INOM
EXAMENSARBETE KEMIVETENSKAP, AVANCERAD NIVÅ, 30 HP
STOCKHOLM SVERIGE 2019,
Performance monitoring of systems for air purification
ANDERS GUSTAFSON
KTH
SKOLAN FÖR KEMI, BIOTEKNOLOGI OCH HÄLSA
Master of Science Thesis in Chemical Engineering
Stockholm, Sweden, 2017
Performance monitoring of systems for air purification
Author:
Anders Gustafson
Examiner:
Klas Engvall
Department of Chemical Engineering
Abstract
Wastewater treatment plants (W W T P ) are often the cause of malodor. The com- pounds which are the main causes of the odor is hydrogen sulfide (H2S), ammonia (N H3), mercaptans (RSH) and volatile organic compounds (V OC) [1]. The odorous air can be analyzed to determine the concentration of the odorants. The odorous can also be analyzed by measuring the odor. The odor is measured, a test panel of people smells the odorous air and determines how many times greater the concentration of the odorants is compared to the odor threshold which is the concentration at which a compound or a mixture is detectable by smell [3].
Measurements were done at three systems for air purification at three different locations, the Vimmerby W W T P , the Alvim W W T P and Renova’s biological waste treatment facility in Gothenburg. The odor was measured at the inlet and the outlet and the concentration of H2S and ozone (O3) were measured at all the sampling points of the systems. The system at Vimmerby consisted of three CIF s followed by an UV-reactor and an AC-reactor. In Alvim there were two system which used U V and AC. The system at Renova consisted of a barrier filter followed by U V and AC.
The system at the Vimmerby W W T P had a conversion rate between 87-97% of H2S. The CIF s had conversions between 50-64% of the H2S. H2S was not detected at any of the other systems. O3 was only detected at Renova where ground level O3 was present at the inlet, 0.16 ppm. The concentration increased to 0.20 ppm after the UV-reactor. The activated carbon could adsorb all the incoming O3.
The odor at the Vimmerby W W T P was determined to 27500 Ou/m3 at the inlet and 19071 Ou/m3 at the outlet. The odor conversion over the system was 31%. The odor conversion at the Alvim WWTP was 99.8%. With an odor of 5490 Ou/m3 at the inlet and 11 Ou/m3 at the outlet. The ingoing air at the system at Renova had an odor of 434 Ou/m3 and was reduced to 36 Ou/m3 at outlet. The odor conversion at Renova was 92%.
To increase the accuracy of the measurements he time between the sampling and measurements should have been minimized. The test panels should also have been larger and the panelists should have been screened in advance, so results from panelists which were over and under sensitive to odors were not included in the final results.
Sammanfattning
Vattenreningsverk ¨ar ofta en k¨alla till d˚aligt lukt. Dem kemiska f¨oreningarna som
¨
ar den huvudsakliga orsaken till d˚alig lukt ¨ar svavelv¨ate (H2S), ammoniak (N H3), merkaptaner (RSH) and l¨attflyktiga organiska f¨oreningar (V OC) [1]. Det g˚ar att m¨ata lukt genom att m¨ata koncentrationen f¨oreningarna som orsakar den d˚aliga lukten. Det g˚ar ¨aven att avg¨ora hur stark lukt ¨ar genom att en testpanel f˚ar lukta p˚a utsp¨adda luktprover. P˚a s˚a vis g˚ar det att best¨amma hur m˚anga g˚anger luften m˚aste sp¨adas ut f¨or att n˚a lukttr¨oskeln. Lukttr¨oskeln definieras som den koncentration som ett ¨amne eller en blandning g˚ar att detektera med hj¨alp av lukt [3].
M¨atningar gjordes vid system f¨or luftrening vid tre anl¨agningar, tv˚a
vattenreningsverk, Vimmerby och Alvim, och Renovas behandlingsanl¨aggning f¨or biologiskt avfall i G¨oteborg. Lukten m¨attes vid inloppet och utloppet av systemen.
H2S och O3 m¨attes vid alla av systemens m¨atpunkter. Systemet i Vimmerby bestod av tre katalytiska j¨arnfilter i serie f¨oljt av en UV-reaktor och aktivt kol. I Alvim fanns det tv˚a system med UV f¨oljt av aktivt kol. Hos Renova bestod systemet av ett partikelfilter f¨oljt av UV och aktivt kol.
Systemet vid Vimmerbys vattenreningsverk hade en oms¨attning mellan 87-97% av H2S. Oms¨attningen av H2S ¨over dem katalytiska j¨arnfilterna var mellan 50-64%.
H2S detekterades inte vid n˚agra av dem andra systemen. O3 detekterades endast vid Renovas anl¨aggning d¨ar markn¨ara O3 fanns i det ing˚aende fl¨odet, 0.16 ppm.
Koncentrationen ¨okade till 0.20 ppm efter U V -reaktorn. Det aktiva kolet klarade av att adsobera allt det inkommande.
Lukten vid systemet vid Vimmerby m¨attes till 27500 Ou/m3 vid inloppet och vid utloppet till 19071 Ou/m3. Oms¨attningen av lukten best¨amdes till 31% ¨over systemet. Oms¨attningen av lukten vid Alvim best¨amdes till 99.8%. Lukten vid inloppet var 5490 Ou/m3 och 11 Ou/m3 vid utloppet. Luften vid inloppet vid systemet hos Renova var 434 Ou/m3 och 36 Ou/m3. Oms¨attningen av lukten hos Renova var 92%.
Exaktheten f¨or m¨attningar kunde ha f¨orb¨attrats om tiden mellan provtagningen och analysen av provet hade minimerats. Testpanelen skulle ocks˚a haft flera deltagare som skulle varit gallrade i f¨orv¨ag s˚a att resultat fr˚an paneldeltagare som var ¨over och ok¨ansliga f¨or lukt inte fanns med i det slutgiltiga resultatet.
Contents
1 Introduction 2
1.1 Odor problems . . . 2
1.1.1 Odor Causing Compounds . . . 3
1.1.2 Hydrogen Sulfide (H2S) . . . 3
1.1.3 Mercaptans (RSH) . . . 4
1.1.4 Volatile Organic Compounds (V OC) . . . 4
1.1.5 Ammonia (N H3) . . . 4
1.1.6 Ozone (O3) . . . 4
1.2 Odor treatments . . . 5
1.2.1 Acid scrubber . . . 6
1.2.2 Ferrosorp . . . 6
1.2.3 Barrier filter . . . 6
1.2.4 Catalytic iron filter (CIF ) . . . 7
1.2.5 Ultra violet reactor (U V ) . . . 8
1.2.6 Activated carbon (AC) . . . 9
1.3 Campaign sites . . . 9
1.3.1 Vimmerby - Wastewater treatment plant . . . 9
1.3.2 Alvim - Wastewater treatment plant . . . 11
1.3.3 Renova - Biological waste treatment facility . . . 12
2 Method 13 2.1 H2S and O3 - Dr¨ager X-am 5600 . . . 13
2.2 Pitot tube . . . 14
2.3 Olfactometer - Scentroid SM100 . . . 14
2.4 Flame Ionization Detector (F ID) . . . 17
2.5 Site considerations . . . 18
3 Results and discussion 18 3.1 Analysis of O3 and H2S . . . 18
3.1.1 Vimmerby . . . 18
3.1.2 Alvim . . . 19
3.1.3 Renova . . . 19
3.2 Olfactory analysis . . . 20
3.2.1 Vimmerby . . . 20
3.2.2 Alvim . . . 21
3.2.3 Renova . . . 22
3.2.4 General thoughts on the olfacometer measurements . . . 22
3.3 F ID . . . 24
3.3.1 Vimmerby . . . 24
3.3.2 Alvim . . . 24
3.3.3 Renova . . . 24
3.3.4 General thoughts on the F ID measurements . . . 25
4 Conclusions 26
Appendix 1 29
Appendix 2 33
Appendix 3 40
Appendix 4 41
Abbreviations
• AC=Activated Carbon
• CIF =Catalytic Iron Filter
• F ID=Flame Ionization Detector
• GC − M S=Gas Chromatography–Mass Spectrometry
• N Ox=Nnitrogen Oxides
• P T F E=Polytetrafluoroethylene
• T HC=Total Hydrocarbons
• RSH=Mercaptans
• U V =Ultra Violet
• V OC=Volatile Organic Compound
• W W T P =Wastewater Treatment Plant
1 Introduction
Exhaust air from Wastewater Treatment Plants (W W T P ) are known to cause odor that are unpleasant for its surrounding area. The main sources of the odor come from volatile organic compounds (V OC), hydrogen sulfide (H2S), mercaptans (RSH) and ammonia (N H3) [1].
