Logic Modules) ............................ 67
Figure 5-6 Station 1 Process Module .............................................................................. 68
Figure 5-7 Schedule of the production plant and capacity of resources ......................... 68
Figure 5-8 Resources Scheduling Rule ........................................................................... 69
Figure 5-9 Record Average Work-In-Process with 5 Replications ................................ 69
Figure 5-10 Record Average Work-In-Process with 48 Replications ............................ 70
Figure 5-11 WIP Validation Average Maximum and Minimum value .......................... 74
Figure 5-12 WIP Maximum and Minimum Validation .................................................. 75
Figure 5-13 Production Rate Validation Average Maximum and Minimum value........ 75
Figure 6-1 Virtual Controller (EmuLogix 5868 ) parameters configuration .................. 87
Figure 6-2 Modules in the RsLogix Emulate 5000 Chassis Monitor.............................. 88
Figure 6-3 Virtual Backplane communication driver ..................................................... 88
Figure 6-4 RsLink Server Connected Elements .............................................................. 89
Figure 6-5 RsLogix5000 Controller Configuration ........................................................ 89
Figure 6-6 Connection Parameters 1756-Generic Module ............................................. 90
Figure 6-7 Connection Properties 1756 Generic Module ............................................... 91
Figure 6-8 RsLogix 5000 Who Active Window ............................................................. 91
Figure 6-9 1789 digital I/O module Data Properties ....................................................... 92
11
LIST OF TABLES
Table 2-1 Specific electricity requirements for Injection Molding process .................... 26
Table 2-2 Main sources of GHG with respect to the effective lifetime .......................... 28
Table 3-1 Emission Factors (Energy generated (kwh)) .................................................. 41
Table 3-2 Emission Factors (Energy Losses (kwh)) ....................................................... 41
Table 3-3 Emissions Factor (Electricity Consumed (kwh)) ............................................ 42
Table 3-4 Emission Factor (Electricity Consumed (kwh)) – Calculation for 2010 ........ 43
Table 3-5 Calculation of the Average, Maximum and Minimum Electricity Generated
Emission Factors (2006-2015) ........................................................................................ 43
Table 3-6 Calculation of the Average, Maximum and Minimum Transmission and
Distribution Losses Emission Factors (2006-2015) ........................................................ 44
Table 3-7 Calculation of the Average, Maximum and Minimum Emission Factors for
Electricity Consumption (2006-2015)............................................................................. 44
Table 3-8 Values for the Emission Factor (Electricity GENERATED) for 2016 ........... 44
Table 3-9 Values for the Emission Factor (Electricity Losses) for 2016 ........................ 45
Table 3-10 Values for the Emission Factor (Electricity Consumed (kwh)) projected to
2016 ................................................................................................................................. 45
Table 3-11 Emission Factor (Electricity Consumed (kWh)), projection, average,
maximum and minimum for the year 2016 ..................................................................... 45
Table 4-1 PLC Input Tag Database ................................................................................. 48
Table 4-2 PLC Output Tag Database .............................................................................. 48
Table 5-1 T-test system validation .................................................................................. 72
Table 5-2 F-Test Two-Sample for Variances.................................................................. 74
12
Chapter 1: Introduction
1.1. Background
In this modern time, competitive globalisation demands manufacturing industry to
respond to constant change and unpredictability. In fact, modern industrial systems are
capable of capturing data in real-time and have the necessity to adjust to changing system
requirements (Tavakoli, et al., 2013). Global competition has shortened the product life
cycles and expects companies to manufacture customised products at low costs with high
quality. The volume of customised products is small and responds to the demand of niche
markets. In order to adapt to these changes, companies have to reconfigure its
manufacturing process and technology making them flexible to develop different types
of products in a short period. With the aim to help increasing this variety, different
techniques to support manufacturing systems have been developing such as IDEF, GRAI-
Grid, simulation, Petri Nets and integrated modelling methods (Hernandez, et al., 2008).
Of these, the simulation technique is used to identify queuing specific manufacturing
problems such as bottlenecks, imbalanced lines, congestions, non-adequate layout among
others by measuring relevant Key Performance Indicators (KPIs) which are essential to
monitor performance and goal realisation within an organisation (Gieskes, et al., 1999)
Modern manufacturing systems seek to establish algorithms for the application of cyber-
physical systems and Industry 4.0 for the total integration of production operations and
business planning and logistics (Monostori, et al., 2016). This intrinsic network
configuration of cyber-physical devices embedded with sensors and actuators aim to
collect and exchange production data for measuring key performance indicators in real-
time through network connectivity. Furthermore, the aforementioned KPIs are suitable to
report costs functions which help managers and directors with strategic decisions and
production planning. This project focus on developing a generic algorithm for measuring
in real-time Key Performance Indicators in a proposed manufacturing system. This
framework includes the real-time data acquisition (DAQ) system from shop floor level
through PLC ladder code programming and user interface which allows controlling and
monitoring the production process. From the acquired production data, KPIs will be
measured through an integration with the next level of Manufacturing Execution System.
13
1.2. Aim and Objectives
The primary aim of this dissertation project is to develop a generic application to extract
Real-Time data from PLCs and historical databases and calculate Key Performance
Indexes (KPI) using instantaneous measurement models with the purpose of improving
the monitoring of performance evaluation and decision-making of manufacturing plants.
This application is a generic framework, and it can be employed in all types of industries
and applications including continuous process and discrete process industries to assess
the performance and productivity of a system in real-time. Therefore, the proposed
method is suitable for continuous output production and finite quantity batch production.
1.2.1. Objectives
1. To provide an architecture solution for the real-time DAQ application by
establishing a Supervisory Control and Real-Time data acquisition structure
including the Programmable Logic Controller and communication software
2. To develop Human Machine Interface (HMI) application software to show KPI
measures in real-time establishing a communication protocol between the system
components.
3. To calculate sensible KPIs measures such as Work-in-Process, Energy
Consumption, GHG emissions, Production Time, Production Rate, Energy
Efficiency, and Number Waiting based on modelling process approach applied to
a proposed manufacturing system.
4. To test and validate the real-time model against a Discrete Event Model
Simulation made in ArenaTM to compare the KPIs metrics obtained in the generic
real-time application with the system performance measurements attained by the
simulated scenario.
1.3. Methodology
The objectives of this project will be accomplished mostly through the development of a
system architecture which allows the real-time measurement of KPIs through the
monitoring of time and quantity data in a manufacturing system. The next chapters fully
describe the DAQ architecture solution implemented and the testing and validation of the
generic application. For example, in order to achieve objective 1, a Supervisory Control
and Real-Time Data Acquisition structure including the Programmable Logic Controller
14
Architecture will be established by configuring and employing logic controller and
communication software such as RsLogix5000, RsLinks and RsEmulate used for real-
time data acquisition modelling. Additionally, Objective 2 will be accomplished by the
development of a LabView user-interface for data presentation and process simulation
and control where an OPC communication protocol between the system components is
established to show KPI measures in real-time. Objective 3, will be realised by proposing
a flow-line production system case to implement and test the DAQ architecture simulating
a discrete event system and gathering relevant production data for computing
comprehensive performance metrics. Objective 4 will be made by validating KPIs metrics
obtained with the real-time model against a Discrete Event System Modelling Simulation
(DESM) developed in Arena software where the simulated model is assessed for an
extended period of iterations.
1.4. Significance
Key Performance Indicators (KPIs) are vital to any business and industrial system. Since
KPIs can be compared with internal targets (e.g. Expected Production Rate or Overall
Equipment Efficiency), or external target (e.g. World class KPIs and Benchmarking
analysis) they provide a measurement of the performance of a company within a period
of time (Özbayrak, 2004). Additionally, KPIs help to determine operational inefficiencies
of a manufacturing system such as bottlenecks, imbalanced lines, congestions, and non-
adequate layout with the purpose of optimising the overall plant performance. By making
consisting efforts to optimise common production variables, usual non-value added
activities (in the Production Value Steam Mapping) that generate waste, and unnecessary
manufacturing costs can be limited and reduced.
The development of real-time data acquisition (DAQ) system to measure KPIs contribute
to the total integration of the different manufacturing levels (shopfloor operations, and
manufacturing planning and control). This integration allows a quick response to
demands and fluctuations of the production systems by gaining real-time feedback on the
state of the system through performance indicators. This structure allows acquiring
accurate information on intensive cost activities (e.g. Equipment Downtime, defective
Finished Products, Rework, Scrap, and excessive Work-In-Process) which have a
significant impact on the aggregated manufacturing costs and delivery time of finished
products. These operational inefficiencies can be measured and controlled in order to
15
improve a manufacturing process, adjust the production capabilities to the fluctuations in
market demand, increasing productivity, energy efficiency, and resource utilisation.
Chapter 2: Literature Review
2.1. Introduction
Manufacturing Key Performance Indicators (KPIs) have been employed in several
industrial systems for assessing the performance of a company in any given period. In
fact, the development of quantitative performance measurements has proven to be highly
useful for managerial decision making as it provides an insight of the efficiency of each
operation compared to the expected value of its competitors. For example, (Ahmada &
Dhafrb, 2002) in “Establishing and improving manufacturing performance measures”
carried out a comparative study of a speciality chemicals plant key performance indicators
with world-class performances values and their process (Benchmarking). The KPIs
chosen for this study were product delivery performance (On-Time-In-Full delivery,
OTIF), Adherence to production plan, Product rate, Quality rate, Availability, Overall
Equipment Effectiveness (OEE). Furthermore, (Özbayrak, 2004) developed a
comparative study of Activity-Based Cost (ABC) estimation in a push/pull manufacturing
system. In this work, a comparison between the material requirements planning (MRP)
and the just in time (JIT) system regarding manufacturing and product costs were carried
out to highlight the difference between the two strategies. The manufacturing and product
cost estimation were based on the ABC methodology which incorporated a mathematical
and simulation model. ABC method associated all activities with costs for product
manufacturing as shown in Figure 2-1. In this way, the ABC results provided a
quantitative approach to determine the cost efficiency of the system and to make
production decision accordingly, and it allows enhancing the production operations with
the highest manufacturing costs.
16
Figure 2-1 Example of an ABC cost method for a manufacturing system
Source: (Özbayrak, 2004)
The results of (Özbayrak, 2004) show that manufacturing costs are directly proportional
to the production planning. In Push system, higher production costs are created by the
large batch sizes increasing the WIP between the workstations. This results in higher
waiting time, less flow of finished product and lower delivery time. On the other hand,
the pull system gives lower manufacturing costs by having smaller batches which
significantly impacts the lead times (delivery time of finished products to the customer)
in terms of set-up operations and machine buffers. However, there is a gap between the
analysis of manufacturing costs of production systems integrating a Push and Pull
combined system by using product and production levelling (“Heinjunka” approach).
This approach means a balance between the two systems to avoid fluctuations in all
aspects of the supply chain including the flow of materials, finished goods inventory, and
work in progress (Hüttmeira & de Trevillea, 2009). In this line, manufacturing key
performance metrics can be linked with cost analysis based on ABC methodology in this
combined Push and Pull Production plan to optimise the general production operation.
(Hernandez, et al., 2008) proposed a methodology based on IDEF model (Integration
Definition for Function Modelling) from a manufacturing data warehouse system to
obtain the performance indicators of any manufacturing plant. The design incorporates
scorecard panels to use KPIs to decide the best actions for continuous improvement and
optimisation. For instance, systems with a non-optimal layout, high buffers between
machines, unbalanced production, overproduction, wastes of time (delays and transport)
or high ratios of defective products can be improved by measurement and control KPI
against expected internal performance objectives and targets.
17
2.2. Manufacturing Execution Systems (MES)
The integration of the Manufacturing Control and Planning with the shop floor core
operations is possible with a computerised Manufacturing Execution System (MES)
which is a bridge between corporate levels of logistics and planning such as Enterprise
Resource Planning (ERP) and the process control level of Supervisory Control and Data
Acquisition (SCADA) (Meyer, et al., 2009 ). With the total integration of the different
manufacturing levels (shopfloor operations and manufacturing planning and control) is
possible to gain real-time feedback on the state of the system through performance
indicators, and cost functions (Brown & Fraser, 2012). This integration allows a quick
response to demands and fluctuations of the production systems which cover research and
development, design, production, logistics and sourcing. A high speed of reaction to
market demands carries a substantial competitive advantage for a company and
subsequent economic growth. Additionally, MES helps to obtain precise information on
intensive cost activities (e.g. labour, scrap, downtime, and tooling) and unnecessary
inventory level (Scholten, 2009).
MES is a vital part of Industry 4.0, Internet of Things and Cyber-Physical Systems (CPS)
since it helps to assess the state of a manufacturing system by automating the
manufacturing execution and control (Monostori, et al., 2016). The aim of MES system
is to respond to constant change driven by global completion which has shortened the
product life cycles and expects companies to manufacture customised products at low
costs with high quality. In order to adapt to these changes, companies have to reconfigure
its manufacturing process and technology making them flexible to develop different types
of products in a short period of time. Such variability is only possible with the accurate
integration of all levels of manufacturing operations.
By gathering real-time data about the production process, an MES system is capable of
observing and upholding the right execution of the manufacturing procedure, observing
and controlling the materials utilised as a part of the generation procedure. By this
approach, MES gives the instruments to examine the information to streamline
productivity providing the instruments to tackle common manufacturing issues and
streamline systems (Brown & Fraser, 2012).
18
The framework for the integration of the manufacturing support functions and control
systems common to all discrete, continuous and batch production processes is described
by the Purdue Enterprise Reference Architecture (PERA model) (Williams & Li, 1990).
Based on the PERA model, various international standard for manufacturing integration
such as “ISA-95-2000: Enterprise-Control System Integration” and “ISO-22400
Manufacturing operations management -KPI” establish several objects models ranging
from Hierarchy, Functional Data Flow, and Operations Activity Model (ISA-95, 2010).
These models provide a structured tasks divisions and information exchange within all
levels of a company. For instance, Figure 2-2 delineates the diverse levels of the utilitarian
chain of command model consisting of the actual production processes (level 0), plant
execution and control (level 1-2), manufacturing output and monitoring (level 3), and
(level 4) business scheduling and logistics (ISO 22400, 2009)
Figure 2-2 Functional Hierarchy PERA model
Source: (ISO 22400, 2009)
2.3. Real-Time Data in Operations Management
Figure 2-2 shows the integration of the aforementioned hierarchy levels. The actual
production process is measured through sensors connected to Programmable Logic
Controls (PLCs) and Supervisory Control and Data Acquisition, SCADA systems for the
monitoring, process monitoring, and process execution. Then, real-time production data
19
is gathered and sent to a MES software which uses this information to evaluate the
performance of the production system through the evaluation of the primary performance
indicators (KPIs) such as Work-in-process (WIP), resource utilization, Energy
Consumption, Overall Equipment Effectiveness (OEE), and Environmental Impact
among others (ISO 22400-2, 2009). Real-Time Discrete Event Simulation Modelling
(DESM) can propose improved scenarios for the optimisation and production
management of the system. In this hierarchy level, KPIs are critical to understanding and
improving manufacturing performance.
Higher level performance factors can be calculated from the system measurable KPIs. In
this way, the performance of the actual process is integrated with an Enterprise Resource
Planning (ERP) which uses real-time databases to trace business resources such as
inventory management, product planning, manufacturing delivery, and the status of the
external supply chain such as sourcing, and sales forecast (Almajali, et al., 2016). ERP
and KPI help managers and directors to make a strategic planning of the manufacturing
production for enhancing the business profitability. For example, there is a relationship
between KPI's models and higher levels of performance factors as shown by the
DuPont Model of financial analysis which is a link to calculate the Return on Equity
which is defined as the returns that investors receive from the firm. The DuPont model
allows explaining how the operational factors impact on a business performance. Figure
2-3 shows a scenario for the investment of an improvement project for a manufacturing
system. The company increased EBIT (Earnings Before Interest and Taxes) by holding
an of 60% result of Overall Equipment Effectiveness (OEE) which is a general
manufacturing KPI measured as a function of Availability, Performance, and
Quality (Leroux, 2010).
20
Figure 2-3 Example of the DuPont Model
Source: (Leroux, 2010)
2.4. Production Performance Metrics
Quantitative manufacturing metrics allow identifying sensible information about the state
of a system which is essential to enhance the production operations of a company. By
acknowledging its state, it is possible to respond to constant change and unpredictability
transforming a traditional mechanic system to a viable system which influence and
modify their environment to their advantage (Mousavi, 2011). There are two primary
groups of production metrics: (1) production performance measures and (2)
manufacturing costs indexes (Groover, 2016). Both of these measures permit tracking and
monitoring part and product costs and identifying common manufacturing problems such
bottlenecks, congestions, imbalanced lines, layout, rework, etc. Additionally, to these
financial and technical performance indicators, human contribution measurement such as
worker efficiency ratio and customer satisfaction can be measured (Zairi M, 1994).
