Printer Friendly

Exploring Engine Oil Reactivity Effects on End Gas Knock in a Direct-Injection Spark Ignition Engine.


End gas knock in spark-ignited (SI) engines is an undesirable combustion phenomenon that can damage the engine. Knock is mitigated in SI engines through a variety of methods, most of which decrease the engine efficiency. Examples are using a reduced compression ratio, late combustion phasing, and fuel enrichment.

A recent study by Amann and Alger [1] showed that differences in the engine oil had a significant effect on the knocking tendency, which was attributed to differences in the engine oil reactivity. The engine oil reactivity was quantified by measuring the derived cetane number (DCN) of engine oil samples diluted in n-heptane. DCN was performed using ASTM D6890 [2], which is intended for measuring the reactivity of diesel fuel, so the dilution was necessary to provide proper flow and spray characteristics for the engine oil samples. The results showed a wide range of reactivity with DCN spanning from 36 to 87. The reactivity from an aircraft lube oil with a hindered ester base stock was significantly lower than those using a more conventional petroleum or synthetic base stock.

When running a single-cylinder engine with four of these oils, Amann and Alger [1] found that there was a significant difference in the phasing of the mid-point of the combustion event (CA50), up to 3[degrees] crank angle (CA), due solely to changes in the oil. Phasing did not always trend with the engine oil reactivity, as the mid-CN oil had the most advanced phasing at a given spark timing. The results showed that the effect of the oil was seen early in the combustion process, increasing the rate of combustion from the timing of the spark to CA50, rather than simply towards the end of the combustion where the burning gases would be near the cylinder walls where the opportunity for interaction with engine oil might be expected to be highest.

Amann and Alger [1] found that because of the differences in combustion duration and knock propensity, the knock intensity was found to be a strong function of engine oil reactivity differences. The combined effect from the accelerated combustion and the engine oil reactivity on knock was that the oil with the lowest reactivity enabled the most advanced combustion phasing. Compared with the engine oil with the worst performance, the low reactivity oil enabled a combustion phasing advance of approximately 4[degrees] CA, which was equivalent to about a 3% decrease in fuel consumption.

This is potentially a very important finding for the goal of reducing petroleum consumption and greenhouse gas emissions. Most high-efficiency technologies are applicable only to new vehicles being sold, and considering vehicle useful life, it can take over a decade for the technology to become ubiquitous and realize its fuel-saving potential. If, however, an improved engine lubricant can be deployed that decreases petroleum consumption in the existing fleet, then the potential benefits of the technology can be realized on a much shorter time-scale.

For the oil to influence the combustion process, though, some of the oil must be consumed during the combustion process. Thus, it is relevant to understand the rate of oil consumption in SI engines. Amann and Alger [1] reported their oil consumption to be 50 g/h on a single-cylinder research engine, although they noted that there was also an oil leak, so this oil consumption number is higher than the amount consumed in the combustion process. Putting this rate of oil consumption into context, however, is not straight forward.

Oil consumption data for spark-ignition engines in the literature is sparse, and the oil consumption data that do exist are reported in a variety of ways (time vs. mileage basis, mass vs. volume basis) and over a variety of engine duty-cycles. By using two assumptions, however, it is possible to convert the data so that comparisons can be made.

Assumption 1: The density of all engine oils is assumed to be 0.855 g/cc.

Assumption 2: Average vehicle speed was 46 miles per hour, consistent with Didot et al. [3]

For comparison purposes, data from four studies in the literature are shown in Figure 1. Three of these studies [3, 4, 5] were conducted to investigate the effect of oil viscosity and volatility on oil consumption. These were not conducted on modern engines, as these references were published a minimum of 25 years before the present study. While each of these studies reported different levels of oil consumption based on oil type, the oil consumption for each engine was averaged across the engine oils that were investigated. The largest of these studies was by Carey et al. [4] and included a total of 43 vehicles with water-cooled engines (model year 1970 through 1974). Many of the vehicles used in this study were employee-owned and oil consumption was recorded for normal use without controlling for the drive cycle and vehicle miles traveled. The studies by Didot et al. [3] and Manni and Ciocci [5] each included two engine studies with controlled simulated driving cycles.

