FOULING OF HEAT TRANSFER SURFACES WITH BIOBASED FEEDSTOCKS
Lappeenranta–Lahti University of Technology LUT Master's Programme in Biorefineries
Turo Tossavainen 2022
Examiners: Adjunct Professor Arto Laari Assistant Professor Kristian Melin
Biorefineries-maisteriohjelma Turo Tossavainen
Lämmönsiirtopintojen likaantuvuus biopohjaisilla syöttöaineilla
Kemiantekniikan diplomityö 2022
87 sivua, 52 kuvaa, 12 taulukkoa ja 3 liitettä
Tarkastajat: Dosentti Arto Laari ja Apulaisprofessori Kristian Melin
Avainsanat: Mäntyöljy, Mäntyöljypiki, lämmönsiirtopinta, likaantuminen, epäpuhtaudet
Diplomityön tavoitteena oli tutkia eri mäntyöljy- ja mäntyöljypikinäytteiden keskinäistä suhteellista likaavuutta lämmönsiirtimen lämpöpinnoilla kaupallisen likaantumistutkimus- laitteen avulla. Näytteiden keskinäisen suhteellisen likaavuuden lisäksi tavoite oli analysoida ja tunnistaa likaavuutta aiheuttavat komponentit tai yhdisteet eri syöttöaineiden osalta.
Syöttöaineina likaantuvuustesteissä käytettiin eri lähteistä hankittuja mäntyöljy- ja mänty- öljypikieriä, sekä näistä eristä eri tavoin pidemmälle prosessoituja näytteitä.
Koeajot suoritettiin pääasiassa kahdella eri virtausnopeudella pyrkien pitämään muut muut- tujat (Sisäänmenolämpötila, lämmityssauvan asetuslämpötila, paine) vakioituina. Lisäksi tutkittiin eri virtausnopeuden ja asetuslämpötilan vaikutusta likaavuuteen valikoitujen yksit- täisten näytteiden avulla.
Koeajojen aikana saadut likakakut analysoitiin IR ja EDS menetelmillä likaantumisen alku- lähteen selvittämiseksi. Koeajon aikana mitatun ulostulolämpötilan laskun avulla voitiin las- kea likaantumisvastukselle numeraalisia arvoja ja näiden perusteella arvioida näytteiden keskinäistä likaavuutta.
Työn kuluessa voitiin vahvistaa näytteiden prosessoinnilla olevan merkittävä kuumien pin- tojen likaantumista hidastava vaikutus. Edelleen voitiin osoittaa että merkittävin likaantu- misvastusta aiheuttava likakerros on kuumalle pinnalle syntyvä helposti irtoava geelimäinen kerros.
Master's Programme in Biorefineries Turo Tossavainen
Fouling of heat transfer surfaces with biobased feedstocks
Master’s thesis chemical Engineering 2022
87 pages, 52 figures, 12 tables and 3 appendices
Examiners: Adjunct Professor Arto Laari and Assistant Professor Kristian Melin
Keywords: Tall oil, tall oil pitch, heat transfer surface, fouling, impurities
Aim for the thesis work was to study relative fouling tendency of different crude tall oil and tall oil pitch samples on heat transfer surface with a commercially available fouling tester.
Further, the aim was to identify components and compounds causing fouling from different feeds.
Studied feeds consisted of different crude tall oil and tall oil pitch batches sourced from different suppliers and further processed samples from previous.
Test runs were conducted mainly with two different flow speeds, while keeping other vari- ables (inlet temperature, heater set temperature, pressure) constant. Additionally, different heater rod set temperatures effect to fouling resistance development speed was studied with individual samples.
Foulant deposits generated during test runs were analysed with IR and EDS to identify com- ponents causing the fouling. The monitored outlet temperature drop was used to calculate fouling resistance values, which was used to compare test samples fouling tendency.
During the work it could be verified that processing of samples is a feasible way to reduce heat transfer surface fouling. Further it could be demonstrated that the layer creating the most of the fouling resistance was a loosely attached gel-like layer that was easy to remove.
This thesis work was done for Neste Oyj between 08/2021 – 02/2022. These have been ex- ceptional times due to Corona pandemic, meaning that all in person interactions and meet- ings needed to be kept at minimum. Still, I feel that I’ve got full support and guidance for all the needed tasks and questions during the work.
I’d like to express my deepest gratitude to Annika Malm and Janne Asikkala, who have been guiding and pushing me constantly towards to better outcome. This project would not have been possible without Neste R&D and Arto Heiska allowing me this fantastic opportunity, or all You wonderful people in analytics and test laboratories in Technology Center.
Roman characters
A surface area [m2]
Cp specific heat capacity [J/kg K]
d diameter [m]
Q heat load [W]
qm mass flow [kg/s]
Rf fouling resistance [m2 K/W]
s tube wall thickness [m]
T temperature [ºC, K]
ΔT temperature difference [ºC, K]
U overall heat transfer coefficient [W/m2 K]
w flow velocity [m/s]
Greek characters
heat transfer coefficient [W/m2 K]
η dynamic viscosity [Pa s]
λ thermal conductivity [W/m K]
ν kinematic viscosity [m2/s]
Subscripts
c cold side
d tube inside diameter
i 1. inner (tube surface) 2. flow in
o 1. outer (tube surface) 2. flow out
rod heater rod
s sample, feed
Dimensionless numbers
Re Reynolds number
=
Pr Prandtl number
=
Nu Nusselt number
=
Abbreviations
ASTM American Society for Testing and Materials, standardizing organization CFA Crude Fatty Acid
CH Hydrocarbons
CTO Crude Tall Oil
C4S Thiophene
CVD Chemical vapor deposition DLC Diamond like carbon
GPC Gel permeation chromatography IC Ion Chromatography
ICP-MS/MS Inductively Couple Plasma – tandem Mass Spectrometry IDID internal diesel injector deposition
LMTD Logarithmic Mean Temperature Difference PECVD Plasma enhanced chemical vapor deposition PHE Plate Heat Exchanger
ppb parts per billion ppm parts per million
SEM/EDS Scanning Electron Microscopy / Energy Dispersive X-ray Spectroscopy UV Ultra Violet-light
TEMA Tubular Exchanger Manufacturers Association ToF-SIMS Time-of-Flight Secondary Ion Mass Spectrometry TOFA Tall Oil Fatty Acid
TOP Tall Oil Pitch
TPD/ Temperature Programmed Desorption /
TPO-MS Temperature Programmed Oxidation - Mass Spectrometry
Table of Contents
Tiivistelmä Abstract
Acknowledgements
Symbol and abbreviations list
1 Introduction ... 3
2 Tall oil processing ... 5
2.1 Crude tall oil ... 7
2.2 Tall oil fractionation ... 9
2.2.1 Tall Oil Fatty Acids ... 11
2.1.2 Tall Oil Rosin ... 13
2.1.3 Tall Oil Pitch ... 15
3 Fouling of heat exchanger surfaces ... 17
3.2 Heat exchangers ... 20
3.2.1 Typical heat exchanger design workflow ... 21
3.2.2 Overall heat transfer coefficient ... 24
3.2.3 Flow conditions ... 27
3.3 Ways to mitigate fouling issues ... 28
3.3.1 Heat Exchanger types used in Heavy fouling applications ... 29
3.3.2 Coatings and modified materials ... 33
3.3.3 Chemical additives and inhibitors ... 36
4 Experimental part ... 37
4.1 Use and data processing of Falex Thermal Fouling Tester FT2 ... 37
4.1.1 Test run parameters ... 40
4.1.2 Heater rod temperature gradient and hotspot ... 45
4.1.3 Test data processing ... 46
4.2 Feed samples ... 48
4.3 Feed sample analysis ... 48
4.4 Test run results ... 53
4.5 Test run deposition analysis ... 72
4.6 Improvement opportunities for future studies with the Falex equipment ... 79
5 Conclusions ... 81
References ... 83
Appendix 1 Test run data charts ... 1
Appendix 2 Images from optical microscopy ... 1
Appendix 3 SEM/EDS Results ... 1
1 Introduction
Crude tall oil (CTO) is a by-product of sulphate pulping. CTO and crude tall oil fractions are 100% bio based chemicals. Global availability of CTO is about 2 000 000 metric tons of which 1/3 is produced in Europe (EU CTO – Added value study, 2016). CTO is refined by several companies in their tall oil refineries to produce value added pine chemicals and some liquid fuel production has also been established that utilizes CTO as a raw material. It is a very complex mixture of different rosin and fatty acids, neutral components, volatile low boiling compounds and impurities (Norlin, 2000; Tikka, 2008). CTO is fractionated in a tall oil refinery to five main fractions; tall oil rosin (TOR), tall oil fatty acid (TOFA), distilled tall oil (DTO), tall oil pitch (TOP) and heads fraction, that is basically all low boiling fatty acids and volatile compounds (Huibers, 2000).
