Analysis and optimization of a dissolved air

flotation process for separation of suspended

solids in wastewater

Oskar Bäck

Natural Resources Engineering, master's

2021

Luleå University of Technology

Department of Civil, Environmental and Natural Resources Engineering

Analysis and optimization of a dissolved air

flotation process for separation of

suspended solids in wastewater

Oskar Bäck

i

Preface This report represents a master thesis within the master program in Natural resource engineering with

focus on water and environmental science at Luleå University of Technology. The study was conducted

in collaboration with Roslagsvatten AB, a Swedish water utility, about Margretelund wastewater

treatment plant’s dissolved air flotation process. The study was held during a period of 20 weeks in the

spring of 2021, corresponding to 30 ECTS.

There are a lot of people I am thankful for helping me through this master thesis, and especially my

supervisor at Luleå university of technology (LTU), Inga Herrmann, for helping me sort through all my

ideas and thoughts and encourage me during these 20 weeks.

I wish to both congratulate and thank my fellow classmates from LTU, and all the discussions we have

had together, through both high and lows. I would also like to direct a special thank you to my supervisor

from Roslagsvatten, Daniel Zetterström, for helping me with everything and anything on-site during the

thesis and always came with a good answer no matter the question.

Thank you, Annelie Hedström, for helping me bring out the most from this thesis as examiner.

I am thankful for the opportunity given to work on this master thesis together with the process division

at Roslagsvatten, and for all the help and support given by everyone at Margretelund wastewater

treatment plant.

ii

Abstract Margretelund wastewater treatment plant (WWTP) operated by the water utility Roslagsvatten AB, was

built in 1956 and is located in Åkersberga town, Stockholm County, Sweden. Margretelund WWTP was

last renovated in 1999, and has been operated with the same physical, chemical, and biological processes

since then. Due to issues with increased phosphorus emissions connected to increased concentration of

effluent total suspended solids (TSS), Roslagsvatten would like to optimize the operation of their dissolved

air flotation (DAF) process and the author was tasked to conduct a study about the subject. The specific

aim of the study was to propose one method for optimization with available means to reduce effluent

TSS concentration during high flow rates for the present DAF process at Margretelund WWTP.

Achieving the aim required an historical analysis of Margretelund WWTP’s DAF process and an

investigation of the effect influent flow rate and effluent recycle rate (ERR) had on effluent TSS

concentration. The increase of effluent TSS was believed to be caused by increased flow rates from

infiltration and inflow (recorded to 32% of total volume the year 2020) affecting the dissolved air flotation

(DAF) process.

The literature study design parameters for a dissolved air flotation process, specifically the recycle flow

pressurization configuration, generated information about which parameters to take into consideration

when optimizing a DAF unit. Analysis of historic effluent measurements at Margretelund showed that

42% of all samples analysed between January 2015 – January 2021 were below 10 mg/l TSS. Each

historical increase of surface load has brought a decreased effluent recycle rate (ERR) and consequently

an increasing percentage of samples exceeding 10 mg/l. A Pearson correlation presented a negative

correlation with both ERR and surface load in relation to effluent TSS concentration. This resulted in

the selection of the experimental factors ERR and surface load to be investigated in this study.

Margretelunds WWTP’s DAF design of ERR being 10-15% and the design surface load of 4 m/h was

the base values for the experimental runs. Increases of ERR percentage was done during the experiment

for four different surface loads (2.5, 4, 5 and 6 m/h), with five steps between 15% up to 35% ERR in

one of the three parallel DAF units in Margretelund WWTP. TSS in the effluent was constantly

monitored using a TSS sensor. Influent TSS was measured at Roslagsvatten’s accredited laboratory in a

24h composite sample with 1 hour for each sub-sample.

The results showed that both the highest and the lowest ERR settings tested provided the lowest average

effluent TSS concentrations. However, a decreased surface load was found to lower effluent TSS

concentration and ERR providing only minor differences within each surface load. Largest surface load

possible was found to be 5 m/h, for an ERR of 15 or 35%. Surface load less than 5 m/h provided a

concentration under 10 mg/l for all ERR setting.

iii

Sammanfattning Margretelund avloppsreningsverk (ARV) placerat i Åkersberga, Stockholms län, byggdes 1956 och drivs

av Roslagsvatten AB. Margretelund ARV har sedan 1956 renoverats vid två tillfällen senast 1999. Samma

reningsprocess för fysisk, kemisk och biologisk rening har använts sedan senaste renoveringen.

Roslagsvatten har haft problem med oönskat tillskottsvatten (motsvarade 32% av total volym 2020) som

har påverkat deras flotationsprocess negativt gällande rening av suspenderat material. Detta har till slut lett

till förhöjda utsläppsvärden av fosfor som finns bundet i det suspenderade materialet. Denna studie har

utförts av författaren på efterfrågan av Roslagsvatten, med syfte att presentera optimeringsåtgärder till

styrning av flotationsprocessen vid höga flöden. För att uppnå målet med studien gjordes en historisk

analys av Margretelunds flotationsprocess samt undersökningar om hur variationer i inkommande flöde

samt recirkuleringsgrad har påverkat koncentration av utgående suspenderat material.

Teori undersöktes och information insamlades angående designparametrar gällande optimering av

flotationsprocesser, mer specifikt en flotationsprocess med recirkulerat trycksatt flöde för avskiljning av

susp. Analys av historiska utsläppsvärden från Margretelund ARV’s flotationsprocess visade på att 42% av

proverna analyserade mellan januari 2015-januari 2021 låg under 10 mg/l för utgående suspenderat

material. Varje historisk ökning av ytbelastning påvisade en minskande recirkuleringsgrad samt en ökande

andel prover som översteg koncentrationen 10 mg/l. Utifrån en Pearson korrelation visades en negativ

korrelationen för både ytbelastning och recirkulationsgrad gentemot koncentration av utgående

suspenderat material. Både recirkuleringsgrad och ytbelastning valdes därför till denna studies

experimentella faktorer. Flotationsprocessen på Margretelund ARV’s var designad för en

recirkuleringsgrad på 10–15% vid ytbelastning på 4 m/h, och valdes som basvärde för experimentet. Fem

olika grader av recirkulation testades för fyra olika ytbelastningar (2.5, 4, 5 och 6 m/h) i intervallet 15–

35% i en av tre parallella flotations bassänger på Margretelund ARV. Koncentration utgående suspenderat

material mättes kontinuerligt med en sensor. Inkommande koncentration bestämdes genom ett dygnsprov

som analyserades av Roslagsvattens ackrediterade laboratorium.

Ett resultat från experimenten var att både den högsta och lägsta inställningen av recirkuleringsgrad visade

de lägsta medelvärdena för utgående koncentration av suspenderat material. Dock, visade resultaten att

en minskande ytbelastning resulterade i lägre koncentrationer av utgående suspenderat material. Vidare

sågs att recirkuleringsgraden enbart hade en låg påverkan på koncentrationerna för varje ytbelastning. Den

högsta möjliga ytbelastningen utan att överstiga 10 mg/l visades vara 5 m/h med recirkuleringsgraderna

15% och 35%.

iv

Table of contents Preface ................................................................................................................................................ i

Abstract .............................................................................................................................................. ii

Sammanfattning ................................................................................................................................ iii

Table of figures .................................................................................................................................. vi

Table of tables ................................................................................................................................... vii

Legend .............................................................................................................................................. vii

1. Introduction ................................................................................................................................ 2

1.1 Purpose of study ........................................................................................................................ 2

1.2 Scope of study ........................................................................................................................... 2

2. Theory ........................................................................................................................................ 4

2.1 Dissolved Air Flotation .............................................................................................................. 4

2.2 History of DAF ......................................................................................................................... 5

2.3 Design parameters ...................................................................................................................... 5

2.3.1 Surface load and particle rise velocity................................................................................... 5

2.3.2 Bubbles and airs solubility ................................................................................................... 5

2.3.3 Gas-to-solid ratio ................................................................................................................ 6

2.3.4 Collision efficiency and flocs type ........................................................................................ 7

2.4 Benefits and disadvantages of DAF compared to sedimentation ................................................... 7

2.5 Phosphorus in bacteria ............................................................................................................... 8

3. Methods .................................................................................................................................... 11

3.1 Margretelund wastewater treatment plant ................................................................................. 11

3.1.1 Catchment area ................................................................................................................. 12

3.1.2 Processes in Margretelund wastewater treatment plant ....................................................... 13

3.1.3 DAF reactor ...................................................................................................................... 13

3.2 Gathering relevant theory and information ............................................................................... 15

3.3 Data collection ........................................................................................................................ 15

3.3.1 Sampling ........................................................................................................................... 15

3.3.2 Analysis of total suspended solids in the influent and effluent of the DAF ........................... 15

3.3.3 Points for influent TSS sampling and effluent TSS monitoring. .......................................... 18

3.3.4 Software ........................................................................................................................... 18

3.4 Experimental design ................................................................................................................ 19

3.5 Analysis of historical DAF data ................................................................................................. 20

3.5.1 Regression analysis ............................................................................................................ 21

3.5.2 Gas – to – solid ratio ......................................................................................................... 21

3.5.3 Amount of phosphorus in wastewater ................................................................................ 21

3.5.4 Processing of historical data ............................................................................................... 22

4. Results .......................................................................................................................................... 24

v

4.1 Experiment results ................................................................................................................... 24

4.1.1 Influent suspended solids concentration ............................................................................. 24

4.1.2 Removal efficiency of total suspended solids ...................................................................... 24

4.1.3 Effluent suspended solids concentration. ............................................................................ 26

4.1.4 Effluent suspended solids concentration over time ............................................................. 27

4.2.5 Temperature and pressure during experimental run ........................................................... 28

4.2 Results from historical data analysis .......................................................................................... 29

4.2.1 Regression analysis ............................................................................................................ 30

4.2.2 Gas-to-solid ratio .............................................................................................................. 31

4.2.3 Amount of phosphorus in suspended solids ........................................................................ 31

5. Discussion ..................................................................................................................................... 33

5.1 Effects of ERR and loading rate on the flotation process .......................................................... 33

5.1.1 Effect of ERR and loading rate on effluent suspended solids and phosphorus concentrations.

................................................................................................................................................. 33

5.1.2 Effect of ERR and surface load on removal efficiency of TSS ............................................ 34

5.1.3 Temperature and pressure during experiment .................................................................... 35

5.2 Analysis of historical data ......................................................................................................... 35

5.2.1 Regression analysis ............................................................................................................ 35

5.2.2 Gas – to – solid ratio ......................................................................................................... 36

5.3 Experimental challenges and sources of errors ........................................................................... 36

5.4 Future investigations ................................................................................................................ 37

6. Conclusions .................................................................................................................................. 40

7. References .................................................................................................................................... 43

8. Appendices ...................................................................................................................................... I

Appendix 1. Enlarged flow scheme of Margretelund wastewater treatment plant .............................. I

Appendix 2. Current situational graphs ........................................................................................... II

Appendix 3. Dates, times and influent TSS value for experiment .................................................. VII

Appendix 4. Critical values for two-tailed test, Pearson correlation.............................................. VIII

Appendix 5. Phosphorus in Margetelund WWTP ......................................................................... IX

Appendix 6. Experiment journal. .................................................................................................... X

vi

Table of figures Figure 1. Effect of air-solids ratio on float concentration and subnatant suspended solids (Source: Wang,

Hung & Shammas, 2005, Physicochemical Treatment Processes, P.444) ............................................... 7 Figure 2. Margretelund WWTP catchment area as described by Roslagsvatten (2015) ........................ 12 Figure 3. Location of Åkersberga town, Sweden (Map from Eniro (2014)) ......................................... 12 Figure 4. Flow scheme Margretelund WWTP, Swedish process description. (Source: Roslagsvatten,

2015) ................................................................................................................................................ 13 Figure 5. Cross section scheme of DAF basin, Swedish description. Source: Roslagsvatten (2015) ...... 14 Figure 6. Teledyne ISCO 6712 sampler, used to take influent samples for TSS analysis. ...................... 16 Figure 7. Sampling bottles and collection bottles used for influent samples .......................................... 16 Figure 8. HACH TSS control unit for monitoring of effluent TSS ..................................................... 17 Figure 9. Effluent point with TSS monitor and overflow in DAF unit ............................................... 17 Figure 10. Influent TSS sampling and effluent TSS monitoring points in Margretelund WWTP's

flocculation and DAF basin. ............................................................................................................... 18 Figure 11. Pump pipe for influent water. ........................................................................................... 19 Figure 12. Control unit for pump used for influent water................................................................... 19 Figure 13. Daily average DAF influent TSS concentration measured during the experiment,

Margretelund WWTP. ...................................................................................................................... 24 Figure 14. Removal percentage of effluent suspended solids at different levels of effluent recycle rate. 25 Figure 15. Influent TSS concentration compared to TSS removal percentage. .................................... 25 Figure 16. 24h average effluent suspended solids concentration with different effluent recycle rates (ERR)

between 15-35%. Subfigures; A: surface load 2.5 m/h, B: surface load 4 m/h, C: surface load 5 m/h, D:

surface load 6 m/h............................................................................................................................. 26 Figure 17. Effluent suspended solids concentration box plot with median and average values with different

effluent recycle rates between 15-35%. Subfigures; A: surface load 2.5 m/h, B: surface load 4 m/h, C:

surface load 5 m/h, D: surface load 6 m/h. ........................................................................................ 27 Figure 18. Effluent suspended solids concentration with different effluent recycle rate (ERR) between

15-35%. Subfigures; A: surface load 2.5 m/h, B: surface load 4 m/h, C: surface load 5 m/h, D: surface

load 6 m/h. ....................................................................................................................................... 28 Figure 19. Temperature during the experimental run see section 3.4 for experimental input for each trial.

