Effect of diesel soot on lubricant oil viscosity

Please download to get full document.

View again

of 10
5 views
PDF
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Effect of diesel soot on lubricant oil viscosity
Document Share
Document Tags
Document Transcript
  Tribology International 40 (2007) 809–818 Effect of diesel soot on lubricant oil viscosity Sam George, Santhosh Balla, Vishaal Gautam 1 , Mridul Gautam  Department of Mechanical and Aerospace Engineering, West Virginia University, Morgantown 26506, USA Received 3 March 2006; received in revised form 8 August 2006; accepted 14 August 2006Available online 29 September 2006 Abstract Soot related lubricant oil thickening is a primary concern for heavy-duty diesel engines. Engines which produce a relatively low level of particulate matter in exhaust emissions show a significant level of soot contamination in the lubricant. This contamination results inlubricant breakdown. The soot contaminates the lubricant and changes the chemical properties resulting in the lubricant ceasing toperform its functions. This causes an increase in viscosity of the engine oil causing pumpability problems. Hence, it is necessary to studythe effects of soot and lubricant oil additives and their interactions on engine oil viscosity.Statistically designed experiments were developed to study the effect of soot contamination on engine oil viscosity. The oil samplesused for the study differed in the base stock, dispersant level, and Zinc Dithiophosphate (ZDP) level. These three variables wereformulated at two levels: Low (  1) and High (1), which resulted in a 2 3 matrix (8 oil blends). Soot was considered as a variable at threelevels: low/0% weight (  1), medium/2% by weight (0), and high/4% by weight (1). This resulted in 24 oil samples, and soot at threelevels helped in determining the non-linear effect of soot on oil viscosity.Experiments were conducted at 40 and 90 1 C to study the effect of the various factors on viscosity with temperature variation. Theresults showed that viscosity of the oil samples increased with increase in soot at both 40 and 90 1 C. The analysis indicated a nonlinearbehavior of viscosity as the amount of soot increased at 40 1 C, whereas a linear variation at 90 1 C.The results obtained were analyzed using the general linear model (GLM) procedure of the statistical analysis system (SAS) package todetermine the significance of variables on viscosity. The statistical analysis system also highlighted the significance of various interactionsamong the variables on viscosity. The statistical analysis results at 40 and 90 1 C showed that the effect of base stock and ZDP levels werenegligible at 40 1 C, whereas the dispersant level and soot level influenced the viscosity of the oil samples at both temperatures. r 2006 Elsevier Ltd. All rights reserved. Keywords:  Diesel soot; Viscosity; Lubricant oil thickening; Base stock; Dispersant 1. Introduction Diesel engines are extensively used in automotivesystems due to their better fuel economy compared toconventional gasoline engines as a result of their higherthermal efficiency. Despite these advantages, diesel enginessuffer from environmental drawbacks such as high levels of exhaust NO x  (oxides of nitrogen), and particulate matter.The major contributors of atmospheric NO x  inventoriesare diesel engines. Some of the key technologies forcontrolling NO x  emissions are controlling fuel injectionsystem parameters, controlling in-cylinder charge condi-tions, exhaust gas recirculation (EGR), and controlling fuelformulation. EGR is one of the more attractive engine-based technologies for reducing NO x  emissions, but thereduction in NO x  is accompanied by an increase inparticulate matter and poor combustion performance.The contamination of lubricating oil by diesel soot is akey factor relating to the increased engine wear. The sootinduced wear mechanism is still not fully understood and amore fundamental knowledge is needed in this area. Sootrelated lubricant thickening is a primary concern of heavy-duty diesel engines. It appears that engines that producerelatively low levels of particulate matter in exhaustemissions show a significant level of soot contaminationin the lubricant resulting in lubricant breakdown. Thislubricant breakdown is one of the major causes of wear in ARTICLE IN PRESS www.elsevier.com/locate/triboint0301-679X/$-see front matter r 2006 Elsevier Ltd. All rights reserved.doi:10.1016/j.triboint.2006.08.002  Corresponding author. Tel.: +13042933111x2312;fax: +13042936689. E-mail address:  Mridul.Gautam@mail.wvu.edu (M. Gautam). 