Information in literature about detailed CTO and TOP composition and especially impurity profiles is scarce. On the other hand, both of the streams are such that their actual composi- tion can vary significantly. Crude tall oil refiner’s collects their raw material (CTO) from several pulp mills and in many cases from several continents. A few variables that affect for overall CTO and TOP quality are the time of the year when the wood is harvested, changes in proportions of wood species fed into the pulp mill, storage time of wood on the wood yard prior chipping and feeding to cooking, variations in used amounts of chemical additives that the pulp mills use etc. Similar situation affects for TOP quality variations. Just the variations of CTO fed into the tall oil refinery affects the TOP quality, possible changes in the tall oil refinery’s internal production modes and process upsets will additionally affect the TOP quality. (Norlin, 2000; Tikka, 2008).
Aim of this work was to identify the impurities or components in CTO and TOP that affect the fouling tendency of heat transfer surfaces of heat exchangers used in tall oil processing and to find ways of mitigating fouling. During this study obtained CTO and TOP samples were analysed in detail to see proportions of main CTO fractions and to identify impurities.
Analysed samples were tested with fouling tendency test equipment to obtain numerical data about fouling behaviour. Main objective for the work was to identify the main components
in CTO and TOP that cause fouling on hot surfaces, to see how process conditions affect fouling rate and if pre-processing can be used to reduce fouling rate.
The thesis work is divided into two parts. In the literature part the process how CTO is sep- arated from black liquor is reviewed, as well as how impurities are carried to CTO and what the typical impurities in CTO and TOP are. Different CTO fractions are also briefly dis- cussed. Typical heat exchanger design workflow is described and most relevant parameters affecting fouling are discussed in more detail. On literature there are limited information available about actual heat exchanger types used in CTO fractionation plants. Literature re- ferred frequently to evaporators used in CTO fractionation, these evaporators transfer also major part of needed energy to distillation process. Evaporators used in CTO distillation are typically thin film and falling film evaporators as these types of evaporators minimize con- tact times of heated medium with hot surfaces and reduce decomposition of valuable com- ponents. Literature about actual heat exchangers used in CTO fractionation process was not identified during searches. However, CTO fractionation plant is known to use tube and shell heat exchangers in the past, when modern best practices recommend to use mainly plate and shell heat exchangers and spiral heat exchangers. Literature on fouling and prevention of fouling is widely available and some potential CTO and TOP fouling mitigation methods are also discussed in this thesis based on general fouling researches. However, literature about CTO and TOP compounds causing fouling and possible fouling mechanisms of them is scarce.
In the experimental part Falex thermal fouling tester configuration, its use and limitations are explained. Fouling tests were carried out with crude and processed feeds and the col- lected measured test data was processed to calculate fouling resistance and overall heat trans- fer coefficient values. In the literature part one potential fouling inhibitor was identified, namely thiophene, hence some testing was done to see if this inhibitor could reduce fouling also with tall oil based feeds.
2 Tall oil processing
Tall oil is a by-product of kraft pulping. During kraft pulping process wood fibres are re- leased and wood extractives are dissolved into cooking liquor. Depending on the geograph- ical area and wood species the yield of wood extractives is 2.5 – 3.5 wt-% per processed ton of wood (Tikka, 2008). The wood extractives disturb the pulping process if not removed from the cooking liquor stream. These can cause especially undesired foaming during pulp washing stage and so they need to be removed from chemical circulation, mainly in the evaporation plant. (Norlin, 2000)
During alkaline cooking process conditions, tall oil components are saponified to sodium salts and the resin and fatty acid salts will dissolve into black liquor. In the evaporation plant black liquor is evaporated and concentrated in several effects, each effect evaporates water out from black liquor and hence liquor density changes. A typical process flow sheet to separate soap from black liquor and further processing of soap to tall oil is presented in Figure 1. When the dissolved resin and fatty acid salts reach the critical concentration, they start to form micelles (Tikka, 2008). Micelles are lighter than condensed black liquor and separates out in the evaporation plant intermediate tanks. Some flocculants may be used to improve separation of resin and fatty acid salts from concentrated black liquor (Huibers, 2000; Tikka, 2008). According to Huibers (2000) two thirds of tall oil soap consist of valued wood extractives and last third is other impurities. These impurities can be captured inside formed micelles, between micelles or between resin and fatty acid salt layers. Captured im- purities consist of black liquor, fibres, lignin, sodium sulfate, sodium sulphide, sodium- and calcium carbonates and other salts (Huibers, 2000; Tikka 2008). Also, remains of used chem- icals (soap flocculants, pulp washing stage anti-foaming agents) can be assumed to be trapped in tall oil soap.
Figure 1 - Tall oil soap collection and acidulation to CTO (modified from Aro and Fatehi, 2007; Tikka, 2008).