Measurements were conducted on three different systems for air purification. Two of the systems were connected to W W T P s. The third system was located at a
biological waste treatment facility. The concentration of H2S and O3 were measured before and after each step of the systems so the performances of the different
methods could be evaluated. A Flame Ionization Detector (F ID) was used to measure the Total Hydrocarbons (T HC) and how it changed between the inlet and the outlet of the systems. An olfactometer was used to measure the odor at the inlet and the outlet of the systems.
1.1 Odor problems
In the surrounding areas of W W T P s there are often problems with odors. The part of a sewage treatment plant which is the largest source of odor is usually situated close to the inlet of sewage water to the plant. Odorants in sewage water is transferred to air through mass transfer when sewage water is in contact with air. The reason why most of the odor comes from the inlet and the outlet is because of the turbulence that is generated at these points. [2] Sewage treatment facilities has several different methods for treating the ingoing water. Among the most common methods are aerobic and anaerobic digestion in an activated sludge. When aerobic digestion is applied the concentration of odorants in liquid phase is usually low because of the biological oxidation that occurs. However, an aerobic process can still be a big cause of odor because of the large surface area that is in contact with the atmosphere. Both aerobic and anaerobic processes lead to the formation of sludge. The formed sludge often lead to the formation of new odorants. [2]
As it often is with odors, the source usually consists of several different molecules which causes the stench. Many of these compounds have very low concentrations which makes it hard to measure them. Instead of using a normal chemical analysis of the contaminated air, sensory methods can be used. This type of analysis is also known as olfactory analysis. There are different ways to perform olfactory analysis.
In some versions, a sample of the air is taken and a panel evaluates the odor values at different dilution ratios. The concentration when a compound is detectable
through smell is called the odor threshold. However, there are problems with relying on human senses like smell. One of the reasons why it is hard to rely on the sense of smell is because everyone is not equally sensitive to odors. Some people are born with a lower sense of smells and other peoples’ sensitivity to a smell is lower because they have been exposed to it repeatedly. Another problem there is when smell is used for measurements is that a person’s perception of a smell decreases when they are exposed to it continuously. This phenomenon is called odor fatigue which means that a person’s perception of an odor will decrease during odor measurements. [3]
One way of quantifying odors is to use a concept called odor intensity. The basics
behind odor intensity is the relationship between concentration and the odor. There are two main models for odor intensity, the Weber-Fechner law (see equation 1) and the Stevens’s law (see equation 2). I stand for the odor intensity, C for the
concentration and a, b, k and n are constants. [2]
I = a ∗ log(C) + b (1)
I = kCn (2)
1.1.1 Odor Causing Compounds
When it comes to W W T P s, odors are mainly caused by V OCs, RSH, N H3 and H2S [4]. In this section, the processes that causes these compounds are described. It also handles what other health issues they can cause.
1.1.2 Hydrogen Sulfide (H2S)
H2S has a strong odor and is therefore a large contributor to odor problems. Its odor is often compared to the smell of rotten eggs [2]. Usually people starts to notice the smell of rotten eggs when the concentration is as low as 0.00047 and 0.0093 ppm [5]
The big difference in odor threshold can have several different reasons which have been mentioned in section 1.1. H2S does not just smell bad, but it also has other effects that can cause health issues for humans and animals. H2S can cause problems for the eyes, the respiratory and the central nervous system. [6] At extremely high concentrations, 1000-2000 ppm, exposure even causes instant death. When H2S is handled, it is important to keep in mind that it is highly flammable, and when it burns toxic compounds such as SO2 is formed. At concentrations of 100-150 ppm H2S is not detectable by smelling. [7]
H2S can be formed through reduction of sulfate (see reaction R1 and R2). [2]
SO42−+ organic matter anaerobic bacteria
−−−−−−−−−−→ S2−+ H2O + CO2 (R1)
S2−+ 2H+ → H2S (R2)
H2S can also be formed through desulphurization of organic compounds such as amino acids. One simplified example of when H2S can be formed can be seen in reaction R3 where cysteine react with H2O and form pyruvic acid, N H3 and H2S. [2]
SHCH2CH2N H2COOH + H2O → CH3COCOOH + N H3+ H2S (R3)
H2S can also be dissociated according to reaction R4. The amount of H2S depend a lot of the acidity of the water. Therefore, the magnitude of an odor problem depends on the acidity of the water in a sewage plant. [2]
H2S H++ HS− H++ S2− (R4)
1.1.3 Mercaptans (RSH)
RSHs are also called thiols and are often a contributor to odor problems. Methyl mercaptan for example has an odor similar too rotten cabbage [8]. When humans are exposed to methyl mercaptan it can cause an increase in blood pressure, hemolytic anemia, methemoglobinemia, coma and death at high concentrations. [9]
1.1.4 Volatile Organic Compounds (V OC)
Emissions of V OCs can come from many different types of sources, some industries that cause these emissions are petroleum refineries, food processors, fiber manufactur- ers, pharmaceutical plants and W W T P s [1]. V OCs is a wide group of compounds which includes several different types of compounds, therefore the effects of them can vary a lot. One of the problems that can come from V OCs is malodor. There are many different methods for treating emissions of V OC such as liquid adsorption, solid ad- sorbents, scrubbing, precipitation, filters, condensation, thermal incineration, plasma biodegradation and photocatalysis. [1].
1.1.5 Ammonia (N H3)
N H3 has an odor which smells like dry urine. Humans can usually detect N H3 at concentration of 50 ppm. But when it is detected with the senses there is problems with odor fatigue in the case of N H3. Therefore, it is desirable to not rely on the human senses for detection. It can also cause respiratory tract irritation and exposure should therefore be avoided. [10]
1.1.6 Ozone (O3)
O3 is a gaseous compound in room temperature with a sharp odor. O3 forms a layer in the stratosphere which play a vital part for life on earth. The role of the O3 in the stratosphere is that it protects the earth from incoming U V -light. The O3 that protects the earth from U V -light is formed by reaction R5 and R6. O3 can also get decomposed by light according to reaction R7. [11]
O2 + U V (< 242nm) → 2O (R5)
O2 + O + M → O3+ M (R6)
O3 + U V or visible → O2+ O (R7)
O3 also exists on the ground level and it is unfortunately toxic. Ground level O3 is formed through reactions R8 and R9. Reaction R10 decompose the formed O3 and keeps the concentration low. [12].
N O2+ hv → N O + O(3P ) (R8)
O(3P ) + O2 → O3 (R9)
N O + O3 → N O2+ O2 (R10)
When V OCs are present radicals can form and lead to the formation N O2 according to reactions R11, R12 and R13. The formed N O2 leads to an increase in the
concentration of O3 through reaction R8 and R9. [12]
V OC + OH → RO2+ products (R11)
RO2+ N O → N O2+ radicals (R12)
Radicals → OH + products (R13)
1.2 Odor treatments
There are several different methods for treatment of odors such as biotrickling filters, bioscrubbers and activated sludge diffusion reactors [3]. In biotrickling filters mi- croorganisms are suspended on materials such as ceramics, resins, plastics and rock.