......................................................................................................................................................... 28 Figure 20. Pressure in pressurized tank during the experimental run. See section 3.4 for experimental

input for each trial. ............................................................................................................................ 29 Figure 21. Three different linear correlations showing relationship between factors chosen from Pearson

correlation ......................................................................................................................................... 30 Figure 22. Gas-to-solid ratio calculated with equation 5 for historical period January 2015-January 2021,

DAF unit Margretelunds WWTP ...................................................................................................... 31

vii

Table of tables Table 1. Air solubility in freshwater (Source: Wang, Hung & Shammas, 2005, Physicochemical

Treatment Processes, P.448) ................................................................................................................ 6 Table 2. Typical dry weight percentage of different compounds in prokaryote bacteria (Source: Metcalf

& Eddy, 2013, Wastewater Enginerring: Treatemt amd Resource Recovery, p. 565) ........................... 9 Table 3. Design loads to Margretelund WWTP based on Swedish EPA’s baseline values (Source:

Roslagsvatten, 2015) ......................................................................................................................... 11 Table 4. Discharge limits Margretelund WWTP (Source: Roslagsvatten 2015) ................................... 11 Table 5. Manufacturers design parameters of the DAF unit in Margretelunds WWTP ........................ 14 Table 6. Experimental design plan ..................................................................................................... 20 Table 7. Collected historically data parameters. .................................................................................. 22 Table 8. Summarized historical data of the percentages of samples exceeding 10 mg/l TSS, and percentage

of ERR for samples both exceeding and not exceeding 10 mg/l TSS in DAF unit in Margretelund

WWTP. Data extracted from aCurve for period Jan 2015- Jan 2021 .................................................. 29 Table 9. Pearson correlations ............................................................................................................. 30 Table 10. Historical yearly average values for TSS, phosphorus concentration and percentage of

phosphorus in effluent TSS ................................................................................................................ 31

Legend Variable Description SI – Unit Abbreviation

NTU Nephelometric Turbidity Unit - NTU

Q Effluent flow rate m3h-1 Effl. flow

Qinf Influent flow rate m3h-1 Infl. flow

Qr Recycled effluent flow rate m3h-1 Rec. flow

CTSS inf Concentration influent total suspended solids mg1L-1 Infl.conc

C TSS eff Concentration effluent total suspended solids mg1L-1 Effl. Conc

A Area m2h-1 -

Vh Surface load m1h-1 Surf.load

℃ Temperature ℃ Temp

m Mass of 1 mg/ml air mg1mL-1 -

a Airs solubility at 1 atm pressure mL1L-1 -

f Fraction of gas dissolution at pressure p, constant value - -

P Pressure N1m-2 -

G/S Gas – to – solid ratio % -

r Pearson Correlation Coefficient - -

n Number Nr -

X Data set - -

Y Data set - -

t Student t-test - -

p Level of Significance % -

1

2

1. Introduction

Located in Åkersberga town, Sweden, is Margretelund wastewater treatment plant (WWTP) with its

receiving waters, Trälhavet, Saltsjön, approximately 300m from shore. Margretelund WWTP is designed

for a population equivalent (PE) of 40 000, with 36 500 PE connected in 2021.

Built in 1956, with two subsequent major renovations, in 1974 and 1999, to meet demands regarding

PE capacity and nitrogen removal. Margretelund WWTP has since 1999 been operated with the same

processes for physical, chemical, and biological treatment of wastewater, with moving bed biofilm reactor

(MBBR) for both biological treatment and a nitrification-denitrification process. Margretelund WWTP

installed a dissolved air flotation (DAF) process in 1999 instead of a more conventional sedimentation

basin for removal of effluent total suspended solids (TSS).

Margretelund WWTP has historically had issues with increased concentration of TSS in effluent water,

caused by increased flow rates and unwanted additional waters. Margretelund WWTP’s design flow rate

is 600 m3/h and increases of up to 32% of the flow rate have been recorded.

Majority of TSS in wastewater is of organic heritage, and organic matter contains the base element

phosphorus, which is a macro nutrient needed for production of new organic organisms, eutrophication

of receiving waters could be a result if the effluent TSS emissions increase. By breaking down organic

matter released into the receiving waters, the amount of available phosphorus in the water increases, and

thus production of organic matter increases. While Margretelund WTTP does have discharge limits for

phosphorus, there are no legal limits for TSS, and it is believed that the DAF process is operated

inefficiently in removal of TSS when spikes of increased flow rates occur.

1.1 Purpose of study The author was tasked by Roslagsvatten AB, the water utility operating Margretelund WWTP, to

conduct a study about their dissolved air flotation process. A study which purpose was to optimize the

present operation design for Margretelund WWTP’s DAF unit with available means was suggested,

including investigation how the DAF unit should be operated during high flow rates to avoid increased

effluent TSS concentrations. The study should provide Roslagsvatten AB with proposed actions to further

optimize the present design of operation of the DAF unit based on a current situational analysis.

The main purpose for this study was to propose at least one method for optimization of operation of

the DAF unit the DAF unit at Margretelund WWTP when increased flow rates occurs. More specifically,

the objectives were to:

• analyse historic flow and TSS data for future proposed actions to the DAF process,

• investigate the effect of influent flow rate and effluent recycle rate (ERR) on the effluent TSS

concentration, and to

• present suggestions on how to optimize the present design of operation of the DAF unit with

available means to minimize effluent TSS discharge.

1.2 Scope of study Theory studied in this master thesis was about dissolved air flotation and differences between the DAF

process to conventional sedimentation. The focus of theory was on recycle flow pressurization

configuration when considering design parameters, since Margretelund WWTP utilize this specific

configuration.

Analysis of historic data was focused on historical design of operation and process efficiency, to evaluate

which factors would provide a proposed impact onto the DAF process. Energy consumption, cost

calculations or physical design calculations were not conducted within the study.

The numbers of experimental factors for this master thesis was determined to two and was

experimentally tested during a period of one month.

3

4

2. Theory

2.1 Dissolved Air Flotation Dissolved Air Flotations (DAF) primary objective is to separate solid-liquid solution through flotation.

This is done by inducing a pressurized, supersaturated solution of gas-liquid mixture into the influent

flow. Pressurized mixture will, when confronted with pressure release in normal atmospheric pressure,

produce gas bubbles that lift total suspended solids (TSS) or colloidal solids to the surface. Bubbles produce

a bubble-particle agglomerate with particulate solids. Through bubbles buoyancy force in liquid, the

density of suspended solids decreases so they rise upwards and float on top off the surface (Metcalf &

Eddy, 2013;Wang, Hung, & Shammas, 2005). This process also works well for separation of oils, dissolved

solutes, heavy- and light solids and grit (Wang, Hung, & Shammas, 2005).

Three more common configurations of DAF processes used for water treatment are the full flow

pressurization, partial flow pressurization and recycle flow pressurization. Difference of these three is how

they pressurize the saturated gas-liquid mixture (Wang, Hung, & Shammas, 2005). Design parameters

deciding what process to utilize are the efficiency in floating different total suspended solids (TSS)

concentration and the surfaceloading rate (Metcalf & Eddy, 2013; Wang, Hung, & Shammas, 2005).

Full flow pressurization

Entire influent feed is pressurized by a pressurizing pump and held in a retention tank before released into

flotation chamber. This process is focused on low surfaceloadings (5–15 m/h) and the highest suspended

solid concentration of the three systems with >800 mg/l of suspended solids. This system is best suited

for water where the suspended particles flocculate rapidly.

Air is induced directly into the feed, to remove any collision impact between a pressurized and

unpressurized flows that occurs in the two other systems. Without any major collision impact, no regard

to shearing strains on to the particle flocs needs to be taken. Coagulating chemicals can be used at the

inlet to further increase the flocculate size, but to also entrap bubbles inside the aggregates resulting in a

strong air to solids bond (Wang, Hung, & Shammas, 2005).

Partial flow pressurization

A portion between 30–50 % of influent water are separated and pumped with a high-pressure pump into

a retention tank for saturation before entering the DAF basin. Remaining influent water is either led by

gravity or a low-pressure pump towards the DAF unit. Pressurized water is induced for production of gas

bubbles from the pressure drop. Partial flow pressurization are suitable in wastewater feeds containing

low concentrations of suspended solids. High shearing force are applied onto the TSS flocs from the high-

pressure pump and the large pressure drop, and may break apart the flocs (Wang, Hung, & Shammas,

2005).

Recycle flow pressurization.

Clarified effluent is recycled back to the influent flow with a effluent recycle rate (ERR) between 10–

120% according to (Metcalf & Eddy, 2013) and ERR of 15–50% according to (Wang, Hung, & Shammas,

2005) of the total flow rate. Recycled effluent is pressurized in a retention tank to 3–6 times the

atmospheric pressure (Metcalf & Eddy, 2013). Pressurized, semi saturated effluent is mixed with the main

influent just before entry point of the flotation chamber (Edzwald, 2010). Gas bubbles produced from

the pressure drop collides with flocculated suspended solids. Reycle flow pressurization have minor issues

with shearing forces from collision of microbubble-particle impact. But it does not have the issues partial

flow pressurization has, since pressurized water are from clarified effluent and no flocculate solids enters

the retention tank. Preliminary addition of chemical coagulation and flocculation is necessary for this

process to achieve bigger floc size with the suspended particle aggregates to attach more microbubbles.

(Wang, Hung, & Shammas, 2005).

5

2.2 History of DAF DAF is a separation process for liquid-solid solutions that until 1920 were used for mineral separation in

mining industry (Edzwald, 2010). At first vacuum-based and full flow pressurised systems were

implemented for clarification of drinking water in the 1920’s (Wang, Hung, & Shammas, 2005).

Dissolved Air Flotation clarifiers during this time had a surface load capacity between 5–15 m/h and

Detention times around 20–25 minutes. Although it was not until in the 1960 that a Scandinavian

company developed a DAF system that laid the foundation of how these systems are used today (Wang,

Hung, & Shammas, 2005; Edzwald, 2010). This new DAF system utilized pressurized recycled effluent

flow as a source to produce air bubbles, and it shortly became a clarification method competing with

settling to become the primary method for surface and wastewater (Edzwald, 2010). From improvements

over the years providing possibilities for production of smaller facilities with increased efficiency came in

mid – 1990, a DAF system based on the Scandinavian method, that could handle surface load up to 15–

30 m/h while reducing the detention times down to 3–5 minutes (Wang, Hung, & Shammas, 2005;

Edzwald, 2010).

2.3 Design parameters When designing and operating a DAF unit, a few important design parameters based on the density of

particle flocs and viscosity of the liquid must be taken into consideration (Metcalf & Eddy, 2013). These

parameters are the concentration of particulate matter, the microbubbles efficiency, air-to-solid ratio,

surface load, particle rise velocity, collision of micro-bubbles and floc and how size distribution

determines efficiency (Wang, Hung, & Shammas, 2005). They are further presented in text below.

2.3.1 Surface load and particle rise velocity

Surface load and rising velocity are both important for DAF and its removal of suspended particles.

Influent feed has a velocity forward through the flotation chamber, so there is a requirement of a fast

enough rising velocity for the particles to be floated. To change the density of the suspended particles

with chemical additives, the rising velocity of the flocs may increase enough before the flocs would reach

the effluent point (Wang, Hung, & Shammas, 2005; Metcalf & Eddy, 2013).

Surface load varies between two different rates depending on DAF design. The first is a loading rate

between 5–15 m/h and used mainly in conventional design. The second DAF design, called highrate

DAF, has been developed since the mid – 1990s and as described in section 2.2, could handle loading

rates between 15–35 m/h (Edzwald, 2010). Detention times in the flotation chamber varies between 3–

60 minutes for both designs, with recycle flow pressurization configuration being within the range of 3-

5 minutes detention time (Wang, Hung, & Shammas, 2005).

