Measuring Diesel Fuel Consumption in a Laboratory Setting Joseph P. Wichlinski, Alexander Taylor, and Gregory Shaver School of Mechanical Engineering, Purdue University Several improvements in diesel engines attribute from additional hardware and calibrations. To continue the improvement of diesel engines, it is necessary to study these increasingly complex systems with accuracy. Specifically, measuring engine fuel consumption requires a system that can account for any fluctuations in fuel properties (e.g., temperature and pressure) to yield accurate results. Current fuel consumption measurement systems on the market are accurate, however they are expensive. This research develops a more affordable design that is capable of measuring fuel consumption at equal accuracy. In this study, an older, retail fuel measurement system was used as a comparison with the newly designed fuel measurement system. A bill of materials was recorded during the building process. For precautionary measures, a controlled amount of water was run through the new measurement system to test safety and functionality. The new measurement system was then attached to a diesel engine, and fuel consumption was measured using the gravimetric method at controlled speeds/torques to test repeatability. Finally, fuel consumption was compared with previously acquired data from the old measurement system to test for accuracy. Results show the new measurement system is over 10 times cheaper than the old measurement system. The old system has a repeatability error of.77%, whereas the new system was found to have a repeatability error of.66%. By providing a more cost-effective product for engine testing, a larger range of researchers can conduct engine testing in a laboratory setting. This creates potential for further improvements to be made in diesel engines. I. INTRODUCTION There are a few known methods for measuring fuel consumption: volumetric gauging, gravimetric gauging, mass flow measuring, and carbon balance measuring [1][2][3]. Volumetric gauging measures the change in volume of fuel over time. Gravimetric gauging measures the change in mass of fuel over time, and is advantageous because it is independent of the fuel consumption rate. Mass flow measuring is more complicated in that it measures the rate of fuel flowing in and out of the engine with a Coriolis meter [2]. A Coriolis meter is a measuring tool that utilizes the conservation of angular momentum in a flowing pipe to calculate the change in mass over time [4]. The most complex method of measuring fuel consumption is using the carbon balance from engine exhaust emissions. This method, as well as the volumetric gauging method, are dependent on many changing factors, which makes this difficult to measure the fuel consumed and reduces the overall accuracy [5]. Of these four methods, gravimetric gauging is one of the more simple and accurate ways of measuring fuel consumption [1][2][4][6][7]. In its most basic form, gravimetric gauging is accomplished by measuring an initial and final mass of fuel in a bucket on a scale over a set unit of time. Taking many of these measurements as fuel is consumed provides accurate and detailed results [1]. Despite the advantage of using this method, there are still several factors that must be considered when designing and implementing a gravimetric fuel measurement system. For instance, not all the fuel injected in the piston cylinder of a diesel engine is used. In modern diesel engines, the fuel is returned to the supply fuel tank [1]. The supply and return pipes in the fuel bucket cause a buoyant force that is seen by the fuel measuring scale. This force depends on the height of the fuel in the bucket, the density of the fuel, and changes with fuel consumption [5][8]. The density of the fuel is dependent on temperature and pressure, and these variables change between the engine fuel intake and spillback [9]. While keeping these factors in mind, it is also important to develop a method for calibrating the fuel measurement system to increase the repeatability and accuracy of experiments [10]. A previously known calibration method for the gravimetric gauging tool involves attaching a known mass to the scale [8]. The gravimetric fuel measuring system currently being used at Purdue University provides measurements with a.77% repeatability error in fuel consumption readings. With some results being as low as a 0.5% decrease in fuel consumption, a definitive conclusion cannot be made due to the high relative uncertainty [11]. II. METHODS In this study, a gravimetric fuel system was designed, built, and implemented in a diesel engine test cell. The new system was designed to be cost effective and tested for functionality, repeatability, and accuracy. A. The Design Process The design of the fuel measurement system was based off of an old fuel measurement system found in a different testing cell. The old fuel measurement system was studied
for an understanding of how it functions. A schematic for the new fuel measurement system was then made. A bill of materials was recorded during the building process, with a goal of keeping the cost lower than the cheapest, $57k alternative. Figure 1: New schematic of Fuel Measurement System Important design applications include having a heat exchanger after the engine, connecting supply and return fuel lines, and using fuel filters as seen in Fig 1. It is necessary to have the heat exchanger after the engine to keep the temperature, and therefore the density, constant. For this reason, thermocouples are placed before and after the engine. The connecting supply and return fuel lines prevent backflow to the pumps, which is monitored by the pressure gauges. Filters are placed throughout the system to prevent any metal shavings or other particles from reaching the fuel bucket, and more importantly, the engine. The manual valves are placed in the system so that the engine can be supplied with fuel straight from the day tank or from the fuel bucket for testing. The emergency valve and proportional valve are to monitor the bucket filling process and prevent overflow. A float switch will also be implemented to monitor overflow. In addition to these features, stainless steel fuel lines were used because of their