Nutrient aqueous solutions goes in to the biotrickling filters and absorbs the odor- ants. The odorants are then consumed by the microorganisms. Bioscrubbers have an absorption towers where incoming odorous air is spray with an aqueous solution.
Then the solution and the absorbed odorants goes to a bioreactor where they are de- graded. Odorous air can also be treated by leading it in to the aeration tank used in the activated sludge process. There the odorous compounds can be degraded by the microorganisms. [3]
CentriAir is a company that has developed different solutions for handling of the compound that causes odor in outgoing air. The type of system which CentriAir makes are customized for what type and the concentration of the unwanted
compounds. Figure 1 is an overview of a possible system where an acid scrubber is the first step followed by a CIF /Ferrosrop, U V and AC. In this section, the different types of cleaning steps that were used at the systems were the field measurements were conducted will be described.
Figure 1: Schematic picture of a system
1.2.1 Acid scrubber
A scrubber can be used to remove several different contaminants from a gas stream.
The principle behind a scrubber is that the ingoing gas is sprayed with water which in contact with a liquid that and the contaminant is solved in the liquid. When the gas contains ammonia, a good option is to use an acid scrubber to remove it. An acid scrubber removes N H3 by absorbing it in the sprayed droplets. Then the adsorbed N H3 react with H2SO4 and from (N H4)2SO4 (see reaction R14). To be able to work properly an acid scrubber must work at low pHs (pH < 4). An acid scrubber can have the capacity to remove well over 90% of the ammonia contaminated gas streams. [13]
2N H3+ H2SO4 → (N H4)2SO4 (R14)
1.2.2 Ferrosorp
Ferrosorp is pellets made on F e(OH)3 that was developed by HeGo Biotec GmbH.
It is used to remove H2S from gas streams. H2S is absorbed according to reaction R15. Then the F e(OH)3 is regenerated when F e2S3 reacts with water and oxygen (see reaction R16). [14]
2F e(OH)3+ 3H2S → F e2S3+ 6H2O (R15)
F e2S3+ 1.5O2 + 3H2O → 2F e(OH)3+ 3S (R16)
1.2.3 Barrier filter
Barrier filter are usually used to remove dust from gas streams. There are different types of barrier filters, but the most common one is called fabric filters. The filters can be made from a lot of different materials such as cotton, glass and Teflon [15]. The gas streams that is cleaned passes through the filter and particulates gets stuck on the ingoing side of filter. With time, more and more particulates accumulate on filter and the pressure drop increases. As the layer of particulates builds up on the filter the filtration increases. However, the filters eventually must be cleaned before the pressure drop becomes too great. It can be done by shaking and vibrating. Another way of doing it is with the help of compressed air. Barrier filters problems with hot particulates which can damage the filter depending on what material the filter is made of. [15]
1.2.4 Catalytic iron filter (CIF )
CIF s were developed to remove H2S from gases. A CIF , has a simple setup where a gas containing H2S is passed through a packed bed with iron. For the process to work is has to be humid in the CIF so rust can form and react with H2S. In prevoius studies it has been shown that CIF s can acive removal rates of over 80% of the incoming H2S [16]. CIF s also have a long potential life span, in some cases even as long as 20 years, which makes it a good option to treat H2S. [16] Figure 2 is a picture of a rusted pall ring which can be used in a CIF .
Figure 2: Rusted pall ring
As mentioned before, rust has to form on iron for it to work. The reactions leading to the formation of rust can be seen blew in reactions R17, R18, R19 and R20. The formation of rust is an electrochemical process where reaction R17 and R18 are the cathodic and the anodic reactions. Then F e2+ and OH− react according to reaction R19. Then F e(OH)2 react with O2 and H2O from air and forms rust (see R20). [16]
4e−+ 2H2O + O2 → 4OH− (R17)
F e → F e2++ 2e− (R18)
F e2++ 2OH− → F e(OH)2 (R19)
F e(OH)2 + 1/2O2+ H2O → F e2O3· xH2O (R20)
The formed rust can then react with H2S according to reaction R21 and form F eS.
H2S can also react and form F e2S3 (see reaction R22). However, F e2S3 is not a stable molecule and quickly starts to decay to F eS (see reaction R23). [16]
F e2O3· xH2O + 3H2S → 2F eS + 3H2O + S (R21)
F e2O3· xH2O + 3H2S → F e2S3+ 3H2O (R22)
F e2S3 → 2F eS + S (R23)
The last part of a CIF process is the regeneration of the rust that were consumed by H2S. The rust is regenerated when F eS reacts with oxygen according to reaction R24 and R25. During these reaction heat is formed. In some instances, the formation of heat has been so great that it has caused explosions [16]. This can happen when too much F eS is in contact with O2. For this to be an issue the concentration of H2S must be greater than 10000 ppm. [16]
4F eS + 3O2 → 2F e2O3+ 4S + heat (R24)
4F eS + 7O2 → 2F e2O3+ 4SO2 + heat (R25)
The free sulfur that forms in reaction R21, R23 and R24, forms a layer which prevents H2S and rust from reacting with each other. This problem is solved by washing the reactor with water regularly. [17]
1.2.5 Ultra violet reactor (U V )
As mentioned before in section 1.1.4, it is important to remove V OCs from gas and liquid streams. The most common methods all have some problems associated with them and most of them are not cost effective. One possible method that can solving this is to use U V as a cleaning step to get rid of V OCs. In many cases U V -light is combined with a photocatalysis like T iO2 [17]. U V -light can also be used to reduces the amount of H2S in a gas stream [18]. To get the best possible cleaning of a gas, U V is often combined with other cleaning steps. [17] For example tests have been conducted where U V was used before a biofilter to treat V OC emissions [19].
When an organic molecule is hit with U V -light, several different photolysis processes can happen like direct photolysis where the U V -light directly breaks down the compound. There is also sensitized photolysis which happens because of transfer of energy through photochemically excited molecules to another molecule. [17]
Another process which photolysis can cause is oxidation which happens via addition or substitution. One way this can happen is by hydroxyl radicals which are formed when O3 absorbs U V -light according to reaction R26 and forms an excited single oxygen atoms. The excited single oxygen atoms can then react water and from hydroxyl radicals (see reaction R27). The hydroxyl radical can then attack an organic molecule according to reaction R28. [17]
O3 + U V (< 310nm) → O(1D) + O2 (R26)
O(1D) + H2O → 2OH· (R27)
R − H + OH· → R · +H2O (R28)
U V -light also causes O3 to form (see reactions R5 and R6 in section 1.1.6). The formed O3 can react with V OCs and break them down. This process is called
ozonation. However, it is a much slower process than photolysis and oxidation where hydroxyl radicals are formed like in reactions R26, R27 and R28. [17]
1.2.6 Activated carbon (AC)
AC can be used to clean both gases and liquids and is often used as a polishing step at the end of a cleaning process. It has a large active surface area that can be larger than 1500 m2/g, which can adsorb compounds in the incoming stream [15]. The surface of AC is mostly hydrophobic and therefore it adsorbs lipophilic substances more easily than other substances. Another factor which affect which compound that will be adsorbed is the boiling point and relative concentration. If the ingoing gas has a relative humidity higher than 65-70%, hydrophobic compounds starts to desorb. To solve this issue the ingoing stream can be dried before it comes in contact with AC.
The main drawback with AC is that is must be replenished after a while, since its ability to adsorb decreases the more it adsorbs. [15]
AC can be produced from oil and coal which have a high bitumen content and organic material such as peat, wood, bone and coconut shells. When AC is produced, these materials is heated at 1000◦C, where they undergo partial oxidation and pores are formed. AC can be impregnated to improve its adsorption properties. [15]
1.3 Campaign sites
The field tests were conducted at three different locations, the Alvim W W T P , the Ren- ova biological waste facility and the Vimmerby W W T P . In this section, an overview of the different facilities is given.