2.3.2 Bubbles and airs solubility

Bubble volume and concentration has been shown to determine the efficiency of the DAF process. With

smaller average bubble size comes higher efficiency gains. The small bubbles, so called micro-bubbles,

are in the range of 10–100 µm with an average value around 40-50 µm (Han, Kim, & Kim, 2007; De

Rijk, Van Der Graaf, & Den Blanken, 1993). De Rijk et.al (1993) showed that a relation exists between

increasing the saturation pressure and flow rate for the effect of decreasing average median size of micro-

bubbles. The effect is however limited, with no noticeable size decrease to micro-bubbles after reaching

a pressure of 6.2 bar or above.

Production of supersaturated water works in retention tanks pressurized between 0.5–3 minutes with

pressure ranging from 170 kpa–650 kpa (Edzwald, 2010; Wang, Hung, & Shammas, 2005). Amount of

dissolved gas into liquid is a function of temperature, with colder water temperature yielding greater

solubility of gas. Although a simpler measurement for dissolved gas saturation are to assume that a gas –

liquid solution is in equilibrium. Then, the concentration of dissolved gas saturation is directly

6

proportional to gage pressure in a retention tank according to Henry’s law (Vallero, 2014; Wang, Hung,

& Shammas, 2005).

Table 1. Air solubility in freshwater (Source: Wang, Hung & Shammas, 2005, Physicochemical Treatment Processes, P.448)

Temperature Volume solubility Weight solubility Density

℃ mL/L mg/L g/L

0 28.8 37.2 1.293

10 23.5 29.3 1.249

20 20.1 24.3 1.206

30 17.9 20.9 1.116

40 16.4 18.5 1.13

50 15.6 17 1.093

60 15 15.9 1.061

70 14.9 15.3 1.03

80 15 15 1

90 15.3 14.9 0.974

100 15.9 15 0.949

Oxygen is the gas mainly used in DAF units, but carbon dioxide and nitrogen gas have been used in DAF

processes (Wang, Hung, & Shammas, 2005).

Bubbles have mainly two mechanisms to make the flocs float, first one is inclusion where the micro-

bubbles are encased into the sludge floc. The second mechanism is adhesion, when micro-bubbles are

adsorbed onto flocs (De Rijk, Van Der Graaf, & Den Blanken, 1993). To improve these two mechanisms

efficiency. De Rijk et.al (1993) suggests keeping the bubble size <100 µm to increase the probability for

inclusion and adhesion for micro-bubbles in flocs. Smaller bubbles make it possible for smaller contact

angle between bubbles and flocs aggregates. Smaller bubbles may be included into a floc more easily than

a bigger bubble (De Rijk, Van Der Graaf, & Den Blanken, 1993). Furthermore, residence time in the

flotation unit is dependent on bubble size. Bubble size affects the rising velocity, with larger bubbles

having increased velocity. Thus, smaller bubbles having lower rising velocity would improve the

possibility of collision between bubble and flocs (De Rijk, Van Der Graaf, & Den Blanken, 1993). If

these bubbles become too big (> 2mm), shearing forces between flocs and micro-bubbles produced due

to high velocity may break the floc at impact.

2.3.3 Gas-to-solid ratio

Gas-to-solid ratio (G/S) is one of the more frequently used methods for improving efficiency of DAF

systems. With a ratio showing the percentage of gas flux in the system compared to flux of suspended

solids (Metcalf & Eddy, 2013). Wang et.al (2005) showed that the ratio should be within an interval of

0.02 to 0.06 for highest concentration of floated concentration (figure 1). Although, a more typical range

seen in various flotation processes from wastewater treatment is between 0.005 to 0.060 (Metcalf & Eddy,

2013).

To increase the gas-to-solid ratio would mean higher concentration dissolved gas per amount of

concentration suspended solids. As mentioned in subsection 2.3.2, an increase of gas volume would lead

to more micro-bubbles per floc. Thus, an increased gas-to-solid ratio would increase the flotation

efficiency (Han, Kim, & Kim, 2007).

7

Figure 1. Effect of air-solids ratio on float concentration and subnatant suspended solids (Source: Wang, Hung & Shammas,

2005, Physicochemical Treatment Processes, P.444)

2.3.4 Collision efficiency and flocs type

Suspended particle floc size is an important factor that affects the collision efficiency of particle –

microbubble collision (Han, Kim, & Kim, 2007). This collision is the process that brings the greatest

impact on efficiency of the DAF system, with smaller floc size requiring smaller micro-bubbles to reach

higher efficiency (Han, Kim, & Kim, 2007).

Edzwlad (2010) assumed that microbubbles flow through the media like sand would in a filter bed,

collecting particles in their path. With larger flocs achieved comes increased collision rate and additionally,

flocs gain increased hydrophobic properties with increased size (Wang, Hung, & Shammas, 2005; Han,

Kim, & Kim, 2007). Suspended solids have a broad size distribution that varies depending on the feed

water composition. Flocculating chemicals can be added to enhance the particle agglomeration for a more

homogeneous size distribution of suspended solids (Metcalf & Eddy, 2013; Wang, Hung, & Shammas,

2005). These chemicals, having a positive net charge, attracts the negatively charged suspended solids for

production of flocs (Wang, Hung, & Shammas, 2005). Two common chemicals used in waste- and raw

water treatment are poly-aluminium chloride and ferric chloride (Wang, Hung, & Shammas, 2005).

An increase in micro-bubble volume concentration was shown by Han et.al (2007) to be an important

factor. Increased volume concentration would provide a greater amount and a wider spread in size

distribution of the micro-bubbles, providing better removal efficiency for the total size range of particle

flocs (Han, Kim, & Kim, 2007).

2.4 Benefits and disadvantages of DAF compared to sedimentation According to Wang et.al (2005), a particle floc coagulated from added chemicals would have a minimum

required detention time of 2–4 hours in a conventional sedimentation unit depending on the surface

load. Metcalf & Eddy (2013) argues however that a sedimentation’s detention time varies between 1.5–

2.5 hours, with typical dimensioning value at 2 hours. While a DAF unit may only require 1–60 minutes

of detention time depending on load and design, which is more beneficial as the DAF would have a

much lower footprint (Wang, Hung, & Shammas, 2005). Detention time differences between the two

8

systems depends on the settling speed. A sedimentation basin is designed to only utilize gravity, and by

adding coagulating chemicals increase the density of flocs, for an increased settling velocity (Wang, Hung,

& Shammas, 2005). A DAF system, on the other hand, uses the micro – bubbles rising velocity, which

is far greater than a particle settling velocity. Because of the higher velocity DAF systems has compared

to a conventional sedimentation, DAF systems can be run with these low detention times.

Velocity differences allows difference in dimensioning of both systems. If both a DAF unit and a

conventional sedimentation where to be designed after a specific surface load and designed flow rate, the

DAF unit would require less surface area and depth than a conventional sedimentation (Wang, Hung, &

Shammas, 2005). DAF systems have been proven by Khiadani et.al (2013) to be more efficient in removal

of turbidity compared to conventional sedimentation with a 25–35% difference within Nephelometric

Turbidity Unit (NTU) ranges of < 20, 30–50 and 90–110. The DAF process had a requirement of less

coagulation additives then what conventional sedimentation required to reach same results. DAF is thus

a process more suitable for removal of low-density particles. but at a cost of more accurate control of the

whole system (Khiadani, Kolivand, Ahooghalandari, & Mohajer, 2013).

According to Svenskt Vatten (2007), a DAF process would be better suited for wastewater containing

a vast amount of smaller flocculated particles, a more efficient process than a conventional sedimentation

basin. Flocs having a density beneath or close to waters density, requires a smaller effort to be floated.

DAF are therefore often used with chemical additives, like coagulant or flocculants, in production of the

flocs (Svenskt Vatten, 2007).

One main disadvantage found to DAF compared to a conventional sedimentation process, is the high

energy demand of producing a pressure of 3-6 atm and of the pumps for recycle flow, generating a greater

cost and larger CO2 emission per cubic meter wastewater treated (Féris et.al, 2000).

2.5 Phosphorus in bacteria Municipal sewage wastewater is in general rich in organic nutrients, especially nitrogen and phosphorus,

which are considered macro nutrients and are essential in forming new organic matter. Prokaryotic cells

utilize these nutrients in wastewater for its production of new cells. With more cells available organic

particles in wastewater start to aggregate and build up flocs of suspended matter flowing with the

wastewater (Metcalf & Eddy, 2013).

Physical removal of phosphorus is needed since phosphorous does not have a gaseous species like

nitrogen. Therefore, phosphorus cannot be removed by evaporation into the atmosphere, but instead

stays bound in particulate or dissolved species in natural systems (Metcalf & Eddy, 2013).

According to Metcalf & Eddy (2013), prokaryotes can be described with the formula C60H87O23N12P,

where phosphorus has a typical dry weight percentage of 2.0 %, in prokaryote bacteria (table 2).

9

Table 2. Typical dry weight percentage of different compounds in prokaryote bacteria (Source: Metcalf & Eddy, 2013,

Wastewater Enginerring: Treatemt amd Resource Recovery, p. 565)

Constituents or element Percent of dry weight

Major cellular material

protein 55.0

polysaccharide 5.0

Lipid 9.1

DNA 3.1

RNA 20.5

Other (sugars, amino acids) 6.3

Inorganic ions 1,0

As cell elements Carbon 50.0

Oxygen 22.0

Nitrogen 12.0

Hydrogen 9.0

Phosphorus 2.0

Sulfur 1.0

Potassium 1.0

Sodium 1.0

Calcium 0.5

Magnesium 0.5

Chlorine 0.5

Iron 0.2

Other trace elements 0.3

10

11

3. Methods

3.1 Margretelund wastewater treatment plant Margretelund wastewater treatment plant (WWTP) located in Österåker municipality in northern

Stockholm County, Sweden, was built in 1956 and renovated during two subsequent periods in 1974

and 1999. Margretelund WWTP has operated the same processes for wastewater treatment since the latest

renovation in 1999, and these processes was still being used during the experimental runs (subsection

3.1.2) (Roslagsvatten, 2015).

There are approximately 36 500 individuals registered in the densely populated areas of Margretelund

WWTP’s catchment area and Margretelund WWTP’s designed max capacity was calculated to 40 000

population equivalents (PE) with a design flow up to 600 m3/h (Roslagsvatten, 2015). Margretelund

WWTP’s designed PE value is based on the Swedish Environmental Protection Agency’s (EPA) baseline

value for biochemical oxygen demand (BOD7) in influent wastewater. The EPA’s baseline assumption

for influent wastewater in Sweden is 70g BOD7/PE (Naturvårdsverket, 2019).

Table 3. Design loads to Margretelund WWTP based on Swedish EPA’s baseline values (Source: Roslagsvatten, 2015)

Parameter SI-unit Value

Design size PE 40 000

Design flow m3h-1 600

BOD7 kg1d-1 2800

Total phosphorous kg1d-1 120

Total Nitrogen kg1d-1 520

Table 4. Discharge limits Margretelund WWTP (Source: Roslagsvatten 2015)

Parameter SI-unit Value Time period Definition

BOD7 mg1L-1 10 Average value per quarter Guide value

BOD7 mg1L-1 10 Average value per year Limit

Total phosphorous mg1L-1 0.3 Average value per quarter Guide value

Total Phosphorous mg1L-1 0.3 Average value per year Limit

Total Nitrogen mg1L-1 15 Average value per quarter Guide value

Total Nitrogen mg1L-1 15 Average value per year Limit

Total Nitrogen % >50% Removal

12

3.1.1 Catchment area

Margretelund WWTP catchment area is seen highlighted within the circle (figure 2). Wastewater is

mainly collected from the town of Åkersberga located in Österåker municipality (center). Although, the

catchment also includes in the north and south edges, smaller areas of Vallentuna municipality (north)

and Vaxholm municipality (south).

Total combined length of sewer piping is 240 km over the total area, consisting of a mix between low

pressure sewers, sea pipeline and self-flow sewers, with the latest having a large majority.

Figure 2. Margretelund WWTP catchment area as described by Roslagsvatten (2015)

Figure 3. Location of Åkersberga town, Sweden (Map from Eniro (2014))

13

3.1.2 Processes in Margretelund wastewater treatment plant

Margretelunds process scheme (figure 4) presents the mechanical, biological and chemical treatment

processes and the flow scheme of the WWTP. The processes are further described below in written text.

Mechanical treatment starts at incoming section of the WWTP with a screen and grit chamber, followed

by pre – sedimentation and in the last step in the WWTP where the water is treated in a dissolved air

flotation (DAF) reactor.

Biological treatment in Margretelund WWTP is conducted in moving bed biofilm reactors (MBBR)

in an aerobic environment. For removal of nitrogen in the wastewater, a pre-denitrification – nitrification

– denitrification process is used in combination with MBBR, see figure 4. Recirculation of water occurs

between the nitrification – denitrification step.

