Reject Water Management at Wastewater Plants in Luleå

(1)

M A S T E R ' S T H E S I S

Reject Water Management at Wastewater Plants in Luleå

François Tissot

Luleå University of Technology MSc Programmes in Engineering

Civil Engineering

Department of Civil and Environmental Engineering Division of Sanitary Engineering

2009:093 CIV - ISSN: 1402-1617 - ISRN: LTU-EX--09/093--SE

(2)

1

Preface

The third year of study at the ENSG (National School of Geology) offers the double chance to realize a master thesis, and to study abroad in a partner university.

As for me, I seized both opportunities.

Following the advices of a friend of mine who was currently doing a master thesis at Luleå Tekniska Universitet (LTU), in the North of Sweden, I considered LTU in my choice. Having studied Environmental and Water Sciences, this university combined an education of high quality in my work field, with a unique life experience, standing at barely 150km from the artic circle.

This report is the summary of my master thesis work, done in Sanitary Engineering, from January to May 2009. It was focused on two wastewater treatment plants – Råneå and Uddebo – in the community of Luleå, and two different results were aimed at.

Taking the opportunity, I would like to thank both the team of the Råneå and the Uddebo plant, for the time they gave me, answering my many questions, changing the settings of the processes to allow me to check every hypothesis I built up, and driving me to and fro the bus station whenever I needed it; Kerstin Nordqvist and Anna-Maria Gustafsson who took care of the analyses of the water samples; Jan-Eric Ylinenpää and Stefan Marklund who supported the project with patience and

interest, giving me the liberty to work as I wanted to, and allowing me to realize lab and pilot scale experiments.

I would particularly like to thank Professor Jörgen Hanæus, Division of Sanitary Engineering at LTU, who was always ready to discuss the results and to help me find a creative way to keep going forward.

(3)

2

Abstract

The following study was conducted during the winter-spring 2009 at two wastewater treatment plants of the Luleå Kommun: Råneå and Uddebo. Several points were studied. For the Råneå plant, the aim of this work was to spot and solve the problem that was disturbing the functioning of the plant and altering the quality of the water rejected into the river. It turned out that a loop in the circulation of the reject water coming from the centrifuge led the way to a very high polymer concentration in the water going back to the entrance of the plant, which seemed to be the disturbing parameter. For the Uddebo plant the aim was to compare the performances of a new dewatering device – a press – to those of the centrifuge. It was found that the press was much more efficient than the centrifuge, using only 0,11kW/h to produce one kg of Total Solids (TS), while the centrifuge used 0,72kW/h to produce the same amount. An investigation about the feasibility of a full scale Freezing-Thawing step was also conducted at the Uddebo plant. It included working on the means to improve the process, in terms of dewatering, spreading, storing and harvesting of the sludge. A proposition of Piled Up boxes allowing a very flexible storage and Freezing-Thawing of the sludge is made at the end of the report.

(4)

3

Table of Tables

Table 1: Facts and numbers about Luleå Kommun ... 6

Table 2: Classification of moisture attachment in wastewater sludge (Marklund 1997)... 10

Table 3: Content of the inflow and outflow water to/from the Råneå plant (year 2008) (values in mg/l) ... 15

Table 4: Yearly results of the wastewater treatment plant of Råneå (in mg/l)... 16

Table 5: Content of the inflow and outflow water to/from the Uddebo plant (year 2008) (values in mg/l. In purple values above the upper limit imposed by the EU regulation)... 18

Table 6: Yearly results of the wastewater treatment plant of Uddebo (in mg/l). ... 18

Table 7: Turbidity (values in NTU) and Secchi depths (values in cm) measured in both sedimentation tanks, and turbidity in the pond after the beginning of the pilot scale precipitation. ... 26

Table 8: Range and mean TS value of the sludge coming from centrifuge and presses. ... 28

Table 9: Summary of the centrifuge and presses comparison... 28

Table 10 : Water content after Freezing-Thawing ... 29

Table 11 : Sludge TS. Samples taken when there was still water and sludge are called wet in opposite to the dry samples taken once the water was drained (five minutes drainage)... 29

Table 12 : Requirements list for a new design of Freezing-Thawing device... 33

Table of Figures

Figure 1: Map of Luleå Kommun... 6

Figure 2: Administration and Corporate Structure of the Luleå Kommun. In red, the Luleå Kommun Company itself... 7

Figure 3: To the left: Details of the thirteen departments of the LTU. (Size or positions of the boxes are not dependent on the importance of the department) To the right: Position and condition of my tutors at Luleå Kommun. ... 8

Figure 4: Drying curve for identifying four different types of water in sludges. (Tsang and Vesilind 1990)... 11

Figure 5: Moisture content versus sludge volume for an initial TS content of 2% and 5%, where 70% of the total are volatile. (Marklund 1997) ... 12

Figure 6: Map of the Råneå plant with detail of the steps and flux inside the plant (blue = water, red = sludge) ... 14

Figure 7: Photography of the centrifuge in place at the Uddebo plant (top) and the presses (botton) from Huber website (for a better view of the press). ... 19

Figure 8: Schematic diagrams of a countercurrent centrifuge (top) and press (bottom) configuration for dewatering sludge, respectively from “Metcalf & Eddy”, and Huber website. ... 20

Figure 9: Aerial picture of the Uddebo plant and the yard used for the Freezing-Thawing pilot scale experiment. ... 22

Figure 10 : Turbidity of the reject waters from centrifuge and both presses. The values displayed are the average values for each measurement (range was about ± 10 NTU). ... 26

Figure 11 : Energy consumption comparison: Press and Centrifuge... 27

Figure 12 : Scheme of the sludge distribution (big pipes) from the sludge mixing tank to centrifuge and presses at the Uddebo plant. The small pipes represent the polymer addition: from mixing tank one in grey, and mixing tank two in rusty red. ... 29

Figure 13 : Detail of the lab scale Freezing-Thawing experiment at Uddebo. 1) Frozen sludge and freezing cylinder 2) Bucket with dry sample and wet sample... 30

Figure 14 : Detail of the pilot scale Freezing-thawing experiment at Uddebo. 1) Pilot area 2) Sludge cover at the end of May 3) Sludge sample... 30

Figure 15 : Sludge TS from the pilot scale Freezing-Thawing experiment at Uddebo... 31

Figure 16: Possible organisation for storage and Freezing-Thawing of the sludge in Piled Up boxes. ... 33

Figure 17: Design of the concept of Piled_Up boxes for sludge Freezing-Thawing:... 34

(6)

5

Foreword

Luleå

Located at the coast of region Norrbotten in northern Sweden, the municipality of Luleå – which includes Nederluleå, Råneå and Luleå City – counts 73,000 inhabitants. It makes it the 26th largest municipality in the country. The City of Luleå itself, founded in 1621 around the medieval church in Gammelstad¹, was moved in 1649 to its present site 10 km nearer the sea, due to the land elevation that had made the bay too shallow for ships to enter.