1.3.1 Vimmerby - Wastewater treatment plant
The Vimmerby W W T P is run by a company called Vimmerby Energi och Milj¨o. The plant takes in wastewater from households and industries. The industries that send their industrial wastewater to the plant are Arla, which has a milk production plant in Vimmerby, and the ˚Abro brewery. The household wastewater comes from households in the Vimmerby municipality, the facility uses coarse screening, pre-settling, bio bed, activated sludge, chemical precipitation and filtration to treat incoming water. They also produce biogas with three reactors with the volumes of 2500, 1400 and 900 m3. [20]
Gas from the Vimmberby W W T P is treated by a system which is schematically described in figure 3 below. It consists of three CIF s in series and a U V -reactor and polishing step with AC. The three CIF s uses pall rings with a diameter of 25 mm.
The total volume of the pall rings in the three reactors are 5 m3 with a total weight of 2.1 tons. The reactors are also fitted with spray pipes that are used for rinsing.
The CIFs were installed on June 3, 2015. The AC is in this case not provided by CentriAir. It comes from a company called Evodor.
Figure 3: Overview of the system in Vimmerby W W T P
Figure 4 is of the CIF s at Vimmerby W W T P . The U V -reactor can be seen in figure 5.
Figure 4: CIF s at Vimmerby W W T P
Figure 5: U V -reactor at Vim- merby W W T P
The AC from Evodor can be seen in figure 6 below.
Figure 6: AC at Vimmerby W W T P
Biogas production can lead to bad odors which is caused by sulphides mercaptans, amines, nitrogen compounds, aldehydes and organic acids. The amount of the odor causing compounds depends on the ingoing substrate, the bacterial composition in the reactor and how to process is managed. [21] From previous measurements it is know that gas contains H2S at concentrations as high as 200 ppm. It also contains RSH and N H3.
1.3.2 Alvim - Wastewater treatment plant
The Alvim W W T P is situated in the Sarpsborg municipality, Norway. At Alvim the incoming water is treated with several different processes, a step screen, sand filter where ironcloride and polymers are added to precipitate phosphor. The water is then cleaned through flocculation and sedimentation. The outgoing sludge from the sedimentation basin then goes through an aerobic reactor where methane is formed and then used to heat up the facility.
The plant has two systems with an U V -reactor followed by an AC (see figure 7).
The old system is located in the old part of the W W T P and handles a volumetric flow rate of 13500 m3/h and has 18 200 W U V -lamps. The U V -reactor in the old system uses 3500 kg of AC. The new system handles a flow of 7000 m3/h and uses 12 200 W U V -lamps. The new system uses 1750 kg of AC. The last time the AC was changed in both systems was in June 2016.
Figure 7: Overview of the system in Alvim A picture of the new system can be seen in figures 8 and 9
Figure 8: The new system at Alvim seen from the inlet
Figure 9: The new system at Alvim seen from the outlet
1.3.3 Renova - Biological waste treatment facility
The Renova facility is located in Gothenburg at the shore of the G¨ota ¨alv river. The facility handles biological waste that comes from food manufacturers, distributors and stores. The incoming waste is grinded, compressed, mixed and filtered into a slurry.
The slurry is then sent away to where it can be used to produce biogas and fertilizers.
Some of the waste is sent away for incineration.
The system at Renova consists of a barrier filter followed by three U V -reactors in parallel with 12 200 W U V -lamps in each of them. After the U V -reactors there are three AC reactors in parallel. The AC has a volume of 11 m3 and weights 5.5 tons.
The last time the AC was changed was on April 27, 2016. The air which goes through the system comes from Renovas ventilation system.
Figure 10: Overview of the system at Renova
Figure 11 was taken at Renova and shows the U V and the AC-reactors.
Figure 11: Picture of the U V and the AC-reactors at Renova.
2 Method
This section describes the equipment that was used and how it was used.
2.1 H
2S and O
3- Dr¨ ager X-am 5600
A Dr¨ager X-am 5600 is a small handheld device which can be used to detect com- bustible gases and organic vapors. It can also detect O2, CO, H2S, CO2, Cl2, HCN , N H3, N O2, P H3 and SO2. Depending on what is to be measured, different sensors must be used. [22]
Dr¨ager X-am 5600 was equipped with an XXS H2S HC sensor for measuring H2S. It can detect H2S at concentrations spanning from 0 to 1000 ppm with a resolution of 2 ppm. Its sensitivity is ±1 % of the value that is measured. However, the sensor can run into problems if the gas contains P H3, N O2 and SO2. An XXS H2S HC sensor is a small electrochemical cell with an anode, cathode and reference electrode [23].
O3 was measured with an XXS Ozone sensor. It can measure O3 in a range of 0-10 ppm. Its resolution is 0.01 ppm and the sensitivity is ±3 % of measured values. [23]
Figure 12: Dr¨ager X-am 5600 measuring device
The Dr¨ager X-am 5600 was used with the X-am 125 external pump unit to measure H2S and O3 at the different sampling points in the systems.
The X-am 5600 did not display the correct concentration of H2S. To overcome that problem the device measured a few known concentrations and a linear plot were adjusted to the data points. X in equation 3 is the measured value and Y is the correct value.
Y = 1.2466X − 1.6144 (3)
2.2 Pitot tube
The velocity of the air can be measured with the help of a pitot tube. A pitot tube has a hole on its tip which is used to measure the total pressure. On the side of the tube there are holes for measuring the static pressure. When the air velocity is measured, the hole for the total pressure is placed in the opposite direct of the flow. The velocity can be calculated with equation 4 which is derived from the Bernoulli equation. P1 is the static pressure, P2 the total pressure, ρ the density of air and v the velocity of air. [24]
v = s
2(P2 − P1)
ρ (4)
The volumetric flow rates of the systems were determined with a pitot tube
connected to a Testo 510 for the different points of the systems. The density of gas was assumed to be to be the same as pure air at the different temperatures. The temperature was measured at all the sampling points.
2.3 Olfactometer - Scentroid SM100
The Scentroid SM100 has an air tank with odorless air that is used to dilute the air that it is analyzed instead of using filters to dilute the air which is analyzed, which is a common method for other olfactometers. The Scentroid SM100 measures Ou/m3 which is defined by equation 5. A basic setup for the olfactometer can be seen in figure 13 below. [25]
D = C/T (5)
Where D stands for odor units, C concentration and T for odor threshold.
Figure 13: Scentroid SM100
Air was sampled with 10 l polytetrafluoroethylene (P T F E) sampling bags at the inlet and the outlet of the systems. The P T F E sampling bags were filled to 90 % of their capacity. The Scentroid SM100 was then used in the lab for the odor analysis.
Before the tests were conducted the olfactory meter was cleaned with nitrogen gas too reduce the amount of V OCs from previous measurements. Sampling was the last thing that were done at the different sites to minimize the time between the
sampling and the olfactory measurements.
The first thing the panelists did during the olfactory tests was to put on the mask and make sure that it fitted tightly. Then the valve on the air tank was opened fully.
The line pressure regulator was then adjusted to the pressure that the manufacturer has calibrated the instrument for, 80 psi (see figure 14). [25]
Figure 14: Line pressure regulator
The next step was to open the secondary shutoff valve. At this point the flow regulator valve was set to zero (see figure 15).
Figure 15: Open secondary shutoff valve
During a few minutes the panelists had to inhale fresh air through the mask so their noses were neutralized. Then the sampling bag was attached to the device and the valve on it was opened. The panelists then started to turn the knob until he or she smelled an odor. The point at which the odor was sensed were noted. The panelists then continued to turn the knob until the next number showed up to make sure that the he or she had felt an odor.