The fourth study by West and Sluder [6] is very recent by comparison, having been published in 2013. This study included 17 vehicle types, with each type having either 3 or 4 replicates for a total of 55 vehicles. All vehicles met U.S. Tier 1 or Tier 2 emission standards and were model year 2000-2009. Vehicles were driven to full-useful life mileage on the Standard Road Cycle, which has an average speed of 46 miles per hour.

There is a significant amount of scatter among the data in Figure 1. Apart from one outlier, oil consumption ranged from near-zero to about 0.8 g/km, with most data under 0.4 g/km. With this dataset, oil consumption is not a function of engine displacement. It is noteworthy that the data point from Amann and Alger [1] is high for many of the older engines, though it is worth reiterating that an oil leak in the experimental system caused a portion of the oil consumption. In contrast, the data from West and Sluder [6] which represents the more modern engines, are significantly lower, all below 0.1 g/km.

The purpose of this study is to determine whether the results of Amann and Alger [1] can be reproduced in an engine with an oil consumption rate that is representative of U.S. Tier 1 or Tier 2 emission-compliant technologies. Three different lube oils were used to investigate whether the engine oil can affect the combustion phasing, and separately, to investigate whether the lube oil can affect the knock propensity at a given CA50 location.

Experimental Approach

Experimental Platform

Experiments were conducted using an engine dynamometer facility at Oak Ridge National Laboratory. The experimental setup was based on a 2007 GM engine (engine code LNF), which is equipped with side-mount direct fuel injection. Prototype pistons were used to increase the compression ratio to 11.85:1 (the stock compression ratio is 9.2:1). The prototype pistons were designed from the stock piston blank and maintain the piston rings, sealing system, and piston skirt design, and are therefore believed to have a minimal impact on oil consumption. In addition, the intake cams were equipped with two different profiles: a low-lift short duration for early intake valve closing operation, and a high-lift long-duration profile for late intake valve closing operation. For a more detailed description of the engine technology, see reference [7]. The modified engine geometry is listed in Table 1 and a photograph comparing the stock and high compression ratio pistons is shown in Figure 2.

The engine was equipped with Kistler piezeoelectric pressure transducers embedded in spark plugs for all four cylinders for indicated mean effective pressure, combustion duration, and knock detection measurements (Kistler 6117BCD25 transducers and Kistler 5010 charge amplifiers). Cylinder pressure was acquired using the National Instruments-based DRIVVEN Combustion Analysis Toolkit software and hardware package. Data at all operating conditions were collected for 1000 consecutive engine cycles.

Engine Operating Conditions

Two nominal engine speed/load conditions were investigated with both high-lift and low-lift cam profiles, for a total of four engine conditions. The low and high lift cam profiles were used because they produced two different combustion profiles, due to differences in turbulence. Details of the engine operating conditions are shown in Table 2. The slower combustion associated with the low lift profile makes it more prone to end gas knock due to the increased time that the unburned reactants are exposed to high temperatures and pressures. Thus, comparing the high and low lift profiles at a common engine fueling rate is of interest from the standpoint of determining the oil impact on knock resistance. Engine operation was performed at a constant fueling rate rather than at a constant power setting. This approach was used because indicated load changes with phasing, and brake load can change with both indicated load and with friction differences from the different engine oils. The fueling rate of 24.7 mg/inj produced a brake mean effective pressure (BMEP) of 7.4 [+ or -] 0.25 bar over the operating conditions investigated in this engine.

At each of the four engine conditions listed in Table 2, ignition timing sweeps were conducted in 1[degrees] CA increments for 6 degrees. The ignition timing at each condition was selected to span the range from a non-knocking condition to a moderately knocking condition. The ignition timing sweeps were conducted in increments of 1 CA degree, and the range of ignition timing that was used for each condition is shown in Table 2. All ignition timing sweeps were performed in triplicate.

Fuel and Engine Oils

All experiments in this study used the same regular-grade 87 anti- knock index (AKI) pump gasoline containing no ethanol. The fuel properties are shown in Table 3.