The municipality of Luleå is the economic, administrative and political centre of Norrbotten. Its activities are a mix of industry, research, education, trade, and services. For instance, the Luleå harbour is the Sweden's fifth-largest harbour and is of particular significance for iron ore from LKAB's mining in Kiruna and Gällivare/Malmberget. During the winter, sea traffic continues at a virtually unchanged rate with the assistance of icebreakers. Furthermore, Luleå Kallax airport makes Stockholm reachable in two hours by one of the fifteen daily flights. At last, Luleå has direct train connections with Stockholm and Narvik.

The climate in Luleå is a sub arctic climate characterized by short cool summers and long cold snowy winters. Despite its extremely northern latitude, the climate is relatively mild compared to other places at similar latitude because of the Gulf Stream. Being so close to the Artic Circle, the city experienced both very short and long periods of sunlight during, respectively, the winter (down to four hours a day) and the summer (up to twenty-two hours a day). During the last fifteen years the average annual temperature has increased by one degree (°C).

The development of the municipality is largely due to Luleå University of Technology (LTU), with 12,000 students in four cities. The University provides education in the spheres of engineering, social sciences, teaching, health sciences, business and theatre. The annual turnover of research is SEK 600 million (approximately EUR 60 million). Close to the University Campus there is the high- tech science park Aurorum accommodating 100 knowledgebased businesses.

Luleå also make profit of an art and conference centre (Kulturens Hus), and the Norrbotten Theatre. Luleå shows a high level of council services, a high level of education (approximately 40 % of the population in the ages 25-64 years have a post-secondary education), as well as a wide range of sports available for the city’s young people.

About 2 kilometres north Luleå, is the campus of the Luleå University of Technology (LTU).

1 The Gammelstad Church Town is a UNESCO World Heritage Site.

(7)

MASTER THESIS – Reject water management at wastewater plants in Luleå – JUNE 2009

L u le å K o m m u n

The Municipality of Luleå can be shortly described as follow (Table 1, Figure 1): Lulea KommunSweden Population31.12.2008: 73,4069,263,872 (2009 census) Area (including water) 2,110 km2 (4,928 km2) 410,817km² (449,964 km²) No. of households 2008: 34,382 Larger population centres 1st Luleå 45,657 inhab, 2nd Gammelstad 4,919, 5th Råneå 1,984

1st Stockholm 1,252,020, 2nd Göteborg 510,492, 3rd Malmö 258,020 Average age 2008: 40.8 years41.0 Average income 2007: Men: SEK1 266,500 Women: 201,500Men 280,800 Women 200,100 GDP per inhabitant 2006SEK 330,000319,000 Rate of millionaires2007: Men: 14 % Women: 10 %Men 20% Women 17% Commuting 2007: To Luleå: 8,080 From Luleå 3,670 Fertility rate2008: 1.75 per woman1,91 Divorce rate2007, women aged 30-59: 23,6 %22,4 % Reported offences 2007: 123 per 1,000 inhabitants143 Passenger cars2007: 484 per 1000 inhabitants464 Table 1: Facts and numbers about Luleå Kommun Figure 1: Map of Luleå Kommun 1 An average conversion rate for the year 2008-2009 is 10 SEK = 1 €

10 km

N

(8)

When it comes to Administration and Corporate structure, the kommun organisation is the following (Figure 2):

Figure 2: Administration and Corporate Structure of the Luleå Kommun. In red, the Luleå Kommun Company itself.

Luleå Kommun employs about 7000 people. The sanitary engineering division counts 55 employees (from which 26 are full time people). This give about 1 employee per 10 000 customers, which is a rather normal figure in EU. The total turnover of the division for the year 2009 was about 115 000 000 SEK, and the investments rounds up at ca 58 000 000 SEK.

The two tutors that supervised my internship inside the Luleå Kommun were:

- Stefan Marklund: Head of the department of Water & Wastewater services in City of Luleå. He has responsibility of all employed personal, all processes and general department economics. When it comes to environmental laws the responsibility is entirely his.

- Jan-Erik Ylinenpää: Works as division manager for all treatment units Water &

Wastewater. He has manpower, operation and economic responsibility.

7 Luleå Kommun

Company

(9)

8

The Luleå University of Technology (LTU)

LTU is the northernmost university of technology in Scandinavia – hence it’s motto “Great Ideas grow better below zero” – and with world-class research and education. The institution strength lies in cooperation with companies and with rest of the world around it. Its contact with companies and society help it develop research and education that satisfy the demands that the world around makes on both the University and its students.

The research is on the whole applied research and in the Faculty of engineering and the Faculty of arts and social sciences. Its annual turnover of approximately EUR 69 million and comprises 70 research subjects in 13 departments (Figure 3 below).

LTU has a staff of 159 professors, 515 teachers and researchers and 641 doctoral students. Research is deeply characterised by multidisciplinary cooperation between the University’s research departments and close interaction with trade, industry and society (as for me, the Luleå Kommun).

Figure 3: To the left: Details of the thirteen departments of the LTU. (Size or positions of the boxes are not dependent on the importance of the department) To the right: Position and condition of my tutors at Luleå

Kommun.

Jörgen Hanaeus Professor

Luleå Kommun LTU

Human Work Sciences

Health Sciences

Business Administration,

Social Sciences

LTU, Skellefteå

Maths

Music and Media Education

Space Science

Civil, Mining and Environmental

Engineering

Languages and Culture

Computer Science and

Electrical Engineering Applied Physics

and Mechanical Engineering

Chemical Engineering

and Geosciences

LTU

Water and Wastewater

Division Manager General Manager

of Water and Wastewater

Services

François Tissot Trainee

Jan-Eric Ylinenpää Stefan Marklund

(10)

9

Introduction

Wastewater treatment is a subtle industry dealing with a raw material whose content is always changing, but has to give a final product of rather constant content.