The olfactometer has different restriction plates that determines the dilution rate of the sample. A plate that dilutes the sample a lot can be used to examine a sample with a strong odor and a plate with a low dilution rate can be used when the odor is weak. The restriction plates can be seen in figure 16. These plates are located
beneath the coupling for the sample bag (see figure 17). Different restriction plates were used depending on how strong the odor was. The correlation between the odor and the position on the knob for the different restriction plates can be found in Appendix 4.
Figure 16: Restriction plates
Figure 17: Location for the restric- tion plates
After each measurement of the samples, a bag with nitrogen was attached to the device and the flow regulator was opened fully, so nitrogen could pass through the tubes.
The P T F E sampling bags were cleaned with nitrogen gas after they were used.
Nitrogen gas were pumped into the bags and were then pumped out of the bags multiple times to ensure that the bags were clean.
2.4 Flame Ionization Detector (F ID)
T HC is a measurement of how much carbon there is in a medium. It does not say what kind of compounds there is in the analyzed medium. It is a unit which can be used to analyze the quality of water, soil and air. One way of determining the T HC is to use a F ID. The F ID that was available at KTH was a Heated total hydrocarbon analyzer model VE 7 (see figure 18).
The working principle for a F ID is that the incoming gas is burned with a mixture or air and hydrogen gas. Carbon compound can form radicals which can react according to reaction R29 and form CHO+ and free an electron. By doing so the current that goes through the F ID will change depending on the amount of hydrocarbons that are burned. By measuring the current the amount of hydrocarbons can be determined. [26]
CH · +O → CHO++ e− (R29)
Figure 18: Heated total hydrocarbon analyzer model VE 7
The F ID was calibrated by attaching an oven bag containing known a concentration of acetaldehyde. To set the zero value for the F ID calibration, an oven bag with pure nitrogen was used.
The samples from the field systems were taken with the P T F E sampling bags.
T HC were measured multiple times for each P T F E sampling bag at different times.
A linear curve was then fitted to the results. By doing so the T HC at the time of the sampling could be approximated. The olfactometer measurements were
prioritized over the FID measurements, which meant that they were only performed if there was enough of the sample left after the olfactometer measurements.
2.5 Site considerations
The fields tests at Vimmerby W W T P took place on January 17 and 18, 2017. The air samples were taken in the afternoon on January 18 at the inlet and the outlet of the system. The U V -reactor at Vimmerby was not working because of a broken magnetron.
Measurements in Alvim took place on January 31, 2017. The U V at the old setup were not working. The air samples were taken in the afternoon at the inlet and the outlet of the new system only.
Measurements at Renova were conducted in the morning on February 1, 2017. The air was sampled before the barrier filter and after the AC.
3 Results and discussion
In this section, the results from the measurements at the different systems are pre- sented.
3.1 Analysis of O
3and H
2S
The results from the measurements with the Dr¨ager X-am-5600 are analyzed in this section. All the data from the measurements can be found in Appendix 1.
3.1.1 Vimmerby
The convertion of of H2S is displayed in figure 19. The system was able to remove at least 87% or more of the H2S even without a working U V -reactor. Even though the conversion rates were high, the concentration of H2S after the AC was still high enough to cause a strong odor (The H2S concentration after the AC was between 6 and 21 ppm). The CIF s perform well under all three of different conditions. The conversion over the CIF s increased from 50% to 62% when the flow rate decreased with 0.1 m3/s from 1.1 to 1.0 m3/s. This indicates that the residence time has a significant impact on the conversion of the CIF s. However, the highest conversion rate over the CIF s was at the highest flow rate, but during those conditions the inlet concentration was higher than during the other measurements. The conversion rate was highest at the highest inlet concentration because there was enough rust on the pall rings, so the amount of reaction sites could compensate for the shorter residence time. The correlation between an increase in inlet concentration and an increased removal of H2S is supported by Boon, AG and Boon, K [27]. All data can be found in tables 2, 3 and 4 in Appendix 1.
Figure 19: Conversion of H2S over the system at the Vimmerby W W T P
3.1.2 Alvim
Neither H2S or O3 were detected at any points at the two systems at Alvim. In the old system, the U V -reactor was not working so it could not form any O3. The absence of O3 at new system could possibly be because the sensor in the Dr¨ager X-am 5600 could not measure at low enough concentrations. The O3 that is formed in the reactor might break down to quickly so the sensor in the Dr¨ager X-am 5600 is not able to detect it.
3.1.3 Renova
The O3 concentration at the different points in the system at Renova can be seen in 1. There is O3 before the U V -reactor which indicates the presence of both V OCs and N Ox in the feed. The reason for this is that the production of ground level O3 is the formation of N O2 by reaction R11, R12. R13 (see section 1.1.6). The results also show that all the O3 which goes into the AC is adsorbed. All measuring data can be found in table 7 in Appendix 1.
Table 1: O3 concentrations at Renova Conc O3 [ppm]
Inlet 0.16
ex-Barrier filter 0.16
ex-U V 0.20
ex-AC 0.00
The Dr¨ager X-am 5600 should have been verified against known concentrations of O3 so the results could be validated and thereby confirm the generation of O3 at Renova and the lack of it at Alvim.
3.2 Olfactory analysis
The results from the olfactory analysis can be found in this part of the report. The measured odors shown in the figures 21, 22 and 20 is the mean values from the odor panels result. All data can be found in Appendix 2.
3.2.1 Vimmerby
The results from the olfactory measurements from the sample from Vimmerby can be seen in figure 20. The average odor at the inlet was 27500 Ou/m3 and 19071 Ou/m3 at the outlet. The odor conversion was 30.6%. The data from the measurements can found in tables 8 and 9 in Appendix 2.
Seven out of the eight panelists experienced an odor of 30000 Ou/m3 from the sample taken at the inlet. When an odor of 30000 Ou/m3 is measured with the Scentroid SM100, the odor is felt at the first mark which shows up when the knob is turned. This means that the odor is 30000 Ou/m3 or higher. Scentroid does not manufacture restriction plates which dilutes the sample more than plate 1, odors stronger than 30000 Ou/m3 could therefore not be measured. To overcome this problem the sample could have been diluted.
The results from four of the panelists were excluded from the overall results, the panelist who conducted the sampling and did all the field test at Vimmerby W W T P for almost two days, the two panelists that had worked with H2S previously and the fourth one which had could not detect an odor from the sample taken at the outlet.
Figure 20: Measured odor at Vimmerby
After each of the measurements from Alvim and Renova the sampling bag was removed and replaced with a bag filled with nitrogen. Then the knob that controls the dilution were opened fully. By doing so the olfactometer was cleaned from remnants from the previous test. By not doing this the remnants of the sample might have caused people to feel the odor too early during the tests on the air sampled at Vimmerby W W T P . Some of the panelists mentioned that the odor sometimes vanished when the dilution decreased and then reappeared later when the
dilution decreased even more. It can also explain that two of the excluded panelists detected a stronger odor at the outlet than at the inlet. All the other panelists detected an odor that was equal or weaker at the outlet compared to the inlet.
3.2.2 Alvim
The results from the olfactometer measurements on the air sampled at the inlet of the new system at the Alvim W W T P can be seen in figure 21. In figure 21 shows that the standard deviation is large compared to the measured odor. Results from four out of nine of the panelists who analyzed the sample taken at the inlet were excluded because of colds. Only one measurement was possible to conduct from the sample at the outlet which showed an odor of 11 Ou/m3 due to the lack of available sample. This means that the odor conversion was 99.8% when people with colds were excluded. All data from the measurements can be found in tables 12 and 13 in Appendix 2.
Figure 21: Measured odor at the inlet of the system at Alvim The results indicates that the system can reduce the odor of the incoming air.
However, the samples taken at Alvim W W T P were analyzed 45-50 hours after they were taken, which increases the uncertainty of the measure odor because of the sample changes in composition between the time of sampling and measurement.