Chemical treatment is conducted in four different steps during the whole process. Coagulants are

added to thicken sludge before dewatering. Flocculants are added both in the grit chamber and in the

flocculation chamber. Phosphoric acids are added to prevent phosphorus shortage in the nitrification

process and external carbon source are added in the last denitrification process.

Figure 4. Flow scheme Margretelund WWTP, Swedish process description. (Source: Roslagsvatten, 2015)

3.1.3 DAF reactor

Margretelunds WWTP had three parallel DAF reactors utilizing recycle flow pressurization configuration

(section 1.1), with a surface area of 50 m2 and a water depth of 4.0 m each. Four wooden walls are located

at the start of the DAF basin (figure 5,). With inlets in different heights to reduce flow rate peaks and

allow a more constant water depth through the DAF reactor. The fourth wall, see figure 5, was tilted to

a certain degree to guide the pressurized flow upward. Micro-bubbles produced during pressure release

was then concentrated to a small area of the DAF basin to increase the bubble volume concentration

(subsection 2.3.4) (Roslagsvatten, 2015). According to the DAF reactors manufacturer, Purac, the design

values for Margretelund WWTP’s DAF process was a surface load of 4 m/h with a percentage of total

flow rate being recycled, called effluent recycle rate (ERR), of 10-15%. Surface load of 4 m/h corresponds

to a flow rate of 200 m3/h. (Roslagsvatten, 2015).

All three DAF basins used one pressurized tank for pressurization and saturation of recycled effluent.

Treated effluent water used as recycled water for the pressure tank was collected by two pumps at the

outlet of the DAF basins, with a capacity of 65 m3/h each (Roslagsvatten, 2015). One pump ran

constantly to refill the pressurized tank, and the second pump started when the first pump was insufficient

to keep a constant volume in the tank. Volume of the pressure tank was 4 m3 with a constant target

14

pressure at 6 bar. If the pressure exceeds 6 bar, a pressure relief valve opened automatically to reduce the

excess pressure.

Two sets of sludge scrapers was installed in each basin, one for the floated sludge blanket and one for the

heavier sludge blanket that settles at the bottom of the basins.

Surface sludge scrapers was run in intermittent mode with an interval of 0-40 minutes between each

start. According to Purac (2000), normal run time value was set to 5 minutes but could manually be

changed within an interval of 0–10 minutes. Bottom sludge scraper started once a day and could be set

to run for 0–120 minutes, with normal runtime at 30 minutes (Purac, 2000; Roslagsvatten, 2015).

The valve regulating ERR opens the nozzle to max capacity momentarily before closing again to set

ERR value at 3 a.m. every night to prohibit build-up of calcium carbonate. This sequence releases a

burst of recycled water into the DAF basin, producing turbulent flow as a result. The turbulent flow can

reduce the DAF process ability to float flocs and it could swirl up sediment from the bottom of the basin.

It was noticed that this re-occurring process increased the effluent TSS concentration at 3 a.m. with a

concentration decreasing slowly over time afterwards. These values where removed from experimental

results and seen as an anomaly due to an unnatural increase of effluent TSS concentration.

Table 5. Manufacturers design parameters of the DAF unit in Margretelunds WWTP

Parameter SI-unit Value

Effluent recycle percentage % 15

Pressurized flow m3h-1 30

Influent flow m3h-1 170

Total flow m3h-1 200

Area m2h-1 50

Surfaceloading rate m1h-1 4

Figure 5. Cross section scheme of DAF basin, Swedish description. Source: Roslagsvatten (2015)

15

3.2 Gathering relevant theory and information Relevant theory and information used for this project was collected from different sources. The literature

used consisted of different scientific articles and relevant books in the subject. Roslagsvatten, the water

utility operating Margretelund WWTP provided a process description of the DAF process from the

manufacturer Purac AB, a floor plan for the DAF unit and self-monitoring program about Margretelund

WWTP.

3.3 Data collection Data used in the thesis was collected through water samples, online monitoring, and extraction of

historical data via computer software’s.

3.3.1 Sampling

During each experimental run samples of influent total suspended solids (TSS) were taken once every

hour with a Teledyne ISCO 6712 Full-Size Portable Sampler that was programmed to do sequence test

over 24 hours, where one sequence was 200ml per bottle and hour. All samples were mixed and then

poured into a 2 – liter bottle to give an estimated average concentration during the latest 24 hours.

The sampler stopped working around hour 15-16 caused by an unknown error and required a hard

reset of the sampler to function again, occurred for two out of 19 experimental runs. This resulted in that

these two samples would not represent the full planned 24 hours, but the collected sample provided

enough water for analysis.

3.3.2 Analysis of total suspended solids in the influent and effluent of the DAF

Analysis for influent TSS in wastewater was conducted by Roslagsvatten ABs accredited laboratory on

site at Margretelund WWTP. The sample was first brought to room temperature at 20 + 2°C before

filtration. The filtration process was done with vacuum filtration in 1.6 µm filter, to later dry the filter in

an oven at 105 + 2°C for minimum 1 hour and maximum 14–16 hours before weighing. The result was

transformed and presented as mg total suspended particles per liter water.

Effluent TSS concentration were continuously analysed with a TSS sc Sensor from HACH. The TSS

sensor was cleaned before every experimental run according to descriptions provided from HACH.

Figure 8 depicts the control unit used for monitoring the effluent TSS concentration, and measured data

was stored and collected using the software aCurve (subsection 3.3.4).

16

Figure 6. Teledyne ISCO 6712 sampler, used to take influent samples for TSS analysis.

Figure 7. Sampling bottles and collection bottles used for influent samples

17

Figure 8. HACH TSS control unit for monitoring of effluent TSS

Figure 9. Effluent point with TSS monitor and overflow in DAF unit

18

3.3.3 Points for influent TSS sampling and effluent TSS monitoring.

Influent TSS sample point was positioned in the second (out of two) flocculation basins discharge point,

at a depth of 0.5 m from the surface (figure 10).

Effluent TSS monitoring point was positioned at the end of the DAF basin at a depth at 0.2 m (figure

10) and before the overflow for treated effluent wastewater (Figure 9).

Figure 10. Influent TSS sampling and effluent TSS monitoring points in Margretelund WWTP's flocculation and DAF basin.

3.3.4 Software

aCurve, a software developed by gemit Solutions AB was used to collect, review, and extract raw data in

real time from Margretelund WWTP’s different control systems. Resolution of data can be manually set

between second, minute, hour, day, or week. Extracted data were further presented in excel, for

calculations and assumptions for this study. Data extracted from aCurve during this project was effluent

TSS, pressurized flow rate, influent flow rate, and effluent flow rate from Margretelund WWTP (table

7).

Software Labware was used to analyse sampling data and lab data from Roslagsvatten ABs accredited

laboratory. Data about concentration total suspended particles and total phosphorous in incoming and

effluent water was extracted from this software for the period of January 2015 – January 2021 to use in

calculations.

Influent TSS sampling point

Effluent TSS sampling point

Waters flow path

Influent TSS sampling point.

Effluent TSS monitoring point.

Waters flow path.

(A). Cross section of Margretelund WWTP’s DAF unit

(B). Overview of Margretelund WWTP’s flocculation and DAF basin

Flocculation basin

DAF basin

19

3.4 Experimental design Based on results from section 4.2, the investigated factors were selected to be the effluent recycle rate

(ERR) and the surface load (equation 1). The ERR was set to vary between different percentages

between 15-35%, with 5% increase for each step. These ERR were tested on four different surface loads,

2.5, 4, 5 and 6 m/h. These four was chosen based on table 10, with the most historically common surface

load being 2.5 m/h, the designed surface load was 4 m/h, and to test if improvements could be achieved

in surface loads above the designed loading rate, 5 and 6 m/h were chosen. The time for each test was

set to 24 hours, allowing the WWTP one hour for stabilizing between flow changes and so daily variations

of incoming SS could be seen for each ERR. See table 8 for full design plan.

To keep a constant influent flow rate in the DAF line that was for the experiment, the section was

sealed off from the distribution path and the influent water was pumped into the experimental line during

the full time-period. Figure 12 depicts the control unit for the pump used for influent water, and the

flow rate was changed by changing the Hz value by trial and error.

Figure 11. Pump pipe for influent water.

Figure 12. Control unit for pump used for influent water.

20

Table 6. Experimental design plan

Experimental design

Test ERR

Pressurized

flow rate

Influent

flow rate

Total flow

rate Area Surface load

nr % m3h-1 m3h-1 m3h-1 m2 m1h-1

1.1 20 60 240 300 50 6

1.2 25 75 225 300 50 6

1.3 30 90 210 300 50 6

1.4 15 45 255 300 50 6

2.1 15 30 170 200 50 4

2.2 20 40 160 200 50 4

2.3 25 50 150 200 50 4

2.4 30 60 140 200 50 4

2.5 35 70 130 200 50 4

3.1 15 18.75 106.25 125 50 2.5

3.2 20 25 100 125 50 2.5

3.3 25 31.25 93.75 125 50 2.5

3.4 30 37.5 87.5 125 50 2.5

3.5 35 43.75 81.25 125 50 2.5

4.1 15 37.5 212.5 250 50 5

4.2 20 50 200 250 50 5

4.3 25 62.5 187.5 250 50 5

4.4 30 75 175 250 50 5

4.5 35 87.5 162.5 250 50 5

3.5 Analysis of historical DAF data Historical data on TSS contents, ERR and flow rates of the WWTP were analysed to investigate if there

were any faults in the current design of operation of the DAF unit and to evaluate the unit’s efficiency.

First, information was collected about the existing DAF unit and what available design parameters that

may be changed without any reconstruction work. This was done by ocular inspection on–site and by

using information about Margretelunds WWTP, the DAF reactor and construction schemes of the DAF

basin available at Roslagsvatten AB.

By assuming a phosphorus concentration of 2% dry weight in Margretelunds TSS, it would correspond

with discharge regulations of 0.3 mg/l phosphorus to an effluent TSS concentration of 15 mg/l.

Calculations and assumptions for the historical data analysis was instead based on a limit of 10 mg/l TSS

to allow room for fluctuating concentration up to 15 mg/l TSS, and a varying phosphorus dry weight

concentration depending on TSS constituents.

Data used for calculations (table 7) was extracted from the software’s LabWare and aCurve, further

described in subsection 3.3.4. From data gathered, surface load and removal percentage of TSS was

calculated with equation 1 and equation 2, respectively.

𝑉ℎ =𝑄𝑟+𝑄𝑖𝑛𝑓

𝐴 (eq.1)

Where 𝑉ℎ is surface load, 𝑄𝑟 is the recycled flow rate in the DAF basin, 𝑄𝑖𝑛𝑓 is the influent flow rate to

the DAF basin and 𝐴 is the cross-section area of the DAF basin.

21

% 𝑟𝑒𝑚𝑜𝑣𝑎𝑙 = 1 − (𝐶𝑇𝑆𝑆 𝑒𝑓𝑓

𝐶𝑇𝑆𝑆 𝑖𝑛𝑓) (eq.2)

Where 𝐶𝑇𝑆𝑆 𝑖𝑛𝑓 is the influent TSS concentration and 𝐶𝑇𝑆𝑆 𝑒𝑓𝑓 is the effluent TSS concentration.

3.5.1 Regression analysis

Pearson correlation was conducted on data chosen from historical data analysis to test how strong

correlations between different design parameters and effluent suspended particle are in Margretelunds

WWTP DAF reactors. Number of tests done was set to 312, with a 310 degree of freedom.

Equation 3 was used in excel to produce the correlation coefficient r. The coefficient may vary between

-1 to +1, indicating a negative or positive correlation, with values closer to zero having low to no

correlation, and values between 0.3-0.5 having low and 0.5-0.7 having moderate correlation (Berman,

2016).

𝑟 = 𝑛(∑ 𝑋𝑌)−(∑ 𝑋)(∑ 𝑌)

√(𝑛 ∑ 𝑋2−(∑ 𝑋)2)(𝑛 ∑ 𝑌2−(∑ 𝑌)2) (eq.3)

Where 𝑛 was the amount of data sets, 𝑋 and 𝑌 were different sets of data that was tested against each

other to see the correlation between them.

Critical values used for comparing Pearson correlation coefficient was used with a two tailed student

t-test and a significance level alpha equal to 0.05, see Appendix 4 for table with critical values. For the

first correlation test, where the number n of data sets exceeds 100, the critical value for 100 was chosen

(Berman, 2016; Siegle, 2015).

To test the null hypothesis of each correlation, equation 5 was used to gain the t – statistic value. The

higher t-statistic one statistic set gains; the stronger correlation exists. Excels built-in statistical function

T.DIST.2T was further used for two-tailed t-test to calculate the probability p for each t – statistic and

compare the null hypothesis against alpha 0.05.

𝑡 =𝑟𝑥𝑦√𝑛−2

√1−𝑟𝑥𝑦2

(eq.4)

Where 𝑡 was the t-statistic value, 𝑟 was the Pearson correlation coefficient calculated from eq.3, 𝑋, 𝑌 and

𝑛 were the same parameters as in equation 3.