1 Summer Intern.  engines. Due to changes in the chemical properties mainlycaused by the accumulation of soot in the engine oil, thelubricant ceases to perform its functions. This results in anincrease in viscosity of the engine oil resulting in pump-ability problems and failure to lubricate the cylinder wallscausing metal-to-metal contact [1]. The factors which canchange or modify the characteristics of the soot surface areexpected to play an important role in controlling theinteractions with soot. Changes in the fuel compositionmay significantly alter the physical structure and surface/bulk chemistry of soot. Hence, it is important to study theinteractions between soot and oil additives in order todevelop high performance diesel engine oils for EGRequipped engines.In the current study, a statistically designed experimentwas developed to study the effects of soot contaminationon engine oil viscosity. The variables that were consideredare the dispersant level, ZDP level and base stock. Thesethree variables were formulated at two levels: Low (  1)and High (1), which resulted in 2 3 matrix (8 oil blends). Thetwo base stocks considered were Group I (which wasassigned   1) and Group II (which was assigned 1). Sootwas considered as a variable at three levels: Low/0%weight (  1), medium/2% by weight (0), and high/4% byweight (1). This resulted in 24 oil samples, and soot at threelevels helped in determining the non-linear effect of soot onoil viscosity.A Wells–Brookfield rotary viscometer was used tomeasure the viscosity of the lubricant oil samples. Thecone and plate geometry of the viscometer offered precisemeasurements of viscosity. Experiments were conducted at40 and 90 1 C to study the effect of the various factors onviscosity with temperature variation.The results obtained were analyzed using the generallinear model (GLM) procedure of the statistical analysissystem (SAS) package to determine the significance of variables and their interactions on viscosity. 2. Background Soot contaminated lubricating oil is one of the majorcauses of diesel engine wear. Soot contamination of thelubricating oil increases the viscosity of the lubricant andcauses pumpability problems. Walls of the combustionchamber fail to get evenly coated with the lubricant andthis ultimately leads to increased wear.Lubricant additives perform a number of diversefunctions. They can be classified into chemically inert andchemically active types. Chemically inert additives improvethe lubricant physical properties and include emulsifiers,demulsifiers, pour point depressants, foam inhibitors, andviscosity modifiers. Chemically active additives interactwith metals to form protective films reducing wear.Chemically active additives include dispersants, detergents,anti-wear and extreme pressure agents, oxidation inhibi-tors, and rust and corrosion inhibitors. Almost allcommercial lubricants contain additives to enhance theirperformances in amounts ranging from less than 1–25% ormore. The function of these additives is to protect metalsurfaces (rings, bearings, gears, etc.), resist oxidation,minimize deposit formation, prevent corrosion and wear,extend the range of lubricant applicability, improve flowcharacteristics, improve lubricant stability and to extendthe lubricant life.The five major wear mechanisms in a diesel engine areabrasion, adhesion, fatigue, corrosion and lubricant break-down. Corrosion and lubricant breakdown involves aseries of chemical reactions that lead to wear whileabrasion, fatigue and adhesion involve mechanical damageof surfaces. For all the above five forms of wear, lubricantcontamination is a predominant driver of wear. Lubricantbreakdown results because of loss of oil properties, such asviscosity, and the accumulation of harmful products of oildegradation. Lubricant breakdown occurs due to severalmechanisms. Generally soot particles carried into thelubricant with blow-by gases combine with anti-wear andviscosity additives in the oil and reduce wear tolerance andincrease viscosity. Lubricant breakdown causes excessivewear due to metal-to-metal contact [1]. Lubricant break-down could also occur if the decomposed products of theanti-wear agents such as ZDP in the oil are preferen-tially adsorbed by soot particles, which would result infailure of anti-wear protection and causing metal-to-metalcontact [2].Soot related lubricant thickening is still not wellunderstood. Earlier works have showed that viscosityincrease was a function of the amount