3.2.3 Renova
The results from the olfactometer measurements from the inlet and the outlet of the system at Renova can be seen in figure 22. The results from four out of eight of the panelists which analyzed the sample taken before the barrier filter were excluded because of colds. Three out of the six panelists which analyzed the sample taken after the AC were also excluded because of colds. The odor conversion over the system was calculated to 91.8%. All data from the measurements can be found in table 16 in Appendix 2.
Figure 22: Measured odor at Renova
3.2.4 General thoughts on the olfacometer measurements
The tank with air which was used with the olfactometer empties quickly during use which means that the panelists was not always able turn up the concentration at their own pace. It also leads to a shorter time for the neutralization of the nose with clean air.
Before the tests the panelists should have been asked if they had eaten or had coffee in the last 30 minutes. If so, they should have had to wait until 30 minutes had passed.
The olfacometer measurements in the lab should instead have been conducted in double blind manner by having a person leading the tests that had no knowledge of where the samples were from, what part of the system they were from or even how they were supposed to smell and not by the person who took the samples. By doing it in a double-blind manner the influence on the panelists from the person who lead the measurements could have been reduced. The panelists should also have been screened in advance so people that are over and under did not participate.
To get more precise results each panelist should have gotten longer time so that restriction plates could have been changed which would have led to more accurate measurements when it was possible.
To avoid olfactory fatigue, no more than two samples should have been analyzed by each panelist during one day. The panelists had to analyze both samples from Alvim W W T P and Renova during the same day.
In figure 25 in Appendix 4, the calibration for the different restriction plates and how strong of an odor each position corresponds to is presented. When the odor is very strong restriction plate one must be used. The difference in odor between the first eight position is very large therefore the uncertainty is very high when air with a very strong odor is tested. This problem increases even more when the sample must be diluted like in the case of Vimmerby. This implies that the Scentroid SM100 should not be used to analyze very odorous air.
One problem that must be thought of when measurements are done with the Scentroid SM100 is the value it gives is based on the odor threshold. This can be problematic since it can vary a lot. According to one source H2S’s odor threshold has been determined to 0.00047 and 0.0093 ppm in different studies. [5]
There were also some issues with the mask. The mask does not fit that well which causes air from the surroundings to be inhaled. The mask is also made of a plastic which has a scent. The scent from the mask and the fact that scents from the surroundings could have interfered with the results measurements.
The compounds in the sampled air react and decompose with time and thereby alters the odor of the air. The olfactometer tests should have been performed in closer proximity to when the samples were taken to get more accurate results.
The big standard deviations in the results from the measurements can be explained by the panelists difference in perception of odors. The panelists should have been screened in advance so no one with a very strong or very weak perception of odors participated in the measurements. If this had been done, the number of outliers would have been minimized, which had resulted in lower standard deviations. To get an even higher degree of accuracy in the results more panelists should have
participated.
Measurements with the Scentroid SM100 has a high degree of uncertainty and gives only very rough values and when it is used to see if the odor decreases over a system.
The calculated conversions should be viewed as indication of a decrees of the odor rather than an exact number.
3.3 F ID
T HC were plotted against time since the sample were taken and a line were fitted to the data points. By doing so, T HC at the time of sampling could be deducted if it was assumed that the T HC decreased linearly with time. All the data can be found in Appendix 3.
3.3.1 Vimmerby
The results from the F ID did not seem reliable when it comes to the air samples taken in Vimmerby. The F ID showed values as high as 10000 ppm for the inlet which is not realistic. The reason for this could be the high concentration of H2S in the sample.
3.3.2 Alvim
The results from the sample taken at Alvim can be seen in figure 23. From these results the T HC at the time of sampling was determined to 24.437 ppm.
Figure 23: T HC measurements, Alvim
3.3.3 Renova
The measurements of T HC on the sample from the inlet at Renova were done 27 and 47.6 hours after the sample was taken. The linear plot that was made can be seen in figure 24. If the T HC decreased linearly the value a the time of sampling was 15.438 ppm.
Figure 24: T HC measurements, Renova
3.3.4 General thoughts on the F ID measurements
The conversion of T HC had been calculated if there had been enough air sampled to do both the olfactory and the F ID measurements. There were only five sampling bags available during the measurements at Alvim and Renova one of the bags could not be used because it contained small amounts of water. The bag that was used to sample the outgoing air at Renova was leaking so it was not possible to measure the T HC at the outlet. There was also not enough air available to measure the T HC more than one time for the outlet at Renova. The reason for this is that the odor was so weak that it was not diluted as much as samples that were more odorous which meant that more of the sample had to be used the olfactory measurements. To solve these problems more air samples should have been taken.
The samples that were taken in Alvim had to wait approximately 24 hours more than the samples from Renova before the first F ID measurements could be
performed. This meant that the error in the approximation of the T HC increased.
To avoid this the field tests should only have been done if it was possible to start the F ID measurements within 24 hours.
When the F ID was calibrated the concentrations of acetaldehyde was never lower than 30 ppm. The results from the F ID shows values that were lower than 30 ppm, therefore the F ID should have been tested for lower T HCs to confirm or discard the results.
The F ID measurements would have benefited from sampling of air at multiple points in the systems so the different steps could have been evaluated. The F ID only measures the T HC it would have been better to use gas chromatography–mass spectrometry (GC − M S) to analyze the content of the streams that goes through the systems. By doing so a better picture of which compounds there are that causes of the odor.
4 Conclusions
• The CIF s at the Vimmerby W W T P had a conversion between 50% and 64%
during the different conditions.The system was not able to remove all the re- maining H2S after the CIF s when the U V -reactor was not working.
• The working U V -reactor at Alvim did not generate any measurable amounts O3 that the available measuring device could detect.
• The O3 concentration increased from 0.16 to 0.20 ppm over the U V -reactor at Renova which was then adsorbed in the AC. The O3in the ingoing feed indicated the presences of V OCs which caused the formation of radicals which lead to the formation of N O2. The formed N O2 then caused the formation of O3.
• Because of lack of restriction plates which can measure strong odors accurately, the Scentroid SM100 should only be used to analyze air with an odor stronger than 3750 Ou/m3.
• A larger group of panelists should have been used to perform the odor tests. The groups should also have been screened in advance to remove people that are very sensitive and very insensitive to odors. By taking these measures the standard deviation would have been lower results would have been more accurate.
• The sampled air’s odors changes with time because the compounds in the samples react and decompose. The results from the odor tests could therefore have been more accurate if they had been done in closer proximity to when the samples were taken.
• The results from the olfactometer measurements had high standard deviations and should be seen more as an indication of a reduction of odor rather than an exact value.
• More measurements of T HC should have been done with the F ID to have more data points to fit a linear plot to. By doing so the approximation of T HC at the time of sampling would have been more accurate.
• T HC measurements should have begun in closer proximity to the sampling to lower the error of the measurement.
• The air should also have been analyzed with a GC − M S to get a better picture of which compounds it is that causes the odor.
References
[1] Suresh T. Gulati Ronald M. Heck, Robert J. Farrauto, editor. Catalytic Air Pollution Control. John Wiley & Sons, Inc, 2012.
[2] P Gostelow, SA Parsons, and RM Stuetz. Odour measurements for sewage treat- ment works. Water Research, 35(3):579–597, 2001.
[3] Raquel Lebrero, Lynne Bouchy, Richard Stuetz, and Ra´ul Mu˜noz. Odor assess- ment and management in wastewater treatment plants: a review. Critical Reviews in Environmental Science and Technology, 41(10):915–950, 2011.
[4] Faruk Dincer and Aysen Muezzinoglu. Odor-causing volatile organic compounds in wastewater treatment plant units and sludge management areas. Journal of Environmental Science and Health Part A, 43(13):1569–1574, 2008.
[5] Jon H Ruth. Odor thresholds and irritation levels of several chemical substances:
a review. American Industrial Hygiene Association Journal, 47(3):A–142, 1986.