After the tall oil soap is separated, it is acidulated to CTO. Traditional process to acidulate tall oil soap is to react it with diluted sulphuric acid at temperature of 93-97 °C. (Norlin, 2000; Tikka, 2008). The overall acid consumption of the process is 200-300 kg per produced ton of CTO. Tall oil soap reaction with sulphuric acid is presented below (from Aro and Fatehi, 2007).
2 R-COO-Na+ + H+2SO-4 2 R-COOH + Na2SO4 (Eq. 1)
To lower the sulphuric acid consumption, pulp mills have adopted carbon dioxide acidula- tion as a first acidulation step to split tall oil soap to CTO. Carbon dioxide is mixed with tall oil soap and water to reduce the soap mixture pH. With this method pH of 7-8 is reached, but the method cannot take the process till completion. The sulphuric acid step is still needed to react the remaining tall oil soap mixture till the pH 3-4 is reached. (Aro and Fatehi, 2017).
Chemical reactions of tall oil soap, when splitting it with carbon dioxide and water, are pre- sented below.
2 R-COONa + CO2 + H2O 2 R-COO-Na+ + H+2CO-3 2 R-COOH + Na2CO3 (Eq. 2)
During the acidulation process micelles are dispersed and formed rosin and fatty acids create reversed micelles that has polar core and hydrophobic outer surface. These reversed micelles start to coalesce and separate as a top layer in the reactor, in case of a batch reactor. Some lignosulfonate based dispersants might be used to speed up the separation. During reversal orientation of micelles, the core captures water and some impurities in the newly formed micelle core. (Huibers, 2000).
2.1 Crude tall oil
Crude tall oil is a complex mixture of different compounds that is produced as a byproduct of chemical pulping. Appearance of CTO is viscous light to dark brown coloured liquid that has a pungent smell caused mainly by sulphur compounds. CTO contains plenty of impuri- ties after acidulation, Water content can be as high as 2-3 %, lignin content 0.9 %, solid sodium sulfate 0.4 % (Huibers, 2000) and ~1 % of low boiling volatiles, mainly turpentines and other gaseous hydrocarbons (Norlin, 2000). According to Huibers (2000) these impuri- ties could be filtered or centrifuged out from CTO not to cause clogging problems down- stream the process. Typical composition of CTO gathered from different geographical areas is presented in Table 1. The acid and saponification numbers are defined by titrating a solu- tion with KOH. The difference of the two numbers is that acid number gives the amount of KOH needed to neutralize the solution, when the saponification number gives amount of KOH consumed to further saponify all carboxylic acids.
Table 1 - Typical CTO composition (modified from Norlin, 2000; Huibers, 2000).
Property Scandinavian
typical mix
United States, pine
Canada, mixed France
Acid number [mg KOH/ g]
138 - 145 165 - 167 128 - 140 165
Saponification number [mg KOH/ g]
160 172 165 172
Volatiles 1 w-% 1 w-% 1 w-% 1 w-%
Fatty acids total 43 – 45 w-% 45 – 46 w-% 38 – 42 w-% 39 w-%
Palmitic 1.5 3 2 2
Stearic 0.5 1 1 1
Oleic 10 20 10 15
Linoleic 17 13 15 11
Pinoleic 5 1 3 1
Arachidic 1.0 0.5 1 0.5
Other 10 6.5 10 9.5
Rosin 29 – 30 w-% 40 – 42 w-% 28 – 30 w-% 50 w-%
Pimaric 2 3 2 5
Palustric 4 7 5 7
Isopimaric 2 4 4 4
Abietic 12 15 10 18
Dehydroabietic 5 4 4 5
Neoabietic 4 6 3 6
Others 1 3 2 5
Neutrals 24 – 27 w-% 12 – 13 w-% 27 – 33 w-% 10 w-%
Moisture < 2 w-%
Ash < 0.2 w-%
Sulphuric acid < 0.02 w-%
Sulfur < 0.3 w-%
Typical CTO consists of 1 w-% of light volatiles, 38 – 46 w-% of fatty acids, 28 – 50 w-%
of rosin, 10 – 33 w-% of neutrals, 2-3 w-% of water and 0.5 – 1 w-% of inorganic impurities.
The neutrals consist mainly of resin and wax alcohols, fatty acid esters, sterols and hydro- carbons. The majority of the mentioned neutrals have either higher or lower boiling point as rosin fraction, except wax alcohols that has similar boiling point (Huibers, 2000). The phys- ical properties of CTO are presented in Table 2. CTO density, viscosity and specific heat vary greatly depending on rosin content (Drew & Propst, 1981).
Table 2 - Typical physical properties of CTO (Norlin, 2000).
Min Max
Boiling point, Atm [C] 180 270
Heat of Vaporization [kJ/kg] 290 330
Specific heat [kJ/kg] 2.1 2.9
Heating value [MJ/kg] 33 38
Density at 20 °C [kg/m3] 950 1020
Viscosity at 70 °C [mm2/s] 25 40
2.2 Tall oil fractionation
Figure 2 below presents a flow chart of a dry tall oil distillation plant. First step in the process is feed dehydration where water and volatiles are flashed out from the feed. Removing low boiling components at this stage allows the downstream process to be operated under very deep vacuum. Dehydration operation takes place at temperature of maximum 200 °C and absolute pressure as low as 3000 Pa (Norlin, 2000).
Figure 2 - CTO Distillation plant flow chart (modified from Knowpulp 2021).
After the dehydration feed is heated in de-pitching stage subsequently with thin film evapo- rators to maximum temperature of 275 °C and pressure is further reduced as low as 800- 1300 Pa. During this step all other fractions than pitch are vaporized and fed to a rosin col- umn to split them further to own fractions in the following process steps (Norlin, 2000). This information agrees with the data that Drew and Propst (1981) have compiled about different tall oil components vapour pressures presented on Figure 3.
Wiped thin film evaporators are generally used in tall oil distillation plants as they minimize fluid contact time on hot surfaces and therefore suit well to prevent decomposition of heat sensitive fatty and resin acids (Bagby et al., 2003; Huibers, 2000).
Figure 3 - Vapour pressures of different CTO components (Drew and Propst, 1981).
2.2.1 Tall Oil Fatty Acids
Tall oil fatty acids (TOFA) are important raw material for bio-based chemicals. TOFA is used for paints and coatings, fuel additives, lubricants, oil field chemicals, surfactants, clean- ers, alkyd resins and dimer acids. (Forchem, 2021; Kraton, 2021.)
TOFA is a mixture of carboxylic acids, mainly oleic and linoleic acid, and typical commer- cial grades have fatty acids content of 90 – 98 wt-%, when balance is rosin and unsaponifi- ables. More detailed composition is presented on Table 3 and most common fatty acid chem- ical structures can be seen on Figure 4. Iodine value represents the amount of unsaturated fatty acids in bulk stream (Knothe, 2002).