As soon as one step of the process is deficient, it is possible that all the plant will show signs of malfunctioning. If the deficient step can not be easily spotted (broken device for instance) it might be quite difficult to find it once the signs of malfunctioning have spread to the whole plant. That what seems to be happening at the Råneå plant, where the sedimentation tank water show a higher turbidity¹, and consequently a higher Phosphorus and Nitrogen content, than expected. This might be a critical issue because once passed through the sedimentation tanks, the water is simply released to the river, without any further treatment.

Traditionally, the dewatering of the sludge at the end of the wastewater treatment process is done by means of mechanical devices (centrifuge, belt filter press or drying bed for instance). Recently, the Uddebo plant purchased two new presses (Huber brand) whose performances are claimed to be comparable to those of one centrifuge (or even better) and added them next to the customary centrifuge.

The climate of northern Sweden, being really cold during the winter, a Freezing- Thawing dewatering process might also be of interest to reduce the sludge volume and increase the sludge Total Solids (TS) content. The yard behind the wastewater plant was used to conduct a first experiment.

Thus, the following study is focused: for the Råneå plant on spotting and solving the problem that is disturbing the functioning of the plant and altering the quality of the water rejected in the river; for the Uddebo plant on comparing the performances of the presses to those of the centrifuge, as well as investigating the feasibility of a full scale Freezing-Thawing step, and find the means to improve the process, in terms of dewatering and spreading / storing / harvesting.

The general context, as well as the detailed aims of the work, are presented in the first section. Then there are the experimental and analytical methods in section two and three. The results and their implications are exposed in section four, while the conclusions are made in section five.

1 Particles lighted up by a beam will scatter this light beam focused on them. The turbidity is then a function of the light reaching the detector. More light reaches the detector if there are many small particles scattering the source beam than if there were few.

(11)

10

I. Purpose of the study I.1. General context

Wastewater treatment is a three step process that turns a single wastewater inflow into three different products: the effluent water that is the clean product which goes to the river, the sludge which can be used in different ways (such as the construction of soil, fertilizers, or composting) and the reject water from the sludge treatment which is brought back to the entrance of the plant and goes through the treatment process once again.

This reject water usually represents about 1% volume of the water inflow of the plant, but its content represents about 25% concentration of this water inflow (N, P, and other elements content). Hence, the reject water is likely to create perturbations in the water treatment process.

Usually, the wastewater treatment process is a succession of three major steps called Mechanical (several screenings), Biological (bacterial activity) and Chemical (flocculation and sedimentation) steps. All of these steps produce different amounts of sludge with low TS (between 1 and 2%). The sludge TS is then increased up to 18-25% thanks to dewatering devices such as thickeners and centrifuges.

The problem for the dewatering of sludge has been subjected to an increased attention during the last 30 years. And in these times when water is becoming more and more a rare resource, it is of importance to come up with an improved sludge treatment, where less water is left to the sludge after dewatering.

The important background to be familiar with when talking about the dewatering has been reviewed by Marklund (1997) in his licentiate thesis work. The main points are spotted below.

Wastewater sludge can be characterized as a mixture of water, dissolved substances and particles. The liquid phase of the sludge is described by different bounding states, corresponding to different bonding strengths or technologies for moisture removal. Several classifications had thus been proposed (cf Table 2):

Source Class

I II III IV Spangler (1960) Gravitational

water

Capillary water

Hygroscopic water

Vesilind (1974) Free water

Floc water

Capillary water

Particle water Smollen (1986) Mechanical

attachment

Physical attachment

Chemical attachment Smollen (1988) Free moisture Immobilized

moisture

Bound moisture

Chemically bound moist.

Tsang &

Vesilind (1990)

Free moisture Interstitial Moisture

Surface moisture

Bound moisture

Table 2: Classification of moisture attachment in wastewater sludge (Marklund 1997).

(12)

11 Spangler (1960) in his reference to soil water classification referred each class of water to the kind of primary force needed for its movement. The types are:

I Gravitational water – water which exists in large pores of soil and which the force of gravity will remove from the soil when conditions for free drainage exist.

II Capillary water – water which is held by cohesion as a continuous film around the soil particles and in the capillary spaces.

III Hygroscopic water – water tightly adhering to solid particles in thin films which can be removed only as vapour.

Vesilind (1974) divided sludge moisture into four categories:

I Free water - water not attached to sludge solids in any way and that can be removed by simple gravitational settling.

II Floc water - water which is trapped within the flocs and travels with them.

Removal is possible by mechanical dewatering.

III Capillary water - water which adheres to the individual particles and can be squeezed out only if these individual particles are forced out of shape and compacted.

IV Particle water - water which is chemically bound to the individual particles.

According to Vesilind’s definition, the free water is the water above the sludge cover settled in a one-litre cylinder. Floc water is the additional volume removed by low to medium centrifuge speed levels (i.e. 6 000 - 8 000 g), while capillary water is removed by gravitational forces from 13 000 g and above. The remaining water is defined as particle water.

Smollen (1986a+b) underlined a basic split in mechanical, physical and chemical attachment, based on the order of bound energy in kJ/kmol. In a later paper Smollen (1988) classified moisture into four fractions, by the use of (i) vacuum filtration, (ii) drying at 30°C and (iii) drying at 105°C. Thus Smollen defines the remaining chemically bound water at 105°C as bound by a powerful linkage that can be broken only by heating above 105°C.

Finally Tsang and Vesilind (1990) used a drying apparatus to dry a thin layer of sludge. By precise weight measurements in a thermal tube dryer they obtained drying curves such as the one shown in Figure 4 below.

Figure 4: Drying curve for identifying four different types of water in sludges. (Tsang and Vesilind 1990)

(13)

12 They classified sludge moisture as follows:

I Free moisture - the moisture removed during the constant rate period of a drying curve. The moisture is not associated with particles and includes void water not affected by capillary forces.

II Interstitial moisture - the moisture removed during the first falling rate period of a drying curve. This is floc moisture when sludge is in suspension, and is present in the capillaries when the cake is formed.