3.5.2 Gas – to – solid ratio

Gas to solid ratio was calculated based on equation 3, to get a visual representation on how Margretelunds

WWTP gas–to–solid ratio (G/S) lays compared to theory (figure 1).

The ratio is calculated from data gathered from aCurve and LabWare during the period of January

2015 – January 2021.

𝐺

𝑆=

(𝑚𝑎)(𝑄𝑟𝑄

)(𝑓𝑃−1)

𝐶𝑠𝑠 𝑖𝑛𝑓 (eq.5)

Where 𝐺 is mass flow rate of gas, 𝑆 is the mass flow rate of solids, 𝑚 is the mass of 1 mg/ml air, 𝑎 is the

airs solubility at 1 atm pressure, 𝑄𝑟 is the recycled effluent flow rate, 𝑄 is the total effluent flow rate, 𝑓 is

a constant value between 0.1-1 for the fraction of gas dissolution at pressure P, assumed to be 0.5. 𝑃 is

the pressure in atm, and 𝐶𝑠𝑠 𝑖𝑛𝑓 is the influent TSS concentration.

3.5.3 Amount of phosphorus in wastewater

Theory proclaims that dry weight of phosphorus in prokaryotic bacteria is up to 2% (section 2.5) and

while wastewater contains an inhomogeneous TSS mixture. Assuming that majority is organic matter, an

estimated phosphorus concentration could be calculated based on concentration of TSS (Metcalf & Eddy,

2013).

22

With 312 datasets of TSS concentrations and phosphorus concentrations collected from Margretelund

WWTP effluent wastewater between January 2015 – January 2021, yearly average values for effluent

TSS, phosphorus and percentage of phosphorus to effluent TSS was calculated (Table 10). These 312

datasets were extracted from LabWare.

3.5.4 Processing of historical data

Chosen data parameters extracted from aCurve and LabWare for the period January 2015 to January 2021

are presented below in table 7.

Table 7. Collected historically data parameters.

Parameter SI-unit

Concentration influent total suspended particle mg1L-1

Concentration effluent total suspended particle mg1L-1

Influent flow rate m3h-1

Effluent flow rate m3h-1

Pressurized flow rate m3h-1

Concentration effluent phosphorus, total value mg1L-1

Resolution of data collected was set to weeks from days and hours to reduce the amount of data sets

down to 312 points. The resolution scale was increased to match different sets of data that had been

sampled during different days but the same week. Some data sets had multiples for the same week, so an

average value was based on these samples to use as one single data point.

Historical concentration of influent TSS was sampled before the grit chamber (figure 4). A calculated

removal of 60% from sample value was done to match the concentration that theoretically would be

achieved after the grit chamber and pre-sedimentation (Svenskt Vatten, 2007).

23

24

4. Results

4.1 Experiment results The setting for the investigated factors off each experimental run may be found in table 6.

4.1.1 Influent suspended solids concentration

Figure 13 presents the sampling results for influent TSS concentration for each day during the experiment.

Highest value was 695 mg/l, lowest was 56 mg/l and average value was 165 mg/l.

Figure 13. Daily average DAF influent TSS concentration measured during the experiment, Margretelund WWTP.

4.1.2 Removal efficiency of total suspended solids

Surface load of 2.5 m/h kept a removal between 96-98% throughout every change of effluent recycle

rate (ERR).

Surface load 4.0 m/h have two peaks at 15% respectively 35% ERR with both values close to 96%

removal. ERR 20-30% are kept almost constant at 92-93% removal.

Surface load of 5 m/h has a highly fluctuating values between 77-92% removal. A removal close to

77% is seen in figure 17 for 25-30% ERR. 5 m/h peaks at 92% removal with 35% ERR.

Surface load of 6 m/h reached a plateau between ERR 20-25% with 92-93% removal, but both

decreasing to 15% ERR and increasing to 30% ERR reduced the removal efficiency.

A influent TSS concentration above 100 mg/l showed to provide an increased TSS removal efficiency

(figure 15).

0

100

200

300

400

500

600

700

Influen

t T

SS c

once

ntr

atio

n [

mg/l]

Date

25

Figure 14. Removal percentage of effluent suspended solids at different levels of effluent recycle rate.

Figure 15. Influent TSS concentration compared to TSS removal percentage.

70%

75%

80%

85%

90%

95%

100%

0

100

200

300

400

500

600

700

1.1

1.2

1.3

1.4

2.1

2.2

2.3

2.4

2.5

3.1

3.2

3.3

3.4

3.5

4.1

4.2

4.3

4.4

4.5

Rem

oval

per

centa

ge

Influen

t T

SS [

mg/l]

Experimental run ID

Influent TSS concentration TSS removal percentage

70%

75%

80%

85%

90%

95%

100%

15% 20% 25% 30% 35%

Rem

oval

per

centa

ge

Effluent recycle rate

2,5 m/h

4 m/h

5 m/h

6 m/h

26

4.1.3 Effluent suspended solids concentration.

Average effluent TSS concentration increases with increased surface load (Figure 16). Both surface loads

of 2.5 and 4 m/h had their average values beneath the limit of 10 mg/l TSS for every ERR tested. While

surface load 5 m/h had two ERR values beneath 10 mg/l, ERR 20 and 35%, and ERR 15, 25 and 30%

exceeding 10 mg/l TSS with only a few mg/l.

Only trend visible between each surface load for changes of ERR, where the concentration increases for

each step in ERR up until ERR 35%.

Surface load 6 m/h (Figure 16.D) has an average concentration reaching 52 mg/l with ERR 20%. All

four surface loads presented in Figure 16 had their lowest average value when their ERR were at highest,

which is 30 or 35%.

With a phosphorus dry weight percentage of 1.86% (table 10), an effluent TSS concentration of 16.13

mg/l would be allowed without exceeding any discharge regulations. Thus, the only surface load

exceeding legal regulations are 6 m/h (figure 16.D)

Designed ERR of 15% is presented low average values for each surface load, except in surface load 5

m/h (Figure 16.C). There it is presented as the median value of the five ERR tested.

Surface load of 6 m/h (Figure 16.D) exceeded the set limit value of 10 mg/l TSS for every ERR tested.

Figure 16. 24h average effluent suspended solids concentration with different effluent recycle rates (ERR) between 15-35%.

Subfigures; A: surface load 2.5 m/h, B: surface load 4 m/h, C: surface load 5 m/h, D: surface load 6 m/h.

(B) (A)

27

Figure 17. Effluent suspended solids concentration box plot with median and average values with different effluent recycle rates between 15-35%. Subfigures; A: surface load 2.5 m/h, B: surface load 4 m/h, C: surface load 5 m/h, D: surface load 6 m/h.

4.1.4 Effluent suspended solids concentration over time

All four surface loads showed that the TSS concentration varies over time (figure.18). Increases between

hour 12-20, except for the surface load of 6 m/h and ERR 20%, where the largest peak is between hour

4-10.

At a surface load of 2.5 m/h, ERR of 20% shows the largest concentration value during the whole

experiment (figure 18.A). ERR 30-35% do not show a concentration that increases over time as strong

as with 15-25% ERR, but instead keeping a more constant concentration (figure 18.A).

ERR 30% was the only ERR to exceed 10 mg/l TSS for several hours at 4 m/h surface load (figure

18.B). 25% ERR exceeds 10 mg/l effluent TSS in hour 18 and was close to the limit 10 mg/l during

several hours.

Every ERR tested during surface load 5 m/h exceeded 15 mg/l TSS during different periods. 35% ERR

provided the lowest average TSS concentration and 20% ERR as second lowest. Both 25% and 30%

ERR had concentrations <25 mg/l TSS.

No ERR tested for surface load 6 m/h had a TSS concentration below 10 mg/l, with 20% ERR

reaching 88.6 mg/l during hour 6. 25% ERR showed strongly fluctuating values between 17-36 mg/l

TSS during the day, but the average was 23 mg/l TSS.

Effluent recycle rate 15% and 30% showed similar concentrations, not having a value higher than 25

mg/l TSS and average values at 15 mg/l TSS.

At surface load 5 m/h and an ERR of 25% (figure 18.C) a sharp increase of effluent TSS concentration

occurred at hour 15. This was caused by an operation technician at Margretelund WWTP, who, by habit,

lowered the ERR in the morning because it was visible turbulent flow in the DAF unit.

28

Figure 18. Effluent suspended solids concentration with different effluent recycle rate (ERR) between 15-35%. Subfigures; A:

surface load 2.5 m/h, B: surface load 4 m/h, C: surface load 5 m/h, D: surface load 6 m/h.

4.2.5 Temperature and pressure during experimental run

The temperature during the experimental run were between 7.5-9.3°C with average temperature 8.3°C.

Figure 19. Temperature during the experimental run see section 3.4 for experimental input for each trial.

6

6.5

7

7.5

8

8.5

9

9.5

1.1 1.2 1.3 1.4 2.1 2.2 2.3 2.4 2.5 3.1 3.2 3.3 3.4 3.5 4.1 4.2 4.3 4.4 4.5

Tem

per

ature

[ ℃

]

Experimental trial id

Temperature

29

The pressure was fluctuating between 5.8-6.0 bar, with a dip down to 5.6 bar during trial 2.3.

Figure 20. Pressure in pressurized tank during the experimental run. See section 3.4 for experimental input for each trial.

4.2 Results from historical data analysis The most efficient historical surface load interval was between 1-2 m/h with only 17% of the samples

(10 out of 57), exceeding the TSS limit of 10 mg/l (Table 8). The higher surface loads analysed proved

to have a much greater percentage of samples exceeding the limitation.

The most noticeable increase was at a surface load of 2-3 m/h, where 55% of the total 312 samples

exceeded the limit (and with this surface load interval containing the most samples evaluated). 61% of the

samples (105 out of 172), exceeded the TSS limit of 10 mg/l for surface load 2-3 m/h.

Further increased surface load presented a lesser number of samples within each step, but also an

increasing percentage of samples exceeding 10 mg/l until reaching 100% at interval 5-6 m/h.

Table 8. Summarized historical data of the percentages of samples exceeding 10 mg/l TSS, and percentage of ERR for samples both exceeding and not exceeding 10 mg/l TSS in DAF unit in Margretelund WWTP. Data extracted from aCurve for period Jan 2015- Jan 2021

5.4

5.5

5.6

5.7

5.8

5.9

6

6.1

1.1 1.2 1.3 1.4 2.1 2.2 2.3 2.4 2.5 3.1 3.2 3.3 3.4 3.5 4.1 4.2 4.3 4.4 4.5

pre

ssure

[bar

]

Experimental trial id

Pressure

Surface

load Flow rate

Number

of samples

Percentage

of total

Number

of samples

> 10 mg/l

Percentage

of samples

> 10 mg/l

Average

effluent

recycle

percentage

Average effluent recycle

percentage with TSS

concentration

> 10 mg/l

Average effluent recycle

percentage with TSS

concentration

< 10 mg/l

m1h-1 m3h-1 No. % No. % % % %

1 < > 7 50 - 350 312 100 183 58,65% 18.15 15.56 21.81

1 < > 2 50 - 100 57 18.24 10 17.54% 25.94 23.93 26.37

2 < > 3 100 - 150 172 55.13 105 61.05% 18.19 16.89 20.23

3 < > 4 150 - 200 46 14.74 34 73.91% 13.49 13.1 14.59

4 < > 5 200 - 250 23 7.37 20 86.96% 12.83 12.54 14.78

5 < > 6 250 - 300 7 2.24 7 100.00% 10.33 10.33 0

6 < > 7 300 - 350 7 2.24 7 100.00% 9.49 9.49 0

> 7 350 0 0 0 0.00% 0 0 0

30

4.2.1 Regression analysis

Effluent TSS concentration have a moderate correlation with both the ERR and with the influent surface

load (r = 0.58 and r = 0.48 respectively, table 9). The negative Pearson correlation coefficient for the

effluent recycle percentage mean that an increasing effluent recycle percentage yields a lower TSS

concentration in the effluent. On the other hand, decreasing the surface load, decreased the effluent TSS

concentration as result.

The absolute value for both ERR and the surface loads correlation coefficient are higher than the

critical value, which indicates a significant linear correlation, and the p-values are >> 0,05 for a two tailed

test meaning that there is a correlation with more than 95% confidence.

Table 9. Pearson correlations

X Y Pearson correlations

coefficient Critical

values T-statistic P-value Confidence R2

Effluent recycle % Effluent

TSS conc. |-0.584| > 0.195 12.68 1.528E-22 95% 0.328

Surface load Effluent

TSS conc. |0.484| > 0.195 9.74 3.599E-16 95% 0.216

G/S Removal

TSS % |-0.584| > 0.195 12.66 1.692E-22 95% 0.320

A moderate linear correlation can be seen in all three subfigures in figure 21, with R2 values between

0.2-0.32 presented both in the subfigures but also in table 9.