(ISO 22400, 2009) describes Key Performance Indicators as “quantifiable and strategic
measurements that reflect an organization’s critical success factors”. The overall purpose
of these metrics is to improve the operations of a company from the lean approach, i.e.,
reducing non-value activities in the value chain and from the corporate perspective of
increasing business profitability reaching the strategic business goals.
Apart from identifying common manufacturing problems, KPIs can be translated to cost
functions with the purpose of helping senior management focus a company´s resources
21
on systems constraints that have the greatest impact on productivity. In most cases, a
manufacturing system is optimised without significant investments in new technology or
machinery (Ahmada & Dhafrb, 2002). For example, KPI measurement assists in the
implementation of Flexible Manufacturing Systems, or Cellular Manufacturing Group
Technology as performance optimisation techniques. Additionally, DESM simulation
assesses the state of the system by measuring KPI and proposing optimised scenarios and
solutions for improvement (Banks et al., 2009) . The real-time instantaneous performance
measurements selected for developing this application are Work-in-Process (WIP),
Number Waiting, Energy Consumption, Energy Efficiency, Production Rate, Waiting
Time, Production Time, and Green-House-Gas (GHG) Emissions.
2.5. Manufacturing Lead Time
As a global indicator of an operating system, the manufacturing lead-time (MLT) delivers
information about total time required to process a part or a given product (Groover, 2016).
MLT comprehends all the operations through the plant, including any unit operations and
non-operational activity. An operational activity refers to the actual transforming process
which gives an Added Value to the finished product. The unit operations can be physical
(change in its shape, dimensions, or adding of components through assembly) or chemical
transformations (change of the matter´s molecules through chemical reactions to form
new products). A non-operational activity is the time spent due to delays, parts being
moved between processing stations, or time consumed in queues (Groover, 2016). Figure
2-4 indicates the main components of MLT for executing an order in a generic
manufacturing system including the actual processing time, setup time of machines, delay
time, transporting and queuing time (ISO 22400, 2009). These components help to
determine the actual Production Rate of the system (amount of finished good processed
within the production time).
22
Figure 2-4 Manufacturing Lead Time components
Source: (ISO 22400, 2009)
The manufacturing lead time is also referred as customer-to-customer time, i.e., the total
time since a customer places an order until the final product is dispatched to the client.
This indicator is of high relevance as MLT directly impacts the Throughput-day dollars
of a company, i.e., the sum of all dollars’ worth of orders that have not been dispatched
multiplied by the late days the stock remains undelivered (Stadtler, 2014). Throughput is
all the money that the system (the company as such) generated through sales. The delivery
performance of a manufacturing system is bounded by inventory and MLT
constraints. Production optimisation techniques seek to measure the impact of a decision
based on the performance of the restriction (Goldratt E., 2004). Inventory control methods
such as reorder point models aims to buffer the systems protecting it against variations
(Woeppel, 2003). This buffer behaves as a risk manager to hold the restriction (inventory
level), and the subsequent throughput-day dollars of the company.
2.6. Work – In-Process
2.6.1. Definition
(Groover, 2016) defines Work-In-Process (WIP) as the amount of parts placed in the
production line that are either being processed or being transferred between processing
operations. That is the overall stock that is in the condition of being changed from raw
material to finished goods. As WIP increases there is more inconvenient for process
management, and unnecessary inventory build-up incurs. Additionally, WIP fluctuations
can alter production schedules, in imbalanced production lines (Lee & Seo, 2016).
23
2.6.2. WIP control models
There have many studies on the performance evaluation of alternative Production
Planning and Control System (PPCS) with the aim of controlling the WIP of the
manufacturing system and its throughput. Alternative PPCS differ from traditional MRP
as the production level is not based on sales forecast but rather in Make-to-Order
Inventory (Lee & Seo, 2016). In that line, (Jodlbauer & Huber, 2008) compared three
types of WIP-controlled pull production systems with constant processing times such as
Kanban, CONWIP (constant work-in-process) and DBR (drum-buffer-rope). A Kanban
system aims to adjust a limit on the amount of WIP between every adjacent pair of
workstations by conveyance the production making the exact quantity for the subsequent
terminal in the previous flowline unit. On the other hand, CONWIP blocks work part
entries at the beginning of the line controlling and maintaining as constant the overall
WIP by communicating the last processing unit to the initial workstation in the sequence
(Lee & Seo, 2016). DBR is a way of sequencing material according to the rate of
production of the bottleneck station (Georgiadis & Politou, 2013). DBR prevents WIP
inventory to rise. The study showed a better response of CONWIP in terms of expected
waiting time and WIP followed by MRP, DBR and Kaban (Lee & Seo, 2016). In
particular, CONWIP which sets a limit on the total WIP for the entire production system
have proved to outperform push system regarding throughput and WIP (Bonvik, et al.,
2000)
An optimal PPCS production aims to monitor, optimise, and minimise WIP. Within a
manufacturing system, WIP is an indirect measure of a system throughput. For a flow
line production line, WIP can represent high inventory costs since it is a hold in stock
which cannot be transformed into profit until the total batch production or total order size
(for continuous flow line production) is dispatched to the next downstream client in the
supply chain (Khojasteh-Ghamari, 2012). Furthermore, excessive WIP inventory leads to
higher resource utilisation, and lower energy efficiency as fluctuations on WIP in
imbalanced systems demand extended busy times and productivity inefficiencies for
bottlenecks stations.
24
2.7. Energy Consumption in manufacturing industries
The inherent production operations in diverse manufacturing industries are energy
intensive. Most transformation process requires in some degree energy inputs to convert
raw materials to output product as depicted in Figure 2-5. These energy inputs come
predominantly from electricity supply and are transformed into useful work, and waste
heat according to the first thermodynamic law of energy conservation and energy
transformation. As many industrial plants employ electricity inputs for their production
machines, fuels combustion is needed at power stations (Gutowski, et al., 2006).
Therefore, assessing the amount of electricity that an industrial plant requires for their
manufacturing operations lead to carbon emissions footprint analysis. Not only the
measurement of energy consumption is necessary to determine environmental
compliance, but power consumption is a primarily component of variable manufacturing
costs as it varies as a function of the proportion output (Groover, 2016). Figure 2-5 depicts
the energy and material inputs and outputs for a manufacturing process.
Figure 2-5 Flow of Energy and Material Inputs and Outputs.
Source: (Gutowski, et al., 2006)
Energy analysis of a manufacturing process show complex energy and material flows
interconnections ranging from the conversion of working materials, auxiliary materials
and fuels (through combustion) into heat, wastes, and emissions (Gutowski, et al., 2006).
Exergy models simplify the energy balance models for a manufacturing process
approximating the potential useful work in the overall supplied power which is achieved
by the interaction between the system and its environment (Sato, 2004) (de Swaan
25
Aarons, et al., 2004). In that line, exergy estimation provides a measurement of the
electricity used in a manufacturing process.
2.7.1. Power Consumption per Resource states
It is important to note that electricity consumption is dependent on the resources state. A
resource state is the condition of operation of each machine at discrete intervals of time.
For this dissertation four resources states are established: Idle, Busy, Standby and Failed.
A piece of equipment is defined as busy when an entity seizes the resource, and therefore
the resource is in processing condition. On the other hand, an idle status means that a
machine is entirely free, waiting for a part to arrive (Kelton, et al., 2015). The standby
state refers to power consumption by an equipment without being unplugged despite
being switched off. Examples of devices which consume standby power are appliances
with "instant-on" capabilities that respond instantaneously to user action without warm-
up delays like motion sensors, light sensors, built-in timers, or security systems and fire
(Ross & Meier, 2001). Finally, when a breakdown takes place, a resource becomes
unavailable, and none of its capacity is used by any entity.
Modern engineering tendencies such as nanotechnology, ultra-precision machining, and
micro manufacturing have low processing rates and high particular electricity
requirements (Gutowski, et al., 2006). Table 2-1 shows the specific electricity
requirements for an Injection Molding process as a function of the rate of material
processed. It is noted that for the electricity changes as a function of the process rate. In
fact, the specific resource state (idle, busy, standby or failed) defines the throughput
(process rate) a machine delivers. For example, the power required when operating at full
processing capacity on the busy state is higher than the power demanded for the idle sate
(waiting condition). Therefore, the variable that determines the energy consumption of a
resource is the throughput on each state. The electricity requirement fluctuates
accordingly.
26
Process Name Power
Required
Process
Rate
Electricity
Required
Injection
Molding
kW cm 3 /s J/cm 3
10.76 3.76E+00 3.41E+03
26.1 9.77E+00 3.21E+03
71.4 5.05E+01 1.96E+03
35.76 1.40E+01 3.09E+03
47.46 2.70E+01 2.30E+03
65.34 4.51E+01 1.99E+03
12.73 7.66E+00 2.20E+03
13.17 1.09E+01 1.75E+03
51.41 4.25E+01 1.75E+03
Table 2-1 Specific electricity requirements for Injection Molding process
Source: (Thiriez, 2005)
2.8. Environmental Impact of Manufacturing Operations
Electricity used in manufacturing operations is generated mainly at thermal power
stations through chemical combustion of fuels. Concerning thermal electricity generation,
there are several processing technologies such as turbo steam (water vapour
thermodynamic cycle); turbo gas (natural gas thermodynamic cycle) and internal
combustion engines (Diesel and Otto thermodynamic cycle) (Turconi, et al., 2013). Each
thermodynamic cycle uses different fossil fuels, among them coal, natural gas, fuel oil,
diesel, oil and bagasse. The primary pollutants from fuel combustion are carbon dioxide
(CO2), carbon monoxide (CO), sulphur dioxide (SO2), nitrogen oxides (NO2), ozone
(O3), partially unburned hydrocarbons and particulate matter (lead, soot, ash and other
toxic metals (Basu, 2010). Additionally, the main environmental problems caused by
these pollutants gases are acid rain, depletion of the stratospheric ozone layer, and the
mentioned global warming effect (IPCC, 2014). While industrial activities regularly
demand electricity, alternative energy production methods, which are friendly to the
environment are investigated. Among them, renewable energies such as hydrogen
production, solar, eolic, geothermal, and biomass are sought renewable sources according
to the strategic resources of each country (Sun, et al., 2012)
80% of the total use of energy on the planet is based on fossil fuels which represent 400
EJ per year (Saidur, et al., 2011). Regarding global electricity generation by source type,
coal is the leading source in the world, reaching 40.6% of the total electrical energy offer
27
which was calculated 21,431 tWh for 2011 (IEA, 2011). Followed by natural gas with
22.2%, hydropower 16%, oil and derivatives with 4.6% while nuclear energy reaches
12.9% (IEA, 2011). Finally, others renewables sources with 3.7% (See Figure.2- 6) (IEA,
2011).
Figure 2-6 World production of electricity by source type
Source: (IEA, 2011)
The world energy model is based on fossil fuels like oil, coal and natural gas, the same
that are large scale emitters of greenhouse gases (GHG), has serious problems of
unsustainability. The Intergovernmental Panel on Climate Change (IPCC) defines GHG
as atmospheric gaseous constitutes that absorb and release radiation at certain
wavelengths inside the Earth´s spectrum of thermal infrared radiation. This property
causes an increment in the ability of the atmosphere to capture and recycle energy emitted
by Earth's surface increasing its temperature (global warming effect) (IPCC, 2014).
Considering this outlook, there has been given significant importance to preserving the
environment limiting GHG emissions. An example of this was the 15th International
Conference on Climate Change held in 2009, where policies and guidelines for all
countries in the world were established once the Kyoto Protocol concluded (Muñoz,
2013). The Kyoto Protocol made on the 1992 United Nations Framework Convention on
Climate Change sought to limit the GHG emissions which cause global warming. Table
2-2 illustrates the principal sources of GHG with respect to the effective lifetime
(estimated duration of the gas in the atmosphere).
Nuclear
Hydroelectricity
Natural Gas
Coal
OilOthers
2011 WORLD ELECTRITCY GENERATION
28
Table 2-2 Main sources of GHG with respect to the effective lifetime
Source: (Forster, et al., 2007)
2.8.1. Greenhouse-Gas Emissions Factors
There are two ways of measuring and recording GHG emissions in industrial activities
by registering emissions at source and by employing event data for estimating the GHG
amount emitted. The continuous emissions monitoring at source uses field sensors that
distinguish the GHG type and measure its concentration. Typical concentration sensors
include infrared sensors, electrochemical gas sensors, and semiconductor sensors (Wali,
2012). These industrial sensors are installed on the stationary emissions units and
connected to SCADA system for continuous detection and monitoring. These industrial
sensors are located at the plant level on the stationary emissions units such as natural gas
fired boilers, gas turbines, oil fired boilers or coal-fired boilers, cement kiln off-gas, or
gasification combined cycles, among others (Campbell, et al., 2000) (Gielen &
Moriguchi, 2003) (IEA GHG, 2002) (Wheeler, 1998)
The second method for measuring GHG is by using industrial activity data (such as
kilowatt-hours of electricity consumed or litres of fuel used) to estimate the GHG
emissions (DEFRA & DECC, 2012). This method employs relevant conversion factors
called “emissions factors” to correlate and transform the production statistics to the GHG
emissions expressed as kilograms of carbon dioxide equivalent (CO2e) released into the
air. For example, emissions factors can calculate the amount of CO2 emitted in mass as a
consequence of burning an amount of oil in a heating boiler (Department of Energy &
Climate Change, 2015). CO2e is a standard unit of measurement that allows the global
warming potential (GWP) of different GHGs types to be compared. The IPCC Fifth
29
Assessment Report established the GWP factors for non-carbon dioxide gases as (GWP
for CH4 = 21[CO2e], and GWP for N2O = 310 [CO2e]) (IPCC, 2014).
In some countries, companies are compelled by law to report their GHG emissions
annually to the local environmental authorities in environmental monitoring reports
containing characterization informs of air pollutants. The purpose is to assurance
regulation compliance, that is limiting all air pollutants and GHG emissions under
stablished maximum allowable concentrations. In the United Kingdom, the Department
of Energy and Climate Change (DECC) and the Department for Environment, Food and
Rural Affairs (Defra) are in charge of developing the Guidelines to Greenhouse-Gas
Conversion Factors for Company Reporting. These guidelines represent the official
government emissions factors which are updated every year according to the annual UK
energy matrix. Figure 2-7 summarizes the Defra methodology to assess the GWP of
different types of GHG. Defra´s Protocol Corporate Standard distributes the broad types
of emissions activities into three main cluster groups. In this way, each activity is itemized
as either Scope 1, Scope 2 or Scope 3 as follows:
• Scope 1 refers to a company´s Direct Emissions as a consequence of their owned
or controlled industrial activities such as fuel combustion, owned transport,
process emissions, and fugitive emissions
• Scope 2 is related to the GHG emissions as consequence of the consumption of
purchased electricity, heat, steam and cooling. These are indirect emissions not
provoked by an organization´s activity but produced as a result of the electricity
generation transmission and distribution.
• Scope 3 clusters other indirect emissions. They are indirect activities not
associated with electricity consumption such as waste disposal, purchased
materials and fuels or transport related activities.
This dissertation project is focused on the measurement of the direct GHG emissions
related to the consumption of purchased electricity under the scope 2. Electricity
generation, and the electricity transmission and distribution as indirect GHG emissions
are considered under scope 3. In this way, all the effects of electricity consumed at each
resource in the production plant for carrying out their manufacturing activities are going
to be related to direct and indirect GHG emissions and GWP assessment.
30
Figure 2-7 Summary of Defra GHG emissions classification
Source: (DEFRA & DECC, 2012)
Figure 2-8 shows the typical structure of the electric grid system which includes the
electrical generation system at power stations, the transmission networks responsible for
transporting the electricity at high voltage to electrical substations, and the distribution
networks to the end user (Department of Energy, 2004). At transmission substations, the
high voltage transmission is converted to local lower voltage form (Short, 2014). The
systems of transmissions and distribution account for energy losses which are relevant to
GHG traceability. For example, a 765 kV line carrying 1000 MW of power can have
losses of 0.5% up to 1.1% (Crawley & Haight, 2013, p. 474) These energy losses need to
be incorporated when the CO2 equivalent emissions for energy consumption of the end
user is calculated.
Figure 2-8 Basic structure of the Electric System
Source: (Department of Energy, 2004)
31
Chapter 3: Methodology
3.1. Introduction
The primary focus of this dissertation project is to develop a generic application to extract
Real-Time data from PLCs and historical databases and calculate Key Performance
Indexes (KPI) using instantaneous measurement models with the purpose of improving
monitoring of performance evaluation and decision-making of industrial activities. The
developed methodology is a generic framework, and it can be employed in all types of
industries and applications including continuous process and discrete process industries.