Crude fatty acid (CFA) is a raw fatty acid stream got from rosin column and it’s fed to heads column for further purification and to separate heads, distilled tall oil (DTO) and TOFA to their own fractions. (Bagby et al., 2003)
Table 3 - Composition of commercial fatty acid grades from tall oil refining (modified from Bagby et al., 2003;
Huibers, 2000).
CFA TOFA
Acid number 190 – 194 197- 198
Iodine value [g I2 / 100g] 131 130
Fatty acids wt-% 90 - 92.8 98 – 98.8
Of which:
Palmitic acid 1.6 0.4
Margaric acid 0.7 0.7
Stearic acid 2.2 2.3
Oleic acid 42.3 46.4
Linoleic acid 34.8 36.3
Linolenic acid 12.7 10.3
Nonadecylic acid 1.1 1.1
Eicosadienoic acid 4.7 2.4
Rosin wt-% 4.5 – 7 1
Unsaponifiables wt-% 2.5 – 4 1.5
Figure 4 - Tall oil fatty acids chemical structures.
2.1.2 Tall Oil Rosin
Tall oil rosin (TOR) is the most valued tall oil fraction and it is typically upgraded further after being fractionated from CTO. During 2009 price of rosin derivatives was around 1600
$ per metric ton on average. Overall TOR production during 1995 was 420 000 ton per year.
Typical composition of TOR can be seen on Table 4. Most common resin acids chemical structures are presented in Figure 5. There are several paths to upgrade distilled rosin and vast amount of uses for modified rosin. Typical modified rosin types are rosinates, rosin esters, maleic acid adduct, phenol-modified rosin, dimerized rosins, hydrogenated rosins, disproportionated rosin, rosin alcohols and rosin acid nitrile. Due to vast amount of possible modification methods and great variation of physical properties of modified rosins they are widely used in very different kind of applications. Typical applications of modified rosins are additives for paper making, adhesives and hot-melt adhesives, glazing agents, inks, oil drilling fluids, metal working fluids, soldering etc. (Fiebach and Grimm, 2000; Class, 2010)
Table 4 – Typical compositions of US and Scandinavian tall oil rosin (modified from Norlin, 2000).
US Scandinavian A Scandinavian B
Acid number [mg KOH/g] 174 173 180
Fatty acids wt-% 2 4 2
Rosin wt-% 92 90 95
Of which:
Pimaric acid 2 3 2
Palustric acid 7 8 7
Isopimaric acid 8 6 6
Abietic acid 40 40 40
Dehydroabietic acid 20 22 23
Neoabietic acid 3 4 4
Softening point [°C] 75 66 75
Unsaponifiables wt-% 6 6 3
Figure 5 - Tall oil resin acids chemical structures.
2.1.3 Tall Oil Pitch
Tall oil pitch (TOP) is the bottom product of CTO distillation and consists mainly of terpene alcohols, fatty alcohols, fatty and resin acid esters, oligomers of them, lignin and sterols. Tall oil producers present typical TOP product values based on combustion use, seen in Table 5.
TOP is mainly sold back to the pulp mills where it is consumed in lime kiln as a renewable
fuel substituting fossil heavy fuel oil (HFO) use. (Holmbom & Erä, 1978; Kurzin et al., 2014)
Table 5 – Typical properties of commercial TOP grades (Kraton, 2017; Forchem 2011) Forchem Fortop 600 Kraton
PF60
Acid number [mg KOH/g] 60 90
Ash wt-% 0.3 < 0.3 Flash point, closed cup [ °C] 150 > 150 Moisture wt-% 0.1 < 0.2
Density at 50 °C [kg/m3] 950 950
Heat of Combustion [MJ/kg] 38 37.6
Sulphur wt-% 0.3 < 0.2 Pour point [ °C] 15 20
Viscosity at 50 °C [mm2/s] 470 100 - 300
Other than renewable fuel use, TOP has some niche applications where it is used as a drilling mud additive, corrosion inhibitor, building material and in road making. Direct incineration of TOP in general is getting less attractive due to governments constantly tightening maxi- mum allowed sulphur content on used fuels. (Kurzin et al., 2014).
3 Fouling of heat exchanger surfaces
In tall oil refining the feedstock needs to be heated up to desired processing temperature.
When heating complex mixtures, like CTO, to high processing temperatures, fouling could cause severe problems on heat transfer surfaces. Fouling could be thought as an additional unwanted insulation layer between fluids that is generated during operation. This unwanted insulation layer is described on calculations as fouling resistance (Rf). Increased fouling re- sistance leads to reduced overall heat transfer process efficiency. Eventually reduced heat transfer efficiency leads to situation where wanted output temperature is not reached or in case of outlet temperature has to be kept constant, wanted output capacity is not reached. In practice this means operability issues and loss of production. Fouling of heat exchanger sur- faces and mitigation of fouling are important issues that need to be taken in account during design work flow. Typical engineering approach to combat this issue is to oversize the heat transfer surface area by +25% …+200%, leading to additional heat exchanger manufacturing costs.
Fouling in general has severe impact on heat exchanger performance and costs raised from fouling issues is evaluated to be roughly 0.15… 0.25% of GDP in industrialized countries (Bansal et al., 2007). To get the feeling about the costs of fouling and what this actually means, according world bank statistical services UK’s, France’s and Germany’s gross do- mestic production was averaging to 3 trillion US$ per country by the time of writing the Bansal et al. article. This means that each of the three countries spent at least 4.5 billion US$
a year just for heat exchanger fouling related issues.
This chapter briefly discusses the engineering subtasks that have the most effect on fouling characteristics of heat exchangers, present a structured approach for fouling mitigation strat- egies and ways to mitigate fouling issue. Further, the most relevant ways to mitigate the fouling of the particular feedstocks studied in this work is discussed.
3.1 Fouling modes
According to Awad (2011) there are six fouling modes generally recognized in literature.
They are:
Particulate fouling
Crystallization or precipitation fouling
Chemical reaction fouling
Corrosion fouling
Biological fouling
Solidification or freeze fouling
Particulate fouling is accumulation of particles from process streams to heat transfer sur- faces. Non-uniform flow speed distribution across the flow channel could promote particu- late fouling. Reducing the amount of particles, e.g. with filtration, reduces particle concen- tration and therefore slower or even eliminate particulate fouling. (Awad, 2011).
Crystallization fouling happens when soluble components e.g. salts reach saturation point locally on liquid stream. This type of fouling requires formation of primary nucleuses on bulk liquid, from where crystallization could start. Reducing amount of potential nucleus (particles, micelles) or reducing local supersaturation (high turbulence) could reduce crys- tallization fouling. (Awad, 2011).
Chemical reaction fouling happens when compounds in process stream react together with heat. Heat transfer surface material could promote chemical fouling acting as a catalyst in this reaction, but is not otherwise participating in foulants generation. (Awad, 2011).
Corrosion fouling is fouling mode when heat transfer surface material is participating on fouling as producing corrosion products acting as foulants. Surface material could react chemically or electrochemically together with fluid. (Awad, 2011).