Three different engine oils with very different compositions were used in this study. The first was a 20W30 baseline oil derived from conventional blendstock used in the ASTM D7589 Sequence VID certification test [8]. The second oil was a 5W30 Mobil 1, which was produced from a fully synthetic blendstock. The final oil was Mobil 254 Jet Oil which was produced from a hindered ester base stock. This set of engine oils represents a wide range of viscosity, shown in Table 4. It also represents a wide range in base stock chemistry, which contribute to significant differences in autoignition characteristics. Gas chromatography/mass spectroscopy (GC/MS) traces for the oils are shown in Figures 3 through 5.

These engine oils were selected because they are similar to the set of oils that represented a wide range of reactivity in the study by Amann and Alger [1]. In that study, the PAO and synthetic oils generally had very high reactivity, with DCN > 70. In contrast, the Jet Oil with a hindered ester base stock had a DCN of 36. Thus, the hindered ester base stock lube oil has a significantly lower reactivity relative to the baseline and synthetic oils.

Engine Oil Flush and Break-in Procedure

The procedure used to change the engine oil between measurements was based on the procedure outlined in the ASTM Sequence VID test [8]. The modified procedure used in this test protocol consisted of a double flush with the high detergent flush oil specified in the Sequence VID test [8] followed by a double flush/triple fill with the target oil. The specifics of the procedure are listed below.

Flush oil 1: Fill with the high detergent flush oil and operate the engine at condition 4 for 30 minutes. Drain.

Flush oil 2: Fill with the high detergent flush oil and operate the engine at condition 4 for 120 minutes. Drain.

Target oil 1: Fill with the target test oil and operate at engine condition 4 for 30 minutes. Drain.

Target oil 2: Fill with the target test oil and operate at engine condition 4 for 30 minutes. Drain.

Target oil 3: Fill the engine with target test oil and break-in for 90 minutes at engine condition 4.

Proceed with experimental campaign.

A catch-and-weigh procedure was used to measure oil consumption using an electronic balance with a capacity of 32 kg and a resolution of 0.1 g. The oil was drained from the engine until the flow of oil from the oil sump had stopped. The oil catch pan was weighed before and after the oil drain. There were no visible oil leaks during this set of experiments.


The results section is broken down into five main subsections. The first discusses the engine test conditions and repeatability. The second presents ensemble averaged engine results while the third characterizes the distribution of the individual combustion cycles. The fourth subsection focuses specifically on the propensity and intensity of knock, and the final subsection discusses engine oil consumption.

Engine Condition Repeatability

Ideally, the intake temperature for all experiments should be held constant, however due to infrastructure and facilities issues, the intake air temperature varied by as much as 7[degrees]C between repeats. This amount of variability had a direct effect on the combustion phasing and thus the knock propensity. For the analysis between the oils, intake temperature operating conditions were matched for the sweeps. The range of variability in intake temperature as well as the data points chosen for the direct comparison between oils (denoted by the larger data markers) is shown in Figure 6. For the comparison, the range of intake manifold temperature variability between oils is low, typically less than 2[degrees]C.

Ensemble Average Results

The relationship between spark timing and CA50 phasing is shown in Figure 7. As expected, advancing spark timing caused the CA50 combustion phasing timing to advance at a nearly 1:1 ratio. Each operating condition had its own relationship between CA50 and spark timing because the low-lift cam required more advanced ignition timing than the high-lift cam for a given CA50 due to the slower combustion event. The high-lift cam could also be operated at a more advanced CA50 combustion phasing than the low lift cams because the faster combustion produced a lower knock propensity.

It is notable that the lubricating oil produces no consistent combustion phasing bias. The CA50 combustion phasing for these ensemble average results do not differ by more than about 1[degrees] CA at any ignition timing, and for engine conditions 1 and 4 the variability is much less, approximately 0.2[degrees] CA. For conditions 2 and 3, which show the larger amount of CA50 variation, it is notable that the baseline oil is the most advanced for condition 2 and the most retarded for condition 3. The lack of a consistent effect indicates that the lube oil composition effect is minor, and the variability that is observed is most likely due to the repeatability of the system. The range of phasing differences are significantly smaller than the impact on CA50 phasing observed by Amann and Alger [1], where CA50 differences between engine oils were as high as 3[degrees] CA.