III Surface moisture - the moisture removed during the second rate falling period of a drying curve. This moisture is held on the surface of the solid particles by adsorption and adhesion.

IV Bound moisture - the moisture not removed during the experiment. This moisture is chemically bound to the solid particles.

From a more practical point of view, these four types of water can be related to four types of dewatering processes (cf Figure 5), even though there is not a perfect correspondence between the type of moisture and the dewatering processes (for instance, the mechanical dewatering do not remove all of the interstitial moisture) (Vesilind 1994).

Figure 5: Moisture content versus sludge volume for an initial TS content of 2% and 5%, where 70% of the total are volatile. (Marklund 1997)

This figure states that 34% TS is a practical upper limit for mechanical dewatering. For further dewatering, use of drying and combustion technique is required. Drying gives a porous matrix with TS up to 80 or 90%, while burning leaves only ashes and all the water is evaporated.

The increase in TS is related to a volume reduction, most noticeable at the first steps of the process where a 50% volume reduction is achieved at only 5-9%TS.

Then mechanical processes bring it down to 15-6%, while burning decreases it to 5- 2% and combustion down to 2-0,5%. These are theoretical remarks and in practice weight reduction is not followed by such a reduction of volume at moisture content below 50-60% as the dried sludge, as well as the incineration residue, becomes porous in its structure.

(14)

13 Economically speaking, the dryer the sludge, the less the volume of sludge, and thus, the cheaper the transport and the easier the handling of the sludge. A good understanding of the sludge treatment, and if possible an improvement of this step, is thus highly desirable.

But nowadays, if the wastewater treatment process is a quite well known practice, the sludge treatment process is still an area where a lot of unknowns are looking for answers.

I.2. Two plants, three study topics

Two places:

The Råneå plant: It cannot handle the external sludge brought during the summer. This extra sludge and/or one of the steps of the process in the plant create perturbations whose effect is visible in sedimentation tanks: high turbidity of the water. This low efficiency of the plant is a subject of concern for the community of Luleå. Moreover, close to the plant, there is a pond that is not used currently. The community wishes to know if it can be turned into a precipitation tank for the reject water from the centrifuge.

The Uddebo plant: A comparison of the presses and centrifuge efficiency (final TS of the sludge, turbidity of the reject water, energy consumption, polymer consumption and Nitrogen/Phosphorus content of the reject water) was conducted.

Moreover the Freezing-Thawing of the sludge is a dewatering method that can easily be developed here in the North of Sweden. Though it’s a quite well known method, the results depend on the sludge obtained at the plant as well as some design considerations like, for instance, the disposition of the sludge (in beds? in thin layers?) and the way to harvest it. The water rejected by this method is also of importance and has to be taken into account given the high elements content it can present.

I.3. Aim of the study

At the Råneå plant: Definition of the conditions in which the plant shows perturbations of the process. Attempt to solve the problem: first without changing the plant functioning, then attempt to find an easily applicable treatment of the reject water from the centrifuge (addition of chemical for instance) that could avoid its recirculation and prevent the perturbation of the plant process. The use of the pond is strongly considered to store and allow the settling of particles in the reject water.

At the Uddebo plant: Comparison of the presses and centrifuge performance for possible generalisation of the presses use. Attempt to determine if the Freezing- Thawing of the sludge can be of interest for the Uddebo plant. An optimum strategy is to be found that could combine low cost production of a low TS sludge, and increase of this TS by natural means such as Freezing-Thawing.

(15)

14

II. Experimental methods

II.1. Råneå:

II.1.1. Short overview of the plant

The city of Råneå is a 2000 inhabitants city, located 30km North East of Luleå (cf Figure 1).The wastewater plant of Råneå is dimensioned for 3800 people, and has been functioning since 1970. The average water inflow to the plant is 978m³/day, with a lower value during the winter (around 600m³/d), a higher value during summer (around 1400m³/d) and top values during snow melting (up to 3995m³/d). The sludge production is around 330m³/y. Five people are working daily at the plant.

The plant schedule is the one of a common plant (cf Figure 6 and Appendix1):

Mechanical, Biological (Aeration on the map) and Chemical (Flocculation on the map) steps, to which is added a dewatering sludge step (by means of two thickeners and two centrifuges in normal time, one only during this study as the other one was under reparation).

Figure 6: Map of the Råneå plant with detail of the steps and flux inside the plant (blue = water, red = sludge)

This figure shows the usual schedule of the plant. The water follows a rather normal path during all the wastewater treatment (Pumping station Mechanical step

Biological step Flocculation Sedimentation tanks and then to the river), while the sludge is collected from the different production places (mainly the sedimentation tanks and the pre-sedimentation tank) and sent to the internal thickener before further treatment. It is then mixed with the sludge from the external thickener which is a mix between the reject water from the centrifuge and the external sludge brought to the plant by trucks. Two arrows figuring the effluent water from both thickeners go to the pumping station. The water coming from the external thickener is the most concentrated and difficult to handle water of the whole plant because it is mainly formed by the reject water from the centrifuge.

Reject Water (to river)

Inflow to plant

Sludge outflow External

sludge

(16)

One should also notice the loop created by the sludge and water circulation between: Storage tank / Centrifuge / Storage tank / External Thickener. Each tour around the loop sees an addition of polymer to the sludge (right before centrifugation). This means that a possibly very high polymer concentration can be reached into the reject water going back to the pumping station, and this may be a disturbing element for the functioning of the centrifuge.

This loop was not indicated on the schedule process, where the reject water from the centrifuge was theoretically going to the entrance pumping station, and was only found late during the work, changing radically the way the plant was working.

When considering the efficiency of the plant process, it is of interest to look at the water contents before and after the plant. (Table 3).