Biological fouling is caused by micro- and macro-organisms sticking and growing on heat exchanger surfaces. Micro-organisms are e.g. algae, fungis and bacteria when macro-organ- isms mean generally significant bigger ones like mussels, clams and vegetation. (Awad, 2011).
Solidification fouling is caused when higher melting point components starts to freeze on subcooled heat transfer surfaces. (Awad, 2011).
Figure 6 – Effect of different fouling modes with different process conditions (from Muller- Steinhagen, 2011)
As seen on Figure 6, multiple fouling modes could participate to overall fouling behaviour simultaneously, one fouling mode could promote or even amplify the others. Increase in flow velocity in general reduces fouling tendency due to increased shear forces on heat trans- fer surface. Increase in temperature reduces the biological fouling, but promotes crystalliza- tion and particulate fouling. Due to this it is extraordinarily difficult task to precisely predict and create models how some fluid mixture will behave fouling wise on varying process con- dition. (Awad, 2011; Muller-Steinhagen, 2011).
Literature about fouling modes and foulants of CTO is scarce. Most closely related topics that could be found on literature and could have some analogy towards fouling caused by CTO are black liquor processing and especially evaporation plant fouling studies. There are plenty of literature about sodium- and sulfate salts scaling of evaporator effects (Aro et al., 2017; Karlsson et al., 2013; Nwaeri et al., 2021). Nwaeri et al. even studied different resin and fatty acids concentrations effect on sodium salt fouling of evaporators. However black liquor evaporators process conditions are quite different from CTO fractionation as they work with mainly aqueus solutions (min 15 w-% water) and in clearly lower temperatures (max ~160 °C). Some literature is available about other bio-oils fouling studies, but as they were mainly focusing fast pyrolysis bio-oils (Javaid et al., 2010; Mazerolle et al., 2021) or natural oils in wastewater treatment application (Ang et al., 2008; Zhao et al., 2016), these are discarded from this thesis.
3.2 Heat exchangers
Heat transfer and different kind of heaters/coolers exist all around us in homes, public and work places. Home equipment like refrigerators, air conditioning, and water heaters all rely on heat transfer to perform in their duties. This kind of services are low temperature and easy, flow mediums are somewhat clean and therefore they are almost invisible equipment in day-to-day life for their users due their low maintenance requirement. However, in indus- trial applications temperatures can be extremely high, flow mediums can be inconsistent and tend to burn, stick or solidify to heat transfer surfaces.
Therefore, industrial heat exchangers are generally specifically designed for their particular service and majority of these heat exchanger units are unique by their process values. During definition and engineering of heat exchangers their life time operability and maintenance work requirements will also be set and mistakes or unknown parameters in design phase can cause very high maintenance and operability costs due excessive fouling or not meeting planned process values and aimed heat duty.
3.2.1 Typical heat exchanger design workflow
Heat exchanger design workflow is certainly an iteration process where preliminary assump- tions in first steps will be evaluated and confirmed in detail during later calculation steps and due to this detailing initial values probably needs to be revised. Heat exchanger type and flow channel dimensions have direct impact on the flow velocities and therefore on flow pattern. This has impact on the overall heat transfer coefficient that might reveal that initially planned overall surface area is not sufficient and whole process needs to be started over.
(Saari, 2010; Shankar et al., 2017).
The following equation is the most fundamental and simplified equation to evaluate needed surface area for planned heat exchanger.
𝐴 =
(Eq. 3)Where A is surface area [m2], Q is heat duty [W], U is overall heat transfer coefficient and ΔT is temperature difference [K] (Nitsche et al., 2016). Heat duty Q can also be expressed as heat and mass balance.
𝑄 = 𝑞
,𝐶
,𝑇
,− 𝑇
,= 𝑞
,𝐶
,𝑇
,− 𝑇
, (Eq. 4)Where qm is mass flow [kg/s] of hot / cold side fluid, Cp is specific heat capacity [J/kg K] of hot / cold side fluid and T is temperature [K] in / out of hot / cold fluid (Saari, 2010). These Equations 3 and 4 are assuming that temperature difference between the hot and cold streams along the exchanger surfaces is constant. This however is rarely the case. (Saari, 2010).
One of the very first engineering decisions, flow directions of hot and cold streams, affect the most for available temperature difference. In co-current flow the temperature difference is at its highest when streams are entering the heat exchanger, but in the outlet of the heat exchanger the temperature difference has got narrower as we see in Figure 7 upper illustra- tion.
Opposite of this is counter current flow where hot and cold streams are flowing against each other and temperature difference change is not so dramatic as in co-current case as we see in Figure 7 lower illustration. In both cases temperature difference of streams keeps changing along the flow path. (Shankar et al., 2017; Saari, 2010).
Figure 7 - Streams temperature differences for co-current and counter current flows (from Mechanical inventions, 2018).
To achieve more precise calculations for heat duty and surface area needed, several more detailed temperature difference equations has been developed to model more precise heat transfer during constantly changing temperature difference. For this purpose, logarithmic mean temperature difference (LMTD) equation has been developed. (Saari, 2010).
𝛥𝑇 =
(Eq. 5)Where ΔTlm is logarithmic mean temperature difference [K], ΔT1 is temperature difference of inlet streams [K] and ΔT2 is temperature difference of outlet streams [K].
For counter current flow ΔT’s will be as presented in Equations 6 and 7.
Δ𝑇 = 𝑇
,− 𝑇
, (Eq. 6)Δ𝑇 = 𝑇
,− 𝑇
, (Eq. 7)For co-current flow ΔT’s will be as presented in Equations 8 and 9.
Δ𝑇 = 𝑇
,− 𝑇
, (Eq. 8)Δ𝑇 = 𝑇
,− 𝑇
, (Eq. 9)More precisely calculated ΔTlm can 𝑏e placed in Equation 3 to substitute ΔT and get more accurate estimate of needed heat exchanger surface area.
In case of multiple pass heat exchangers even more complex ways to estimate temperature difference between streams and correction factors for LMTD needs to be used for modelling heat duty more precisely. (Shankar et al., 2017)
3.2.2 Overall heat transfer coefficient
As shown in Equation 3, overall heat transfer coefficient is expressed as U. This representa- tion is, however, condensed and simplified equation from the several different components, or heat resistant layers seen in Figure 8, that affect for overall heat transfer resistance.
Figure 8 - Thermal resistance layers on fouled tube surfaces (Nitsche et al., 2016).
In the following equation the overall heat transfer coefficient U has been split to its actual components and it reveals how complex the overall heat transfer coefficient is. (Nitsche et al., 2016)
=
+
+ + 𝑅
,+ 𝑅
, (Eq. 10)Where h and c are heat transfer coefficients of the hot / cold fluid [W/ m2 K], s is wall thickness of the heat transfer surface [m], λ is thermal conductivity of the heat transfer wall material [W/ m K] and Rf,h and Rf,c are the fouling resistance values of the hot / cold heat transfer surface sides [m2 K/ W].