The ensemble average cylinder pressure and heat release rate for select ignition timing conditions are shown for Condition 1 through 4 in Figures 8 through 11, respectively. The most advanced ignition timing for all conditions shows evidence of end gas knock, as indicated by the abrupt spikes in the heat release rates, whereas the most retarded ignition timing shows no evidence of knock. The cylinder pressure trace and heat release rates are similar for all engine oils, with no consistent shift in the combustion event based on oil type.

Individual Cycle Analysis

The previous section illustrates that there are no significant engine oil-based differences in the ensemble averages of combustion phasing, cylinder pressure, or heat release rate. Because there are significant cycle-to-cycle variations in spark-ignited engines, an analysis of the individual cycles was also performed to assess whether the oil type affected the distribution of combustion phasing, combustion duration, or knocking intensity. Examples of the variability of cylinder pressure typical for this SI engine are shown in Figure 12 for conditions 3 and 4. It should be noted that these were stable engine conditions, with a coefficient of variation (COV) of the gross indicated mean effective pressure (IMEPg) of 1.19 for engine condition 3, and 1.22 for engine condition 4. Some individual engine cycles exhibit a significant amount of knock while other cycles appear to have relatively weak or retarded combustion by comparison. Thus, to assess whether the oil influences the combustion process, we must assess not only the average condition, but the range of the distribution associated with the condition.

To assess the cycle-to-cycle variability, it is useful to understand the three distinct phases of combustion in an SI engine, as described by Heywood [9]. The first of these, quantified here by the ignition to CA5 duration, is the flame development phase, in which a highly-turbulent thick flame develops from a spherical flame kernel established by the spark discharge. The second phase, quantified by the CA5 to CA50 duration, is the rapid burning phase, in which the developed turbulent flame propagates across the combustion chamber to the near-wall region. The third phase is the flame termination, and it occurs after the flame front is near the wall and propagation is limited.

The first stage of combustion, described as the early flame kernel growth development process and represented by the spark-to-CA5 duration, is shown in Figure 13 at the same spark timings illustrated in Figures 8 through 11. The figures show the distribution of the spark-to-CA5 distribution from 1000 consecutive engine cycles, where the point at the middle of the box is the mean value, the box itself represents the inner quartile range (inner 50% of the distribution), and the whiskers represent the 1% and 99% points of the distribution (inner 98% of the distribution). The duration for each of the statistical parameters of the distribution are shown as data labels.

At a given engine condition, changing the ignition timing has minimal effect on the spark-to-CA5 duration. This finding agrees with a previous work from Szybist and Splitter [10] that showed that while flame speed decreases with pressure and increases with temperature, flame speed is nearly constant during the compression stroke as the temperature and pressure effects offset one another. It is notable that changing the cam profile imparts large changes in the flame kernel development process, with the low lift profile increasing both the duration of the spark-to-CA5 and the width of the spread in distribution. The high-lift cam creates a significantly faster spark-to- CA5 duration, and with a given cam profile, the increasing engine speed increases the spark-to-CA5 duration.

At a given engine condition and ignition timing, the variation in the spark-to-CA5 duration with different engine oils is minimal. In all cases, the spark-to-CA5 durations for the different engine oils were within 0.6[degrees] CA, and were within 0.3[degrees] CA of each other for most cases. The oil-to-oil changes in inner quartile range duration are comparable to the variation of the mean value, meaning that the cycle-to-cycle variability of the combustion event is not affected by the engine oil. These results are typical of good day-to-day repeatability in the absence of any changes to the engine configuration, and as was discussed previously, the intake manifold temperature changed as much as 2[degrees]C between oil tests. Finally, there is no consistent trend for which oil is associated with the longest or shortest spark-to-CA5 duration: Synthetic Oil is associated with the shortest spark-to-CA duration for Condition 1, the Baseline Oil provides the shortest duration for Condition 2, and the Jet Oil provides the shortest duration for Condition 3. Thus, no consistent oil impact on the spark-to-CA5 duration is observed. This finding is counter to the finding from Amann and Alger [1] where the early stages of the combustion process were statistically affected with changing engine oil composition.