and size of the sootparticles. It is believed that a fraction of the soot reachesthe engine oil sump via blow-by or oil on the cylinderwalls resulting in increase of oil viscosity [3]. Increase inviscosity in turn causes oil pumpability problems, whichfurther results in insufficient lubrication and metal-to-metal contact.Bardasz et al. [4] statistically designed experiments toinvestigate the nature of soot, its role in oil viscositygrowth, and to study the interactions involved withadditives that inhibit viscosity growth. The matrix wasdesigned to examine the effects of engine type, mode of operation, and the oil formulation. Mack EM6-285 andGM 6.2L engines operating both under high speed andhigh torque conditions were used in their study. The factorsselected for the study were two levels of dispersant basicityand two levels of detergent sulfonate. The tests wereconducted at high torque and high speed. They concludedthat the performance of all matrix samples exceeded thestandard test limits under either high torque or high speedconditions. They found that the increase in viscosity atboth 40 and 90 1 C depended on the percent soot in thedrain, and the particle size of soot, and correlated directlywith the surface area of soot dispersed in the oil. They alsofound that drain samples from all engine tests wereessentially Newtonian. They observed that soot suspen-sions in the drain oils obtained for every engine test werecolloidally stable. ARTICLE IN PRESS S. George et al. / Tribology International 40 (2007) 809–818 810  Bardasz et al. [5,6], in a later study, performedexperiments to determine the portion of lubricant viscositygrowth that was related to bulk oil oxidation versus sootcontamination. A statistically designed experiment wasdeveloped to examine the effects of dispersant level,dispersant type, anti-oxidant level, and detergent metaltype on average roller follower shaft wear and viscositygrowth. It was concluded that oil thickening was notcaused by oxidation, but was a result of the amount of sootpresent and/or soot particle interaction.Akiyama et al. [7] studied the phenomena of abnormalcylinder wear in EGR equipped diesel engines. Theyconcluded that the cylinder wear of a diesel engineequipped with EGR increases at low temperatures andsuggested that the abnormal wear may be due to corrosionof cast iron. Corrosion of cast iron is due to formation of sulfuric acid formed when condensed water reacts with thecombustion SO x  (oxides of sulfur). However, this may notbe the primary reason for engine wear as the sulfur contentin diesel fuels has been reduced to 0.05wt%.Cadman and Johnson [8] studied the effect of EGR onengine wear using analytical ferrography technique. Thecollected oil samples from the engine were analyzed formetal wear debris using analytical ferrography technique.A 15% EGR showed a significant increase in the concen-tration of the wear particles. Equilibrium concentrationswith 15% EGR were ten times higher than normal baselinelevels. They also believed that soot acts as an abrasive toremove the anti-wear surface coating provided by theadditives in the lubricant.Kim et al. [9] conducted experiments using a statisticallydesigned oil test matrix to investigate both oil viscosity anddiesel engine oil additive components. They investigatedthe effect of oil formulations on diesel engine valve trainwear. They concluded that laboratory wear tests couldproperly differentiate the anti-wear performance providedby different engine oils and that an anti-wear additive filmmust form on the metal surface to reduce wear. The anti-wear properties of the diesel engine oil could be improvedby increasing the ZDP concentration. They suggested thatimproved specifications were needed, as the existing dieselengine oil specifications were not adequate to protect everyengine. 3. Experimental equipment and procedure 3.1. Wells Brookfield rotary viscometer A Wells Brookfield rotary viscometer was used for thecurrent study. The cone and plate geometry of theviscometer offered precise measurements and it requiredonly a small volume of sample to perform the tests. TheWells Brookfield viscometer can be driven at six differentspeeds of rotation and is a precisely rotating torque meter.The small volume of oil sample in the cup offers aresistance to the cone rotating upon a flat surface. The oilsample cup was removed after every experiment and wascleaned with acetone for the next experiment. A schematicof the Wells Brookfield viscometer set-up is shown inFig. 1 [10]. 