Figure 21. Three different linear correlations showing relationship between factors chosen from Pearson correlation

(A) (B)

(C)

(A) (B)

(C)

31

4.2.2 Gas-to-solid ratio

Gas-to-solid ratio (G/S) was varying between 0.008 to 0.131 with no distinct seasonal variation, with an

average value at 0.047 (figure 22.

Figure 22. Gas-to-solid ratio calculated with equation 5 for historical period January 2015-January 2021, DAF unit Margretelunds

WWTP

4.2.3 Amount of phosphorus in suspended solids

The percentage of phosphorus in dry weight effluent TSS has an average percentage of 1.74% over the

period of January 2015-January 2021 (table 10).

Yearly average phosphorus concentration never exceeded the discharge limit of 0.3 mg/l (table 10),

except 2021, which was only sampled during 4 out of 52 weeks of the year.

Table 10. Historical yearly average values for TSS, phosphorus concentration and percentage of phosphorus in effluent TSS

Year Average effluent

TSS concentration

Average phosphorus

concentration

Average dry weight

percentage of phosphorus in

effluent TSS

yr mg1l-1 mg1l-1 %

2015 19.89 0.24 1.14%

2016 19.76 0.27 1.35%

2017 9.00 0.19 2.19%

2018 10.11 0.19 2.18%

2019 9.92 0.25 2.73%

2020 11.97 0.22 2.23%

2021 19.75 0.42 2.15%

2015-2021 14.34 0.25 1.74%

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

2014 2015 2016 2016 2017 2017 2018 2018 2019 2019 2020 2021

G/S

32

33

5. Discussion

5.1 Effects of ERR and loading rate on the flotation process

5.1.1 Effect of ERR and loading rate on effluent suspended solids and phosphorus

concentrations.

ERR and surface load were the two factors to be tested during this experiment, to test if a decreased

average effluent TSS concentration could be achieved, and as a result reducing the phosphorus

concentration in effluent water organically bound in the TSS.

Phosphorus concentration

Wastewater TSS are a heterogenous mass with a broad size distribution between 1-100 µm

supracolloidal particles and >100 µm settleable particles (Azema et.al, 2002). Margretelund WWTP’s

measured TSS could have a varying content of TSS for each day depending on the influent wastewater.

Margretelund WWTP´s wastewater TSS is product from a mixture of sludges both produced in the

WWTP and transported with influent wastewater. The mixture consist of chemically flocculated sludge,

biological sludge, or inorganic solids, with every part containing a varying amount of phosphorus

(Lidström, 2013).

Huacheng et. al (2011) showed that a phosphorus concentration in biological sewage sludge had a

varying dry weight percentage of 0.97-1.74% phosphorus, while in section 2.5, it is presented from theory

that 2% of prokaryotic bacteria’s dry weight are phosphorus (Metcalf & Eddy, 2013). Average percentage

of phosphorus in Margretelunds WWTP historic data corresponded to 1.74% (table 10) and are in the

middle of both literature values. Although, phosphorus in sewage sludge are non-volatile (Huacheng

et.al, 2012), Margretelund WWTP’s total phosphorus concentration cannot be assumed to be in a solid

species only. Phosphoric acid added into the wastewater (subsection 3.1.2), and old organic sludge starting

a hydrolysis process, could be two sources of dissolved phosphorus in the measured total concentration

(Särner, 2007).

To base Margretelunds phosphorus concentration on TSS concentration was assessed to not provide a

correct value, it could be both overestimated and underestimated depending on the sewage sludge

composition. Too many different sources of phosphorus, both from TSS changes and external sources,

are not considered during the analysis. Although, it can provide a sufficiently accurate estimate

phosphorus concentration from basing it on concentration of effluent TSS.

Effluent suspended solids concentration

The experiment included both the design parameters (4 m/h and 10-15% ERR), increasing and

decreasing different predetermined ERR and surface loads (table 6). It was noticed that a general decrease

of surface load would provide a lower average effluent TSS concentration, as seen in figure 18. Thus, the

hypothesis (based on historical data analysis and the regression analysis) that by increasing the ERR, a

reduction of effluent TSS concentration would follow, were proven wrong.

Haarhoff & van Vuuren (1995) mentioned from a comparison study over clarification DAF process (as

used in Margretelund) between South Africa, Finland, Britain, and the Netherlands, that clarification

plants did generally perform well for all sites, while they also follows a narrower band of design parameters

then DAF plants used for sludge thickening. Zone geometry of the reaction zone, where micro-bubbles

and flocs collide, was considered most crucial for overall success with clarification. The ERR used in

studied sites ranged from 6% to 30% (Haarhoff & van Vuuren, 1995), with Finland having the same

climate as Sweden and presenting an ERR of 30% in the study, the range of ERR tested in this thesis

are assessed reasonable although the hypothesis was proven wrong.

All four surface loads showed that with increasing ERR, a higher average effluent TSS concentration

was achieved up until reaching 35% ERR. At 35% ERR, a decrease of average TSS concentration

compared to ERR of 15% with the same surface load. This trend with high and low ERR settings was

34

noticed for every surface load tested. Results in figure 18.A shows that the settings with lowest average

effluent TSS concentration are in surface load 2.5 m/h and an ERR of 35%. Followed by the design

setting of 15% ERR for all (2.5, 4, 6 m/h) but surface load 5 m/h (Figure 18.C).

That both the lowest and the highest ERR setting (15 and 35%) proved to be the two better ERR

options does not follow the hypothesis based on theory, that increased ERR would reduce the effluent

TSS concentration. Although, influent TSS concentration and influent TSS constituents have varied

between every experimental run (figure 13), and with changing influent conditions, variations in effluent

TSS concentration are a possible outcome. Most representable experimental result would be gained from

keeping influent TSS conditions constant throughout the entirety of the experiment. However, a WWTP

will most likely never achieve a constant influent TSS, so more experimental runs are needed for every

setting tested during this experiment to gain more representable average value in results.

The surface load of 6 m/h (figure 18.D) was not tested with ERR 35% due to limitations within the

recycling system. An ERR of 35% would have resulted in a flow rate of 105 m3/h, approximately 80%

peak efficiency of what the two recycling pumps could achieve. Although, a too high recycled flow rate

can result in turbulent flow in the DAF unit that may decrease the ability to float flocs and swirl up settled

sediment from the bottom of the basin. So, to increase the ERR to 35% was deemed providing

unnecessary risks for gaining increased effluent TSS concentration.

5.1.2 Effect of ERR and surface load on removal efficiency of TSS

Removal efficiency for suspended solids proved to be functioning well for the designed surface load and

below (2.5 and 4 m/h). With all four surface loads tested presenting TSS removal of >90% for one or

more ERR setting (figure 14). While changes in ERR showed no obvious effect on removal efficiency

(figure 14), ERR can theoretically decrease the removal efficiency due of its negative correlation with

G/S ratio and removal efficiency (equation 5). However, experimental results showed that by matching

the fluctuating influent TSS concentration with removal efficiency (figure 15) it was proven that the

experimental runs with higher influent TSS concentration provide a higher TSS removal efficiency for

every surface load tested (2.5, 4, 5 and 6 m/h), without any influence of ERR.

Although, the scope of this thesis does not include any research on the efficiency of the flocculation

basins and the size variation of flocs transported into the DAF process. Odegaard (1995) presented that a

difference between utilizing a chemical flocculation process prior to a sedimentation and DAF exists.

Odegaard (1995) showed that a DAF process benefits from a more turbulent flowrate created by an

intense stirring in the flocculation basins, to achieve a smaller mean floc size. However, a residence time

up to 25-30 minutes may be necessary, while the rotational speed of stirrers should be kept twice as high

compared to flocculation for a conventional sedimentation process.

R. Arnold et.al (1995) also presented from a pilot study where a full-scale DAF clarifier pilot used with

dimensioned parameters close to that of Margretelund WWTP DAF process. R. Arnold et.al (1995) used

a designed surface load of 4.9 m/h and an ERR of 20% reached an TSS removal of 82-84% from a

combination of coagulant and flocculant chemicals in the flocculation. Margretelund WWTP’s TSS

removal proved to be more efficient then showed in this study. However, coagulants were added to

increase the dry solids percentage in sludge before dewatering and it could be of interest to study the

effect from adding coagulants into Margretelund WWTP’s wastewater during high flow rates.

With the observations that TSS removal efficiency is seemingly mostly affected by the influent TSS

concentration (figure 16), the studies presented by Odegaard (1995) and R. Arnold et. al (1995) provides

incitement to conduct further research onto the flocculation process positioned prior to the DAF unit at

Margretelund WWTP (figure 4).

Surface load 6 m/h had, during experimental run 1.1 and 1.2 (ERR 20 and 25%), an unusual large

amount of influent TSS (figure 14). Following the correlation found with figure 15, a presumably good

removal percentage could be achieved although the effluent TSS concentration were relatively high.

35

5.1.3 Temperature and pressure during experiment

Saturation of oxygen into water is dependent on temperature, pressure, and organic content (Ingri, 2011).

The recycled water comes from effluent water and thus treated from organic matter, so oxygens saturation

is determined from pressure and temperature for the DAF process at Margretelund WWTP.

The pressure in the pressurization tank was set to remain at constant 6 bar during the experiment, but

pressure variations occurred (Figure 20). Pressure ranging from 5.6-6.0 bar were observed over the

experiment. The fluctuating pressure was believed to be caused by changes to the constant water volume

normally present in the pressurized tank from increases of ERR. A lower water volume would leave

more volume to be filled with gas, and the system would then be dependent on the efficiency of the air

compressor. Experimental run 2.3 (figure 20) was affected by a planned power outage during a 4-hour

period, causing the air compressors and recycling pumps to be turned off. Pressurized flow continued

during the power outage, lowering the water volume inside the pressurized tank, and thus reducing

pressure.

Temperature ranged between 7.5-9.3°C, changing with the outdoors weather. Snowmelt had just

started in the start of the experiment and continued throughout the full period, with heaviest period

during first two weeks. Warmer temperature allows less air to be saturated into water and would require

a larger mass flow rate of gas to achieve the same volumetric saturated amount as in colder water (table

1).

Saturation of oxygen are important for G/S ratio (equation 5), though no correlation between

temperature and pressure changes were noticed in the experimental results (Figure 19-20). Temperature

difference of 1.8°C is believed to do low difference for the G/S ratio compared to changes of pressure or

ERR.

5.2 Analysis of historical data Efficiency of Margretelund WWTPs DAF unit (table 7), evaluated for the time-period of January 2015

– January 2021, shows that the DAF unit is underachieving in removing TSS for 183 out of 312 (58%)

historical samples. However, with the designed surface load of 4 m/h, the DAF unit should have the

potential to remain under 10 mg/l for 275 out of 312 (88%) of the historical samples analysed.

In subsection 2.1.3, it is presented that the DAF unit is designed for a surface load of 4 m/h, with an

ERR between 10-15%. However, the DAF unit cannot function within its designed parameters and in

need of some operative changes (table 8). Most noticeable factor are the different average ERR for each

surface load, and difference between samples within the limit and for those exceeding the limit value

(table 8). It is presented that the ERR where samples exceed the limit value are a few percentages lower

than for those samples within the limit value, and that the average ERR are decreasing by every step of

surface load. Increasing the ERR above 20%, from the designed 10-15%, the DAF unit’s efficiency would

theoretically improve.

Based on the changes of ERR in table 7 for the WTTPs DAF unit, historically occurring reduction

of ERR provided an increased number of sample percentage exceeding the limit value. It was believed

to be caused by a few different things but assumed mainly on two. The first cause being the pressurized

flow rate was operated on a constant value based on design parameters (4 m/h with ERR 15%),

corresponding to a recycle flow rate of 40 m3/h, and was uncapable to change unless manually done.

Another cause could be that the effluent recycle system could be under dimensioned, disallowing it to

keep a predetermined ERR value following increases of surface load.

5.2.1 Regression analysis

In table 9, both the ERR and the G/S show a negative Pearson coefficient r based on historical data,

with ERR showing negative correlation with effluent TSS concentration and the G/S ratio towards

36

effluent TSS removal efficiency. Surface load was positively correlated with effluent TSS concentration.

From the correlations it can be concluded that, if the ERR is increased or the surface load decreased, a

reduction of effluent TSS concentration would be theoretically achieved with a confidence of 95%.

While the experiment investigated ERR and surface load as the factors to test, G/S ratio proved a

moderate negative correlation from historical data worth discussing. Equation 5 used to calculate the G/S

ratio, is based on both pressure and the ERR to change its mass flow rate of gas, and the mass flow rate

of solids are based on the influent TSS concentration.