The second phase of combustion, described as turbulent flame propagation and represented by the CA5-to-CA50 duration, is shown in Figure 14 for the same operating points and the same statistical representations as Figure 13. As with the early flame kernel growth, the largest differences are noted in comparing operating conditions. The high lift cam configuration causes the CA5-to-CA50 duration to be much shorter at both engine speeds, and it causes a tighter distribution. This is likely the result of a higher level of turbulence associated with a higher intake valve lift. It can also be observed that the combustion duration decreases at the more advanced combustion phasing (earlier spark timing). This is the expected result and has been reported previously [10]. Combustion duration increases with reduced turbulence scale, and reduced turbulence scale is associated with more advanced combustion phasing [11, 12].

At a given engine condition and ignition timing, minimal variation in the CA5-to-CA50 duration is observed when the lube oil is changed. For all conditions, the variability in the average CA5-to-CA50 duration is 0.7[degrees] CA or less when the oil is changed, and is typically within 0.3[degrees] CA. The fluctuation in the inner quartile range is similar to that of the mean, indicating that the cycle-to-cycle variability does not change with engine oil. As with the spark-to-CA5 duration, this result can be considered good day-to-day engine repeatability. Thus, no oil impact on the CA5-to-CA50 duration is observed.

The distribution of CA50 location, which is the cumulative result of the spark-to-CA5 duration and the CA5-to-CA50 duration, is shown in Figure 15 for each condition. As shown in Figure 7, the average CA50 combustion phasing advances approximately 1[degrees] CA for every CA of spark advance. At a given engine operating condition and ignition timing condition, the average CA50 combustion phasing is within 1.1[degrees] CA for the three engine lubricants in all cases, but is more typically within 0.5[degrees] CA. Additionally, the oil that produces the most advanced CA50 combustion phasing is different for the different operating conditions. For the 2000 rpm operation (conditions 2 and 4), the Jet Oil produces the most advanced phasing. However, for 1500 rpm operation (conditions 1 and 2) the Jet Oil produces the middle or the most retarded phasing. The small magnitude of the CA50 changes combined with the lack of consistency on which oil is associated with the more advanced phasing indicates that the small differences associated with the combustion event cannot be attributed to the changing engine oil.

When the distributions shown in Figure 15 are tested for normality at a 95% confidence level, most distributions are confirmed to be normal. Thus, standard deviation can be used as the characteristic width of the CA50 distribution, as is shown in Figure 16 for the mid combustion phasing cases from Figure 15 (note that the two cases with the cross-hatched fill failed the normality test at the 95% confidence level). It can be observed that there is a distinct difference between the low lift intake cam (conditions 2 and 4) and the high lift intake cam (conditions 1 and 3). The high lift intake cam has a much narrower distribution of CA50, with a standard deviation of nominally 1.5 to 1.8[degrees] CA. The low lift cam has a much wider distribution with a nominal standard deviation distribution from 2.5 to 2.6 CA. The wider CA50 distribution for the low lift cam can be attributed to a slower and longer combustion process.

It is noteworthy that in Figures 15 and 16 there are differences that can readily be attributed to the intake cam and the combustion phasing, but as was the case with the ensemble averages, there are no clear oil effects. The minor differences that are present are not consistent between oil types. The Jet Oil has the largest distributions for Conditions 1 and 2, but the narrowest distributions for Conditions 3 and 4.

The results in this subsection show that oil-specific effects cannot be detected on the ensemble average combustion event. Additionally, no oil-specific effects could be detected in the early phases of the combustion event: the early flame kernel development process (spark-to-CA5 duration) and the turbulent flame propagation (CA5-to-CA50 duration), nor were clear oil-specific effects on the CA50 combustion phasing distribution detected.

Knock Propensity and Knock Intensity

The previous section illustrated that changing the oil did not change the early phases of the combustion process leading to CA50. However, this finding does not provide an indication as to whether the knock propensity changes with oil type at a given combustion phasing, or whether the intensity of the knocking events change. To quantify the magnitude of knock, the peak-to-peak knock intensity (KI) is calculated from the pressure signal for individual cycles. The KI distributions are shown in Figure 17, and as expected, the KI is higher at the more advanced combustion phasing. Unlike the distribution of CA50, the KI distributions are not normal distributions.