3.2. Viscosity measuring setup A schematic of the experimental setup that was used forconducting viscosity tests is shown in Fig. 2. Theexperiments were conducted at 40 and 90 1 C. Silicone oilwas used for the bath to attain temperatures above 100 1 Cquickly and easily. The viscometer cup had a jacket aroundit for flow of heating oil that transferred heat across the jacket wall to the oil sample. The pump maintained aconstant flow rate of heated silicone oil around the jacket,thus maintaining a constant sample oil temperature duringviscosity measurements. There was tremendous heat loss atelevated temperatures and this affected the viscosity values.The entire inlet and outlet tubes along with the jacket were ARTICLE IN PRESS Fig. 1. Schematic of Wells Brookfield viscometer. Brookfield ViscometerHigh temperature bathPumpFlow of heated oil Fig. 2. Schematic of the experimental setup for viscosity tests. S. George et al. / Tribology International 40 (2007) 809–818  811  completely insulated in order to reduce the heat losses. Inaddition, a heater was used to keep the surroundingtemperature of the air warm to prevent heat loss due toconvection. The silicone oil bath had to be maintained atgreater than 100 1 C to obtain a temperature of 90 1 C in the jacket around the viscometer cup. Two thermocouples, oneeach for the inlet and outlet of the jacket were used to monitorthe temperature during the course of the experiments. 3.3. Design of experiment If the designed experiment consists of more than one factor,the factors can influence the response individually or as acombination. In order to take care of such responses, anappropriate statistical model needs to be designed anddeveloped to determine the effects of the various factors andthe interactions between them. The factors that were taken intoconsideration for this study were base stock, dispersant level,and zinc dithiophosphate level. The three factors were tested attwo levels, High (1) and Low (  1). The lubricant compositionmatrix is shown in Table 1. The eight samples in Table 1 were tested at three levels of soot. The amount of soot for level Low(  1) was 0wt% or no soot contamination, for Medium (0),2wt%, and for High (1), 4wt%. This resulted in 24 samplesfor the viscosity tests. The Group I base stock was assigned avalue of    1 and the Group II base stock, a value of 1. Thefactor-level combinations for the designed experiments areshown in Table 2. To most accurately simulate actuallubricants, other components such as anti-oxidant, viscosityindex improver, calcium, magnesium, detergent, rust inhibitor,anti-foamant, and pour point depressant were kept constant.These functional intermediates are generally present in allcommercial lubricants. The 24 samples produced from thefactor-level combination for the designed experiments weretested on the Wells Brookfield viscometer and the resultsobtained were analyzed statistically using the GLM procedureof the SAS package. The SAS analysis gave the individualeffects and interactions of the three variables, base stock,dispersant, and ZDP with soot. 3.4. Randomization In any experimental design, all the factors that affect theresponse must be taken into consideration. But this is notthe case always and there is a possibility that somefactor might be neglected. In order to average out allthese uncertain factors in the experiment, the test runswere completely randomized. In a randomized design,all the factor-level combinations in the experiment includ-ing the repeat tests were randomized. In this study, allthe 24 oil samples were randomized first and then theviscosity tests were conducted according to the randomsequence. 3.5. Soot collection and oil sample preparation Soot was not isolated from the lubricant as lubricantadditives are known to absorb on the surface of soot,which would either influence or suppress its nascent surfacechemistry [1,11–13]. Instead, soot was collected from thewalls of a chamber that was placed in the exhaust line of aCaterpillar 3304 engine. The chamber was placed close tothe exhaust manifold, so that the soot collected wasrepresentative of soot that would enter the oil sump. Sootwas carefully scraped of the walls of the chamber after itwas allowed to build up over several hours of operation of the engine in the test cell facility at the Engines andEmissions Research Laboratory at West Virginia Univer-sity.The preparation of a stable soot suspension is achallenging task because the density of the soot particles(approx. 1.8 g/cm 3 ) is higher than the density of oil(approx. 