Therefore, the proposed method is suitable for performance evaluation and improvement
in continuous output production and defined quantity batch production.
This chapter outlines the discrete event models employed to calculate sensible KPIs
measures such as Work-in-Process, Energy Consumption, GHG emissions, Production
Time, Production Rate, Energy Efficiency, Number Waiting, and Waiting Time.
Furthermore, the chapter provides the architecture solution for the real-time data
acquisition application, the systems parameters and data sources to be integrated with the
HMI interface to calculate the real-time performance measures and make predictive
simulation according to simulated scenarios.
The chapter starts describing the Supervisory Control and Real-Time Data Acquisition
structure including the Programmable Logic Controller Architecture, logic controller and
communication software employed such as Databrige, RsLogix5000, RsLinks and
RsEmulate used for real-time data acquisition. Then it explains the communication
protocols of the system components and the HMI application software to show KPI
measures in real time. Afterwards, the modelling process approach is explained with the
selected performance measurements applied to a proposed manufacturing system.
3.2. Supervisory Control and Real-Time Data Acquisition structure
The proposed application is programmed to collect direct input from field equipment, and
industrial programmable logic controllers through Input and Output Sensors and power
controls switches in each workstation of a production line system. The acquired data can
be sent from the field equipment to the PLC through temporary storage buffer and I/O
system bus (Bolton, 2012). The input data is received by the PLC program panel and the
32
input/output unit. As this is a generic approach, the only needed feature for the field
equipment and PLC software is to have communication capabilities such as Open
Platform Communications standard (OPC) to communicate with a computer/server
(Hong & Jianhua, 2006). The OPC Data Access (OPC DA) specification is used to read
and write real-time data. This protocol states the communication of the real-time plant
data between the PLC control devices. Figure 3-1 shows a typical OPC connection
scenario with a single server- secondary software connection on a single computer. The
PLC hardware communication protocol is transformed into OPC protocol by the OPC
server software. This connection allows an OPC client software such as a HMI software
to connect to the industrial controller. The OPC secondary software uses the OPC server
to receive data from or send commands to the PLC or field equipment (Cogent Systems
Inc, 2010)
Figure 3-1 OPC process control
Source: (Cogent Systems Inc, 2010)
3.2.1. Programmable Logic Controller Architecture
This project aims to make a generic industrial application to acquire real-time data
through the PLC programming. The ladder programming which reassembles a typical
manufacturing plant is embedded in the RAM memory and the permanent storage for the
operating system is stored in the ROM memory. The input and output channels have
temporary buffer stores for the purpose of information transmission (Bolton, 2012). The
typical architecture of the PLC used in this project is shown in Figure 3-2 consisting of
the major components such as the central processing unit, memory and input/output
interfaces. All information and data are transmitted via a bus system and is sent from the
CPU to memory and input/output units. The bus system includes control commands,
address routes, input data collection and output data execution (Bolton, 2012). The central
processing unit (CPU) controls and executes the program logic for all components of the
33
PLC according to the frequency set from 1 to 8 MHz located as the operational speed of
timing and synchronisation of all PLC elements (Bolton, 2012).
Figure 3-2 PLC architecture
Source: (Bolton, 2012)
3.2.2. PLC programming and communication software
Allen Bradley control software is used for the implementation of the proposed model as
this dissertation focuses on the PLC programming code and the UI development for
measuring real-time KPI. The project is created by configuring PLC emulation software
such as RsLogix Emulator 5000 and the PLC control software Rslogix5000. RsLogix
Emulator 5000 software was employed to emulate the function of the PLC controller
without the real hardware and hence test the HMI application with simulated I/O modules.
RsLinks manages the communications between the PLC controllers and the HMI. The
PLC ladder program developed in RsLogix 5000 is used by the PLC to interpret the input
signals and operate the program outputs accordingly to the embedded code (Allen-
Bradley, 2016). All steps for configuring RsEmulate with RsLogix5000 through RsLinks
application are outlined in Appendix A.
3.2.3. Databridge
Databridge software uses the Extract-Transform-Load methodology for real-time data
acquisition, database recording, and mathematical computations (On-Control Inc, 2015).
34
This means that Databridge collects the real-time variables of the industrial controllers in
a defined module, transforms these variables using mathematical calculations, and
records the output result into real-time databases. Databridge supports different
communications protocols, and the actions to be executed in a module are defined by the
algorithms depending on the data type. For example, the OPC module extracts directly
from the OPC server the PLC tags, i.e., the particular route each I/O system has within
the controller programming.
This generic application extracts input data from industrial controllers or historical
databases. Among the communication network protocols supported by the ETL
application modules are comma-separated values (CSV), MODBUS, Ethernet/IP. The
output variables can be loaded to real time databases such has Structured Query Language
(SQL), Simple Object Access Protocol (SOAP), DF1 (Allen Bradley RS232 interface
modules) (On-Control Inc, 2015)
3.2.4. Human Machine Interface (HMI)
According to (Jander, et al., 2012) study of a methodological framework to evaluate the
human-machine interaction (HMI) readiness in system development for task-critical
environments, the primary objective of a user interface is that it can communicate
information through it into a system. Once this communication is achieved, the next goal
is to develop such communication in the easiest and most convenient way possible for the
characteristics of the user who uses the service. Considering this approach, of the
numerous possible interfaces designs (user-centred design, activity-oriented design,
scenario-based design) (Oppermann, 2002), the activity oriented-design was chosen as
this project seeks to develop a generic Real-Time Discrete Event Application to Measure
KPI in any industry type. This measurement is achieved by measuring relevant production
metrics in any production lines.
The HMI application is developed using LabView software of National Instruments Inc.
LabView software applications are varied and important, such as data acquisition and
signal processing, instrument control, automated testing and validation systems, and
monitoring systems and automatic control (National Instruments, 2013). The KPI models
are programmed using graphical representations of functions to control the front panel
objects from the real-time input data. This method is a graphical programming language,
characterised by using icons that allow visual programming from a data stream allowing
the user to focus on the process and not in programming codes.
35
This user interface displays all conditions, state and relevant parameters of the system
including the state of resources (idle, busy, standby, failure). Not only that, but the UI
displays all the time each device spends in each state. From this information, KPIs are
calculated and displayed. Apart from assessing the state of the resources, the UI shows
the inputs and outputs counters of parts going through each machine in the flow line
production line as numeric indicators. Also, all part sensors and electrical switches can
be active and deactivate from this user interface.
3.3. Classification of systems parameters
It is important to understand the underlying relationship among industrial KPIs with the
purpose of developing a system approach to derivate comprehensive performance metrics
which accurately assess the overall state of a manufacturing system. In fact, industrial
KPIs are not independent measures of performance, but they contain intrinsic reciprocal
relationship between all activities and factors that impact the efficiency and performance
of a production system. By making consisting efforts to optimise common production
variables, usual non-value added activities (in the production Value Steam Mapping) that
generate waste and unnecessary manufacturing costs can be limited and reduced.
Generally, non-value added activities comprehend Energy, Availability, Quality and
Performance Losses.
In fact, (Kang, et al., 2015) developed a hierarchical structure study of KPIs with the
purpose of inferring pairwise dependencies among performance metrics in a
manufacturing system. By knowing these relationships between KPIs, it is possible to
determine the common supporting elements from which the calculation of the KPIs is
based on. These supporting elements are measures of time and quantity which are
monitored on machines, orders or workers at the production level. Additionally, KPIs are
grouped into different categories depending on the disclose of system performance
features. For example, by measuring the production time on the machines and by
measuring the quantity of production in the system, the KPI Production Rate, and Energy
Consumption are obtained as Basic KPI which grants information about a single feature
of performance operation (Energy, and Throughput). From these two basic KPIs, the
Energy Efficiency Indicator is obtained as a comprehensive KPI which gives an overall
assessment of the production system, and it is based on the measures of the basic KPIs at
the supporting elements of the manufacturing system.
36
Based on the above methodology, the development of Basic and Overall KPIs for this
dissertation is shown in Figure 3-3.
Figure 3-3 Methodology for KPI categorization
Adapted from: (Kang, et al., 2015)
3.4. Work-In-Process
According to (Little, 1961) study in Queuing Theory, the quantity of parts located in a
factory at a discrete time is directly proportional to the rate at which these parts are
processed through the plant multiplied by the time the parts spend in the facility. This
formula is constrained to steady-state conditions meaning that the initial uncertainty given
by abnormal operating conditions at the start of the production sequence such as readiness
of resources, the scarce flow of raw materials, are eliminated in a flow-line production
plan. The initial transient state is eliminated when all line production resources achieved
the desired cycle time.
Liilte´s Work-in-Process formula is defined by the following equation:
𝐿 = 𝑊 ∗ 𝜆 (3-1)
Where:
37
L= the expected number of units in the system, parts
Λ= processing rate of units in the system, parts/min
W= expected time that a unit spends in the system, min
From this definition, (Groover, 2016) correlates L (the expected units in the system at a
discrete time) with a factory Work-in-Process (the quantity of parts being processed or
between processing operations, at a given time). In this way, the processing rate of units
(Λ) is now (Rh) the hourly plant production rate (parts/hours), and the expected time a
part spends in the system (W) is indicated as the average manufacturing lead-time (MLT,
hours). Equation 2 shows this relationship
𝑊𝐼𝑃 = Rℎ ∗ 𝑀𝐿𝑇 (3-2)
Where:
WIP= Work-in-Process in the plant, parts
Rh = hourly plant production rate, parts/hours
MLT= average manufacturing lead time, hours
The hourly plant production rate accounts for all operations to produce a specific part,
and the set of production rate of the plant´s resources as (Groover, 2016):
𝑅𝑝𝑝ℎ = ∑𝑅𝑝𝑖∗𝑓𝑖
𝑛𝑜
𝑛𝑖=1 (3-3)
Where:
Rph = average hourly plant production rate, parts/hours;
Rpi = production rate of machine i when processing part style j, parts/hour;
no = the number of operations required to produce part style j,
fi = the fraction of time that machine i is processing part style j.
MLT incorporates all the operating time (the time a part spends in a resource being
processed known as Value Added Time) and non-operation time (the non-Value Added
time which a part spends on queues, being transferred, or handled) The non-operating
time also comprehends equipment availability (the time probability a resource is able to
operate before failure occurs) (Groover, 2016). For a flow-line mass production, a part
must go through all the processing units one at a time according to the processing
sequence. Equation 3 shows MLT for a flowline mass production:
38
𝑀𝐿𝑇𝑖 = ∑ (𝑇𝑠𝑢 + 𝑄𝑗𝑇𝑐𝑖 + 𝑇𝑛𝑜𝑖)𝑛𝑜𝑗𝑖=1 (3-4)
Where:
MLT = average manufacturing lead time, min
𝑇𝑠𝑢= Setup time for operation i on part j, min
𝑄𝑗= quantity of part j in the batch (for job shop floor production Q=1), parts
𝑇𝑐𝑖 = cycle time for operation i on part or product j, min/pc
𝑇𝑛𝑜𝑖= non-operational time associated with operation i, min
Given that the cycle time for a flow-line mass production line is the minimum time a part
spends on each resource of the line; this metric considers all operations in each machine
for making a work unit. Furthermore, the particular setup time for each activity at a single
resource is also counted. Therefore, MLT is defined as the time between start and
completion of a part of the line. As this project seeks to develop a generic KPI
measurement model for a broad range of industrial activities, the effects of setup time,
and handling time in the WIP analysis can be simplified by setting input and output
sensors on all machines in the production line, assembly station and warehouse. By doing
so, all parts being processed and all parts waiting in a queue. Thus, it will measure all
parts currently in the system. This approach is in accordance with (Kelton, et al., 2015)
pg.111 where a simplified WIP expression defined as the total number of parts in the
system, for any given time, WIP is the number in queue plus the number of parts being at
a processing operation. Equation 3-5 shows WIP calculation:
𝑊𝐼𝑃 = ∑ (𝑃𝑃𝑀𝑖 + 𝑁𝑊𝑄𝑖)𝑛𝑖=1 (3-5)
Where:
PPMi = parts in process at machine i, parts
NWQi = Number Wating in Queue i, parts
n =total number of workstations
Equation 3-5 can be simplified as the difference between the number of parts entering the
system minus the number of components exiting the system. Thus, the measure of WIP
at any production time will be given by the overall quantity of parts in the system. In this
approach, WIP is measured by parts sensors at the input and output of the production line.
Each sensor is linked to a counter function in the PLC programming code.
WIP=N_In - N_Out; (3-6)
39
WIP= Work-In-Process for each entity type, parts
N_In= number of entities that entered the system for each entity type, parts
N_Out= number of entities that left the system for each entity type, parts
3.5. Electricity Consumption
Considering a production line as a series of processing steps within each manufacturing
resource, the electricity requirements for all the processing operations and the different
states of each equipment will regularly change within the production period. In order to
correctly assess the electricity consumption of a manufacturing system, it is necessary to
track the time each resource was in each of these States describe in 2.7.1 with the purpose
of reporting the required statistics. The biggest energy requirement in a manufacturing
equipment is to start up the process and maintain the equipment in the idle state
(Gutowski, et al., 2006). The power requirements to take a resource from the inactive to
the idle state and then sustain an operating condition is modelled by Equation 3-7. Here,
the overall power consumption is proportional to the quantity of material being processed,
and the idle power.
𝑃 = ∑ 𝑃𝑜 𝑖𝑛𝑖=1 + 𝑘�̇� (3-7)
Where:
P = total power requirement, kW
Po = idle power of each resource, in kW
�̇� = rate of material processing, cm3/sec,.
k = process constant energy, kJ/cm3
n= number of resources in the production line (Gutowski, et al., 2006)
The idle power is given by the equipment features which can be found on the technical
manufacturer specifications. The constant energy rate is provided by the particular
operation taking place representing the amount of energy needed in the process. The
above model considers the power requirements as pure exergy that can be tracked down
to fuel consumption at power plant generation (Gutowski, et al., 2006). Equation 3-8
defines the electricity consumption fluctuation depending on the equipment state, and the
time spent in each particular state as (EERE, 1999):
𝐸 = ∑ 𝑃𝑖×𝑡/60𝑛𝑖=1 (3-8)
40
Where:
E = Overall electricity consumption in the production line, kilowatt-hours (kWh)/day
P= Power consumption in each state, W
t= time the resource is operating in each state (idle, busy, standby), minutes
N= number of available resources in the production line.
The power consumption is considered as nominal power. This is the maximum power
demanded by a machine under normal use at each state. This approach considers each
resource to withstand the amount of power demanded by the manufacturing process.
However, due to fluctuations in current, overuse, or in situations other than the design
specifications, the actual power can differ from the nominal, being higher or lower
(Atkins & Escudier, 2013) .
3.6. Methodology for Calculating the 2016 GHG Emission Factor
The official UK government methodology "Defra´s Guidelines to GHG Company
Reporting" establishes that to estimate the amount of GHG emissions in an industrial
production sector, it is necessary to collect relevant activity data related to the plant
production operations. For example, this activity related data can be the amount of
electricity used or fuel consumed, and then multiply it by an (emission) conversion factor
as expressed by Equation 3-9:
GHG emissions [kgCo2e] = activity data [kWh] x emission factor [kgCO2e/kWh] (3-9)
The above equation shows the calculation of all significant GHGs emissions combined
(kg CO2e per electricity consumption). The factors are then divided into separate
emissions factors for each gas (kg CO2e of CO2/CH4/N2O per electricity consumption)
which aggregate to the total amount of kg CO2e emitted per electricity consumed
(DEFRA & DECC, 2012)
This section shows the methodology for calculating the emissions factor for the year 2016
based on data from electricity generated factors, transported and distributed electricity
factors, and electricity consumption factors from previous years.