Råneå wastewater plant 2008

Date Inflow sample Outflow sample

BOD7 CODCr Tot-P NH4-N Tot-N susp BOD7 CODCr Tot-P NH4-N Tot-N susp

8-9 jan 144 320 4,9 20 30 168 5 26 0,082 21 21 <5

23-24 jan 120 320 4,8 20 36 128 4 21 0,098 14 16 <5

6-7 feb 97 230 4,3 22 30 99 4 21 0,069 21 24 <5

19-20 feb 104 240 4,9 23 34 90 5,8 28 0,069 21 25 <5

5-6 mar 178 260 5,1 27 32 120 6,1 19 0,072 26 25 <5

17-18 mar 280 4,9 24 36 145 5,6 22 0,098 23 27 <5

Quarter 1 129 275 5 23 33 125 5 23 0,08 21 23

1-2 apr 277 5,2 27 39 198 7 34 0,12 18 22 <5

16-17 apr 130 233 3,8 18 30 140 7 25 0,16 17 23 6,2

28-29 apr 46 210 2,7 4,9 10 121 4 25 0,14 4,8 8,1 8,2

12-13may 32 94 1,2 1,2 8,1 20 3 22 0,094 3,6 5,9 <5

28-29may 60 148 1,2 10 17 51 12 40 0,23 10 15 12

11-12 jun 96 190 2,8 13 21 101 6 29 0,075 13 15 <5

23-24 jun 173 369 4,2 20 28 166 7 30 0,093 20 20 <5

Quarter 2 116 207 3 13 22 114 7 29 0,13 12 16

8-9 jul 91 160 3,4 18 24 3,6 24 0,065 15 17

23-24 jul 140 400 4,1 16 25 5 29 0,078 14 16

4-5 aug 151 320 4,5 23 33 134 5 31 0,1 21 23 <5

19-20 aug 94 230 2,9 14 21 141 3,4 30 0,15 14 15 6,6

3-4 sep 76 170 3,1 17 24 60 3,1 28 0,1 17 17 <5

15-16 sep 190 399 5,1 23 33 219 5,0 30 0,082 20 23 <5

30sep-

1oct 120 255 4,6 29 38 92 5 30 0,088 25 29 <5

Quarter 3 123 276 4 20 28 129 4 29 0,09 18 20

15-16 oct 130 236 4,1 24 33 79 7 33 0,11 22 23 <5

27-28 oct 64 220 3,6 20 27 70 5 32 0,079 19 20 <5

11-12 nov 83 210 4 27 35 67 6 32 0,11 26 28 <5

26-27 nov 74 280 4,8 33 40 83 6 34 0,12 33 36 <5

9-10 dec 120 250 5,9 27 35 75 7 37 0,23 25 29 6,9

22-23 dec 110 190 3,5 21 31 66 6 31 0,076 19 23 <5

Quarter 4 97 231 4 25 34 73 6 33 0,12 24 27

Table 3: Content of the inflow and outflow water to/from the Råneå plant (year 2008) (values in mg/l)

15

(17)

16 BOD7 is the acronym from Biological Oxygen Demand and is a measure of the oxygen consumption by micro-organisms in the water. CODCr stands for Chemical Oxygen Demand and is a test used to measure the oxygen equivalent of the organic material in wastewater that can be oxidized chemically using dichromate in an acid solution. Thus, BOD represents the micro-organism content of the water, while the COD represents the total oxygen consumption in the water.

To see Table 3 a little clearer, annual mean values were calculated, compared to upper limit authorised values and efficiency of the process was calculated as percentage of removal. Put under a shorter form, the results of the plant are the following (Table 4):

BOD7 CODCr Tot-P NH4-N Tot-N Susp

Inflow 116 249 4,0 20 29 110

Outflow 5,5 29 0,11 19 21 <5

Limit Min 70% Min 75% < 0,5

Efficiency 94% 87% 97% 5% 28% 95%

Table 4: Yearly results of the wastewater treatment plant of Råneå (in mg/l).

The plant shows seems to be really efficient for BOD, COD, Tot-P and Suspended solids reduction, while the NH4 and Tot-N reduction is low, even inexistent. This is only normal because the design of the plant is made without any Nitrogen treatment step. At present, the latest proposal from the highest EU-court indicates that no special N-treatment is necessary at coastal wastewater treatment plants from the city of Gävle (350km north of Stockholm) and upwards. A new EU regulation should come up in the next few years, but it should not concern northern Sweden and Finland at all.

In places where N-removal is regulated, depending on the plant size the maximum allowable effluent Tot-N level is between 10 and 15 mg/l. Nowadays around 55 treatment plants in Sweden (connection above 10 000 people) have N- removal technology. When considering the Uddebo plant, the Nitrogen values remain acceptable, even without any N-removal step in the plant.

II.1.2. Sampling

At the Råneå Wastewater Treatment plant (WWTP), several samplings and measurements were made, all following the same protocol.

As the plant has no automatic sampling device, the sampling was made by the mean of a bucket of ten litres (or a cup when the opening was too tight). The flow was calculated by measuring the time needed to fill up the bucket, thus the values are precise only to ±0,6m³/h. Turbidity, pH, and temperature were the basic parameters measured every time.

II.1.3. Attempts to identify and solve the problem of the plant

Several tracks were followed to answer this question.

First of all, an overall checking of the basic parameters (pH, T, Turbidity, and flows) of the main steps of the process was done: reject from the internal thickener, the external thickener, both sedimentation tanks, the reject going into the pumping station at the entrance of the plant (likely to influence the process for it is highly concentrated water) and the reject from the centrifuge.

(18)

17 All parameters were found constant, and no problem (not even the problem of high turbidity in the sedimentation tanks that was previously observed) was detected.

But a difference of turbidity in between the two sedimentation tanks was noted. Thus it was decided to focus on these two tanks, and on the reject water from the centrifuge, to try to assess if the pond could be of any use in a precipitation step of this water.

Precipitation experiments were then conducted on the reject water from the centrifuge, as well as the checking of the turbidity of the sedimentation tanks and the flocculation tank. The first one, in order to determine if a pilot scale test of a precipitation step using the pond was possible, the second one, to follow the evolution of the turbidity as a function of the reject water characteristics. Experiments on the reject water were made by mean of addition of Aluminium Sulphate, Ferric Chloride, variation of pH, addition of polymers (non-ionic, cationic and anionic) or combinations of them (Appendix 3).

At this step of the study, it was most likely that the centrifugation was the weak process because of the high variation in turbidity and water content of its reject.

Thus, focus was made on every aspect of it: polymer mixing, maturation, addition, sludge inflow, outflow and running time. At this occasion, the circulation of the reject water from the centrifuge and the loop (cf Figure 6) was found out.