Although G/S ratio negatively correlates with effluent TSS removal efficiency and ERR negatively

correlates with effluent TSS concentration, based on equation 5. Theoretically increase of ERR would

provide a decrease of effluent TSS concentration, but also a decrease to TSS removal efficiency from the

increased G/S ratio. Correlation between G/S ratio and effluent TSS removal efficiency was proven

incorrect by the experimental run (Figure 15), where the only factor found to affect removal efficiency

during the experiment, was the concentration of influent TSS (subsection 5.1.2).

5.2.2 Gas – to – solid ratio

Margretelunds historical G/S ratio has its average value within the theoretical interval of 0.02-0.06 (figure

1), but the value is heavily fluctuating from close to 0.008 to 0.13. No seasonal variations of the historical

G/S ratio were observed (Figure 22). Variations of historical G/S ratio was caused by parameters that

cannot be controlled, such as the concentration of influent TSS or the total flow rate through the WWTP.

Margretelund WWTP has historically had a G/S ratio exceeding the theoretical intervals upper limit of

0.06, with no presumed benefit on the flotation process according to Metcalf & Eddy (2013). However,

the high G/S ratio utilized may have been an unnecessary energy demand with an increased cost for

Roslagsvatten AB as result. This might have been avoided by monitoring influent TSS concentration in

the flocculation basin and changing either pressure or ERR to aim for a G/S ratio within the interval

0.02-0.06.

The G/S ratio was not one of the chosen factors of this project to be controlled during experimental

runs, see section 3.4, and further investigation will be necessary to control the effect G/S ratio may have

on Margretelund WWTP’s DAF unit.

5.3 Experimental challenges and sources of errors • Influent sampling worked as expected, although more influent samples for each day (instead of

only one) would have provided more representative information about daily variations. These

extra samples could not be taken and analysed due to Covid-19 interfering, since the author was

not allowed inside Roslagsvatten’s laboratory to do the analysis himself. Instead, the employed

lab workers had to do the analyses, but lack of time resulted in only one sample a day was

accepted.

• Effluent water samples for TSS analysis was not collected and analysed in Roslagsvatten’s

laboratory due to Covid-19. As mentioned in subsection 3.3.2, the effluent concentration for

suspended solids was collected with a TSS sc Sensor from HACH (figure 8) providing sampling

data to be extracted from aCurve. Monitored data was useful to daily variations of effluent TSS,

however, the effluent TSS data could not be used to assume daily variations of influent TSS.

Validation of extracted data against a TSS analysis sample would have given more reliable results.

For instance, information would have been achieved on if the sensor were drifting in value or

not, and so, in need of calibration.

37

• Sludge scrapers in the pre-sedimentation malfunctioned at the beginning of the experiment (1.1

to 1.2) due to high accumulated volumes of sediment, while having unusually high flow rate due

to the start of snowmelt. The sediment accumulation was caused by two pumps used to pump

settled sludge from the pre-sedimentation basin was old and worn down, and thus inefficient to

handle the amount of sludge settling. Without a fully functioning pre-sedimentation, the influent

TSS concentration where six times larger the first day than during the rest of the experiment

(figure 16). The experiment got postponed a few days until the issues with the pre-sedimentation

where solved. The first two experimental runs 1.1 and 1.2 (table 8) was not redone due to time

limitations even though they present an influent TSS concentration six to three times higher than

the average TSS for subsequent trials (Figure 16). The increased TSS concentration may present

a nonrepresentative value for the experiment, especially for trial 1.1, with unusually large effluent

TSS compared to every other experimental run.

• The air compressor for the pressurized tank malfunctioned during experimental run 1.1 and 2.1,

resulting in reduced pressure in the pressurized tank. The change of pressure decreased the G/S

ratio (equation 5), resulting in reduced mass flow rate of gas and lowering the production of

micro-bubbles. Although, no noticeable affect was seen from this on effluent TSS concentration

(Figure 16.B and Figure 16.D). Experimental run 1.1 was affected by abnormally high influent

TSS concentration.

5.4 Future investigations This experiment has investigated two out of several identified factors that theoretically influences the

reduction of TSS in the DAF unit at Margretelund WWTP. Limitations had to be done due to a time

limit, and chosen factors was assumed to present the biggest impact based on current situational analysis,

section 4.2 and regression analysis, subsection 3.5.1 and subsection 4.2.1.

The factors that were not tested but still might have proved useful are presented below, with a small

description on what issues these factors might prevent or solve.

The surface scrapers are run in an intermittent mode (subsection 2.1.3) and based on observations by

the author, a thick floated sludge blanket was built-up during the experiment, covering the entire DAF

basin. One specific observation was, that increased surface load, both from influent flow rate and

pressurized flow rate, rapidly generated a floated sludge blanket. During the periods with thick sludge

blanket covering the basin, larger flocs of suspended solids were observed to follow the treated effluent

water discharge. A possible reason could be that the surface sludge scrapers being run too seldom.

Decreasing downtime between runs or increase the runtime of the scrapers may mitigate the build-up of

floated sludge.

Changes in flocculation and coagulation of influent TSS were not tested in the experiment as done in

the study by Odegaard, (1995). It is of interest to study if changes in dosage of flocculating chemicals or

rotational speed of the stirrers to increase detention time in the flocculating basin, per cubic meter

wastewater. If these changes could provide variations on average size of built-up flocs. Different average

sizes of flocs entering the DAF unit would have different density and capabilities to be floated. Increased

size would theoretically increase collision efficiency (subsection 1.3.4) while reduced size would have a

density close to or beneath that of water and may be floated with less effort.

Control of G/S ratio from change of pressure in the pressurization tank or temperature in recycled

water, while leaving the ERR constant, would be interesting to study how different G/S ratios would

affect the DAF unit (equation 5). Testing either of these factors would require a better control of influent

TSS concentration and a system that would react to changes in TSS to decrease or increase the pressure

38

from a set ratio. Temperature changes could provide changes to saturation of air into a liquid (table 1),

with increasing water temperature resulting in decreased saturation of oxygen.

The nozzle for pressurized water at pressure release point in the DAF unit was not controlled if they

are correctly placed or designed during the experiment. The DAF process was built in 1999 and since

then, there may exist more suitable nozzles for Margretelund WWTP. Due to technological advancement

between 1999 and 2021.

39

40

6. Conclusions Historical data analysed provided information that the DAF unit has been inefficient for all surface loads

but the lowest (1-2 m/h) between January 2015 – January 2021. With the historical data and regression

analysis, effluent recycle rate (ERR) was found to be one important factor that showed a negative

correlation towards effluent TSS concentration. The ERR showed a declining percentage for every

increasing step of surface load analysed, while the percentage of samples exceeding the limit of 10 mg/l

increased with increased surface load.

The cause for declining ERR percentage is assumed to be faults in design of operation of the DAF

unit paired with under-dimensioned recycling pumps. The pressurized flow rate has historically been

operated within a flow rate interval of 25-30 m3/h, seemingly operated with constant pressurized flow

rate and not from keeping a constant ERR. Although, experimental results showed that the two recycling

pumps for treated effluent water are under-dimensioned if the pressurized flowrate exceeds 43 m3/h. The

flow rate of 43 m3/h corresponds to an ERR of 21.5% with surface load 4 m/h, 10 percent units above

the DAF units design parameters, or 14.3 % ERR for surface load 6 m/h.

The experimental results show that increasing surface load provided a steady average increase of

effluent TSS concentration. Despite of this, Margretelund WWTP’s DAF unit showed to efficiently

handle a surface load up to 5 m/h for all ERR values tested. 3 out of 5 ERR settings resulted in effluent

TSS concentration beneath 10 mg/l and the other two exceeded with very little margin. While

Margretelund WWTP is designed for a surface load of 4 m/h, this provided information that it has

potential to be operated at flow rates above its design.

Effluent recycle rate (ERR) proved to show no noticeable impact in effluent TSS concentration based

on the average experimental results and there was different trends in how effluent TSS concentration

varied across different ERR. The highest ERR setting (30 or 35% depending on surface load) provided

the lowest average effluent TSS concentration, with the lowest setting (15%) having the second lowest

average for 3 out of 4 surface loads tested (2.5, 4 and 6 m/h). Based on experimental results, the potential

of ERR to decrease effluent TSS concentration probably exists since the highest setting provided the

lowest average effluent TSS concentration. However, due to variations in influent TSS, every

experimental run had different conditions and one test for each ERR is assessed too be insufficient.

Further testing with ERR is needed to gain a more representative result.

Temperature and pressure are theoretically important for the gas-to-solid ratio. However, the

differences recorded during the experiment provided no noticeable connection to either increased or

decreased effluent TSS removal efficiency.

Proposals to Roslagsvatten on how to gain an efficient DAF process at Margretelund WWTP and lower

the effluent TSS emissions with present design of operation:

• Roslagsvatten should aim to remain a max capacity surface load of 5 m/h for an average effluent

TSS concentration of 10 mg/l.

• Roslagsvatten should remain within an ERR of minimum 15% for all surface loads.

• Roslagsvatten should during high surface loads (> 5 m/h), aim to reduce influent TSS

concentration into the DAF process with the help of chemical additives or other means.

• Roslagsvatten should during warm temperatures, increase either pressure or ERR to mitigate the

reduced gas-to-solid ratio generated from the reduced saturation of oxygen into water.

Proposed process improvements to Roslagsvatten to increase the efficiency of Margretelund WWTP’s

DAF process are:

• To change the two pumps recycling treated effluent water to the pressurized tank. Their capacity

are limited to 130 m3/h, and if Roslagsvatten are to keep a minimum ERR of 15% for all surface

loads, a flow rate capacity of 135 m3/h is required for ERR of 15% with surface load 6 m/h.

41

• Examine the pressure release nozzles, if the ones installed and used during the experiment are the

nozzles producing the highest efficiency for micro-bubble production from pressure release.

• Increase the surface area of the DAF basin to reduce the average surface load.

The factors chosen not to be studied during this study but could be of interest for Roslagsvatten to test

at Margretelund WWTP, before any investigations for proposed improvements or reconstruction of the

DAF process are conducted:

• Changes in run- or downtime of surface sludge scrapers. To study if increased removal of the

floated sludge blanket paired with both increased ERR setting, and with designed value of 15%,

could provide a reduced effluent TSS concentration with increasing surface loads.

• Change of dosage volume of flocculating chemicals per cubic meter wastewater. This would

provide information about how different dosages could generate a size and density variation of

built-up flocs and how the ability to float changes.

• Additive of coagulating chemicals into the DAF basin. To study if the ability of flocs to be floated

changes if the flocs become more compact (dense), especially during periods with high influent

TSS concentration.

• Varying the detention time in the flocculating basins from changing the stirrer’s rotational speed.

Could generate a size and density variation of flocs, whit increased detention time resulting in

larger average size of the TSS flocs.

• Different gas-to-solid ratio should be tested, by changing either ERR or pressure with changes

in influent TSS concentration, to evaluate gas-to-solid ratios and how they would affect the DAF

process.

42

43

7. References

Arnold, S. R., Grubb, T. P., & Harvey, P. J. (1995). Recent applications of dissolved air flotation pilot

studies and full scale design. Water Science and Technology, 31(3-4), 327-340.

Azema, N., Pouet, M.-F., Berho, C., & Thomas, O. (2002). Wastewater suspended solids study by optical

methods. Colloids and Surfaces, 131-140.

Berman, J. J. (2016). Chapter 4 - Understanding Your Data . In J. J.Berman, Data simplification, Taming

Information with Open Source Tools (pp. 135-187). Baltimore : Elsevier Inc.

De Rijk, S., Van Der Graaf, H., & Den Blanken, J. (1993). Bubble size in flotation thickening. Water

research, 28(2), 465-473. doi:https://doi.org/10.1016/0043-1354(94)90284-4

Edzwald, J. (2010). Dissolved air flotation and me. Water Research, 44(7), 2077-2106.

doi:https://doi.org/10.1016/j.watres.2009.12.040

Eniro. (2014). Aerial photograph . Eniro Sverige AB.

Féris L, A., Gallina, S., Rodrigues, R., & Rubio, J. (2000). OPTIMIZING DISSOLVED AIR

FLOTATION DESIGN SYSTEM. Brazilian Journal of Chemical Engineering, 17(4-7).

doi:10.1590/S0104-663220000004000019

Haarhoff, J., & van Vuuren, L. R. (1995). DESIGN PARAMETERS FOR DISSOLVED AIR

FLOTATION IN SOUTH AFRICA. Water science technology, 31(3-4), 203-212.

Han, M., Kim, T., & Kim, J. (2007). Effects of floc and bubble size on the efficiency of the dissolved air

flotation (DAF) process. Water Science and Technology, 56(10), 109-115. Retrieved 2 8, 2021,

from https://ncbi.nlm.nih.gov/pubmed/18048983

Huacheng, X., Hua, Z., Liming, S., & Pinjing, H. (2012). Fraction distributions of phosphorus in sewage

sludge and sludge ash . Waste Biomass Valor, 355-361. doi:10.1007/s12649-011-9103-5

Ingri, J. (2011). FRÅN BERG TILL HAV - en introduktion till miljögeokemi (1:2 ed.). Luleå:

Studentlitteratur AB.