The following tables show the data collection on emission factors of the last decade
(2006-2015). The information gathering was carried out in order to obtain a representative
41
sample thereof, by investigating official literature sources as "GOV.UK" cited in the
reference (Department for Business Energy & Industrial Strategy, 2016)
Electricity Generated (kwh)
UK Grid Electricity Year
kg CO2 kg CH4 kg N2O
2006 0,47033 0,00021 0,00283
2007 0,46359 0,00022 0,00291
2008 0,49263 0,00022 0,00322
2009 0,49054 0,00024 0,00303
2010 0,48219 0,00026 0,00286
2011 0,44917 0,00027 0,00261
2012 0,45706 0,00028 0,00267
2013 0,44238 0,00029 0,00281
2014 0,49023 0,00033 0,00369
2015 0,4585 0,00035 0,00334
Table 3-1 Emission Factors (Energy generated (kwh))
Source: (Department for Business Energy & Industrial Strategy, 2016)
T&D- UK electricity (kwh)
UK Grid Electricity Year
kg CO2 kg CH4 kg N2O
2006 0,04487 0,00002 0,00027
2007 0,03621 0,00002 0,00023
2008 0,03831 0,00002 0,00025
2009 0,03884 0,00002 0,00024
2010 0,03883 0,00002 0,00023
2011 0,03838 0,00002 0,00022
2012 0,03611 0,00002 0,00021
2013 0,03783 0,00002 0,00024
2014 0,04287 0,00003 0,00032
2015 0,03786 0,00003 0,00028
Table 3-2 Emission Factors (Energy Losses (kwh))
Source: (Department for Business Energy & Industrial Strategy, 2016)
According to (DEFRA & DECC, 2012) methodology to calculate and report GHG
emissions, it is established that the emission factor for Consumed Electricity is expressed
by the following equation:
42
Emission Factor (Electricity CONSUMED) =
Emission Factor (Electricity GENERATED) + Emission Factor (Electricity
LOSSES T&D) (3-10)
For instance, the overall 2010 kgCO2e emissions factor for consumed electricity in terms
of equivalent kilograms of CO2 is determined as:
Emission Factor (Electricity CONSUMED) 2010 kgCO2e= 0, 48219 + 0, 03883
Emission Factor (Electricity CONSUMED) 2010 kgCO2e= 0,521020 kg CO2e/kWh
Table 3-3 summarize the results Electricity Consumed (Emission Factor) obtained by
computing Equation 3-11 for the years 2006 to 2015 with respect to all GHG types
Electricity Consumed (kwh)
UK Grid Electricity Year
kg CO2 kg CH4 kg N2O
2006 0,515200 0,000230 0,003100
2007 0,499800 0,000240 0,003140
2008 0,530940 0,000240 0,003470
2009 0,529380 0,000260 0,003270
2010 0,521020 0,000280 0,003090
2011 0,487550 0,000290 0,002830
2012 0,493170 0,000300 0,002880
2013 0,480210 0,000310 0,003050
2014 0,533100 0,000360 0,004010
2015 0,496360 0,000380 0,003620
Table 3-3 Emissions Factor (Electricity Consumed (kwh))
The Total Factor of Greenhouse Gases (GHGs) was calculated by the algebraic sum of
the different types of GHG factors for each year, as shown below with an example for the
year 2010:
TOTAL GHG Emission Factor (Electricity CONSUMED) 2010 kgCO2e =
Emission Factor (Electricity CONSUMED) 2010 kgCO2 + Emission Factor
(Electricity CONSUMED) 2010 kgCH4 + Emission Factor (Electricity
CONSUMED) 2010 kgN2O (3-11)
43
TOTAL GHG = 0.521020 + 0.000280 + 0.003090.
TOTAL GHG = 0.524390 kg CO2e
Electricity consumption (kwh)
UK Grid Electricity Year
kg CO2 kg CH4 kg N2O TOTAL GHG kg CO2e
2010 0,521020 0,000280 0,003090 0,524390
Table 3-4 Emission Factor (Electricity Consumed (kwh)) – Calculation for 2010
According to the researched data, the statistics for the year 2016 are partially with
Electricity Generated (emission factor). Therefore, it is necessary to calculate the Energy
Loss factor) and Electricity Consumed factor projected to 2016. This computation was
performed by the use of average, and the maximum and minimum functions with respect
to the emission factors data range from 2006 to 2015 as shown on Table 3-5.
Electricity generated (kWh)
UK Grid Electricity Year
kg CO2 kg CH4 kg N2O TOTAL GHG kg
CO2e
2006 0,47033 0,00021 0,00283 0,47337
2007 0,46359 0,00022 0,00291 0,46673
2008 0,49263 0,00022 0,00322 0,49608
2009 0,49054 0,00024 0,00303 0,49381
2010 0,48219 0,00026 0,00286 0,48531
2011 0,44917 0,00027 0,00261 0,45205
2012 0,45706 0,00028 0,00267 0,46002
2013 0,44238 0,00029 0,00281 0,44548
2014 0,49023 0,00033 0,00369 0,49426
2015 0,4585 0,00035 0,00334 0,46219
AVERAGE 0,469662 0,000267 0,002997 0,47293
MAXIMUM 0,49263 0,00035 0,00369 0,49608
MINIMUM 0,44238 0,00021 0,00261 0,44548
Table 3-5 Calculation of the Average, Maximum and Minimum Electricity Generated
Emission Factors (2006-2015)
T&D- UK electricity (kWh)
UK Grid Electricity Year
kg CO2 kg CH4 kg N2O TOTAL GHG
kg CO2e
2006 0,04487 0,00002 0,00027 0,04516
2007 0,03621 0,00002 0,00023 0,03646
2008 0,03831 0,00002 0,00025 0,03857
2009 0,03884 0,00002 0,00024 0,0391
2010 0,03883 0,00002 0,00023 0,03908
2011 0,03838 0,00002 0,00022 0,03863
2012 0,03611 0,00002 0,00021 0,03634
2013 0,03783 0,00002 0,00024 0,03809
2014 0,04287 0,00003 0,00032 0,04322
44
2015 0,03786 0,00003 0,00028 0,03816
AVERAGE 0,039011 0,000022 0,000249 0,039281
MAXIMUM 0,04487 0,00003 0,00032 0,04516
MINIMUM 0,03611 0,00002 0,00021 0,03634
Table 3-6 Calculation of the Average, Maximum and Minimum Transmission and
Distribution Losses Emission Factors (2006-2015)
Electricity consumption (kwh)
UK Grid Electricity Year
kg CO2 kg CH4 kg N2O TOTAL GHG kg
CO2e
2006 0,515200 0,000230 0,003100 0,518530
2007 0,499800 0,000240 0,003140 0,503190
2008 0,530940 0,000240 0,003470 0,534650
2009 0,529380 0,000260 0,003270 0,532910
2010 0,521020 0,000280 0,003090 0,524390
2011 0,487550 0,000290 0,002830 0,490680
2012 0,493170 0,000300 0,002880 0,496360
2013 0,480210 0,000310 0,003050 0,483570
2014 0,533100 0,000360 0,004010 0,537480
2015 0,496360 0,000380 0,003620 0,500350
AVERAGE 0,508673 0,000289 0,003246 0,512211
MAXIMUM 0,5331 0,00038 0,00401 0,53748
MINIMUM 0,48021 0,00023 0,00283 0,48357
Table 3-7 Calculation of the Average, Maximum and Minimum Emission Factors for
Electricity Consumption (2006-2015)
As mentioned above, there is no official data for the 2016 Electricity Consumed factor
thus the 2016 projection for the Electricity Consumed and Energy Losses are carried out.
However, there is official data for the Electricity Generated factor for the 2016 UK Grid
which is shown Table 3-8.
Electricity Generated (kwh)
UK Grid Electricity Year
kg CO2 kg CH4 kg N2O TOTAL GHG kg CO2e
2016 0,40957 0,00039 0,00209 0,41205
Table 3-8 Values for the Emission Factor (Electricity GENERATED) for 2016
Source: (Department for Business Energy & Industrial Strategy, 2016)
In order to conduct the projection to 2016 of the Energy Consumed (emissions factor),
first it is considered an average emission factors (Energy Loss) for the years 2006 to 2015
as tentative values for 2016.
45
Electricity Generated (kwh) T&D LOSSES- UK electricity (kwh)
UK Grid Electricity Year
kg CO2 kg CH4 kg N2O TOTAL GHG
kg CO2e kg CO2 kg CH4 kg N2O
TOTAL GHG kg CO2e
AVERAGE 0,469662 0,000267 0,002997 0,47293 0,039011 0,000022 0,000249 0,039281
2016 0,40957 0,00039 0,00369 0,41205 0,039011 0,000022 0,000249 0,039281
Table 3-9 Values for the Emission Factor (Electricity Losses) for 2016
Taking into account the results shown in Table 3-9, it is proceeded to calculate the Energy
Consumed (emission factor) for the year 2016, by using Equation 3-11 for all types of
GHG and its total. The results are presented in in Table 3-10.
Electricity Consumption (kwh)
COMPONENTS kg CO2 kg CH4 kg N2O TOTAL GHG kg CO2e
PROJECTION 2016 0,448581 0,000412 0,002339 0,451331
Table 3-10 Values for the Emission Factor (Electricity Consumed (kwh)) projected to
2016
Finally, the estimate of the average, maximum and minimum values projected for 2016
of each of the components of Greenhouse Gases (GHGs) and its total is made. The results
obtained are the result of calculating the values obtained previously in Table 9 averaged
with the values "PROJECTION 2016", an example of calculation is presented below
Projection 2016 Kg CO2 = 0.448581 Kg CO2 /kWh
Average (2006-2015) 2016 Kg CO2 = 0.508673 Kg CO2 /kWh
Thus:
Average 2016 Kg CO2 = (0.448581 + 0.508673) /2= 0.478627 Kg CO2 /kWh
Electricity Consumption (kwh)
COMPONENTS kg CO2 kg CH4 kg N2O TOTAL GHG kg CO2e
PROJECTION 2016 0,448581 0,000412 0,002339 0,451331
AVERAGE 2016 0,478627 0,000351 0,002793 0,481771
MAXIMUM 2016 0,490841 0,000396 0,003175 0,494406
MINIMUM 2016 0,464396 0,000321 0,002585 0,467451
Table 3-11 Emission Factor (Electricity Consumed (kWh)), projection, average,
maximum and minimum for the year 2016
46
Chapter 4: Implementation
4.1. Introduction
This chapter explains the design and development of the real-time Discrete Event
Application to measure Key Performance Indicators. The implemented system
comprehends two main components: The PLC programming, configuration and
communication made in Rslogix5000, RsEmulate, and RsLinks, and the User-Interface
(UI) developed in Databridge and LabView software for data presentation and process
simulation and control. A hypothesised manufacturing system is proposed to create the
controller logic, acquire real-time data, simulate the discrete event system, and measure
relevant performance metrics described in chapter 3. The proposed production system
scenario to implement and test the generic application is explained as follows:
• A flow-line production system comprised of four material processing units
(workstations), an assembly station, and a quality control unit. The type of
operation is a flow-line sequential mass production. The UI incorporates Power
ON/OFF buttons for all resources in the scheme as well as Part Sensors to detect
the quantity of material going through the production line.
• The pre-empted failure and schedule rule are set for all resources in the system
meaning that the production will stop at the exact moment unplanned stops occur.
Additionally, the uptime, downtime and scheduled capacity are specified by the
final user of the application and can be modified from the UI.
The sequential material flow going from Machine 1 through the Warehouse is described
in the next flowchart (Figure 4-1). At each workstation two sensors are located for
measuring the quantity of parts entering and exiting a single material processing units.
47
Machine 1
Queue 1
Machine 2
Queue 2
Machine 3
Queue 3
Machine 4
Input Parts Mach
1
Output Parts
Mach 1
Input Parts Mach
2
Output Parts Mach
2
Input Parts
Mach 3
Output
Parts Mach
3
Input Parts
Mach 4
Output Parts
Mach 4
Queue 4
Machine 5
Queue 5
Assembly Machine
Queue 6
Warehouse
Input Parts
Mach 5
Output Parts Mach
5
Input Part
Assembly
Output Part
Assembly
Input Warehouse
Output Warehouse
Figure 4-1 Flow Line Processing Sequence
In Rslogix 5000, the memory allocation method for the controller is defined by tag
databases (Allen-Bradley, 2016). The Input and Output modules relating to the PLC tags
are described by Table 4-1 and 4-2.
PLC Input Tag Name
3:I.Data[1].0 Reset Parts Counters
3:I.Data[1].1 Power On Mach 1
3:I.Data[1].2 Power Off Mach 1
3:I.Data[1].3 Input Parts Mach 1
3:I.Data[1].4 Output Parts Mach 1
3:I.Data[1].5 Input Parts Mach 2
3:I.Data[1].6 Power On Mach 2
3:I.Data[1].7 Power Off Mach 2
3:I.Data[1].8 Output Parts Mach 2
3:I.Data[1].9 Input Warehouse
3:I.Data[1].10 Output Warehouse
3:I.Data[1].11 Power On Assembly
3:I.Data[1].12 Power Off
Assembly
3:I.Data[1].13 Input Part Assembly
3:I.Data[1].14 Output Part
Assembly
3:I.Data[1].15 Reset Timers
3:I.Data[1].16 Input Simulation
3:I.Data[1].17 Stop Simulation
3:I.Data[1].18 Power On Mach 3
3:I.Data[1].19 Power Off Mach 3
3:I.Data[1].20 Input Parts Mach 3
3:I.Data[1].21 Output Parts Mach 3
3:I.Data[1].22 Power On Mach 4
3:I.Data[1].23 Power Off Mach 4
48
3:I.Data[1].24 Input Parts Mach 4
3:I.Data[1].25 Output Parts Mach 4
3:I.Data[1].26 Power On Mach 5
3:I.Data[1].27 Power Off Mach 5
3:I.Data[1].28 Input Parts Mach 5
3:I.Data[1].29 Output Parts Mach 5
3:I.Data[1].30 Stop Failures
3:I.Data[1].31 Start Failures
Table 4-1 PLC Input Tag Database
The following table shows the PLC Output Tag Database for the ladder program
PLC Output Tag Name
0.0 Machine 1
0.1 Machine 2
0.2 Machine 3
0.3 Machine 4
0.4 Machine 5
0.5 Assembly
Table 4-2 PLC Output Tag Database
4.2. Programmable Logic Controller ladder code
The PLC ladder code starts with the measurement of the Planned Production Time
establishing this period as the available line operating time the system is scheduled to
work according to the plant planned shift in a day. Any planned stops during the
production time are not included in this measurement since the system will be in the non-
operative state and the entire production line will be switched off.
In order to measure the available time, both a Retentive Timer (RTO) and a Count Up
Counter (CTU) are employed. RTO timer is applied to track the time when an instruction
is on or off while keeping track on a retentive base (Allen-Bradley, 2016). In this case, a
new tag named "Available-Seconds" is created in the RTO controller. The Present Value
instruction is set as 60000 milliseconds specifying the value the timer has to attain before
the controller triggers the Done bit. Once the "Availabe_Time_Seconds" Done bit is
active, the next rung incorporates a CTU counter to measure the "Available-Minutes" tag.
The CTU instruction increments its accumulated value (ACC) at each false to true
transition, increasing its stored value by one count (Allen-Bradley, 2016). A maximum
value of 8751 minutes is specified in the CTU Present Value for when the ACC value
49
reaches the defined present value, the done bit in the CTU is activated stopping the count.
These steps are shown in Figure 4-2.
To reset the RTO accumulated value, a reset instruction (RES) with the same address
"Available-time-seconds" is included. This action will generate a 1-minute cycle
continuously increasing the CTU accumulated value for each minute the production line
is running. Figure 4-2 shows the RTO and the CTU ladder code for the "Available-time-
seconds" and "Available-time-minutes" tags.
Figure 4-2 Line available operating time code
In the next rung, a “normally open contact” (XIC) connected to the PLC input
“3:I.Data[1].1” is placed as a start switch to power ON Machine 1. Likewise, a “normally
closed" instruction (XIO) linked to the PLC input “3:I.Data[1].2” works as a stop switch
to power OFF Machine 1. In this application, (XIC) and (XIO) are input commands
analogous to relay contacts that can be triggered from the RsEmulate I/O module or from
the LabView user interface. Additionally, an “Output Energize instruction” (OTE)
connected to the PLC output tag “Machine 1” (PLC output 3:O.Data[0].0) is placed in the
program. OTE output is analogous to a coil relay that can energised an equipment (Allen-
Bradley, 2016). When the power on button is triggered, the contact XIC is closed, and
hence current to OTE is sent. Machine 1 will then be powered on by energising the motor
contactor coil. Another XIC input with the same address “Machine 1” as the OTE has
been placed in parallel latching the circuit. Once the power on switch is released, the XIC
associated with it goes back to a normally-open position, but the motor of Machine 1 will
continue to be energised because the contact in parallel bypasses the current to the OTE.
Additionally, another normally closed contact addressed as the “Mean Time between
Failures” (MTBF) of Machine 1 is placed to interrupt the current to the OTE. This action
50
will power off Machine 1 when a breakdown occurs. All failures in each machine of the
production line are established from the LabView user-interface in terms of uptime and
downtime. Figure 4-3 shows the start and stop ladder logic for Machine 1. The similar
logic is applied to power on/off the remaining equipment on the line.