A pilot scale test of the precipitation of the reject from the centrifuge was also tried after a possible negative feedback had been identified on the reject water circulation in the existence of the loop. In this purpose, the reject water was not recirculated anymore but was simply sent to the pond behind the plant after addition of roughly 20mg/l of Aluminium. Thus, no more heavily polymer concentrated water was sent to the entrance of the plant, and the turbidity in the sedimentation tanks was watched carefully to see any effect of the loop breakage.

II.2. Uddebo:

II.2.1. Short overview of the plant

The city of Luleå and its agglomeration counts 73 000 inhabitants. The wastewater treatment plant of Uddebo treats more or less the water of the whole city of Luleå, as well as some of the closest small cities. The wastewater plant of Uddebo is dimensioned for 91 000 people, and has been functioning since 1968. The average water inflow to the plant is 24 400m³/d, and top values during snow melting (up to 58 480m³/d). The sludge production is around 7000m³/y. The budget of the plant is around 14 000 000 SEK/y and five people are working daily at the plant.

The plant schedule is rather similar to the one of Råneå, except for the size and numbers of tanks that are very different. The Uddebo plant also counts a moving bed biofilm reactor, converting BOD into CO2 and cell mass, and also having a nitrifying effect. But its main deviation from the Råneå plant is the existence of a digester of the sludge prior to centrifugation or pressing which allows delivering more sludge from external private users to the plant. The digester realizes the stabilization of the sludge, which means it: 1) reduces pathogens, 2) eliminates offensive odours and 3) reduces or eliminates the potential for putrefaction. (Appendix 2)

As for the Råneå plant, the efficiency of the Uddebo plant is shown by analyses of the influent and effluent water (cf Table 5) and the annual mean values calculated as percentage of removal (cf Table 6).

(19)

18 The increase in Tot-N content of the water is most likely to be a contribution from the external sludge brought to the plant. This sludge is sent to the digester, and once the sludge has been dewatered, the reject water, containing the N goes back to the entrance of the plant. This way it increases the N content of the clear water.

Uddebo wastewater plant 2008

Date Inflow sample Outflow sample

BOD7 CODCr Tot-P NH4-N Tot-N susp BOD7 CODCr Tot-P NH4-N Tot-N susp

2-3 jan 199 414 6,1 26 37 248 31 76 0,44 33 37 28

15-16jan 110 190 3,2 29 36 139 52 102 0,62 32 39 29

30-31 jan 220 441 5,6 27 35 248 60 118 0,8 28 30 32

13-14 feb 210 396 5,8 23 38 205 42 86 0,33 33 39 4,8

25-26 feb 230 370 5,5 23 35 200 39 79 0,44 26 28 18

11-12mar 191 450 6,0 20 40 200 40 87 0,20 27 30 10

26-27 mar 282 500 6,5 23 33 285 36 81 0,22 27 31 11

Quarter 1 206 394 6 24 36 218 43 90 0,44 29 33 19

10-11 apr 213 416 6,1 24 43 221 32 80 0,43 26 30 22

21-22 apr 153 320 3,8 15 27 160 27 72 0,35 22 26 17

6-7 maj 71 200 2,6 9,9 17 95 8 36 0,21 10 16 5,4

21-22 maj 92 190 3,5 16 23 90 18 52 0,25 21 24 13

3-4 juni 110 230 3,3 17 25 100 23 75 0,49 20 29 28

17-18 juni 184 432 5,1 19 31 271 12 53 0,26 19 29 18

Quarter 2 137 298 4 17 28 156 20 61 0,33 20 26 17

1-2 juli 200 540 5,0 17 37 280 17 75 0,41 17 31 26

16-17 juli 220 440 5,2 16 17,0 63 0,53 20

29-30 juli 130 360 4,7 15 29 10,0 40 0,36 11 23

11-12 aug 153 350 4,0 19 33 198 11 46 0,56 19 26 29

27-28 aug 190 470 4,7 19 31 229 6,6 38 0,37 12 22 18

9-10 sep 190 430 5,1 17 35 257 13 44 0,16 12 25 7,7

23-24 sep 236 390 5,4 20 35 244 25,0 57 0,21 25 33 21

Quarter 3 188 426 5 18 33 242 14 52 0,37 17 27 20

8-9 okt 60 200 2,7 9 16 130 34 102 0,47 25 31 40

21-22 okt 120 280 4,3 18 33 154 21 61 0,36 20 25 18

3-4 nov 150 490 5,0 21 44 210 25 88 0,53 34 42 31

19-20 nov 160 450 5,2 22 38 190 37 116 0,84 30 38 40

2-3 dec 260 420 5,7 24 39 190 34 77 0,42 33 38 17

15-16 dec 270 570 5,7 24 43 320 25 74 0,31 32 39 19

Quarter 4 170 402 5 20 36 199 29 86 0,49 29 36 28 Table 5: Content of the inflow and outflow water to/from the Uddebo plant (year 2008) (values in mg/l. In

purple values above the upper limit imposed by the EU regulation).

BOD7 CODCr Tot-P NH4-N Tot-N Susp

Inflow 177 382 5 20 33 203

Outflow 27 72 0,42 24 30 21

Limit Min 60% Min 75% < 0,5

Efficiency 83% 79% 91% -20% 9% 89%

Table 6: Yearly results of the wastewater treatment plant of Uddebo (in mg/l).

(20)

19

II.2.2. Presses/Centrifuge Comparison

The centrifuge (NX 4500 Alpha Laval) and the presses (Huber Technology, Figure 7) were compared regarding several parameters: final TS of the sludge, turbidity of the reject water from the dewatering devices, energy consumption, polymer consumption and Nitrogen/Phosphorus content of the reject water. The samples were taken right at the entrance and exit of the devices and put into clean beakers.

On the pictures the scale is given by the devices themselves: the centrifuge is about 3 meters long, and the presses are about 4 meters long.

Figure 7: Photography of the centrifuge in place at the Uddebo plant (top) and the presses (botton) from Huber website (for a better view of the press).

Dewatering part

Sludge outflow Motor

Break

Sludge inflow

Reject water (hidden by foot of

centrifuge)

Dewatering part Sludge outflow

Motor Sludge inflow

Reject water Pipes bringing washing water

(hidden by foot of press) Reactor

(21)

20 The principles of functioning for the centrifuge and the press are quite similar (Figure 8). The main difference is the rotation velocity of the screw inside the device (and the point of addition of the sludge inside it).