The wall thickness of the tube or plate s and thermal conductivity of wall material λ in Equa- tion 10 are set by decision about construction material. Thermal conductivities of different construction material are well known and values are readily available on heat exchanger design and simulation software material libraries (TEMA, Ansys, Aspen+).
The rest of the components in this equation have to be either determined experimentally or estimated from the available data and considered if the other conditions are similar so as- sumptions can be valid. In case of heat transfer coefficients (h and c), they need to be calculated from chosen flow velocity, heat exchanger flow channel dimensions and fluid properties. To be able to solve heat transfer coefficients for each fluid, three numbers Re, Pr and Nu have to be solved. Re is the dimensionless number characterizing flow conditions.
Pr is dimensionless number describing momentum and heat diffusivity. Nu is solved based on two previous ones with empirical formula created based on flow conditions and fluid service, heating or cooling. (Saari, 2010; Nitsche et al., 2016).
𝑅𝑒 =
(Eq. 11)Where Re is Reynolds number, w is flow velocity [m/s] and ν is kinematic viscosity [m2/s]
and dh is hydraulic diameter [m].
𝑃𝑟 =
𝐶𝑝(Eq. 12)
Where Pr is Prandtl number and η is dynamic viscosity [Pa s].
𝑁𝑢 = 0.023 𝑅𝑒
.𝑃𝑟
(Eq. 13)Where Nu is Dittus-Boelter correlation of Nusselt number and n is a constant for Prandtl number having value 0.4 on heating service and 0.3 on cooling service. Red is Reynolds number for tube side fluid. Nu expressed with Dittus-Boelter correlation is valid when fluid is inside of tube, Pr number is in range of 0.6… 160, Red >10 000 and tube length per diameter is more than 10. When tube and shell heat exchangers are used, most fouling fluid is typically placed inside of tubes as flow pattern is more predictable and stable compared to shell side. Tube side is also easier to clean if only tube side is clogged and shell side has clean fluid that does not cause fouling. In this case for cleaning is only needed to remove head(-s) and tubes can be cleaned. If shell side would be clogged whole tube bundle would be needed to be removed for cleaning.
Overall heat transfer coefficient of each fluid can eventually be solved using equation.
𝛼 =
(Eq. 14)Where di is tube inside diameter [m].
Key for estimating the overall heat transfer coefficient, at least with some accuracy, is to know the flow velocities of both fluids based on preliminary chosen heat exchanger type and flow channel dimensions and to have data of fouling factors of each fluids. These can be determined experimentally in lab, from existing heat exchangers on similar service or ac- quired from literature. (Saari, 2010; Nitsche et al., 2016).
3.2.3 Flow conditions
Selected heat exchanger, more precisely dimensions of flow channels, define the flow ve- locity and flow characteristics in channels. The higher the flow velocity, the higher is the turbulence and thus the shear stresses on the heat transfer surface. Rohr et al., (1998) com- piled a chart from other researcher’s work on how fluid velocity affects wall shear stress, seen in Figure 9. Increasing fluid flow velocity is directly proportional to the wall shear stress, both in laminar (intermittent line) and turbulent (continuous line) flow conditions.
Figure 9 - Wall shear stress relation to fluid velocity (from Rohr et al., 1998).
Increasing shear stress on heat transfer surface leads to increasing shear stress on fouling layer and in case of turbulent flow region, increasing flow rate also increases turbulence and eddy currents leading to reduced fouling and improved heat transfer coefficient and thus to better heat exchanger performance. (Kukulka & Leising, 2009)
3.3 Ways to mitigate fouling issues
Heat exchangers fouling mitigation has been studied extensively and many holistic and well- structured mitigation strategies and ways how to approach the issue has been developed. In general fouling mitigation can be divided into three main categories that all affect the overall fouling behaviour of the heat exchanger at different stages of the heat exchanger life cycle.
Figure 10 below presents these main stages and some typical ways to affect fouling behav- iour during each stage.
Figure 10 - Fouling mitigation ways (modified from Müller-Steinhagen et al., 2011).
Previous chapters have described the characteristics that has to be considered or calculated for every new heat exchanger designed. On next chapters it is discussed more detail on some optional aspects that could be taken into account during design of heat exchangers and would potentially have great effect on overall fouling behaviour of new heat exchanger.
3.3.1 Heat Exchanger types used in Heavy fouling applications
Oil and gas industry is very conservative towards design changes and deviations from solu- tions that has been agreed as an industry standard practice. TEMA class R - shell and tube heat exchanger design standard has been the de facto standard for heat exchanger construc- tion types in refinery applications for the past. While oil & gas industry has relied on tube and shell heat exchangers as their work horses for decades, other industries have developed many other interesting heat exchanger types to cope with demand of wide variety of heavy fouling applications.
Several researchers have pointed out that plate heat exchangers (PHE) have enhanced flow characteristics (high flow speed meaning higher turbulence and shear rates) and also physical size of PHE is smaller, compared to tube and shell heat exchangers. Table 6 presents the fouling resistance difference between PHE and tubular heat exchanger during other ways similar process conditions. (Awad, 2011; Shankar et al., 2017).
Table 6 – Comparison of typical fouling resistance values with PHE and Tubular HE (Compiled from Awad, 2011)
Rf [m2 K / kW]
Fluid PHE Tubular type HE
Soft water 0.018 0.18 - 0.35
Cooling water 0.044 0.18 - 0.35
Seawater 0.026 0.18 - 0.35
River water 0.044 0.35 - 0.53
Lube oil 0.053 0.36
Organic solvent 0.018 – 0.053 0.36
Steam 0.009 0.18
As Awad pointed out in his paper, PHE’s fouling resistance values with similar fluids are typically 1/10th of similar tube and shell heat exchangers values. Drawback for PHE’s in the earlier when having gasketed structure have been lower allowable working pressure and
temperature ratings compared to tube and shell heat exchangers. Typically, with heavy foul- ing fluids PHE’s have been gasketed type that allows operators to strip them in pieces during maintenance for cleaning. Typically, PHE’s have dozens of corrugated plates that are stacked together and gaskets are used as sealing between the plates. Numerous gasketed surfaces are always a risk of some degree and this has caused operational concerns about mechanical integrity towards gasketed plate heat exchangers. (Shankar et al., 2017)
Figure 11 – Fully welded and cleanable PHE (From Alfa Laval, 2021)
A modern PHE made for heavy fouling and extreme process conditions is seen in Figure 11.
This PHE is fully welded, easily cleanable (4-sided detachable covers and wide channel di- mensions) and has wide selection range of wetted construction materials (Alfa Laval, 2021).