The distribution in KI shown in Figure 17 encompasses a range of CA50 phasing for a given ignition timing, as was shown in Figure 15. Within the CA50 distribution grouping, the combustion cycles that have advanced phasing have an increased tendency to knock. This is illustrated in Figure 18 for three ignition timing conditions. However, even at a given ignition timing and combustion phasing, only a fraction of the engine cycles result in knock. Further, between the ignition timing conditions, there is a significant amount of overlap in the CA50 combustion phasing.

To assess the knock propensity at a given CA50 combustion phasing in a holistic manner, data from all ignition timing conditions and all intake temperature conditions (as shown in Figure 6) were combined for a given engine operating condition and engine oil. Data were then binned by CA50 combustion phasing. These bins were used to show the fraction of engine cycles that exceed threshold values of knock intensity for each condition. The results are shown in Figure 19.

Engine conditions 1 and 2 show KI threshold values of 100 and 200 kPa. For engine condition 1, all oils behave similarly at a KI threshold of both 100 kPa and 200 kPa. For engine condition 2, there is considerably more variability. A much higher fraction of engine cycles for the Jet Oil register as having a KI [greater than or equal to] 100 kPa. However, a KI of 100 kPa is very light knock, and at a KI threshold of 200 kPa, the Jet Oil shows similar behavior to the baseline oil.

Engine conditions 3 and 4 show KI threshold values of 100 and 300 kPa. For condition 3, all oils show similar behavior at the 100 kPa threshold, with the Jet Oil demonstrating approximately 5% more cycles above this threshold. However, as with Condition 2, when these same oils are evaluated at a higher KI of 300 kPa, the Jet Oil is comparable to the Baseline oil. Condition 4 is the only operating condition for which there is a consistent trend in KI threshold. The Jet Oil here is seen to increase the fraction of cycles at KI thresholds of both 100 kPa and 300 kPa. Thus, while a trend may be perceptible, it is worth noting that the magnitude of the oil-to-oil differences are very small relative to a change in CA50 combustion phasing of 2 CA.

Thus, a change in KI at some engine operating conditions cannot be ruled out entirely. However, the increased KI is not consistent across all engine operating conditions, and the magnitude of the differences are small relative to the changes that occur with combustion phasing changes. Finally, the engine oil that is associated with the increased KI is the low reactivity engine oil. Amann and Alger [1] reported KI reductions of as much as 5 bar (500 kPa) with this oil. Thus, while the data cannot completely rule out any oil-based effect on KI, the findings in this study show KI differences between oils to be roughly an order of magnitude smaller than previously reported.

Oil Consumption

For the present study, the oil consumption was measured using a catch-and-weigh method. The maximum oil consumption for the three engine oils was 150 g over a period of 20 hours when the engine consumed approximately 106 kg of fuel. Using the same methodology and assumptions described in the introduction, this oil consumption rate translates to about 0.1 g/km. This oil consumption rate is in-line with U.S. Tier 1 and Tier 2 emissions-compliant vehicles [6], and is significantly lower than that of older vehicles [3, 4, 5]. This oil consumption rate is as much as a factor of 7 below that in the study by Amann and Alger [1], again noting that they experienced an oil leak that was partly responsible for this level of oil consumption.

For engine oil type to influence the combustion process and knock propensity, it is necessary for the oil to interact with the fuel and air mixture through oil consumption. While this interaction does occur, the level of oil consumption is much lower for U.S. Tier 1 and Tier 2 emissions-compliant engines than it was for older technology engines, thereby reducing the opportunity for oil to significantly impact the combustion. With the level of oil consumption observed in the present study, no clear trends could be observed with regards to oil effects on the early stages of the combustion process, CA50 phasing, cyclic variability, knock propensity at a given phasing, or KI magnitude.

However, because the results from Amann and Alger [1] illustrate a much larger influence of engine oil on combustion and knock, it can be concluded that the oil consumption levels in that study are sufficient to impact combustion performance to a much larger extent. Thus, the reduction in oil consumption that has occurred in the last 25+ years has included the additional benefit of reducing the impact of engine oil reactivity on the combustion process itself.