Figure 8: Schematic diagrams of a countercurrent centrifuge (top) and press (bottom) configuration for dewatering sludge, respectively from “Metcalf & Eddy”, and Huber website.

II.2.3. Freezing-Thawing experiments

Freezing-Thawing experiments were conducted both on lab scale and on pilot scale.

In lab scale:

Sludge was sampled from different points (cf Figure 12) to investigate the possible effects of polymer mixing on sludge freezing. The experiments were conducted at the University Laboratory.

(22)

21 To simulate a 1D freezing process, the sludge samples were frozen in tubes made of porous heat insulating material, open at the top and closed at the bottom by a wood cut, on which they stood.

As the TS of the sludge is low (between 2 and 5% TS), the freezing height of sludge was approximated to the freezing height of water, which, according to Marklund (1990) follows the root square law:

S k

h = × −

_.

Where h = ice thickness (cm) k = degree-day coefficient

S = sum of frost days multiplied with the mean daily temperature (°C) for each day.

For the northern part of Sweden k is between 2.2 (including the effect of an insulating snow layer) and 3,6 (snow free ice layer). To be sure that the freezing will be completed according to theoretical times, k was taken equal at 2 in this study.

Similarly, thawing height is expressed as:

V k z= '×

Where z = actual depth of thawing (cm) k’ = degree-day coefficient

V = sum of non-freezing days multiplied with the mean daily temperature (°C) for each day.

A usual value (cf. Marklund) for k’ in north of Sweden is 4cm V .

A freezing temperature of -7/-8°C was chosen. The winters in north of Sweden being most likely colder, this temperature was set to show what would happen during a relatively warm year. To decrease the time needed to freeze all the sludge (h=35cm, which would have required around 40 days of freezing), layers of 5 cm were added every 24 hours for 7 days.

Once the sludge was entirely frozen, the samples were put out of the isolation pipes and placed in clean buckets in a fridge at temperature of 7°C, until entire melting. It accidentally happened that the fridge temperature was changed during melting, freezing the sample after almost complete thawing. A second thawing took place, at 7°C for one day, and then at 4°C for conservation.

Water samples were taken for Nitrogen, Phosphorus and CODcr analyses, while sludge samples were taken for TS analyses.

In pilot scale:

A total cover of about 25cm was spread on the yard behind the plant (cf blue – green area on Figure 9) in two sequences: first the sludge is spread, then time is given for the sludge to freeze, before addition of a second layer on the first one. Only two layers were spread. This way of doing had the drawback that the tractor had to drive on the first layer to spread the second one. The second, flat area (red area on the picture) was used to store the sludge. The sludge pile could be as high as 2 or 3 meters from place to place on this area.

(23)

22

Figure 9: Aerial picture of the Uddebo plant and the yard used for the Freezing-Thawing pilot scale experiment.

Drains were made to collect the water into the pond at the back of the yard when the melting occurred. The sludge will then be analysed in terms of TS values.

50m Freezing/

Thawing experiment

area Storage

area

(24)

23

III. Analytical methods III.1. Sampling

The sampling was made using different kinds of beakers and cups, cleaned before each sampling, and rinsed with deionised water after each use. Turbidity, pH, and temperature measurements were run with a 2100N Turbidimeter HACH, and a pH meter coupled to a thermometer. When the same cup or beaker was used several times a day at the plant, it was carefully rinsed between each measurement. The pH meter calibration was regularly checked to avoid analytical errors. Depth in sedimentation tanks was measured using a Secchi disk (cf. p.25).

III.2. P, N, COD analyses

The Phosphorus, Nitrogen, COD and BOD analyses were performed at the Uddebo WWTP respectively following the protocols:

- Tot-P - SS-EN ISO 6878:2005-1 Determination of total phosphorus after peroxidisulfate oxidation

- Tot-N - SS 028131-1 and SM 4500 NO3-B - BOD7 - SS-EN 1899-1

- CODCr - Lange cuvette test

Where SS-EN stands for “Swedish standards institute” and SM for

“Standard methods”

The Phosphorus and Phosphate analyses were also performed using a LCK 349 LANGE kit, to get a quick estimation of the water content. The limitations of this technique were 0,05-1,50 mg/l Phosphate and 0,15-4,50 mg/l Phosphorus.

III.3. Precipitation experiments

Precipitation experiments were performed according to the following procedure:

the samples (one or two litres, depending on the case) were put in clean beaker (for the one litre ones) or bucket (for the two litres’ ones). The precipitating solution was then added by mean of a pipette, and in the few seconds following this addition, mixing began thanks to a mini-flocculator. The mixing was a 15 minutes process: 30 seconds of fast mixing (speed above 500rpm), 4 minutes of fast mixing (140rpm), 5 minutes of slow mixing (110rpm), and 6 minutes of very slow mixing (80rpm). The solution was then poured into a new beaker where settling was carried out.

During the settling process, samples were taken at constant depth in the solution, and turbidity was measured by mean of a 2100N Turbidimeter HACH.

III.4. TS analyses

When TS analyses were performed directly at the plant, one sample was analysed at the time, using a Sartorius MA 45.

When TS analyses were performed at the lab, several samples (up to thirty) were analysed at the same time. The samples were placed in clean, hot and dry china cup and weighted before to be put in the oven at 105°C over night. The cups were then put one hour in desiccator beakers, and the samples were weighted afterwards.

(25)

24

IV. Results and Discussion IV.1. Råneå’s problem

Several tracks were followed to answer this question.

IV.1.1. Overall checking

An overall checking of the basic parameters (pH, T, Turbidity, and flows) of the main steps of the process was done, and all parameters were found constant (cf Appendix 4). Only a difference in turbidity between the two sedimentation tanks was noted.

IV.1.2. Precipitation experiments

In an attempt to prevent the reject water from the centrifuge to disturb the process in the whole plant, precipitation experiments were conducted on this water by means of addition of Aluminium Sulphate, Ferric Chloride, variation of pH, addition of polymers (non-ionic, cationic and anionic) or combinations of them (Appendix 3). It turned out that the turbidity (and thus the content) of the reject water varied a lot (by a factor 30 for the extreme values) and could vary by of factor 9 from day to day. The precipitation results were not satisfying anymore when the turbidity of the reject water was too high (above 150-200 NTU), whatever the combination tried for precipitation.