Figure 12 – Counter current flow spiral heat exchanger (From Zohuri, 2017)
A counter current spiral heat exchanger is seen in Figure 12. Spiral heat exchangers have small physical size compared to heat transfer area and their typical flow characteristics are high flow speed leading to high turbulence. This leads to higher heat transfer coefficients and high shear rates minimizing the needed surface area and reducing fouling rate. Spiral construction also naturally compensates different thermal expansion of hot / cold sides.
(Shankar et al., 2017).
Figure 13 – Fluidized bed heat exchanger (From Pronk et al., 2009)
Fluidized bed heat exchangers are developed for harsh fouling conditions. General arrange- ment of a fluidized bed heat exchanger is presented in Figure 13. Key principle is to mix some particles in the fluid before it is fed into the heat exchanger and in contact with the heat transfer surfaces. While passing heat transfer surfaces, particles rub and collide with heat transfer surfaces while cleaning the heat transfer surfaces from deposits. After the flow has passed the heat exchanger, particles are separated from the process stream and circulated back to the inlet side of the heat exchanger through a downcomer. The particles fed to the process stream also improve the fluidized bed heat exchanger's heat transfer coefficient. The fluidized solid particles breakup the boundary layer improving the heat transfer coefficient even at low-fluid velocities. Due this even 8-times higher heat transfer coefficients can be reached compared to a similar heater without fluidized bed particles. Particles are typically made of steel, ceramics or glass. Typical applications for fluidized bed heat exchangers are the ones with heavy fouling e.g. desalination plants, geothermal recovery, waste water pro- cesses and refineries. (Kang et al., 2011; Pronk et al., 2009)
3.3.2 Coatings and modified materials
The objective for surface modification with coatings or other ways is to significantly reduce or diminish the fouling and in some cases corrosion. Main mechanism to achieve this object is to lower surface energy and thus reduce contact angle of solid-liquid interface. (Lee &
Woonbong, 2017; Müller-Steinhagen et al., 2011)
Coatings can be a very cost-effective and fast way to reduce fouling. Most basic coatings do not require special machinery and some epoxy coatings could be even self-applied. Kukulka
& Leising presented in their study comprehensive evaluation of different coating options and cost impact. Following Table 7 summarizes their work comparing coating options. (Ku- kulka & Leising, 2009)
Table 7 – Coating type comparison (compiled from Kukulka & Leising, 2009).
Film PTFE Based PPG E-coat Heresite Epoxy
Max operation temp [°C] 260 204 204 135
Typical thickness [μm] 25.4 25.4 76.2 100
Thermal conductivity [W/ m K] 0.25 0.7 0.86 0.53
Heat transfer coefficient [kW/ m2 K] 0.10 0.04 0.09 0.19
Thermal resistance [m2 K/ kW] 9.8 27.6 11.3 5.3
Price [$ / m2] 1210 170 920 24
As seen on Table 7, typical coating systems has moderate maximum allowable operation temperature and in the other hand, the required coating layer thicknesses itself acts as an insulation layer when estimating overall heat transfer coefficient for the heat exchanger. Ad- ditional thermal resistance of these coating films are 5.3… 27.6 m2 K/ kW top of possible fouling layer. This type of coatings are also prone for damages during cleaning and mainte- nance.
More sophisticated methods to lower material surface energy are focusing on modifying the very top layer of the metallic material itself. This approach has some clear benefits compared
to coatings. These have similar or near similar mechanical properties as has the parent metal.
This means that mechanical wear resistance during cleaning or maintenance is similar as parent metal and allowable operation temperatures are similar as parent metals, reaching well over 400 °C. (Müller-Steinhagen and Zhao, 1997; Mattox, 2018).
Santos et al. studied modified stainless steel (316L grade) surfaces in dairy industry more recently (2017). Their most interesting methods are discussed here and some of the charac- teristics of different methods are compared in Table 8. They evaluated and tested ion im- plantation and turbulent ion implantation with SiF3+ and MoS2+ as an ion source, plasma enhanced chemical vapor deposition (PECVD) with diamond like carbon (DLC) coating, sputtering with DLC coating and autocatalytic Ni-P-PTFE coatings methods. According to their study chemical vapor deposition (CVD) and galvanic coating (autocatalytic) methods have low adhesion to base material. Sputtering adhesion depends on desired coating, DLC has good adhesion properties also with sputtering method. Ion implantation is very interest- ing and promising method to modify materials surface energy and significantly lower fouling rate. This method however is very slow process for full size heat exchanger manufacturing, as implantation of 30 cm2 area with SiF3+ to density of 1 x 1017 ions/ cm2 required 10 hours of time (Müller-Steinhagen and Zhao, 1997, 3321-3332). Turbulent ion implantation has advantage to be able to modify also more interesting shaped objects, e.g. tube internal sur- faces and valves, it can also produce much thicker deposition layer than other methods. (San- tos et al., 2004).
Table 8 – Comparison of surface modification techniques (Santos et al. 2004).
Ion Implantation
Turbulent ion Implantation
Sputtering PECVD
Adhesion Very good Very good Low – Good Low - Moderate
Layer thickness [μm] 0.2 2 - 100 2 1 - 2
Difficult geometries No Yes Yes, outer surface No
Contact angle, water [ °] 31 42 67 70
Surface roughness [nm] 24 26 30 28
In Table 8 difficult geometries describe 3-dimensional objects or possibility to modify tubes internal surfaces. Water contact angles are measured by sessile drop method on stainless steel material delivered at 2R surface condition. Non-modified reference material had sur- face roughness of 30 nm and 24° contact angle with water. In Table 8 for both ion implan- tation and turbulent ion implantation SiF3+-bombardment values was used. For sputtering and PECVD methods DLC-coating values was used. Information about these surface modi- fication techniques costs is scarce, but according IAEA (2000), costs for ion implantation can vary between 1 000 – 10 000 US$ / m2. This information was gathered mainly from semiconductor industry, so fair assumption about ion-implantation costs in heat exchanger manufacturing applications might be > 10 000 US$ /m2.
Above mentioned methods and research of them have focused mainly on hydrophobicity of materials and therefore none or very limited amount of information was provided about how these surface modifications interact with oils. Lee and Hwang (2017) demonstrated easily scalable and simple path to modify titanium surface to reach super-oleophobic conditions after presented treatment.
Figure 14 - Contact angles of different droplets on super-oleophobic Ti-surface (Lee and Hwang, 2017)
Key to reach this condition was to etch and anodize the material surface so that material hierarchical profile was micro horn-like structure with nanopores, as they described the sur- face appearance. On Figure 14 is seen that all tested fluid droplets had contact angles greater than 150° with modified titanium surface. (Lee and Hwang, 2017)
3.3.3 Chemical additives and inhibitors
Numerous commercial fouling inhibitors are developed and readily available to combat against specific substances causing fouling, e.g. salt dispersants, FeS dispersants, anticoking additives, polymer dispersants and slurry antifoulants. (Chimec, 2021)
In their study, Stephenson et al. (2015) tested high sulphur content (3.5 wt-%) fossil oil at high metal temperature (540 °C) and concluded that thiophene can potentially act as an anti- fouling agent to reduce fouling. Main target of the study was to reduce forming of metal sulphates, especially preventing iron sulphate to deposit on heat transfer surface. The team experienced as high as 20-fold fouling rate reduction with 5.7 vol-% dose of thiophene when testing it with austenitic heat transfer surface (AISI 316). With such high dosing of light hydrocarbon mixed to heavy fraction, one could suspect thiophene to start acting as a solvent in general.