Figure 4-3 Start/Stop logic code for Machine 1
Once Machine 1 is energized, the program starts measuring the equipment Operating
Time referred as the time a Machine is powered ON and available for equipment
processing. Since the previously defined XIO instruction for MTBF, shuts down the
resource, the Operating Time excludes any Down Time caused by unplanned
maintenance, and breakdowns. A retentive timer RTO measures each 60 seconds cycle
(Present value) while sending the signal of each completed bit to a CTU counter to rate
the accumulate value in minutes the equipment is powered on. Additionally, a XIC entry
associated with an Input Part Sensor for the PLC input address “3:I.Data[1].3” gives the
signal each time a part (material being manufactured through the production line) enters
Machine 1. The amount of parts incoming “Workstation 1” are then tallying by a CTU
counter that keep the accumulate value of each trigger of the “Input Part Sensor Mach 1”
(see Figure 4-4).
Figure 4-4 Available time and Input sensor CTU for Machine 1
51
When a part enters a resource, the equipment varies its state from an idle to a busy
condition. To account this busy time, an auxiliary OTE addressed with a program defined
tag named “Busy-Mach-1” is created. This auxiliary OTE will be triggered if Machine 1
is powered on and a part has been detected by the “Input Parts Sensor Mach 1” (PLC
entry 3:I.Data[1].3) as shown in Figure 4-5. The resource in this Workstation will change
its state again once the processing time is done and the part leaves the work-terminal.
Hence, an XIO (evaluate if open) instruction associated with the “Output Part Sensor
Mach 1” (PLC entry 3:I.Data[1].4) is allocated to interrupt the sequence of the auxiliary
OTE “Busy-Mach-1”.
Furthermore, all the time a machine is not in a Busy state, it will be in an Idle condition
if there are no unexpected failures or unplanned stops (standby mode). Therefore, an XIO
command linked to the auxiliary OTE “Busy_Mach_1” is associated with another
auxiliary OTE named “Idle_Mach_1”. Additionally, if the “Busy” OTE is triggered the
program will then start measuring the time the resource is on a Busy state by employing
a 60 seconds cycle RTO timer and a CTU counter for measuring the Busy time in minutes.
Also, for logging the quantity of parts that leave Machine 1, each time the “Output Part
Sensor Mach 1” is triggered it will be recorded by a CTU counter linked to a user defined
tag named “Output Machine 1” (see Figure 4-5)
Figure 4-5 Busy Time and CTU output for Machine 1
52
As total Wok-In-Process of the production line accounts for all parts that are being
processed or being transferred between machines, the difference between the parts that
exit Workstation 1 and the parts that enter it gives the number of parts being processed
by the resource. Therefore, to measure the Parts-in-Process of Machine 1, a Subtract
command is allocated to assess the difference between the accumulated value of
“Input_Machine_1.ACC” and the accumulated value of the “Output_Machine_1.ACC”.
Both of these tags are associated with the CTU counters previously defined for the “Input
Part Sensor Mach 1” and the “Output Part Sensor Mach 1” (see Figure 4-6). This approach
to calculating Parts- in-Process is repeated for the rest of the machines on the production
line.
Figure 4-6 Parts in Process Machine 1
All the time a resource is not in the busy state or the idling condition, then the machine is
on Standby mode (Inactive state) if there are no unexpected breakdowns due to equipment
failure. Therefore, an auxiliary OTE command associated with a user defined tag named
“Inactive_Mach_1” is activated in the program if the PLC entry 3:I.Data.[1].2 (“Power
off Mach 1”) is triggered and the PLC output “Machine 1” is not energised. Additionally,
the resources change its state again from Inactive to Idle when powered on. Therefore, an
XIO instruction linked to the PLC input “Power On Mach 1” will stop activating the
auxiliary OTE “Incative_Mach_1” and the auxiliary OTE “Idle Mach 1” will be active
each time the resource is not working (Busy mode). Once a part departures Workstation
1, it is transferred to the following station in the processing sequence and wait for any
additional time for the downstream Machine 2 to be Idle. Hence, to account for this Queue
Time, an auxiliary OTE named “Aux_Queue 1” is activated if the “Output Parts Sensor
Mach 1” is triggered. The auxiliary OTE “Queue 1” is deactivated when a part enters
Workstation 2 and triggers the sensor “Input Parts Mach 2”. Consequently, an XIO
instruction linked to the “Input Parts Mach 2” (PLC entry 3:I.Data.[1].5) interrupts the
auxiliary OTE “Queue_1” (See Figure 4-7)
53
Figure 4-7 Inactive State for Machine 1 and Queue 1 OTE
To establish the amount of parts that are waiting in the Queue 1, a Subtract command is
employed to calculate the difference between the accumulated value of “Output Machine
1. ACC” and “Input Machine 2. ACC”. The first ACC value is associated with the CTU
counter for the “Output Part Sensor Machine 1” (PLC entry 3.I.Data.[1].4) and the second
value is linked to the CTU counter “Input Part Sensor Machine 2” (PLC entry
3.I.Data.[1].5). this approach to computing the Number Waiting is replicated for the rest
of the Queues in the system considering the output counter of the upstream workstation
and the input counter of the downstream station (see Figure 4-8).
Figure 4-8 Number Waiting Queue 1 and Parts in Process Machine 2
Finally, an additional Input routed as “Reset Timers” (PLC input 3:I.Data.[1].15) are
linked to reset instructions for all RTO timers including the busy and operating mode in
minutes and seconds. Similarly, a XIC input named as “Reset Part Counters” (PLC entry
3:I.Data.[1].0) is set to reset all parts counter in the system (See Figure 4-9)
54
Figure 4-9 Reset Timers and Counter
4.3. Key Performance Indicators User Interface
For developing the user interface in LabView, first Booleans controls are placed in the
Front Panel. These controls are Push Buttons related to all PLC inputs as described in
section 4.1 including Power On and Power Off switches for all five Workstations, and
Assembly unit, sensors of parts being processed through the production line as Input and
Output for all the Terminals and Warehouse, Reset Timers and Counters. The connexion
between the PLC and the user interface is made through Data Socket Transfer Protocol
(DSTP) employing an OPC server explained in chapter 3.2. The LabView control Data
Binding properties are changed for the Booleans controllers. For example, to associate
the Machine 1 “Power On” PLC entry 3:I.Data.[1].1 with the LabView push button, the
Databinding properties of the LabView Boolean controller are changed using the DSTP
protocol routed to the OPC server path: opc://localhost/RSLinx OPC
Server/[Intento_5]Local:3:I.Data[1]/1 (see Figure 4-10)
55
Figure 4-10 User Interface OPC server connection
4.3.1. Production Rate per Hour
This section shows the measurement of system production rate, namely the number of
Manufactured Products (finished goods) over the available line production time.
Considering, the sequence of production described in section 4.1, the manufactured
products are defined as the number of parts that exit the system, i.e., have gone through
all the resources in the production line, and enter the warehouse for storage. Therefore,
the production rate per hour of the system is calculated considering the accumulated
values of the PLC tags “Input Warehouse.ACC” and “Available-time-minutes" tags
described by the following formula:
RealPR =Input_W
AvaTime/60 (4-1)
Where: RealPR= production rate, parts/hour
AvaTime= Planned Production Time, minutes
Input_W= Input Warehouse, parts
56
To represent the above measurement in the user interface, the Input Warehouse sensor
(PLC input 3:I.Data[1].9) is connected to a numeric control in LabView Block Diagram
using the OPC server route:
“opc://localhost/RSLinxOPCServer/[Intento_5]Program:MainProgram.Input_Warehous
e.ACC”
In this way, the accumulated value of the PLC program tag “Input Warehouse” is
displayed in the user interface. Additionally, a Formula Node is introduced to evaluate
formula 4-1 using as inputs the “Input Warehouse” tag and as Output variable the KPI
“Production Rate per hour” (see Figure 4-11). The result of the formula is then displayed
in a Numeric Indicator in LabView Front Panel.
Figure 4-11 Production Rate per Hour Formula Node and Front Panel Indicator Display
4.3.2. Number Waiting
For each Queue in the system, the “Number Waiting” tag described in 4.2 is routed to
several numeric controllers in the LabView Front Panel employing the OPC server
connection. Then, to display the performance metric "Number Waiting" a Build Array
option is employed to concatenate the different Number Waiting KPIs (of Queue 1 to
Queue 6) to a 1- dimensional array. Then, for displaying purpose, the 1-dimensional
grouped data is linked to a second Build Array followed by a 2-dimensional transpose
table. In this way, the clustered data is converted to a 2-dimenasional array and then
connected to a Number to Fractional String element for formatting options where the
digit’s precision is added. The correspondent results are shown in the LabView Front
Panel by a compiled 2-dimensional table as shown in Figure 4-12.
57
Figure 4-12 Number Waiting Table Array
4.3.3. Work-In-Process Calculation
As described in section 3.4 the measure of WIP at any production time will be given by
the overall quantity of parts in the system, i.e, the difference between the number of parts
entering the system (going in Machine 1) subtracted the number of components exiting
the system (entering the Warehouse). For this purpose, LabView numeric controls are
linked to the tag “Input Parts Sensor Mach 1” (PLC entry 3:I.Data[1].3) and the tag “Input
Warehouse” (PLC input 3:I.Data[1].9) which as described in section 4.2 these tags
correspond to the accumulated of CTU counters associated with sensors at the input and
output of the production line. Thus, these variables are used as inputs in a Formula Node
for evaluating Equation 3-6 as shown in Figure 4-13.
Figure 4-13 Work-in-Process Formula Node
58
Additionally, the amount of parts being processed at each resource is given by the PLC
tag “Parts in Process” described in section 4.2. This tag relates to Subtract instructions
which measure the difference between the accumulated value of CTU counters previously
defined for the PLC tags “Input Part Sensor” and the “Output Part Sensor” for every
machine. Therefore, LabView numeric controls are routed to the OPC server to address
the PLC tags “Parts in Process”. Then, for displaying purposes the PLC tags “Parts in
Process”, “Input Parts” and “Output Parts” for every Machine in the system are clustered
using Build Array, 2-dimensional Transpose, and Number to Fractional String elements
to form a compiled 2-dimensional table as shown in Figure 4-14.
Figure 4-14 Parts being Processed in the System Table Array
59
4.3.4. Electricity Consumption
To measure the Electricity Consumption for all machines in the system, arrays of 32-bit
floating point numbers for Busy, Idle and total Energy Consumption are incorporated in
the Formula Node. To evaluate these metrics, it is necessary to input the Power
Consumption data for all resources as described by Equation 3-8. To make a generic
assumption, this application considers the Idle Power Consumption as a quarter of the
Busy Power Consumption. In LabView sliders type numeric controllers are placed in the
Front Panel and linked to a new PLC tag created in RsLogix “Power Consumption
Machine 1”. This action is repeated for the rest of equipment on the production line. In
this way, the selected value in the user interface is recognised by the PLC controller
program. Additionally, to assess the time a machine is in the Idle state the following
formula is included which is a function of the PLC tags “Busy time” and the “Available
time” previously defined in section 4.2 for each resource in the system.
𝐼𝑑𝑙𝑒 𝑡𝑖𝑚𝑒 = ∑ 𝑂𝑝𝑒𝑟𝑎𝑡𝑖𝑛𝑔 𝑇𝑖𝑚𝑒 − 𝐵𝑢𝑠𝑦 𝑇𝑖𝑚𝑒𝑛𝑖=1 (4-2)
Where:
n= number of machines in the production line
Idle, Busy and Operating Time, minutes
Next, the following formulas derived from Equation 3-8 are included in the LabView
Formula Node:
𝐸𝐶𝐵 = ∑ 𝑃𝑐𝑜𝑛𝐵𝑢𝑠𝑦×𝑡𝑖𝑚𝑒𝐵𝑢𝑠𝑦
60
𝑛𝑖=1 (4-3)
𝐸𝐶𝐼 = ∑ 𝑃𝑐𝑜𝑛𝐼𝑑𝑙𝑒×𝑡𝑖𝑚𝑒𝐼𝑑𝑙𝑒
60
𝑛𝑖=1 (4-4)
𝐸𝐶 = ∑ 𝐸𝐶𝐼 + 𝐸𝐶𝐵 𝑛𝑖=1 (4-5)
Where:
ECB= Electricity Consumption Busy state, kWh
PconBusy= Power Consumption in the Busy State, W
ECI= Electricity Consumption Idle state, kWh
PconIdle= Power Consumption in the Idle State, W
EC= Total Electricity Consumed, kWh
Figure 4-15 shows the Formula Node with the above Equations added for all resources in
the system and the construction of a 2-dimensional Build Array Table to display the
indicators in the LabView Front Panel.
60
Figure 4-15 Electricity Consumption (Total, Idle and Busy State) Formula Node
4.3.5. Unit Consumption
Unit Consumption (UC) refers to the amount of energy employed for the overall
production output measured at discrete intervals during the Planned Production time. UC
is an efficiency ratio of the electricity consumed for making the overall the quantity of
produced parts described by Equation 4-6:
𝑈𝐶 =𝐸𝐶
𝐼𝑛𝑝𝑢𝑡_𝑊 (4-6)
Where:
UC= Unit Consumption of the system, kWh/part
EC= Total Electricity Consumption, kWh
Input_W= Finished Manufactured Products (Input Warehouse), parts
To evaluate Equation 4-6 in the Formula Node, the Total Energy Consumed described in
section 4.3.4 is divided by the PLC tag “Input Warehouse” (PLC input 3:I.Data[1].9)
which as explained in section 4.2 correspond to the accumulated of CTU counter
associated with the overall output of the production line. The Unit Consumption metric
is then displayed in Labview Front Panel using a numeric indicator as shown in Figure 4-
16.
61
Figure 4-16 Unit Consumption Formula Node
4.3.6. Green Houses Gas Emissions
Figure 4-17 presents the values of emission factors (UK electricity consumption (kWh)),
in the years 2006-2015 according to the official methodology described in section 3.6
Figure 4-17 Electricity Consumed Factor in the years 2006 to 2015
Source: (Department for Business Energy & Industrial Strategy, 2016)
Figure 4-18 indicates the results obtained in the evaluation of the 2016 projected values
of emission factors (electricity consumption (kWh)) for the different types of GHG
emissions as described in section 3.6.
0.000000
0.100000
0.200000
0.300000
0.400000
0.500000
0.600000
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
ELECTRICITY CONSMUED FACTOR Kwh VS YEARS 2006 TO 2015
kg CO2 kg CH4 kg N2O TOTAL GHG kg CO2e
62
Figure 4-18 Electricity Consumed Factor 2016
Source: The Author
Figure 4-19 summarizes the 2016 projection results obtained for the average, maximum
and minimum emission factors (electricity consumption (kWh)) as described in section
3.6.
Figure 4-19 Average, Maximum, Minimum and 2016 projection Energy Consumed Factor.
Source: The Author
Considering the above results (projected emissions factors for 2016) and Equation 3-9,
the following equations are employed to calculate the GHG emissions regarding
Electricity Consumption in the proposed manufacturing system
𝑘𝑔𝐶𝑂2 = 𝐸𝐶[𝑘𝑊ℎ]×0.448581 [𝑘𝑔𝐶𝑂2
𝑘𝑊ℎ] (4-7)
0.000000
0.100000
0.200000
0.300000
0.400000
0.500000
kg CO2kg CH4
kg N2OTOTAL GHG kg
CO2e
2016 PROJECTION ENERGY CONSUMED FACTOR kWh
kg CO2 kg CH4 kg N2OTOTAL GHG kg
CO2e
AVERAGE 2016 0.478627 0.000351 0.002793 0.481770
MAXIMUM 2016 0.490841 0.000396 0.003175 0.494401
MINIMUM 2016 0.464396 0.000321 0.002585 0.467451
PROJECTION 2016 0.448581 0.000412 0.002339 0.451332
0.0000000.1000000.2000000.3000000.4000000.5000000.600000
2016 ELECTRICITY CONSUMED FACTOR (kWh) (AVERAGE, MAXIMUM, MINIMUM)
63
𝑘𝑔𝐶𝐻4 = 𝐸𝐶[𝑘𝑊ℎ]×0.000412 [𝑘𝑔𝐶𝑂2𝑒
𝑘𝑊ℎ] (4-8)
𝑘𝑔𝑁𝑂2 = 𝐸𝐶[𝑘𝑊ℎ]×0.002339 [𝑘𝑔𝐶𝑂2𝑒
𝑘𝑊ℎ] (4-9)
𝐺𝐻𝐺[𝑘𝑔𝐶𝑂2𝑒] = 𝑘𝑔𝐶𝑂2 + 𝑘𝑔𝐶𝐻4 [𝑘𝑔𝐶𝑂2𝑒] + 𝑘𝑔𝑁𝑂2 [𝑘𝑔𝐶𝑂2𝑒] (4-10)
Where:
EC=total electricity consumed (kWh)
GHG= total greenhouses emissions measured in equivalent CO2 emissions (kgCO2e)
The above equations are then evaluated in the Formula Node using the total Energy
Consumption described in section 4.3.4. Next, a 2-dimensional array table (6 digits’
precisions) is created to display the metrics in the Font Panel (see Figure 4-20).