During those precipitation experiments, extra sludge was brought to the plant to study its effect on the process. It appeared that extra sludge, increasing the concentration of the reject water, disturbed the usual functioning of the plant, but that it was not the only parameter, as high turbidity values in both sedimentation tanks were observed even without any extra sludge brought to the plant.

Moreover, the high turbidity observed in the sedimentation tanks was not following any cycle and seemed to happen randomly. It was concluded that most probably the problem was coming from the plant itself, rather than from an external source (like heavily loaded water in the pumping station on Monday mornings after gathering of the weekend water for instance).

IV.1.3. Centrifuge checking

At this step of the study, it was most likely that the centrifugation was the weak process because of the high variation in turbidity and water content of the reject.

Thus, focus was made on every aspect of it: polymer mixing, maturation, addition, sludge inflow, outflow and running time (Appendix 5).

Quickly, the mixing and the maturation of the polymer was ruled out. These two steps are assured by a TOMAL Polymer make up unit PolyRex, which allows a mixing time and a maturation time of one day each, which is way beyond what is necessary to reach an efficient polymer mixing (usually from one to three hours).

The polymer addition is made by means of a pump, running always at the same speed, and delivering a rather constant amount of solution to the sludge going to the centrifuge.

Focus was thus made on the inflow of sludge to the centrifuge. For this occasion, the real nature of the sludge influent was found: a mix between the sludge from both thickeners. When these sludges were examined, the sludge from the

(26)

25 internal thickener showed a TS of 1,38-4,40%, while the one from the external thickener was no more than 0,05-0,80%. This explained the rather low TS of the sludge going to the centrifuge; around 2,5%.

This centrifuge checking also allowed the understanding of the circulation of the reject water from the centrifuge, which was a loop (cf Figure 6) implying addition of polymer at each tour. When this loop was pointed out, it was thought that a possibly high concentration of polymer in the reject water could affect the processes of the plant. It was then decided to break this loop and see what would be the effect on the plant, especially on the turbidity in the sedimentation tanks.

Note: All along the checking of the sedimentation tanks both the depth (using a Secchi disk¹) and the turbidity (using the turbidimeter) were measured. This was the occasion to check that the Secchi disk, sometime considered as a non scientific device, was actually really well working. When put on a depth = f (turbidity) diagram, the plot of the data gave a tendency curve with a coefficient of correlation of R² = 0,71 for both tanks (Appendix 5).

IV.1.4. Pilot scale precipitation

In order to break the loop the reject water was not recirculated anymore but was simply sent to the pond behind the plant after addition of roughly 20mg/l of Aluminium. Thus, no more heavily polymer concentrated water was sent to the entrance of the plant, and the turbidity in the sedimentation tanks was watched carefully to see any effect of the loop breakage.

This experiment had a double aim: first see if the loop was the weak point of the plant process, and second, try a pilot scale precipitation of the reject from the centrifuge into the pond.

One week was given to the process to reach equilibrium after what one week of measurement was done (cf. Table 7).

The first observation that can be made is that after the breaking of the loop, no more high turbidity values were observed. Even though accurate measurements were not made every day, checking was made by the operators of the plant that the Secchi depth was not shorter than 150cm. The same checking was made for 3 weeks, during which only 1 depth shorter than 150cm were found.

Three weeks is obviously a too short time to be sure that the loop was the problem disturbing the functioning of the plant, but still, the long Secchi depths observed in both sedimentation tanks are along the lines of this hypothesis.

The pond turbidity was measured at the end of the pipe sending the reject water to the pond. Turbidity before (3 first values) and after (3 last values) the sending of the reject water from the centrifuge has been measured. As the pond contains 400m³ of water, it will take about a month to replace all the clear water by the reject water and nothing can be assumed yet about the efficiency of the precipitation process. It will be necessary to keep checking the turbidity in the pond for at least the 2 next months in order to determine the real impact of the precipitation try. If it turns out that the precipitation is not efficient enough, sending of the reject water from the centrifuge to the pumping station should be tried, but with care to avoid any loop effect.

1 The Secchi disk is a circular disk used to measure water transparency (usually in oceans and lakes).

The disk is slowly lowered down in the water. The depth at which the pattern on the disk (or the disk itself) is no longer visible is taken as a measure of the transparency of the water. This measure is known as the Secchi depth and is related to water turbidity.

(27)

26 Day NTU reject Turbidity Sedi 1 Turbidity Sedi 2 Depth 1 Depth 2 NTU Pond

20th March 170 4,5 2,2 80 135

23rd March 180 9,8 4,9 60 90

24th March 170 7 5,6 70 80

26th March 180 5 3,9 90 110

31st March 1160 4,2 4 110 120

2nd April 1800 10 7 70 80

3rd April 872 10 9,6 60 70

14th April 240 1,8 1,3 170 200

15th April 160 1,6 1,6 190 200

16th April 350 4,4 4,2 120 130

17th April 925 2,5 2,5 160 160

20th April 1445 4,2 3,6 100 120

21st April 1095 3,4 3,1 110 130

22nd April 130 145

24th April 9,5 8,5 80 90

4th May Breaking of the loop

8th May 180 7,8 8,9 130 130 18

13th May 145 3,5 3,1 170 180 12

14th May 3 2,7 160 190 15

18th May 390 3 1,55 170 200 13/18

19th May 355 2,7 2,7 170 190 15/20

20th May 550 3 3 160 160 15/20

Table 7: Turbidity (values in NTU) and Secchi depths (values in cm) measured in both sedimentation tanks, and turbidity in the pond after the beginning of the pilot scale precipitation.

IV.2. Presses/Centrifuge comparison at Uddebo

A first comparison was made looking at the turbidity of the reject waters and their nitrogen content (the phosphorus content had not been found because it was over the LANGE detection limit). Results are shown in Appendix 6 and Figure 10.

Turbidity

0 100 200 300 400 500

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 min

NTU . Centrifuge

Press_1 Press2

Figure 10 : Turbidity of the reject waters from centrifuge and both presses. The values displayed are the average values for each measurement (range was about ± 10 NTU).

Reject Water Management at Wastewater Plants in Luleå

M A S T E R ' S T H E S I S