An experimental study was conducted in a spark-ignited direct-injection engine to determine the extent to which oil reactivity impacts combustion phasing and knock propensity. Conclusions are as follows:

* CA50 phasing was found to be dependent on spark timing and operating condition, but no dependency on engine oil type was observed.

* The duration of the initial phases of combustion, both on a mean and on a cycle-to-cycle basis, were found to be dependent on spark timing and operating condition. No dependency on engine oil type was observed.

* Peak-to-peak knock intensity was found to be dependent on combustion phasing, but showed no correlation with engine oil type.

* Oil consumption for the test engine was equivalent to about 0.1 g/km, similar to levels measured from late-model light- duty vehicles.

The results from this study stand in contrast to the findings of Amann and Alger [1], who reported a significant influence of engine oil on the combustion process and knock propensity in a research engine with an oil consumption rate more representative of older-technology engines. In contrast, the engine oil consumption in the present study is a factor of seven lower, and is in-line with U.S. Tier 1 and Tier 2 emissions-compliant engines, and as a result, there are fewer opportunities for the engine oil to interact with the combustion process.

As a result of the very low oil consumption rates, the opportunity for the engine oil to interact with the fuel-air mixture in the combustion chamber is very small. Based on the results in this study, we concluded that engine oil reactivity does not affect knock-limited operation in more modern engines (U.S. Tier 1 emissions-compliant or newer), even over the wide range of viscosity investigated. Thus, while there is no apparent need to consider a possible detriment associated with an inferior engine oil, there is also no opportunity to improve engine efficiency by using a lower reactivity oil to reduce engine knock. It is worth noting that the conclusions of this study apply to conventional end-gas knock only, and do not consider engine oil impacts on stochastic preignition, which is a distinctly different phenomenological event.

Contact Information



The authors appreciate the support of the U.S. Department of Energy, particularly Kevin Stork of the Vehicle Technologies Office. Grateful acknowledgment is also due to Sam Lewis at ORNL for GC/MS analysis of the oil samples.


ATDC - After top dead center

AKI - Anti-knock index

BMEP - Brake mean effective pressure

bTD[C.sub.f] - Before firing top dead center

CA - Crank Angle

CA5 - Crank angle at the timing of 5% heat release

CA50 - Crank angle at the timing of 50% heat release

CN - Cetane number

COV - Coefficient of variation

DCN - Derived cetane number

DI - Direct injection

GC/MS - Gas chromatography/mass spectroscopy

IMEPg - Gross indicated mean effective pressure

KI - Knock intensity

PAO - Paraffinic-aromatic-olefinic

SI - Spark ignited


[1.] Amann, M. and Alger, T., "Lubricant Reactivity Effects on Gasoline Spark Ignition Engine Knock," SAE Int. J. Fuels Lubr. 5(2):760-771, 2012, doi:10.4271/2012-01-1140.

[2.] ASTM, "Standard Test Method for Determination of Ignition Delay and Derived Cetane Number (DCN) of Diesel Fuel Oils by Combustion in a Constant Volume Chamber," ASTM D6890, West Conshohocken, PA, 2016.

[3.] Didot, F., Green, E., and Johnson, R., "Volatility and Oil Consumption of SAE 5W-30 Engine Oil," SAE Technical Paper 872126, 1987, doi:10.4271/872126.

[4.] Carey, L., Roberts, D., and Shaub, H., "Factors Influencing Engine Oil Consumption in Today's Automotive Engines," SAE Technical Paper 892159, 1989, doi:10.4271/892159.

[5.] Manni, M. and Ciocci, G., "An Experimental Study of Oil Consumption in Gasoline Engines," SAE Technical Paper 922374, 1992, doi:10.4271/922374.

[6.] West, B. and Sluder, C., "Lubricating Oil Consumption on the Standard Road Cycle," SAE Technical Paper 2013-010884, 2013, doi:10.4271/2013-01-0884.

[7.] Hoyer, K., Moore, W., and Confer, K., "A Simulation Method to Guide DISI Engine Redesign for Increased Efficiency Using Alcohol Fuel Blends," SAE Int. J. Engines 3(1):889-902, 2010, doi:10.4271/2010-01-1203.