[6] OSHA. Osha fact sheet, hydrogen sulfide (H2S). 2005.
[7] Occupational Safety & Health Administration. Hydrogen sulfide, https://www.osha.gov/SLTC/hydrogensulfide/hazards.html, 2016-02-24.
[8] Center for Disease control and prevention (CDC). Methyl Mercaptan (CH3SH) CAS 74-93-1; UN 1064.
[9] Center for Disease control and prevention (CDC). Occupational safety and health guidline methyl mercaptan, 1988.
[10] Public Health England. Ammonia toxicological overview, 2015.
[11] Christos Zerefos, George Contopoulos, and Gregory Skalkeas. Twenty Years of Ozone Decline: Proceedings of the Symposium for the 20th Anniversary of the Montreal Protocol. Springer Science & Business Media, 2009.
[12] S. Ted Oyama. Chemical and catalytic properties of ozone. Catalysis Reviews Science and Engineering, 2007.
[13] RW Melse and NWM Ogink. Air scrubbing techniques for ammonia and odor reduction at livestock operations: Review of on-farm research in the netherlands.
Transactions of the ASAE, 48(6):2303–2313, 2005.
[14] Ferrosorp, http : //www.f errosorp.de/english/, 2016 − 03 − 02.
[15] Per Olof Persson, editor. Cleaner production - strategies and technology for envi- ronmental protection. KTH - Royal Institute of Technology, 2011.
[16] D Brooker, I Evanson, and A Shammay. Rusty iron–a cost-effective odour control upgrade for overloaded systems. 2012.
[17] Madhumita Bhowmick Ray. Photodegradation of the volatile organic compounds in the gas phase: a review. Developments in Chemical Engineering and Mineral Processing, 8(5-6):405–439, 2000.
[18] Lan-Yan Xia, Ding-Hong Gu, Jing Tan, Wen-Bo Dong, and Hui-Qi Hou. Photoly- sis of low concentration H2S under UV/VUV irradiation emitted from microwave discharge electrodeless lamps. Chemosphere, 71(9):1774–1780, 2008.
[19] Gholamreza Moussavi and Madjid Mohseni. Using uv pretreatment to en- hance biofiltration of mixtures of aromatic vocs. Journal of hazardous materials, 144(1):59–66, 2007.
[20] Vimmerby energi och milj¨o, http : //vimmerbyenergi − miljo.se/, 2016 − 04 − 22.
[21] Anna Schnurer and Asa Jarvis. Microbiological handbook for biogas plants.
Swedish Waste Management U, 2009:1–74, 2010.
[22] Dr¨ager. Specialists for your safety portable gas detection from dr¨ager.
[23] Dr¨ager. Dr¨agerSensor & Portable Instruments Handbook, 2015.
[24] The engineering toolbox, http : //www.engineeringtoolbox.com/air − density − specif ic − weight − d600.html, 2016 − 04 − 25.
[25] IDES Canada Inc. Sentroid model SM100, operating and maintenance manual, 2013.
[26] Daniel C Harris. Quantitative chemical analysis. Macmillan, 2010.
[27] AG Boon and K Boon. Catalytic-iron filters for effective and low-cost treatment of odorous air. Water and Environment Journal, 13(3):189–194, 1999.
Appendix 1
Appendix 1 contains the results from the measurements with the Dr¨ager X-am 5600 at the different locations.
Table 2: Results from measurements with Dr¨ager X-am 5600 at Vimmerby W W T P Inlet ex-CIF 1 ex-CIF 2 ex-CIF 3 ex-U V ex-AC
Temperature Average [◦C] 20.2 12.2 8.9 5.1 17.6 9.1
Sample 1 20.2 12.2 8.9 5.1 17.5 8.9
Sample 2 20.1 12.2 8.8 5.1 17.4 9.0
Sample 3 20.4 12.1 8.9 5.2 17.8 9.5
∆P [Pa] 566 524 553 666 553 342
Sample 1 568 525 576 668 580 330
Sample 2 558 538 531 690 536 336
Sample 3 572 508 552 641 544 360
Flow rate [m3/s] 1.5 1.4 1.5 1.6 1.5 5.8
Residence time [s] - 1.2 1.1 1.0 - -
H2S average [ppm] 197.8 119.7 88.1 71.5 71.5 18.3
Sample 1 195.3 120.6 88.1 70.7 70.7 18.3
Sample 2 200.3 120.6 88.1 70.7 70.7 18.3
Sample 3 198.8 118.1 88.1 73.2 73.2 18.3
O3 0 0 0 0 0 0
Sample 1 0 0 0 0 0 0
Sample 2 0 0 0 0 0 0
Sample 3 0 0 0 0 0 0
Table 3: Results from measurements with Dr¨ager X-am 5600 at Vimmerby W W T P Inlet ex-CIF 1 ex-CIF 2 ex-CIF 3 ex-U V ex-AC Temperature Average [◦C] 20.9 16.8 12.2 9.8 27.8 12.7
Sample 1 19.0 16.8 12.5 9.7 28.0 12.5
Sample 2 22.0 16.8 12.3 9.8 27.7 12.9
Sample 3 21.8 16.9 12.8 9.8 27.8 12.6
∆P [Pa] 294 275 275 249 265 87
Sample 1 294 290 250 243 278 98
Sample 2 296 260 290 248 260 79
Sample 3 292 274 284 256 257 85
Flow rate [m3/s] 1.1 1.1 1.0 1 1.1 0.6
Residence time [s] - 1.6 1.6 1.7 - -
H2S average [ppm] 167.1 123.0 104.8 83.9 78.2 20.8
Sample 1 165.4 123.0 103.1 83.2 78.2 20.8
Sample 2 167.9 123.0 105.6 83.2 78.2 20.8
Sample 3 167.9 123.0 105.6 85.6 78.2 20.8
O3 0 0 0 0 0 0
Sample 1 0 0 0 0 0 0
Sample 2 0 0 0 0 0 0
Sample 3 0 0 0 0 0 0
Table 4: Results from measurements with Dr¨ager X-am 5600 at Vimmerby W W T P Inlet ex-CIF 1 ex-CIF 2 ex-CIF 3 ex-U V ex-AC Temperature Average [◦C] 21.1 14.0 12.5 10.5 30.3 7.3
Sample 1 21.1 16.1 12.4 10.5 30.4 7.2
Sample 2 21.0 14.5 12.5 10.5 30.2 7.3
Sample 3 20.9 11.5 12.6 10.6 30.3 7.4
∆P [Pa] 220 221 234 355 142 124
Sample 1 200 205 229 342 127 130
Sample 2 220 217 232 350 139 118
Sample 3 240 240 241 374 160 124
Flow rate [m3/s] 1.0 0.9 1 1.2 0.8 0.7
Residence time [s] - 1.8 1.7 1.4 - -
H2S average [ppm] 158.8 88.1 67.4 60.7 50.7 5.9
Sample 1 158.0 88.1 68.2 60.7 50.7 5.9
Sample 2 162.9 88.1 68.2 60.7 50.7 5.9
Sample 3 155.5 88.1 65.7 60.7 50.7 5.9
O3 0 0 0 0 0 0
Sample 1 0 0 0 0 0 0
Sample 2 0 0 0 0 0 0
Sample 3 0 0 0 0 0 0
Table 5: Results from measurements with Dr¨ager X-am 5600 at the new system at Alvim
Inlet ex-U V ex-AC Temperature Average [◦C] 21.3 22.2 23.1
Sample 1 21.4 22.3 23.1
Sample 2 21.3 22.1 23.1
Sample 3 21.3 22.1 23.1