Khiadani, M., Kolivand, R., Ahooghalandari, M., & Mohajer, M. (2013). Removal of turbidity from

water by dissolved air flotation and conventional sedimentations systems using poly aluminium

chloride as coagulant. Desalination and Water Treatment, 52, 985-989.

doi:10.1080/19443994.2013.826339

Lidström, V. (2013). Vårt Vatten (Vol. 2). Lund: Svenskt vatten.

Metcalf, & Eddy. (2013). Wastewater Engineering: Treatment and Resource Recovery. Boston:

McGraw-Hill Education.

Naturvårdsverket. (2019). Vägledning om Naturvårdsverkets förskrifter (NFS 2016:6) om rening och

kontroll av utsläpp av avloppsvatten från tätbebybbelse. Stockholm: Naturvårdsverket.

Odegaard, H. (1995). Optimization of flocculation/flotation in chemical wastewater treatment. Water

Science and Technology, 31(3-4), 73-82.

44

PURAC AB. (2000). Drift och skötselinstruktioner, Margretelund ARV. 6-29.

Roslagsvatten. (2015). Egenkontroll Margretelund reningsverk. Åkersberga: Roslagsvatten.

Siegle, D. (2015, 2 24). Level of significance for two-tailed test. Neag School of Education - University

of Conneticut.

Svenskt Vatten, A. (2007). Avloppsteknik 2, Reningsprocessen. Stockholm: Svenskt vatten, AB.

Särner, E. (2007). Biologisk fosforavskiljning med hydrolys av returslammet och utan anaerob volym i

huvudströmmen. Torsås: Svenskt Vatten AB.

Vallero, D. (2014). Fundamentals of Air Pollution (5 ed.). Duke university, Durham, USA: Academic

Press. doi:https://doi.org/10.1016/C2012-0-01172-6

Wang, L., Hung, Y.-T., & Shammas, N. (2005). Physicochemical Treatment Processes. Lenox Institute

of Water Technology, Lenox, MA: Humana Press.

45

I

8. Appendices

Appendix 1. Enlarged flow scheme of Margretelund wastewater treatment

plant

Figure A1.1. Enlarged flow

scheme of Margretelund

WWTP, Swedish

descriptions.

II

Appendix 2. Current situational graphs

Figure A2.1. Effluent TSS concentration at 1-2 m/h surface load for historical data, January 2015 – January 2021

Figure A2.2. Effluent TSS concentration at 2-3 m/h surface load for historical data, January 2015 – January 2021

0

5

10

15

20

25

1.4

8

1.5

1

1.6

5

1.6

8

1.6

8

1.7

4

1.7

5

1.8

2

1.8

3

1.8

4

1.8

7

1.8

7

1.8

7

1.9

0

1.9

2

1.9

3

1.9

4

1.9

5

1.9

6

1.9

6

1.9

7

1.9

8

1.9

8

1.9

9

1.9

9m

g/l

m/h

0

5

10

15

20

25

30

2.0

0

2.0

1

2.0

3

2.0

6

2.0

9

2.1

1

2.1

4

2.1

6

2.2

0

2.2

1

2.2

4

2.2

8

2.2

9

2.3

2

2.3

5

2.4

0

2.4

3

2.4

5

2.4

7

2.5

0

2.5

1

2.5

5

2.5

6

2.5

7

2.6

0

2.6

5

2.6

8

2.7

3

2.8

0

2.8

4

2.8

8

2.9

3

2.9

8m

g/l

m/h

III

Figure A2.3. Effluent TSS concentration at 3-4 m/h surface load for historical data, January 2015 – January 2021

Figure A2.4. Effluent TSS concentration at 4-5 m/h surface load for historical data, January 2015 – January 2021

0

5

10

15

20

25

30

35

40

3.0

0

3.0

2

3.0

4

3.0

6

3.0

6

3.0

7

3.1

1

3.2

3

3.2

5

3.2

9

3.2

9

3.3

8

3.4

9

3.5

3

3.6

2

3.6

4

3.7

0

3.7

9

3.8

9

3.9

2m

g/l

m/h

0

10

20

30

40

50

60

70

80

90

100

4.0

0

4.0

0

4.0

3

4.0

4

4.0

8

4.1

2

4.1

3

4.1

7

4.2

1

4.2

3

4.2

5

4.2

9

4.3

1

4.4

7

4.5

0

4.5

6

4.5

7

4.5

9

4.6

2

4.6

5

4.6

7

4.8

2

4.8

4m

g/l

m/h

IV

Figure A2.5. Effluent TSS concentration at 5-6 m/h surface load for historical data, January 2015 – January 2021

Figure A2.6. Effluent TSS concentration at 6-7 m/h surface load for historical data, January 2015 – January 2021

0

5

10

15

20

25

30

35

40

45

50

5.0

8

5.1

4

5.2

9

5.3

3

5.4

9

5.6

1

5.8

5

mg/

l

m/h

0

5

10

15

20

25

30

35

40

45

50

6.0

1

6.0

3

6.0

4

6.0

6

6.2

7

6.9

6

mg/l

m/h

V

Figure A2.7. TSS removal effciency in different effluent recycle percentages with surface loads 1-6 m/h, January 2015 - January

2021.

Figure A2.8. TSS removal effciency in effluent recycle percentages of 5-30% with surface loads 1-6 m/h, January 2015 - January

2021.

70.00%

75.00%

80.00%

85.00%

90.00%

95.00%

100.00%

105.00%

110.00%

0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% 35.00%

TSS

rem

ova

l

ERR

70.00%

75.00%

80.00%

85.00%

90.00%

95.00%

100.00%

105.00%

110.00%

4.00% 9.00% 14.00% 19.00% 24.00% 29.00%

TSS

rem

ova

l

ERR

VI

Figure A2.9. TSS removal effciency in effluent recycle percentages of 15-30% with surface loads 1-6 m/h, January 2015 - January

2021.

Figure A2.10. TSS removal effciency in historical samples with effluent recycle percentages of 20-35% with surface loads 1-6

m/h, January 2015 - January 2021.

70.00%

75.00%

80.00%

85.00%

90.00%

95.00%

100.00%

105.00%

110.00%

14.00% 16.00% 18.00% 20.00% 22.00% 24.00% 26.00% 28.00% 30.00%

TSS

rem

ova

l

ERR

70.00%

75.00%

80.00%

85.00%

90.00%

95.00%

100.00%

105.00%

110.00%

19.00% 21.00% 23.00% 25.00% 27.00% 29.00% 31.00% 33.00% 35.00%

TSS

rem

ova

l

ERR

VII

Appendix 3. Dates, times and influent TSS value for experiment Table A3.1. Dates, times and influent TSS value for experiment

Date start Date stop Time start Time stop Infl. Conc

1.1 2021-02-23 2021-02-24 13,00 13,00 695,38

1.2 2021-02-24 2021-02-25 13,00 13,00 317,5

1.3 2021-02-27 2021-02-28 13,00 13,00 137,25

1.4 2021-02-28 2021-03-01 13,00 13,00 88,1

2.1 2021-03-01 2021-03-02 13,00 13,00 106,06

2.2 2021-03-02 2021-03-03 13,00 13,00 124,75

2.3 2021-03-03 2021-03-04 10,00 10,00 121,31

2.4 2021-03-05 2021-03-06 12,00 12,00 138,24

2.5 2021-03-06 2021-03-07 13,00 12,00 146,3

3.1 2021-03-07 2021-03-08 12,00 10,00 190,48

3.2 2021-03-08 2021-03-09 13,00 8,00 162,42

3.3 2021-03-09 2021-03-10 10,00 9,00 196

3.4 2021-03-10 2021-03-11 10,00 10,00 198,33

3.5 2021-03-11 2021-03-12 10,00 8,00 164,38

4.1 2021-03-12 2021-03-13 10 12 56,58

4.2 2021-03-13 2021-03-14 12 11 59,85

4.3 2021-03-14 2021-03-15 11 11 61,67

4.4 2021-03-15 2021-03-16 10 10 55,58

4.5 2021-03-16 2021-03-17 11 9 117,16

VIII

Appendix 4. Critical values for two-tailed test, Pearson correlation Table A4.1. Critical values for two-tailed test (Source: Siegle, (2015))

df = n -2 Level of Significance (p) for two-tailed test

df .10 .05 .02 .01

1 .988 .997 .9995 .9999

2 .900 .950 .980 .990

3 .805 .878 .934 .959

4 .729 .811 .882 .917

5 .669 .754 .833 .874

6 .622 .707 .789 .834

7 .582 .666 .750 .798

8 .549 .632 .716 .765

9 .521 .602 .685 .735

10 .497 .576 .658 .708

11 .476 .553 .634 .684

12 .458 .532 .612 .661

13 .441 .514 .592 .641

14 .426 .497 .574 .623

15 .412 .482 .558 .606

16 .400 .468 .542 .590

17 .389 .456 .528 .575

18 .378 .444 .516 .561

19 .369 .433 .503 .549

20 .360 .423 .492 .537

21 .352 .413 .482 .526

22 .344 .404 .472 .515

23 .337 .396 .462 .505

24 .330 .388 .453 .496

25 .323 .381 .445 .487

26 .317 .374 .437 .479

27 .311 .367 .430 .471

28 .306 .361 .423 .463

29 .301 .355 .416 .456

30 .296 .349 .409 .449

35 .275 .325 .381 .418

40 .257 .304 .358 .393

45 .243 .288 .338 .372

50 .231 .273 .322 .354

60 .211 .250 .295 .325

70 .195 .232 .274 .303

80 .183 .217 .256 .283

90 .173 .205 .242 .267

100 .164 .195 .230 .254

IX

Appendix 5. Phosphorus in Margetelund WWTP

Figure A5.1. Historical percentage of phosphorus in dry weight effluent suspended particles, Margetelund WWTP, January 2015

– January 2021

Figure A5.2. Historical phosphorus concentration, seasonal variations over 6 years, Margretelund WWTP, January 2015 –

January 2021

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

6.00%

7.00%

8.00%

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56

Dry

wei

ght

of T

SS [

%]

week

TheoreticalTOT-P value

TOT-P 2015

TOT-P 2016

TOT-P 2017

TOT-P 2018

TOT-P 2019

TOT-P 2020

TOT-P 2021

0

0.2

0.4

0.6

0.8

1

1.2

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56

mg/l

week

TOT-P limit

TOT-P 2015

TOT-P 2016

TOT-P 2017

TOT-P 2018

TOT-P 2019

TOT-P 2020

TOT-P 2021

X

Appendix 6. Experiment journal.

Below is listed the experimental runs with corresponding date when things did not go as expected and

thoughts from the author of what happened and some possible results of that action.

Trial 1.1. 23-24/02/2021

Snowmelt just started, and the sludge scrapers is not working in pre-sedimentation, resulting in very high

flow rates and total suspended solids (TSS) throughout the wastewater treatment plant.

One out of two air compressors stopped working sometime during the night, resulting in low to no

pressure for the pressurized recycled effluent.

Recirculation in the nitrification has been running during the trial, disrupting the constant flow rate, so

influent flow rate =/= effluent flow rate for trial 1.1.

Trial 1.2. 24-25/02/2021

The sludge scrapers are still not functioning, so unusually high concentrations of influent SS is noticed.

Postponing further trials until pre-sedimentation is functioning again.

Trial 2.1. 1-2/03/2021

No pressure in pressurized tank during the morning (8-12 am) the 2nd of mars. Air compressor is once

more the fault.

Trial 2.3. 3-4/03/2021

Planed power outbreak that was mentioned the same day as occurring, resulting in lost test result between

10am-2pm. Trial started once more when power was back.

Trial 2.5. 6-7/03/2021

Influent SS sampler stopped working from unknown reason during hour 14 of 24, resulting only in half

a sample. Still water enough to collect for analysis. Calibrated the sampler before starting it once more.

Trial 3.2. 8-9/03/2021

A lot of floated sludge with surface load 2.5 m/h, one thought is if the sludge scrapers are inefficient.

Bigger flocs, 5-10cm diameter, of SS seeps through with effluent water.

Trial 4.1. 12-13/03/2021

One of the other parallel DAF units (line 1) uses a pressurized flow of 60 m3/h, when it normally should

be around 30 m3/h. May have affected the saturation level in the pressurized tank if water levels are too

low.

Trail 4.2. 13-14/03/2021

Influent sampler stopped working form unknown error at hour 19. Still water enough to collect for

analysis. Calibrated the sampler before starting it once more.

Trial 4.3. 14-15/03/2021

The WWTP mechanic reduced the pressurized flow rate from expected 62 m3/h too ~20 m3/h the

morning on the 15th of mars, was not noticed until 11 am. Disrupted 4-5 hours during the trial.