0.88g/cm 3 ). Soot particles generally tend toagglomerate if the dispersant does not perform its functionand this makes it very difficult to prepare a stable sootsuspension artificially. The sedimentation of the sootparticles is possible if the sample is stored for a longperiod of time before performing the tests. In order toavoid this and to prepare stable soot suspensions, aprocedure proposed by Ryason et al. [14] was followed.The procedure was also used in prior studies at WestVirginia University [15–17]. The step-by-step proce-dure that was followed for the preparation of all thesamples is given in the Appendix. An ounce (29.57cm 3 ) of the oil sample was measured using an electronic weighingmachine and was poured into a glass vial. Soot was alsoweighed using the electronic weighing machine and wasadded into the glass vial to obtain the required soot-oilsample. 4. Results and discussion The tests were performed at 40 and 90 1 C, and eachsample was tested four times and the results obtained wereaveraged and this average was used for the statisticalanalysis. Randomization of the oil samples was completedinitially, and the tests were then performed. The resultsfrom the viscosity tests are shown in Table 3. The averageviscosity, standard deviation, and the coefficient of variance are presented in the table. ARTICLE IN PRESS Table 1Lubricant composition matrixBlend number Base stock Dispersant level ZDP levelWVU 397 Group I (  1) Low (  1) Low (  1)WVU 398 Group I (  1) High (1) High (1)WVU 399 Group I (  1) Low (  1) High (1)WVU 400 Group I (  1) High (1) Low (  1)WVU 401 Group II (1) Low (  1) Low (  1)WVU 402 Group II (1) High (1) High (1)WVU 403 Group II (1) Low (  1) High (1)WVU 404 Group II (1) High (1) Low (  1) S. George et al. / Tribology International 40 (2007) 809–818 812  4.1. Viscosity test analysis and results at 40 1 C  The viscosity tests data consists of the average viscosityvalues (in centistokes, 1cSt ¼ 1mm 2 /s) of the four trials onthe Wells Brookfield viscometer for each oil sample. Fig. 3shows the variation of viscosities for all the samples testedat 40 1 C. The error bars represent two standard deviations( 7 2 s ). It is clear from the plot that soot has a non-lineareffect on viscosity of lubricants. Comparing the viscositiesof the same blends with different levels of soot indicated ARTICLE IN PRESS Table 2Factor-level combinations for the statistical analysisSamplenumberBasestockDisp.levelZDPlevelSootlevelBasestock  disp.Basestock  ZDPBasestock  sootDisp.  ZDPDisp.  sootZDP  soot1   1   1   1   1 1 1 1 1 1 12   1 1 1   1   1   1 1 1   1   13   1   1 1   1 1   1 1   1 1   14   1 1   1   1   1 1 1   1   1 15 1   1   1   1   1   1   1 1 1 16 1 1 1   1 1 1   1 1   1   17 1   1 1   1   1 1   1   1 1   18 1 1   1   1 1   1   1   1   1 19   1   1   1 0 1 1 0 1 0 010   1 1 1 0   1   1 0 1 0 011   1   1 1 0 1   1 0   1 0 012   1 1   1 0   1 1 0   1 0 013 1   1   1 0   1   1 0 1 0 014 1 1 1 0 1 1 0 1 0 015 1   1 1 0   1 1 0   1 0 016 1 1   1 0 1   1 0   1 0 017   1   1   1 1 1 1   1 1   1   118   1 1 1 1   1   1   1 1 1 119   1   1 1 1 1   1   1   1   1 120   1 1   1 1   1 1   1   1 1   121 1   1   1 1   1   1 1 1   1   122 1 1 1 1 1 1 1 1 1 123 1   1 1 1   1 1 1   1   1 124 1 1   1 1 1   1 1   1 1   1Table 3Viscosity of the oil samples at 40 and 90 1 C at the different soot levelsSample ID Sampletemp. ( 1 C)0% soot 2% soot 4% sootAvg. visc.(cSt)Std. dev. COV(%)Avg. visc.(cSt)Std. dev. COV (%) Avg. visc.(cSt)Std. dev. COV(%)WVU 397 40 69.67 1.99 2.86 81.07 3.65 4.50 151.52 2.01 1.33WVU 398 40 81.79 1.52 1.86 95.45 2.60 2.73 182.95 3.01 1.64WVU 399 40 73.84 1.73 2.35 85.61 3.11 3.64 177.84 3.46 1.94WVU 400 40 89.18 2.27 2.55 96.20 1.75 1.82 202.27 2.27 1.12WVU 401 40 77.25 3.00 3.88 90.91 3.01 3.31 185.61 3.11 1.68WVU 402 40 89.74 2.80 3.12 108.34 1.81 1.67 202.27 3.73 1.84WVU 403 40 58.31 2.84 4.86 83.34 2.55 3.06 126.13 1.97 1.56WVU 404 40 90.12 2.23 2.47 101.52 1.84 1.81 200.74 2.58 1.28WVU 397 90 14.77 0.57 3.85 17.42 1.31 7.53 24.05 1.43 5.94WVU 398 90 16.67 0.87 5.21 20.08 0.33 1.63 26.89 0.33 1.22WVU 399 90 14.58 0.33 2.25 17.61 0.57 3.23 23.86 0.98 4.12WVU 400 90 17.05 1.14 6.67 17.05 0.57 3.33 18.37 0.33 1.79WVU 401 90 14.01 0.33 2.34 15.15 0.66 4.33 19.70 0.33 1.67WVU 402 90 16.48 0.57 3.45 20.08 0.87 4.32 21.78 0.33 1.51WVU 403 90 13.83 0.33 2.37 16.86 0.33 1.95 19.32 0.57 2.94WVU 404 90 16.86 0.33 1.95 18.37 0.33 1.79 20.26 0.33 1.62 S. George et al. / Tribology International 40 (2007) 809–818  813
Similar documents
View more...
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x