[8.] ASTM International, "Standard Test Method for Measurement of Effects of Automotive Engine oils on Fuel Economy of Passenger Cars and Light-Duty Trucks in Sequence VID Spark Ignition Engine," ASTM D7589-16, ASTM International West Conshohocken, PA, 2016.

[9.] Heywood, J.B., Internal Combustion Engine Fundamentals (New York, McGraw-Hill, 1988).

[10.] Szybist, J. and Splitter, D., "Effects of Fuel Composition on EGR Dilution Tolerance in Spark Ignited Engines," SAE Int. J. Engines 9(2):819-831, 2016, doi:10.4271/2016-01-0715.

[11.] Zhang, Y., Jesch, D., Oakley, J., and Ghandhi, J., "High Resolution In-Cylinder Scalar Field Measurements during the Compression and Expansion Strokes," SAE Technical Paper 2013-01-0567, 2013, doi:10.4271/2013-01-0567.

[12.] Heim, D.M., Jesch, D., and Ghandi, J., "Size-Scaling Effect on the Velocity Field of an Internal Combustion Engine, Part II: Turbulence Characteristics," International Journal of Engine Research 15(2), 2014, doi:10.1177/1468087413501316.

Jim Szybist, Oak Ridge National Laboratory

Brian West, Oak Ridge National Laboratory


Received: 02 Aug 2017

Revised: 16 Sep 2017

Accepted: 20 Sep 2017

e-Available: 07 Mar 2018


Knock, Oil reactivity


Szybist, J. and West, B., "Exploring Engine Oil Reactivity Effects on End Gas Knock in a Direct-Injection Spark Ignition Engine," SAE Int. J. Fuels Lubr. 11(1):2018, doi:10.4271/04-11-01-0002.

This manuscript has been authored by UT-Battelle, LLC, under Contract No. DE-AC0500OR22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others to do so, for the United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (
TABLE 1 Engine geometry.

Number of         4
Bore x Stroke    86 x 86 mm
Connecting rod  145.5 mm
Displacement      2.0 L
Compression      11.85:1
Fuel injection  Direct injection,
system          side- mounted

TABLE 2 Engine operating conditions investigated.

                 Condition   Condition   Condition   Condition
                 1           2           3           4

Engine           1500        1500        2000        2000
Fueling            24.7        24.7        24.7        24.7
Lambda [-]          1.00        1.00        1.00        1.00
Cam Profile      High        Low         High        Low
Spark              12-17       17-22       17-22       26-31

TABLE 3 Fuel properties.

Oxygenates ASTM D5599 (%v)     <0.1
Reid vapor pressure, ASTM      13.13
D5191 (psi)
10% distillation point, ASTM   97
D86 ([degrees]C)
30% distillation point, ASTM  144
D86 ([degrees]C)
50% distillation point, ASTM  205
D86 ([degrees]C)
70% distillation point, ASTM  253
D86 ([degrees]C)
90% distillation point, ASTM  316
D86 ([degrees]C)
RON, ASTM D2699(-)             90.2
MON, ASTM D2700(-)             83.9
Anti-Knock Index (AKI)         87.1
Octane S (-)                    6.3
LHV, ASTM D240 (MJ/kg)         43.454
[lambda] = 1 AFR (-)           14.70
C, ASTM D5391 wt. (%)          86.49
H, ASTM D5391 wt. (%)          14.06
O, ASTM D5599 wt (%)           <0.1
Specific gravity, ASTMD         0.7289

TABLE 4 Viscosities of engine oils.

                Viscosity   Viscosity
                cSt @ 40    cSt @ 100
                C           C

Baseline Oil    102.0       12.1
Synthetic Oil    61.7       11.0
Jet Oil          26.4        5.3
COPYRIGHT 2018 SAE International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2018 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Szybist, Jim; West, Brian
Publication:SAE International Journal of Fuels and Lubricants
Article Type:Technical report
Date:Feb 1, 2018
Previous Article:Investigations on Spark and Corona Ignition of Oxymethylene Ether-1 and Dimethyl Carbonate Blends with Gasoline by High-Speed Evaluation of OH*...
Next Article:Impact of Siloxanes in Biomethane on the Performance of a CNG Vehicle.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters