23
André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

Embed Size (px)

Citation preview

Page 1: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

André Augustinus10 September 2001

DCS Architecture Issues

Food for thoughts and discussion

Page 2: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 2André Augustinus

Disclaimer

What will be presented is not a proposal to be accepted or rejected.

It is merely a collection of (personal) ideas. It is meant to make you start thinking and

trigger discussions. With the outcome of these discussions we could start to design the full Detector Control System.

We are eager to hear your comments!

Page 3: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

André Augustinus10 September 2001

André’s ideas on DCS

A rather personal view

and probably somewhat DELPHI biased

Page 4: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 4André Augustinus

Introduction

From the ALICE Controls Coordination mandate:

The DCS should allow centralised operation• By a single operator

• Not necessarily a DCS expert

• From a central control room

… to have the Detector Controls System (DCS) ready for exploitation by the end of 2005, allowing to control and operate the experiment (from a central operator workplace) during all modes of operation …

Page 5: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 5André Augustinus

Introduction

A central (non-expert) operator should• at all times be informed about anomalies in the

experiment• be able to give the necessary commands to run

the experiment efficiently (minimal downtime)

A detector (expert) operator should• have a detailed view of the detector/system• be able to give any expert commands

Page 6: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 6André Augustinus

The DCS, global view

Hierarchical system• Central DCS• Detector CS’s and/or sub-detector CS’s

Connected to external systems• Gas, Electricity etc.• Magnet, DAQ, LHC Accelerator etc.

Page 7: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 7André Augustinus

The DCS, global view

Gas, Electricity, …Magnet, DAQ, LHC

ITS, TPC, TRD,TOF, Muon, …

ITS-SPD, ITS-SDD, ITS-SSDMuon-Track, Muon-Trig

Page 8: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 8André Augustinus

The DCS, global view

Central DCS collects the status of all detectors and external systems

Through the Central DCS the operator will give generic commands to all or a set of detectors

The Central DCS can perform pre-programmed operations (automated operation)

Page 9: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 9André Augustinus

The DCS, global view

A Detector CS consists of a collection of sub-system controls

The sub-system controls is controlling a collection of equipment or devices that is logically grouped together• HV, LV, Cooling, …

The sub-system controls is the interface to the hardware of the detector

Page 10: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 10André Augustinus

The DCS, global view

CentralDCS

TPC DetectorDCS

HV LV Cooling

HVcrate 1

HVCrate 2

LVcrate

Coolingsystem

T monitor

Tsensors

Page 11: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 11André Augustinus

Central operation, an example

Operator issues command “get ready for physics”• Get ALICE ready for datataking

• Ramp up HV, switch on or off equipment, change operational parameters.

Command is issued through an operator interface (PVSS) to the Central DCS

Central DCS dispatches the command to Detector CS’s

Page 12: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 12André Augustinus

Central operation, an example

Upon receipt the Detector CS will issue commands to sub-system controls

• “Ramp up” to HV, “Switch on” to LV, “load physics” to temperature monitoring

The sub-system controls will perform the necessary action on the hardware

• Send a command to CAEN to initiate a ramp

At this moment the command execution has finished

Page 13: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 13André Augustinus

Central operation, an example

The operator should get feedback on the execution of the command• Should know when command is finished

• (Has reached the hardware)

• Should get feedback from the hardware• (The hardware is doing/has done something)

Page 14: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 14André Augustinus

Central operation, an example

CAEN received command and starts ramping It reports this back to Detector CS: “I’m

ramping up” and the Detector CS will detect a change from “off” to “ramping up”

Based on this new information it will recalculate its own state and change from “not ready” to “HV ramping up”

This state is seen by the Central DCS and shown to the operator

Page 15: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 15André Augustinus

Background

Detector CS: operations logic• Programmed commands• Generic devices• Programmed logic, maybe templates/framework

Sub-system controls: hardware details• One sub-system controls per type of device• “device driver”

Page 16: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 16André Augustinus

External Systems

Information from external systems is essential• For display to operator• For archiving• To trigger actions

In principle, all information from any external system is available, via software

Crucial events should be backed up by hardwired interlocks

Page 17: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 17André Augustinus

Dividing the system

Each detector should, if needed, be able to run independent from central operation• Commissioning, debugging, calibration, …

Partitioning Propagation of commands and events should

be handled properly in case of partitioning Who decides who has control

Page 18: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 18André Augustinus

The DCS, global view

CentralDCS

ITSDCS

TRDDCS

TOFDCS

HMPIDDCS

• Partitioning

Page 19: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 19André Augustinus

Archiving and Logging

Archiving• Store the value of a monitored parameter for later

retrieval (time stamped)• Performance• Post Mortem analysis• For use with offline data analysis

Logging• Store all events that occurred in the system

• Commands given (what, by whom)• Anomalies, errors, alarms

Page 20: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 20André Augustinus

Configuration

Aim for a uniform system, that only needs to be configured to the needs of the user• Provide an ALICE DCS framework• Configurable, nothing “hard coded”• Configuration data should be stored in some

database

Page 21: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 21André Augustinus

More ideas

Central DCS and Detector CS could be very conveniently implemented as Finite State Machines (will be (is) part of PVSS)

User interfaces (user defined) can be connected at any point in the system

Access control and access rights can be applied at any point in the system

Page 22: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 22André Augustinus

More ideas

One could envisage automated operation:• Routine operation

• LHC declares stable beam conditions• Prepare DCS and DAQ for physics• Start datataking when DCS is ready• Pause datataking when DCS detects serious problem• Ramp down at beamdump

• Apply standard corrective actions• Ramp up a tripped channel

Page 23: André Augustinus 10 September 2001 DCS Architecture Issues Food for thoughts and discussion

10 September 2001DCS Workshop 23André Augustinus

Summary

Meant to trigger reflection and discussion We will produce a discussion document

elaborating on the issues presented here Come to a definition if the Alice DCS

architecture