∆P [Pa] 20 15 133
Sample 1 18 15 132
Sample 2 22 15 134
Sample 3 21 16 132
Flow rate [m3/s] 2.9 4.5 5.3
Residence time [s] - - -
H2S average [ppm] 0 0 0
Sample 1 0 0 0
Sample 2 0 0 0
Sample 3 0 0 0
O3 0 0 0
Sample 1 0 0 0
Sample 2 0 0 0
Sample 3 0 0 0
Table 6: Results from measurements with Dr¨ager X-am 5600 at Alvim’s old system Inlet ex-AC
Temperature Average [◦C] 17.0 17.9
Sample 1 16.9 17.9
Sample 2 17 17.9
Sample 3 17.1 17.9
∆P [Pa] 25 74
Sample 1 26 80
Sample 2 32 70
Sample 3 17 73
Flow rate [m3/s] 2.9 11.5
Residence time [s] - -
H2S average [ppm] 0 0
Sample 1 0 0
Sample 2 0 0
Sample 3 0 0
O3 0 0
Sample 1 0 0
Sample 2 0 0
Sample 3 0 0
Table 7: Results from measurements with Dr¨ager X-am 5600 at Renova
Inlet ex-Filter ex-U V reactor 3 ex-AC reactor 3 ex-AC
Temperature Average [◦C] 16.1 16.1 16.8 16.2 16.2
Sample 1 16.0 16.1 16.9 16.2 16.2
Sample 2 16.1 16.1 16.7 16.2 16.3
Sample 3 16.1 16.1 16.9 16.3 16.2
∆P [Pa] 23 34 28 148 69
Sample 1 16 35 30 145 69
Sample 2 23 34 28 148 68
Sample 3 29 33 26 150 70
Flow rate [m3/s] 3.0 29.6 7.3 9.7 31.6
Residence time [s] - - - 1.1 1.0
H2S average [ppm] 0 0 0 0 0
Sample 1 0 0 0 0 0
Sample 2 0 0 0 0 0
Sample 3 0 0 0 0 0
O3 0.16 0.16 0.20 0 0
Sample 1 0.16 0.16 0.20 0 0
Sample 2 0.16 0.16 0.20 0 0
Sample 3 0.16 0.17 0.20 0 0
App endix 2
Appendix2containsresultsfromtheolfactorymeasurementswithScentroidSM100. σ=qX (xi−µ)2/n(6) Table8:MeasuredodorattheinletofthesystematVimmerbyWWTP DatesampTimesamDatemeasTimemeasElapsedtime[h]Ou/m3 AverageOu/m3 Panelist118-Jan15:3019-Jan15:0823.6333022569 Panelist218-Jan15:3019-Jan15:2623.97500 Panelist318-Jan15:3019-Jan15:3424.130000 Panelist418-Jan15:3019-Jan15:4524.310000 Panelist518-Jan15:3019-Jan15:5324.430000 Panelist618-Jan15:3019-Jan16:0024.530000 Panelist718-Jan15:3019-Jan16:1224.710000 Panelist818-Jan15:3019-Jan16:1724.830000 Panelist918-Jan15:3019-Jan16:2324.930000 Panelist1018-Jan15:3019-Jan16:3225.030000 Panelist1118-Jan15:3019-Jan16:4225.230000 Panelist1218-Jan15:3019-Jan16:5225.430000Table9:MeasuredodorattheoutletofthesystematVimmerbyWWTP DatesampTimesampDatemeasTimemeasElapsedtime[h]Ou/m3 AverageOu/m3 Panelist118-Jan14:5319-Jan17:1026:31500016279 Panelist218-Jan14:5319-Jan17:1326:330000 Panelist318-Jan14:5319-Jan17:0526:230000 Panelist418-Jan14:5319-Jan17:4326:82 Panelist518-Jan14:5319-Jan16:2825:6569 Panelist618-Jan14:5319-Jan17:1526:415000 Panelist718-Jan14:5319-Jan16:5826:130000 Panelist818-Jan14:5319-Jan17:0026:130000 Panelist918-Jan14:5319-Jan17:2026:56000 Panelist1018-Jan14:5319-Jan18:1627:415000 Panelist1218-Jan14:5319-Jan17:3526:77500
Table 10: Odor at the inlet at Vimmerby when outliers were excluded Odor [Ou/m3] Frequency µ F req ∗ (x − µ)2 σ
10000 1 2750 306250000 5972
30000 7 6250000
Table 11: Odor at the outlet at Vimmerby when outliers were excluded Odor [Ou/m3] Frequency µ F req ∗ (x − µ)2 σ
6000 1 19071 170862245 9973
7500 1 133897959
15000 2 33153061
30000 3 358301020
Table12:MeasuredodorattheinletofthenewsystematAlvim DateSampTimesampDatemeasTimemeasElapsedtime[h]Ou/m3 AverageOu/m3 Panelist131-Jan14:1502-Feb11:2945.227003467 Panelist231-Jan14:1502-Feb13:4747.53750 Panelist331-Jan14:1502-Feb14:0747.93750 Panelist431-Jan14:1502-Feb14:2048.12 Panelist531-Jan14:1502-Feb14:3348.32 Panelist631-Jan14:1502-Feb15:4049.47500 Panelist731-Jan14:1502-Feb15:5049.67500 Panelist831-Jan14:1502-Feb16:0049.86000 Panelist931-Jan14:1502-Feb16:1049.92 Table13:MeasuredodorattheoutletofthenewsystematAlvim DatesampTimesampDatemeasTimemeasElapsedtime[h]Ou/m3 AverageodorOu/m3 Panelist131-Jan14:0502-Feb11:5245.81111
Table 14: Odor at the inlet at Alvim when outliers were excluded Odor [Ou/m3] Frequency µ F req ∗ (x − µ)2 σ
2700 1 5490 7784100 1957
3750 1 3027600
6000 1 260100
7500 2 8080200
Table15:MeasuredodorattheinletofthesystematRenova DatesampTimesampDatemeasTimemeasElapsedtime[h]Ou/m3 AverageodorOu/m3 Panelist101-Feb12:0202-Feb16:3028.52272 Panelist201-Feb12:0202-Feb16:5028.82 Panelist301-Feb12:0202-Feb17:2829.42 Panelist401-Feb12:0202-Feb16:4028.6280 Panelist501-Feb12:0202-Feb12:1024.1390 Panelist601-Feb12:0202-Feb17:1029.1390 Panelist701-Feb12:0202-Feb17:2029.3390 Panelist801-Feb12:0202-Feb17:0129.0721 Table16:MeasuredodorattheoutletofthesystematRenova DatesampTimesampDatemeasTimemeasElapsedtime[h]Ou/m3 AverageOu/m3 Panelist101-Feb12:1602-Feb18:0429.8230 Panelist201-Feb12:1602-Feb17:5929.728 Panelist301-Feb12:1602-Feb17:4729.545 Panelist501-Feb12:1602-Feb12:0023.760 Panelist701-Feb12:1602-Feb17:5529.719 Panelist801-Feb12:1602-Feb17:4029.428
Table 17: Odor at the inlet at Renova when outliers were excluded Odor [Ou/m3] Frequency µ F req ∗ (x − µ)2 σ
280 1 434 23778 150
390 3 5861
721 1 82254
Table 18: Odor at the outlet at Renova when outliers were excluded Odor [Ou/m3] Frequency µ F req ∗ (x − µ)2 σ
19 1 36 278 18
28 1 59
60 1 592
Appendix 3
This part contains the results from the measurements with the F ID.
Table 19: F ID inlet at Alvim
Date samp Time samp Date meas Time meas Elapsed time [h] T HC [ppm]
31-Jan 14:15 02-Feb 14:49 48.6 22.8
31-Jan 14:15 03-Feb 11:35 69.3 22.1
Table 20: F ID inlet at Renova
Date samp Time samp Date meas Time meas Elapsed time [h] T HC [ppm]
01-Feb 12:02 02-Feb 15:00 27.0 14.0
01-Feb 12:02 03-Feb 11:38 47.6 12.9
Appendix 4
Figure 25: Calibration table for the Scentroid SM100 [25]