Figure 4-20 GHG emissions Formula Node
64
Chapter 5: Testing and Validation
5.1. Introduction
This chapter describes the methodology to test and validate the proposed model to measure
real-time KPIs. To accomplish this objective, an ideal manufacturing system is considered,
and three independent runs of the selected application were carried out to test and register
indicators of performance such as WIP, GHG Emissions, Energy Consumption, Energy
Efficiency, Production Rate, Number Waiting and Waiting Time. Then, the validation of the
system was made by comparing the results obtained in the Real-Time PLC application with
a Discrete Event System Modelling (DESM) developed in Arena. Additionally, for carrying
out the DESM simulation, suitable calculations of the number of replications was conducted
for the simulated model in an extended period of iterations.
The simulated model comprehends all processes and workstations described in section 4.1
for the selected manufacturing system case. Furthermore, DESM is applied to distinguish
where the system can be improved by gathering performance statistics of the system and test
improvement scenarios simulated in Arena software. These actions provide a comprehensive
report of the performance factors for optimising the manufacturing system, and they allow to
compare the results obtained in the simulated KPIs against the metrics found by employing
the Real-Time model.
5.2. Testing of the Real-Time Model
In order to test the model, three manual independent trials of 15, 30 and 45 minutes
respectively were carried out. In this way, all input and output part sensor for the
Machines are activated manually from the LabView user-interface according to the
selected manufacturing system described in 4.1. To implement these trials, ideal operating
conditions for the production line are considered, where equal distribution of processing
times, transfer times and arrivals of parts is created. The purpose if this scenario is to have
an effective balance between all workstations, eliminate queues in the system and to
maintain a constant work-in-process (CONWIP). The operational conditions of the
manufacturing system are described next:
• Constant and Equal Processing Times for all workstations in the system (Machine 1
to Machine 5 and Assembly unit) set to 1 minute
65
• Equal Transfer Time between all Workstations for the parts in the system set to 1
minute
• Constant Time between Arrivals of entities to the system entering the first
Workstation in the production line set to 1 minute.
• Two scenarios to measure the impact of the resources availability in the KPI
measurements are proposed: the first scenario with 100% availability of the resources,
and constant scheduled capacity of the resources during the production period. The
second scenario with a Time Based Failure of MTBF=1 minute and MTTR=0.1
minute. The failures of the resources follow the pre-empt rule.
5.3. Discrete Event System Simulation
This section describes the steps to simulate the generic real-time KPI model using Arena
software for Discrete Event Modelling. By simulating the system in DESM software, it
is possible to measure the impact of all operating factors in performance metrics for every
process and compare the simulated key performance measures with the metrics obtained
by employing the real-time application. Furthermore, DESM can deploy statistics of
alternative scenarios for optimisation of the manufacturing system. The simulation logic
of the Ideal Manufacturing System described in section 5.2 is presented in Figure 5-1.
Figure 5-1 Discrete Event Modelling Logic
First, the model considers the arrival pattern for entities received in the system as Constant
type stream defined by a continuous value of 1-minute time between arrivals. A create
module is used for recreating this arrival pattern of one order at a time as shown in Figure
66
5-2. The number of entities that enter the system at a given time with each arrival is
defined by the order sizes (measured by the number of components). It follows a constant
distribution of 1 entity per arrival. Additionally, the first creation of the orders of parts
for processing is set to start at the same time as the simulation of the system starts, so it
does not have a delay time.
Figure 5-2 Create Parts Module
Then, a Record Module is used in order to collect the statistics of all components
generated before they enter the system for conversion into final products. The Entity
Statistics record type is selected in this Record Module as shown in Figure 5-3.
Additionally, once the parts leave the final Assembly station, all the Entity Statistics is
recorded before the final dispose of the entities of the system which simulates the storage
of the finished goods in the Warehouse.
Figure 5-3 Entry of Parts Record Module
The parts depart from the process and are transferred to the downstream Workstation in
the manufacturing system. For this purpose, a Station module called Station 1 is
incorporated to define the location where component handling occurs (see Figure 5-4).
After a resource release an entity, a route module is included to transfer the parts to the
67
specified stations in the production line sequence. A constant delay time of 1 minute is
set as the transfer time to the next station as shown in Figure 5-4.
Figure 5-4 Route and Transfer time between Stations
Afterwards, the processing sequence defined via the routing in the production line is
included. All five Stations in the sequence are modelled using the same Station-Process-
Route logic (as shown in figure 5-5). In this way, a component arrives at each Station and
is processed by the machine. If the resource is busy, the component queues for the
machine to be idle. At each machine stations, the highest priority is given to the earliest
orders (First Input First Output queuing rule). Then, the component is sent to its next step
in the processing sequence when the machine finishes processing it.
Figure 5-5 Station-Process-Route (Production Line Logic Modules)
The Action of each Process Module is defined as Seize-Delay-Release, and a single
capacity machine resource is incorporated to each process. The Expression for the
Processing Time uses a constant distribution of 1 minute assigned equally for the rest of
resources in the system. The delay time is considered as the core processing time and sent
to the Value Added Time of the manufacturing system (see figure 5-6)
68
Figure 5-6 Station 1 Process Module
The operating schedule for the production line is defined in the Schedule Module. An
assumption of one shift of 17 hours in a day is made. Consequently, the number of hours
the real system operates under Replication Parameters (Hours Per Day) is 17 hours. This
implies that the average utilization for a fully utilised machine is 17/17= 100% as shown
in figure 5-7. A value capacity of 1 is set for each resource.
Figure 5-7 Schedule of the production plant and capacity of resources
The scheduling rule for all the resources in the system is set to Pre-empt meaning that the
resources will stop their processing operations when the defined shift time is done. All
69
resources in the system are associated with the defined schedule and capacity defined
previously as shown in Figure 5-8.
Figure 5-8 Resources Scheduling Rule
5.4. Verification and Validation of the system
In order to verify the results that have been obtained through simulation (shown in
Appendix C), it is obligatory that they undergo a process of validation and verification.
Verification involves testing if the model has been implemented correctly and program
debugging. Whereas, Validation refers to the assessment whether the right model has
been built or not by comparing the simulated model with the real-time model. First, the
necessary number of replications are calculated to obtain a good reliability and accuracy
of the simulated model and the performance indicators. Then, statistical analysis is
performed to compare Arena's output against the real-time data.
5.4.1. Increasing the Confidence Interval for Terminating Conditions
First, the DESM simulation was carried out under five replications, 1 day of replication
length, and 17 hours per day. The Output Summary values (Appendix C) of all five
replications exhibit a half width “calculated” for the variables of Entities, Process,
Resources and Queues. The average number of the Work-In-Process has been used as
analysis data to calculate the appropriate tolerance level. After five replications, the
sample WIP mean is 11.9791. The obtained half width is 0.02 under a 95% interval of
confidence (average value obtained ± half width). (see Figure 5-9)
Figure 5-9 Record Average Work-In-Process with 5 Replications
70
In order to increase the accuracy of the model, the tolerance level of the number of
replications (n) that reduces the half width interval to its half is calculated by iteration
using (Kelton, et al., 2015, p. 284) formula:
𝑛 ≅ 𝑛0 ℎ0
2
ℎ2 (5-1)
Where: 𝒏𝟎 is the number of the initial replications, 𝒉𝟎 is the obtained half width, and h
is the half of the obtained half width (Kelton, et al., 2010)
𝑛 ≅ 11.9791 0.022
0.012
𝑛 ≅ 48
It takes a total number of 48 replications to achieve a better exactitude of the simulation.
At the same time, using this replication number a better value of Half Width is obtained
for what concern to the chosen variable (WIP). This result means that there is a good
reliability of the results from the modelling approach using 48 replications as the output
values are independently distributed under a 95% interval of confidence.
Figure 5-10 Record Average Work-In-Process with 48 Replications
5.4.2. Validation of the collected data in Real-Time and Simulated Data using T-Test
One of the statistical tools that is used for the validation of the model is the T-test which
is used to compare the means of data with the variation in data of two independent
samples, i.e., to test whether or not there is a statistically significant difference between
the two samples assuming unequal variances (Devore, 2012). Thus, two-sample t-test is
used as there is a difference of sample size. The following procedure (Devore, 2012) is
used to verify the Arena's output against the collected data in real-time:
Step 1: Specify the hypotheses. The hypothesis to be tested is:
- Ho the real-time model measure of performance = the simulated system measure of
performance
71
- Hα the real-time measure of performance ≠ the simulated system measure of
performance
Ho: μ1 - μ2 = c vs. Ha: μ1 - μ2 ≠ c (two-sided test) (5-2)
Where:
c= hypothesized difference in the means
μ= medians of the samples (Parameters of interest)
Step 2: Calculate the t-Statistic value according to Ho:
𝑡 =μ1̅̅̅̅ −μ2̅̅̅̅ −𝑐
√𝑆𝑝2(1
𝑛1+
1
𝑛2)
(5-3)
Where:
𝑆𝑝2 =(𝑛1−1)𝑆12+(𝑛2−1)𝑆22
𝑛1+𝑛2−1 (5-4)
n=sample size
s2= sample standard deviation
Step 3: Choose a level of significance α=0.05 (confidence level 95%)
Step 4: Calculate degrees of freedom (dof)
𝑑𝑜𝑓 = n1 + n2 − 2 (5-5)
Step 5: Calculate the t-critical value using t-Distribution Table, i.e., t-values as function
of degrees of freedom and significance level (t α/2, dof)
If t-Stat < + (t-Critical) accept Ho. Otherwise, reject Ho
5.4.2.1. Validation of Work-In-Process
A sample of 10 replication results is gathered from the simulated model regarding the
output variable Total Work-in-Process. Similarly, the PLC Real-Time model was run
three times under 15, 30 and 45 minutes, and the total WIP of the application was
registered. The simulation data that is obtained using Arena is compared with the real-
time data using the Labview HMI to prove the validity of the PLC model. The following
table shows the t-Test: Two-Sample Assuming Unequal Variances
72
Number of Replications
Total WIP by Replication in
Simulation
Total WIP by Observation in Real-Time
1 11.954 12
2 11.977 12
3 11.984 12
4 11.988
5 11.99
6 11.992
7 11.993
8 11.994
9 11.994
10 11.995
WIP Simulated
System Real Time
System
Mean 11.9861 12
Variance 0.0001581 0
Observations 10 3
Hypothesized Mean Difference 0
Df 9
t Stat -3.495818411
P(T<=t) one-tail 0.003383946
t Critical one-tail 1.833112933
P(T<=t) two-tail 0.006767893
t Critical two-tail 2.262157163
Table 5-1 T-test system validation
As shown from the table 5-1: - (t-Critical) < t-Stat < + (t-Critical) -2.26 < -3.49 < 2.26
The null hypothesis is verified. Therefore, the outcomes of the simulated models reflect
the real time model. There is no significate difference between the means of both models.
5.4.3. Validation of Collected data in Real-time and Simulated Data using F-Test
Another statistical tool that is used for the validation of the model is F-test, which
compares the variances of data samples. Hence, two-sample f-test is utilised as there is a
difference of sample size. The following procedure (Devore, 2012) is used to verify the
Arena's output against the collected data by the real-time model:
Step 1: Specify the hypotheses. The hypothesis to be tested is:
- Ho the real-time model measure of performance = the simulated system measure of
performance
73
- Hα the measure of performance ≠ the simulated system measure of performance
Ho : μ1 - μ2 = c vs. Ha : μ1 - μ2 ≠ c (two-sided test)
Where: c= hypothesized difference in the means.
Step 2: Choose a level of significance α=0.05
Step 3: Calculate the test statistic (on the basis that samples variances are unknown and
may not be the same). The test statistic is given by the formula (Devore, 2012):
𝐹 =𝑠𝑥
2
𝑠𝑦2 (5-6)
Sx and Sy are sample variances given by Equation 5-7:
𝑠𝑥2 =
∑ (𝑋𝑖−�̅�)2𝑛𝑖=1
𝑛−1 (5-7)
Where: �̅� and �̅� are sample means
Step 4: Find the p-value which represents the probability area in the tails of the
distribution with the calculated degrees of freedom
Step 5: State the conclusion: Once the p-value is known, it is compared to the significance
level.
- If the p-value is ≤ α Ho is rejected. Otherwise, Ho is accepted
The above procedure is applied to test the dispersion of variances to compare the
performance indicators of the real-time model with the simulated model. A sample of 10
replication results gathered from the Arena simulated model is measured against a set of
real-time model results (under three independent runs). The following table shows the F-
Test Two-Sample for Variances
74
WIP Simulated System Real-Time
System
Mean 11.9861 12
Variance 0.0001581 0
Observations 10 3
Df 9 2
F 65535
P(F<=f) one-tail
F Critical one-tail 19.38482572 Table 5-2 F-Test Two-Sample for Variances
As shown from the table 5-2: F > F-Critical (65535> 19.38). The null hypothesis is
verified; therefore, the outcomes of the simulated models reflect the real-time application.
There is no significate difference between the variances of both model for measuring
system performance
5.4.4. Graphical Validation of the Performance Indicators
This section carries out a graphically validation of the real-time KPIs with the simulated
KPIs obtained from 48 replications runs. Figure 5-11 compares the simulated WIP from
high and low average values and the real-time measured values. For each resource, the
vertical bar shows the maximum average and minimum average, whereas the green points
represent the measured values obtained from the real-time simulation. The WIP value fell
into the simulated average ranges for all the simulated iterations.
Figure 5-11 WIP Validation Average Maximum and Minimum value
11.93
11.94
11.95
11.96
11.97
11.98
11.99
12
12.01
WIP
WIP VALIDATION
Maximum Avg Minimum Avg Real-Time Measure
75
Similarly, the maximum and minimum WIP values through the entire simulation period
give the range of variation of the WIP represented by the vertical bar in Figure 5-12. This
range of variation represents the period when the system starts producing until the steady
conditions are reached giving a constant Work-in-Process in the production system. The
real-time WIP value falls under the range of the simulated WIP maximum and minimum
values.
Figure 5-12 WIP Maximum and Minimum Validation
Figure 5-13 makes a similar comparison but considering the Production Rate (Number
Out obtained in a 1-hour period of production). The real-time measure falls within the
ranges determined by the maximum and minimum average of the simulated model.
Figure 5-13 Production Rate Validation Average Maximum and Minimum value
0
2
4
6
8
10
12
14
WIP
WIP VALIDATION
Maximum Minimum Real-Time Measure
0
50
100
150
200
250
300
350
Number Out
Production Rate Validation
Maximum Avg Minimum Avg Real-Time Measure
76
Chapter 6: Conclusions
6.1. Meeting the Research Objectives
This section shows how this dissertation effectively accomplished its objectives;
1. To provide an architecture solution for the real-time DAQ application by establishing
a Supervisory Control and Real-Time data acquisition structure including the
Programmable Logic Controller and communication software
• This objective was attained by configuring the PLC emulation software RsLogix
Emulator5000 to communicate with the PLC control software Rslogix5000 through a
RsLinks Server as described in Appendix A. RsLogix Emulator 5000 software was
employed to emulate the function of a PLC controller without the real hardware and
test this application with simulated digital I/O modules. The PLC ladder program was
developed in RSLogix5000 as described in section 4.2 to control and acquire data from
a simulated manufacturing plant outlined in section 4.1.
2. To develop Human Machine Interface (HMI) application software to show KPI
measures in real-time establishing a communication protocol between the system
components.
• This objective was achieved in chapter 4.3 by programming the LabView user-
interface where the UI controllers are routed to the OPC server in order to be
linked to the PLC program tags determined in section 4.2
3. To calculate sensible KPIs measures such as Work-in-Process, Energy Consumption,
GHG emissions, Production Time, Production Rate, Energy Efficiency, and Number
Waiting based on modelling process approach applied to a proposed manufacturing
system.
• This objective was also accomplished in chapter 4.3 by programming the
LabView UI where the KPIs models outlined in Chapter 3 were employed for
calculating and reporting the performance metrics. For this purpose, the data
acquired from the PLC control program in section 4.2 was routed to the UI for the
77
simulated manufacturing plant where the mentioned KPIs were calculated as
shown in Chapter 4.