4 Experimental part
Experimental part was performed at Neste Corporation laboratories located in Kilpilahti Technology Center. Also all analyses were performed at in-house laboratories at Technology Center analytics laboratories.
4.1 Use and data processing of Falex Thermal Fouling Tester FT2
Fouling tests were carried out with Falex thermal fouler tester FT2 that is specifically de- signed for studying fouling tendency of different fluids. Test configuration of the fouling tester was decided to be “once through” where fresh feed is ran through the test equipment only once. The test rig was configured to have a feed reservoir from where to pump the test sample through the tester and eventually to the product reservoir. Other option would have been continuous operation of the tester where sample is fed from the feed reservoir through the testing equipment and back to feed reservoir to allow longer test runs. As the feed reser- voir volume is 1000 ml, this limits the test durations based on feed rate in current configu- ration. The overall appearance of the equipment can be seen in Figure 15. The feed reservoir has also an agitator that prevents solid particles to settle on the bottom and keeps the sample homogenous.
The test rig has tracing and insulation on each of the reservoirs and all pipelines. All com- ponents have also temperature monitoring possibility so the fouling tester can hold constant temperature from the feed reservoir, through the pump and pipelines to the heating block thus minimizing feed temperature variations.
Figure 15 - Falex thermal fouling tester, feed circulating setup (Falex Corporation, Operation and maintenance manual version 2.0)
Actual fouling behaviour during testing is seen from the measured temperatures, as the outlet temperature starts to decrease when fouling starts to take effect. After the test, run data is processed to gain an actual fouling resistance value using other known parameters e.g. con- stant volume/mass flow of tested medium. During testing, the tested fluid is pumped into the heater tube assembly with constant volumetric flow and as density of tested fluid is known, mass flow can be calculated. The heater tube assembly is basically a precisely manufactured rod that electrical current is heating. A thermocouple sensor is constantly monitoring to keep it in constant temperature. Details of heater rod assembly can be seen on Figure 17. The heated fluid flows around the rod in an outer pipe. The temperature of the feed entering the heater chamber is constantly monitored as is also the outlet temperature and based on these measured values fouling resistance and fouling factor can be eventually calculated.
Figure 16 – Detailed heater rod assembly
The mass of the heating rod is measured prior the test to obtain the reference mass of the clean rod. For this purpose Mettler Toledo AG analytical balance with accuracy of 1 mg and with capacity of 220 g was used.
Prior the actual test, the equipment runs a pressure test and warms up the reservoirs and pipelines to target temperature and after this flushes and fills the system with fresh feed.
After the equipment’s tightness is checked and the pump, heater chamber and pipelines are filled the test can be started and heating of the rod together with data collection will start.
The test itself is fully automatic and the equipment will run the test till the preliminary de- fined run time and cools down the equipment after the test run. After cooling down the
equipment, the rod is removed from the heater rod assembly, the reservoirs are emptied, the equipment is cleaned and samples collected.
4.1.1 Test run parameters
The test run parameters are presented in Table 9. For the test runs, two basic parameter sets were used. A shorter, 150 min (2.5 h) test run, was performed with 4 mm/s of fluid flow velocity over heater rod. The longer, 1080 min (18 h) test run was performed with 1 mm/s of fluid flow velocity over heating rod. Both fluid velocities were in all cases laminar flows.
Table 9 – Test run parameter
Parameter Test run time 150 min Test run time 1080 min
Heater rod temp [ °C] 330 330
Pump speed [ 0 -100%] 12% 3%
Flow speed over heater rod [mm/s] 4 1
Lines tracing temp [ °C] 85 95
Feed reservoir temp [ °C] 90 90
Receiving reservoir temp [ °C] 90 90
Feed tank mixer speed [1/min] 80 80
System pressure [bar(g)] 15 15
The main test run parameter was the heater rod target temperature and fluid flow velocity.
Secondary parameters, that mainly affect how hot the tested sample will enter the actual heating rod, were feed reservoir target temperature, pipelines heat tracing target temperature and pump casing heat tracing target temperature.
The last parameters that can be set are the overall pressure level during test and receiving reservoir target temperature. These last parameters have least effect on overall test results.
The receiving reservoir target temperature was set high enough to prevent compounds from collected liquid to form precipitates. The overall pressure was set higher than vapour pres- sure of possible low boiling CTO components to prevent them from vaporizing and causing
pressure spikes. The test equipment had an over pressure safety interlock that would cause equipment to abort the test if pressure would increase during test.
Table 10 – Test plan with key parameters
Run Run
ID Feed Feed ID Rod Temp [°C] Flow velocity
[mm/s] Purpose of run
1 - Standard - 375 Rod temp gradient
2 - Standard - 375 Rod temp gradient
3 R453 TOP 2135 330 4 Reference
4 R450 TOP 2135 330 1 Reference
5 R460 TOP 2135 330 1 Repetition
6 R465 TOP 2135 330 1 Insulation effect
7 R448 TOP 2137 330 4 Reference
8 R451 TOP 2137 330 1 Reference
9 R462 TOP 2137 330 1 Repetition
10 R480 TOP 2137 330 1 Interim. cleaning
11 R481 TOP + C4S 2137 330 1 Thiophene effect
12 R447 CTO 2508 330 4 Reference
13 R454 CTO 2508 330 1 Reference
14 R461 CTO 2513 330 1 Reference
15 R463 CTO 2520 330 1 Reference
16 R446 CTO 2546 330 4 Reference
17 R449 CTO 2546 330 4 Repetition
18 R452 CTO 2546 330 1 Reference
19 R455 PROC2-CTO 2546 330 1 Processing 2
20 R464 PROC1-CTO 2546 330 1 Processing 1
21 R468 CTO 2546 300 1 Lower temperature
22 R470 CTO 2546 270 1 Lower temperature
23 R469 CTO 2546 240 1 Lower temperature
24 R474 CTO 2546 330 1 Interim. cleaning
25 R472 TOP 2615 330 1 Reference
26 R479 PROC-TOP 2615 330 1 Processing 1
Test plan is seen on Table 10. Test’s done with higher flow velocities (R446 – R448), it was observed that oil inlet temperature to heater rod assembly was creeping during the test and overshot the target oil inlet temperature (90 °C) being 95… 96 °C after the test was ran for 1.5 h.
For this reason, in the following test run (R449) the inlet pipeline heat tracing was lowered to 85 °C to see if this inlet temperature overshot was caused by the inlet line heat tracing