4. To test and validate the real-time model against a Discrete Event Model Simulation
made in ArenaTM to compare the KPIs metrics obtained in the generic real-time
application with the system performance measurements attained by the simulated
scenario
• This objective was accomplished in chapter 5.4 by validating the KPIs obtained
using Arena with the KPIs obtained using the real-time model. The results indicate
that there is no significant statistical difference between the means and variances
of the sets of KPIs of both models and that the real-time measure falls within the
range of average maximum and minimum value defined by the simulated model
for an extended number of replications
6.2. Results and Findings
This project considered a comprehensive calculation of the GHG emissions factor to
relate the Electricity Consumption of manufacturing plants to GHG emissions on the
basis that all electricity consumed by the production system can be traced down to its
GHG emissions at power plant generation sources. By considering this approach, it
was necessary to incorporate the energy losses corresponding to Transmission and
Distribution lines in addition to the electricity generation emissions. In this way, the
Energy Consumption indicator complies with the overall energy balance which
includes energy generation, T&D and consumption by the final user. As described in
section 3.6 the GHG emissions factor for electricity consumption were calculated and
projected to the current year. For this purpose, the sum of the emission factors of each
component: KgCO2, Kg CH4; kg N2O was established to be equal to the Total
Greenhouse Gases (GHGs) expressed in equivalent KgCO2e. The range of data
analysed was over a 10-year period (2006-2015) for projecting average values which
were used to calculate the 2016 projected values. The total GHGs emission factors
were projected to a maximum of 0.494401 KgCO2e, and an average of 0.481770
KgCO2e. The projected 2016 value was 0.451332 kg CO2e, while the minimum value
calculated was 0.467451 KgCO2e. It is noted that the minimum value was bigger than
the projection. This outcome was obtained because the minimum value was the result
of the arithmetic average of the actual value of the projection 2016 and the minimum
78
obtained in the years 2006-2015. Furthermore, analysing each GHG type, it is
observed that the most significant emissions factor with respect to the total GHGs in
KgCO2e are given by the kgCO2 emissions while a small fraction relates to kgCH4
and kg N2O. In this way, by calculating all types of GHG emission, it is possible to
quantify all gaseous components which carry global warming potential caused by any
industrial activity.
This project demonstrated that time and energy intensive activities can be limited by
measuring improvement procedures against the WIP indicator to assess the state of
the system productivity. This reduction was observed in section 5.2 where the testing
stage of the real-time model attained three independent trials which were manually
activated from the user-interface. These tests considered ideal operational conditions
to verified the application, gathered real-time metrics of performance, and validated
the collected data against a simulated scenario in which the system was replicated by
48 iterations. For this purpose, a line balancing method was adopted to limit the
variability of the system and simplify the correlations between both models. In this
way, an effective balance between the processing operations in all Workstations was
accomplished. As shown in Appendix C, this balance was reflected by the constant
Number Waiting obtained in the queues of the system and by maintaining a constant
work-in-process (CONWIP) operation throughout the production time as the
difference between Number-In and Number-Out for 48 replications. CONWIP was
achieved by employing constant and equal processing times for all workstations in
the system, equal transfer time for the parts being moved between all resources,
constant time between arrivals of entities entering the first Workstation in the
production line, and constant pre-empted scheduled capacity of all machines in the
production system. These defined conditions allowed to monitor and minimise WIP.
Further analysis of this metric, show WIP as a Global KPI which reflects the overall
state of the entire system as it is based on several intrinsic relationships between
supporting measures elements of quantity and time measurement. In fact, it is
observed that WIP was a direct measure of the system productivity as excessive WIP
inventory leads to higher resource utilisation, lower Unit Consumption (energy per
production) as in imbalanced systems, fluctuations in WIP demand extended busy
times and productivity inefficiencies for bottlenecks stations. For example, the
proposed CONWIP simulation results originated zero Number Waiting in the
79
resources queues (shown in Appendix C) having no effective bottlenecks once the
uncertainty of the start of production was passed and the system entered in a steady
state.
This project incorporated novel KPIs such as Unit Consumption which was
considered as a comprehensive performance indicator to measure the energy
efficiency of the manufacturing plant based on single features KPIs of energy
(Electricity Consumed) and product throughput (Production Rate). Thus, it is noted
that Unit Consumption displayed the ratio of utilisation of resources for the
manufacturing operations as it measures the amount of energy needed as external
input to the system for the system to produce the desired throughput. Further analysis
showed that Unit Consumption was an alternative method to measure the productivity
and optimisation of the production line as any improvements in the manufacturing
system would be reflected by the Unit Consumption metric. For example, it is
observed that procedures that increase the productivity of the process operations such
as reduction of electricity consumption or increasing of production output delivered
lower values of the Unit Consumption metric. Therefore, it is concluded that the
impact of the improvement and optimisation activities can be assessed against this
energy efficiency metric to sustain actions that have the highest impact on the overall
productivity of the manufacturing plant.
The successful validation shown in section 5.4. of a comprehensive KPI such as WIP
provided a good reliability to the entire model because it showed that the measures of
time and quantity in which the computing of these metrics was based, and the single
features KPIs were correctly modelled to reflect the performance of the system. The
chosen variables to validate the real-time model against the discrete event simulation
were performance metrics based on measures of quantity and time. In this way, by
monitoring these elements at discrete intervals in the production line, several Basic
KPIs were obtained which reflect a single feature of the state of the system. In this
application, the obtained Basic KPIs include Parts in Process, Number Waiting,
Available Production Time, Production Rate, Available Resource Time, Busy
Resource Time, Idle Resource Time, Electricity Consumption. These metrics show
relevant information about the conditions of operation of resources, energy, queues in
the system, and throughput of products. From these single indicators, further
computing allowed to obtain Global KPIs which gathered these basic indicators of
performance and highlighted the state of the production system using a single
80
comprehensive KPI. For example, this application obtained the global KPIs Unit
Consumption, Work-in-Process, and GHG emissions as shown in Chapter 4.
This project considered the electricity requirements for all the processing operations
to change frequently within the production period. This assumption meant the
necessity of tracking the time each resource was in the different possible states to
correctly assess the Electricity Consumption of a manufacturing system. As shown in
section 4.3 this project tracked the all operative states of the system. Because of this
action, it was possible to identify the process and elements responsible for the biggest
energy demands and related GHG emissions. From this information, a distribution of
the Energy Consumption indicator among the various states of the resources was
obtained which built an accurate identification of the energy requirements of the
manufacturing plant. Furthermore, it was possible to attribute GHG emissions to the
Idle, Busy or Standby modes as shown in section 4.3. Therefore, it is concluded that
GHG emissions can be considered as global productivity indicators which are based
on energy metrics for reflecting the overall performance of a system
6.3. Future Work
The further expansion of this project may include the integration of other KPIs
especially Overall Equipment Effectiveness (OEE) to compare it with Unit
Consumption both of them Global KPIs which reflect the overall efficiency of the
system. OEE measures the operating efficiency of the production line based on the
computing of production losses relating to equipment downtime (Availability losses),
Idling and Minor Stops (Performance losses), Production Rejects (Quality losses).
The energy efficiency given by the Unit Consumption (energy per production) metric
can be compared against the OEE elements. In this way, common sources of
performance, availability and quality losses can be accounted as sources of energy
waste in a manufacturing plant. Additionally, these causes of failures can be
extrapolated to an average energy cost and value of wasted energy totally integrating
basic and global KPIs with financial indicators.
Similarly to the above point, Greenhouse gas (GHG) models can be compared with
OEE elements to determine common causes of performance, availability and quality
process-related GHG emissions. For example, breakdown and repair, or quality
81
control GHG emissions. In this way, GHG metric can be fully integrated as an overall
efficiency metric of a production system.
Apart from identifying common manufacturing problems, KPIs may be translated to
cost functions with the purpose of managerial allocation of resources on systems
constraints that have the greatest impact on productivity. Since each KPI have its own
units and ways of measurements, it is necessary to translate each indicator to a cost
function for operational and management strategies. This work is beneficial to any
industrial systems mainly manufacturing companies to optimise the performance of
the business and increase profitability leading to a competitive market advantage. A
cost function for each indicator could be analysed to find the optimised each operation
within the company. The analysis of the KPI costs functions will help to determine
the effectiveness of the production planning of a company and help to make strategic
decisions with the purpose of manage the variety of the products and process, increase
profitability, reduce waste and unnecessary manufacturing costs leading to a
competitive advantage to any organisation. From this analysis, proposals for
improvements and enhanced scenarios which will lead to new simulated models
related to the specified performance measures. Additionally, the best cost function for
each KPI could be found based on Activity Costs Models resulting in an optimised
manufacturing process to achieve a balance between operations and costs of
manufacturing.
82
REFERENCES
Ahmada, M. & Dhafrb, D., 2002. Establishing and improving manufacturing
performance measures. Robotics and Computer Integrated Manufacturing, Volume
18, p. 171–176.
Allen-Bradley, 2016. Logix5000 Controllers General Instructions Reference, s.l.:
Rockwell Automation Publication 1756-RM003Q-EN-P.
Almajali, D. A., Masa'deh, R. & Tarhini, A., 2016. Antecedents of ERP systems
implementation success: a study on Jordanian healthcare sector. Journal of
Enterprise Information Management, 29(4), pp. 549-565.
Atkins, T. & Escudier, M., 2013. A Dictionary of Mechanical Engineering. First ed.
Oxford: Oxford University Press.
Basu, P., 2010. Biomass Gasification and Pyrolysis: Practical Design and Theory. First
ed. Oxford: Elsevier Science.
Bolton, W., 2012. Mechatronics: a multidisciplinary approach. Sixth ed. Harlow:
Pearson Education.
Bonvik, A. M., Dallery, Y. & Gershwin, S. B., 2000. Approximate analysis of
production system operated by a CONWIP/finite buffer hybrid control policy.
International Journal of Production Research, p. 2845–2869.
Brown, B. & Fraser, J., 2012. Unleash manufacturing visibility, flexibility and speed
Realistic modelling enables a new generation of plant software for Critical
Manufacturing SA. Cambashi Limited.
Campbell, P., McMullan, J. & Williams, B., 2000. Concept for a competitive coal fired
integrated gasification combined cycle power plant. Fuel, 79(9), pp. 1031-1040.
Cogent Systems Inc, 2010. OPC DataHub, Cogent Real-Time Systems Inc. [Online]
Available at: http://www.opcdatahub.com/WhatIsOPC.html
[Accessed 28 07 2016].
Crawley, G. & Haight, R., 2013. The World Scientific Handbook of Energy. First ed.
South Carolina: World Scientific Publishing.
de Swaan Aarons, J., van der Kooi, H. & Shankaranarayanan, K., 2004. Efficiency and
Sustainability in the Energy and Chemical Industries. New York, NY, USA.: CRC
Press.
DEFRA & DECC, 2012. Guidelines to Defra / DECC’s GHG Conversion Factors for
Company Reporting: Methodology Paper for Emission Factors, s.l.: AEA for the
Department for Environment, Food and Rural Affairs and the Department of
Energy and Climate Change (DECC).
Department for Business Energy & Industrial Strategy, 2016. Collection - Government
emission conversion factors for greenhouse Gas Company reporting. [Online]
83
Available at: https://www.gov.uk/government/collections/government-conversion-
factors-for-company-reporting,
[Accessed 04 08 2016].
Department of Energy & Climate Change, 2015. Greenhouse gas reporting - Conversion
factors 2015, s.l.: Government emission conversion factors for greenhouse gas
company reporting and Energy and climate change: evidence and analysis.
Department of Energy, 2004. Final Report on the August 14, 2003 Blackout in the
United States and Canada: Causes and Recommendations. s.l.:U.S.-Canada Power
System Outage Task Force, United States Department of Energy.
Devore, J. L., 2012. Probability and Statistics for Engineering and the Sciences. Eight
ed. Boston: Brooks/Cole.
EERE, 1999. Estimating Appliance And Home Electronic Energy Use, Office of
Energy Efficiency and Renewable Energy. [Online]
Available at: http://energy.gov/energysaver/estimating-appliance-and-home-
electronic-energy-use
[Accessed 27 08 2016].
Forster, P. et al., 2007. Changes in Atmospheric Constituents and in Radiative Forcing.
In: Climate Change 2007: The Physical Science Basis. Contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel on
Climate Change. Cambridge: Cambridge University Press.
Georgiadis, P. & Politou, A., 2013. Dynamic Drum-Buffer-Rope approach for
production planning and control in capacitated flow-shop manufacturing systems.
Computers & Industrial Engineering, 65(4), p. 689–703.
GHG, I., 1999. The Reduction of Greenhouse Gas Emissions from the Cement Industry.
PH3/7, p. 112.
Gielen, D. & Moriguchi, Y., 2003. Technological potentials for CO2 emission reduction
in the global iron and steel industry. International Journal of Energy Technology
and Policy, 1(3), pp. 229-249.
Gieskes, J. F., Boer, H. & Baudet, F., 1999. CI and performance: a CUTE approach.
International Journal of Operations & Production Management, , 19 (11), p. 1120–
37.
Goldratt E., C. J., 2004. The goal. Monterrey: Granica.
Groover, M. P., 2016. Automation,Production Systems, and Computer-Integrated
Manufacturing. s.l.:Pearson.
Gutowski, T., Dahmus, J. & Thiriez, A., 2006. Electrical Energy Requirements for
Manufacturing Processes. Lueven, s.n.
Gutowski, T., Dahmus, J. & Thiriez, A., 2006. Electrical Energy Requirements for
Manufacturing Processes, Lueven: 13th CIRP International Conference of Life
Cycle Engineering .
84
Hernandez, Perez-Garcia & Vizan, 2008. An integrated modelling framework to support
manufacturing system diagnosis for continuous improvement. Robotics and
Computer-Integrated Manufacturing, 24(2), p. 187–199.
Hlupic V, P. R., 1999. Guidelines for selection of manufacturing simulation software.
IEE Trans , p. 31:21–9.
Hong, X. & Jianhua, W., 2006. Using standard components in automation industry: A
study on OPC Specification. Computer Standards & Interfaces, 28(4), p. 386–395.
Hüttmeira, A. & de Trevillea, S., 2009. Trading off between heijunka and just-in-
sequence. International Journal of Production Economics, p. 501–507.
IEA GHG, 2002. Building the Cost Curves for CO2 Storage. PH4/9, July, 48, p. 48.
IEA, 2011. World Energy Outlook, Paris: International Energy Agency, OECD/IEA.
IPCC, 2014. IPCC Fifth Assessment Report, s.l.: Intergovernmental Panel on Climate
Change (IPCC).
ISA-95, 2010. ANSI/ISA-95.00.01-2010 (IEC 62264-1 Mod) Enterprise-Control
System Integration, s.l.: International Society of Automation.
ISO 22400, 2009. ISO 22400-2 Manufacturing operations management — Key
performance indicators —Part 2: Definitions and descriptions of KPIs.
s.l.:International Organization for Standardization.
ISO 22400-2, W., 2009. ISO/WD 22400-2 Manufacturing operations management —
Key performance indicators —Part 2: Definitions and descriptions of KPIs. s.l.:s.n.
Jander, H., Borgvall, J. & Ramberg, R., 2012. Towards A Methodological Framework
For Hmi Readiness Evaluation. Stockholm , Swedish Defence Research Agency,
Stockholm University.
Jodlbauer, H. & Huber, 2008. Service level performance of MRP, Kanban, CONWIP
and DBR due to parameter stability and environmental robustness. International
Journal of Production Research,, 46(8), p. 2179–2195..
Kang, N., Zhao, C. & Li, J., 2015. Analysis of Key Operation Performance Data in
Manufacturing Systems. 2015 IEEE International Conference on Big Data.
Kelton, D., Sadowski, R. & Zupick, N., 2015. Simulation with Arena. Sixth ed. New
York: McGraw Hill.
Khojasteh-Ghamari, Y., 2012. Developing a framework for performance analysis of a
production process controlled by Kanban and CONWIP. Journal of Intelligent
Manufacturing, 23(1), p. 61–71.
Lee, H. & Seo, D., 2016. Performance evaluation of WIP-controlled line production
systems with constant processing times. Computers & Industrial Engineering, p.
138–146.
85
Leroux, M., 2010. Choosing KPI’s Relevant to your Business Objectives. Global
Marketing – Collaborative Production Management.
Little, J. D. C., 1961. A Proof for the Queuing Formula: L = λW. Operations Research ,
pp. 383 - 387.
Meyer, H., Fuchs, F. & Thiel, K., 2009 . Manufacturing Execution Systems: Optimal
Design, Planning, and Deployment. New York: McGraw Hill.
Monostori, L., Ka´da´, R., Bauernhansl & Kondoh, S., 2016. Cyber-physical systems in
manufacturing. CIRP Annals - Manufacturing Technology, Volume 65, p. 621–
641.
Mousavi, A., 2011. Introduction To Systems Modelling & Simulation. [Online]
Available at: http://www.brunel.ac.uk/~emstaam/
[Accessed 2016 08 25].
Muñoz, V., 2013. The ecuadorian energy matrix. Universidad Nacional de Loja,
Empresa Eléctrica Regional Sur S.A, Loja, Ecuador,, pp. 5, 6, 7.
National Instruments, C., 2013. LabView Users Manual. [Online]
Available at: http://www.ni.com/pdf/manuals/320999e.pdf
[Accessed 10 08 2016].
On-Control Inc, 2015. On Control Technologies. [Online]
Available at: http://www.oncontrol-tech.com/index.html
[Accessed 2016 08 26].
Oppermann, R., 2002. User-interface Design. International Handbooks on Information
Systems, pp. 233-248.
Özbayrak, M., 2004. Activity-based cost estimation in a push/pull advanced
manufacturing system. International Journal of Production Economics, p. 49–65.
Rockwell Automation, 2010. Getting Results with RSLogix Emulate5000, s.l.:
Rockwell Automation Software Reference Manual.
Ross, J. & Meier, A., 2001. Whole-House Measurements of Standby Power
Consumption. Energy Efficiency in Household Appliances and Lighting , pp. 278-
285 .
Saidur, R. et al., 2011. A review on biomass as a fuel for boilers. Renewable and
Sustainable Energy Reviews,, 15(5), pp. 2262-2289.
Sánchez, T., 2012. Generation with renewable energy, s.l.: Latinamerican Energy
Organization, OLADE.
Sato, N., 2004. Chemical Energy and Exergy – An Introduction to Chemical
Thermodynamics for Engineers. New York: s.n.
86
Scholten, B., 2009. MES guide for executives: why and how to select, implement, and
maintain a manufacturing execution system. s.l.:Research Triangle Park, NC:
International Society of Automation..
Short, T. A., 2014. Electric power distribution handbook.. Boca Raton: CRC press..
Stadtler, H., 2014. Supply Chain Management: An Overview. Supply Chain
Management and Advanced Planning, pp. 3-28.
Sun, Z. et al., 2012. Research and development of hydrogen fuelled engines in China,.
Elsevier, International Journal of Hydrogen energy, Volume 37, pp. 664-681.
Tavakoli, S., Mousavi, A. & Broomhead, 2013. Event Tracking for Real-Time Unaware
Sensitivity Analysis (EventTracker). IEEE Transactions On Knowledge And Data
Engineering, 25(2).
Thiriez, A., 2005. An Environmental Analysis of Injection Molding. Cambridge, MA,
USA: Massachusetts Institute of Technology, Project for M.S. Thesis, Department
of Mechanical Engineering .
Turconi, R., Boldrin, A. & Astrup, T., 2013. Life cycle assessment (LCA) of electricity
generation technologies: Overview, comparability and limitations. Renewable and
Sustainable Energy Reviews, Volume 28, p. 555–565.
Wali, R. P., 2012. An Electronic Nose to Differentiate Aromatic Flowers using a Real-
Time Information-Rich Piezoelectric Resonance Measurement. Procedia
Chemistry, Volume 6, pp. 194-202.
Wheeler, F., 1998. Solving the heavy fuel oil problem with IGCC technology.. Heat
Engineering, 62(2), pp. 24-28.
Williams, T. & Li, H., 1990. PERA and GERAM—enterprise reference architectures in
enterprise integration. Information Infrastructure Systems for Manufacturing II,
Volume 16, pp. 3-30.
Woeppel, M., 2003. Manufacturer guide to implement theory of restrictions. Cuenca:
Monsalve Moreno.
Zairi M, 1994. Benchmarking the best tool for measuring competitiveness.
Benchmarking Quality Management Technology, p. 11–24.
87
Appendix A: RsLogix 5000, RsLogix Emulate 5000, RsLinks
Configuration
The implementation of this project starts with the programming of Allen Bradley systems
which are carried out configuring the simulation software RsEmulate 5000 and the control
software for Allen Bradley PLCs such as Rslogix5000 and RsLinks. RsLogix
Emulator5000 software was employed to simulate the function of a PLC without the real
hardware and hence do forward-thinking debugging. This project focuses only on the
PLC code and UI development.
First, when opened the RsLogix Emulate5000 shows the Chassis Monitor which is a
software application that permits to configure simulated I/O modules. In the Chassis
monitor, a simulated processor is loaded in an empty chassis slot. The type of virtual
processor selected is: EmulLogix 5868 Controller. The configuration parameters of this
virtual controller are: Version 20, Start-up Mode: Remote Program, Memory Size: 3072
KB, Periodic Save Interval: 10 minutes (See Figure 6-1)
Figure 6-1 Virtual Controller (EmuLogix 5868 ) parameters configuration
88
In the slot 3 of the RsEmulate Chassis Monitor, the digital module I/O is added using the
following parameters: Virtual 1789-SIM 32, Point I/O Simulator. The function of this
digital I/O module is to simulate digital inputs and outputs to the virtual PLC (EmuLogix
5868 controller). The modules are the showed in the Cassis Monitor (see Figure 6-2)
Figure 6-2 Modules in the RsLogix Emulate 5000 Chassis Monitor
The configuration of RsLinks is employed to manage the communication between the
controllers and HMIs. In RsLinks a communication driver is created as the way that
RsLinks will be linked to RsLogix Emulate5000. In the Configure Driver option, a new
driver is chosen with the following parameters: Virtual Backplane (SoftLogix58xx, USB)
name: “Intento_5” (see Figure 6-3) (Allen-Bradley, 2016).
Figure 6-3 Virtual Backplane communication driver
89
In the RSwho option, it is possible to see the all elements of the Virtual Chassis including
the virtual processor (RsLogix5000 Emulator) and the module of digital I/O (1789-Sim
32 I/O Simulator) as shown in Figure 6-4. These elements are now connected through
the RsLinks server
Figure 6-4 RsLink Server Connected Elements
Once the configuration of RsLogix Emulate5000 and RsLinks is ready, a new Project is
created in the PLC programming software Rslogix5000 in which, it is possible to upload
and download the Virtual Chassis through the RsLinks server. When creating the project
type, the virtual Controller RSlogix Emulate 5000 is selected, the version of the controller
in Rslogix5000 is the same as the version entered in the emulator (Version 20). The
Chassis Type is “1756 10-Slot ControlLogix Chassis” as shown in Figure 6-5 (Allen-
Bradley, 2016).
Figure 6-5 RsLogix5000 Controller Configuration
90
.
Once the controller is selected, RsLogix 5000 automatically creates the project
characteristics. In the Controller Organizer window, the Virtual Backplane is selected,
and a new module is added from which the “1756-Generic Module digital I/O” is selected.
This digital I/O module replicates the digital I/O module previously defined in the
emulator. The module properties are configured as: Name: “Digital I/O”, Assembly
Instance: Input=1, Output=2, Configuration=16. Size: Input=2, Output=1,
Configuration=0 (see Figure 6-6) (Allen-Bradley, 2016)
Figure 6-6 Connection Parameters 1756-Generic Module
In the Connection tab of the Module Properties, the RPI (requested packet interval) must
be placed in 50 otherwise it will not work the module digital I/O. The manufacturer gives
those values which must be followed to work with the simulated PLC (Rockwell
Automation, 2010) (see Figure 6-7).
91
Figure 6-7 Connection Properties 1756 Generic Module
In the RsLogix5000 Communications Menu, the “Who Active” option is selected. Here,
the RsLogix5000 program is connected online with the RsLinks server. For this purpose,
from the Who Active menu, the Rslogix5000 Emulator is selected and then first the Set
Project Path button is pressed followed by the Download button (see Figure 6-8).
Figure 6-8 RsLogix 5000 Who Active Window
Once the virtual PLC is running online, the Controller Tags option shows all the tags
created by the simulated I/O module. To test that the connection between RSLogix5000
and RsLogix Emulate 5000, the inputs of the RsLogix5000 ladder code can be activated
92
from the digital I/O module in the emulator Chassis Monitor (Rockwell Automation,
2010). Here, the digital I/O is selected and from the Module Properties option it is possible
to toggle on or off the inputs of the digital I/O Module as shown in Figure 6-9.
Figure 6-9 1789 digital I/O module Data Properties
93
Appendix B: PLC Ladder Code Command Lines
Commands
Lines Code Output
0 Available_time_seconds
1 Available_time_minutes
2 Machine 1
3 Mach1_On_Seconds
4 Mach1_On_Minutes
5 Input_Machine_1
6 Busy_Mach_1
7 Idle_Mach_1
8 Mach1_busy_time_seconds
9 Mach1_busy_time_minutes
10 Output_Machine_1
11 Parts_in_Process_Mach1
12 Aux_Queue_1
13 Machine 2
14 Mach2_On_Seconds
15 Mach2_On_Minutes
16 Input_Machine_2
17 Num_Waiting_Q_1
18 Busy_Mach_2
19 Idle_Mach_2
20 Mach2_busy_time_seconds
21 Mach2_busy_time_minutes
22 Output_Machine_2
23 Parts_in_Process_Mach2
24 Aux_Queue_2
25 Machine 3
26 Mach3_On_Seconds
27 Mach3_On_Minutes
28 Input_Machine_3
29 Num_Waiting_Q_2
30 Busy_Mach_3
31 Idle_Mach_3
32 Mach3_busy_time_seconds
33 Mach3_busy_time_minutes
34 Output_Machine_3
35 Parts_in_Process_Mach3
36 Aux_Queue_3
37 Machine 4
38 Mach4_On_Seconds
39 Mach4_On_Minutes
40 Input_Machine_4
41 Num_Waiting_Q_3
42 Busy_Mach_4
94
43 Idle_Mach_4
44 Mach4_busy_time_seconds
45 Mach4_busy_time_minutes
46 Output_Mach_4
47 Parts_in_Process_Mach4
48 Aux_Queue_4
49 Machine 5
50 Mach5_On_Seconds
51 Mach5_On_Minutes
52 Input_Machine_5
53 Num_Waiting_Q_4
54 Busy_Mach_5
55 Idle_Mach_5
56 Mach5_busy_time_seconds
57 Mach5_busy_time_minutes
58 Output_Mach_5
59 Parts_in_Process_Mach5
60 Aux_Queue_5
61 Assembly
62 Assembly_On_Seconds
63 Assembly_On_Minutes
64 Input_Assembly
65 Num_Waiting_Q_5
66 Busy_Assembly
67 Idle_Assembly
68 Asse_busy_time_seconds
69 Asse_busy_time_minutes
70 Output_Assembly
71 Parts_in_Process_Asse
72 Aux_Queue_6
73 Input_Warehouse
74 Num_Waiting_Q_6
75 Output_Warehouse
76 Inventory_Level
95
96
97
98
99
100
101
102
103
104
Appendix C: ARENA Simulation Results
Gabriel Barriga - License: STUDENT
Summary for Replication 48 of 48
Project: Real Time KPIs Run execution date : 9/ 7/2016
Analyst: Washington Barriga Model revision date: 9/ 7/2016
Replication ended at time : 48960.0 Minutes
Base Time Units: Minutes
TALLY VARIABLES
Identifier Average Half Width Minimum Maximum Observations
_____________________________________________________________________________________
_____________
Parts.VATime 6.0000 .00000 6.0000 6.0000 48949
Parts.NVATime .00000 .00000 .00000 .00000 48949
Parts.WaitTime .00000 .00000 .00000 .00000 48949
Parts.TranTime 6.0000 .00000 6.0000 6.0000 48949
Parts.OtherTime .00000 .00000 .00000 .00000 48949
Parts.TotalTime 12.000 .00000 12.000 12.000 48949
Process 5.Queue.WaitingTime .00000 .00000 .00000 .00000 48953
Process 6.Queue.WaitingTime .00000 .00000 .00000 .00000 48951
Process 1.Queue.WaitingTime .00000 .00000 .00000 .00000 48961
Process 2.Queue.WaitingTime .00000 .00000 .00000 .00000 48959
Process 3.Queue.WaitingTime .00000 .00000 .00000 .00000 48957
Process 4.Queue.WaitingTime .00000 .00000 .00000 .00000 48955
DISCRETE-CHANGE VARIABLES
Identifier Average Half Width Minimum Maximum Final Value
_____________________________________________________________________________________
______________
Parts.WIP 11.998 (Insuf) .00000 13.000 12.000
Resource 1.NumberBusy 1.0000 (Insuf) .00000 1.0000 1.0000
Resource 1.NumberScheduled 1.0000 (Insuf) 1.0000 1.0000 1.0000
Resource 1.Utilization 1.0000 (Insuf) .00000 1.0000 1.0000
Resource 2.NumberBusy .99996 (Insuf) .00000 1.0000 1.0000
Resource 2.NumberScheduled 1.0000 (Insuf) 1.0000 1.0000 1.0000
Resource 2.Utilization .99996 (Insuf) .00000 1.0000 1.0000
Resource 3.NumberBusy .99992 (Insuf) .00000 1.0000 1.0000
Resource 3.NumberScheduled 1.0000 (Insuf) 1.0000 1.0000 1.0000
Resource 3.Utilization .99992 (Insuf) .00000 1.0000 1.0000
Resource 4.NumberBusy .99988 (Insuf) .00000 1.0000 1.0000
Resource 4.NumberScheduled 1.0000 (Insuf) 1.0000 1.0000 1.0000
Resource 4.Utilization .99988 (Insuf) .00000 1.0000 1.0000
Resource 5.NumberBusy .99984 (Insuf) .00000 1.0000 1.0000
Resource 5.NumberScheduled 1.0000 (Insuf) 1.0000 1.0000 1.0000
Resource 5.Utilization .99984 (Insuf) .00000 1.0000 1.0000
Resource 6.NumberBusy .99980 (Insuf) .00000 1.0000 1.0000
Resource 6.NumberScheduled 1.0000 (Insuf) 1.0000 1.0000 1.0000
Resource 6.Utilization .99980 (Insuf) .00000 1.0000 1.0000
Process 5.Queue.NumberInQueue .00000 (Insuf) .00000 .00000 .00000
Process 6.Queue.NumberInQueue .00000 (Insuf) .00000 .00000 .00000
Process 1.Queue.NumberInQueue .00000 (Insuf) .00000 1.0000 .00000
Process 2.Queue.NumberInQueue .00000 (Insuf) .00000 .00000 .00000
Process 3.Queue.NumberInQueue .00000 (Insuf) .00000 .00000 .00000
Process 4.Queue.NumberInQueue .00000 (Insuf) .00000 .00000 .00000
COUNTERS
105
Identifier Count Limit
_____________________________________________________________
Record 1 48961 Infinite
Record 2 48949 Infinite
OUTPUTS
Identifier Value
_____________________________________________________________
Parts.NumberIn 48961.
Parts.NumberOut 48949.
Resource 1.NumberSeized 48961.
Resource 1.ScheduledUtilization 1.0000
Resource 2.NumberSeized 48959.
Resource 2.ScheduledUtilization .99996
Resource 3.NumberSeized 48957.
Resource 3.ScheduledUtilization .99992
Resource 4.NumberSeized 48955.
Resource 4.ScheduledUtilization .99988
Resource 5.NumberSeized 48953.
Resource 5.ScheduledUtilization .99984
Resource 6.NumberSeized 48951.
Resource 6.ScheduledUtilization .99980
System.NumberOut 48949.
ARENA Simulation Results
Gabriel Barriga - License: STUDENT
Output Summary for 48 Replications
Project: Real Time KPIs Run execution date : 9/ 7/2016
Analyst: Washington Barriga Model revision date: 9/ 7/2016
OUTPUTS
Identifier Average Half-width Minimum Maximum # Replications
_____________________________________________________________________________________
______________
Parts.NumberIn 24991. 4163.5 1021.0 48961. 48
Parts.NumberOut 24979. 4163.5 1009.0 48949. 48
Resource 1.NumberSeized 24991. 4163.5 1021.0 48961. 48
Resource 1.ScheduledUtilization 1.0000 .00000 1.0000 1.0000 48
Resource 2.NumberSeized 24989. 4163.5 1019.0 48959. 48
Resource 2.ScheduledUtilization .99982 9.1733E-05 .99804 .99996 48
Resource 3.NumberSeized 24987. 4163.5 1017.0 48957. 48
Resource 3.ScheduledUtilization .99964 1.8347E-04 .99608 .99992 48
Resource 4.NumberSeized 24985. 4163.5 1015.0 48955. 48
Resource 4.ScheduledUtilization .99945 2.7520E-04 .99412 .99988 48
Resource 5.NumberSeized 24983. 4163.5 1013.0 48953. 48
Resource 5.ScheduledUtilization .99927 3.6693E-04 .99216 .99984 48
Resource 6.NumberSeized 24981. 4163.5 1011.0 48951. 48
Resource 6.ScheduledUtilization .99909 4.5867E-04 .99020 .99980 48
System.NumberOut 24979. 4163.5 1009.0 48949. 48
Simulation run time: 0.27 minutes.
Simulation run complete.
Recommended