corrosive resistance to diesel fuel. Figure 2: Fuel Bucket Design The bucket used to hold the fuel was designed to dampen wave formation, and therefore prevent fluctuations in measuring mass. This was achieved by separating the intake and outtake sides of the bucket with two metal sheets and allowing a small open section for passage. B. Measuring Fuel Consumption The designed system measures the weight of the bucket and its containing fluid. After taking the weight of the bucket into account, the weight of the fluid in the bucket is measured and converted to mass by dividing by the acceleration of gravity, or 9.81 m/s 2. These measurements are taken over constant time intervals, allowing the calculation for fuel consumption rate to be made: Fuel Consumption Rate = Fuel mass time However, measuring fuel mass requires certain factors to be taken into account. C. Factoring in Corrections It is commonly known that the density of a liquid is dependent on its temperature and pressure. The density of diesel fuel increases with increasing pressure and decreasing temperature (and vice versa). Because the fuel bucket and pipe lines have a fixed volume, this creates fluctuations in the measured fuel weight. The pressure in the fuel bucket is constantly under atmospheric pressure. The temperature, however, is susceptible to changes due the much warmer fuel returning from the engine spillback. As mentioned earlier, a heat exchanger must be placed on the fuel return pipe to reduce the temperature and therefore keep a constant fuel density. The pipes that are submerged in the fuel bucket also affect the weight of fuel measured. The submerged pipes create a buoyancy force that adds to the measured weight force. This buoyancy force and weight of fuel measured are both functions of the fuel height in the bucket. Therefore, the following buoyancy correction equation can be derived: mass fuel = mass measured (kg) CSA pipe(m 2 ) CSA bucket (m 2 ) mass measured (kg) + ρ fuel ( kg m 3) CSA pipe(m 2 ) height constant (m) Variables: CSA = cross-sectional area height constant = height between floor of bucket and bottom of pipe ρ = density D. Functionality Testing Water was placed in the fuel bucket and was run through the fuel measuring system. In this case, the system was not connected to an engine, but was set up as a closed loop (connected to itself). Initially the weight of water measured is expected to decrease until a steady flow of water is running through the pipes. With a constant amount of water running through the system, the weight measured in the bucket was analyzed for any fluctuations. Fluctuations in weight indicated any leaks in pipes or major pressure changes throughout. E. Steady-State Fuel Consumption The gravimetric fuel measuring system was connected to a 6.4L diesel engine. Diesel fuel was placed in the fuel bucket to begin, and the engine was set to run at 800rpm / 50ft-lbs to make sure that the fuel measurement system was supplying fuel correctly. The engine was then set to run at 1600rpm / 157 ft-lbs, a steady state repeat point at which data was gathered. Weight measurements of the fuel in the
bucket were taken over 120 second and 240 second time logs with a 100 Hz measuring frequency. It is important to wait until all pipes are flowing at the start and end of the measurements to ensure the only measured change in fuel mass is from engine consumption. The mass of the fuel is then calculated from the weight measurement every.01 seconds. This test was done multiple times to ensure repeatability, and the results were compared with the older gravimetric fuel measurement system with the same engine and engine conditions. III. RESULTS A. Measurements With the engine running at a steady state of 1600 rpm / 157 ft-lbs, the new fuel measurement system measured an average fuel consumption rate of 2.4037 g/s. At the same speed and torque on the same engine, the old purchased fuel measurement system measured an average fuel consumption rate of 2.3652 g/s. Both values are found from the 240 second logging average. Figure 3 and Figure 4 show one of the five tests with the new fuel measurement system that were taken over 240 seconds. Figure 4: Average Fuel Consumption Rate The repeatability error is measured as: max or min outlier average average 100 where the numerator is the largest difference between the outlier and the average fuel consumption rate. The percent error could not be calculated because there was no known theoretical value. For this reason, the average fuel consumption rate was taken to be the theoretical value, making the value a percent error in repeatability. As seen in Figure 5, the repeatability error for the new fuel measurement system was found to be 0.66% over five data sets, taken from a moving average over 500 data points in each set. Under the same engine conditions and logging period, the old fuel measurement system had a repeatability error of 0.77% over six data sets. With a logging time of 120 seconds, the repeatability error in the new fuel measurement system was found to be 0.99% over thirteen data sets. Figure 3: Change in fuel mass over time error =.99% error =.66% error =.77%
Figure 5: error for different performed tests Because of the lower repeatability error, the compared values referred to in this paper are associated with the 240 second log. The average break specific fuel consumption (BSFC) for the new fuel measurement system is 1522.813g (kw hr) with a standard deviation of 9.290. The average BSFC for the old fuel measurement system is 1498.742g (kw hr) with a standard deviation of 7.264. B. Cost According to Figure 6, the cost of the new fuel measurement system refers strictly to the parts purchased for the system contained inside the fuel cabinet, seen in Figure 7. Item Price Proportional Valve & electrical unit $ 1,044.00 Solenoid Valve $ 201.60 Fuel Filters & Housing $ 336.90 DIN valve adapter $ 9.75 Electric Motor $ 215.00 Manual Valves $ 166.00 Fuel Bucket $ 707.00 Storage Cabinet $ 495.95 Casters $ 63.26 Fluid pump $ 153.75 Framing $ 107.28 Heat exchanger $ 378.15 Pipes, adapters, and hoses $ 2,180.14 Load Cell $ 400.00 Thermocouples $ 62.00 Stainless Steel Tubing $ 290.79 Bucket hanger $ 12.38 Braided Tubing $ 163.22 Load Cell signal conditioner $ 210.00 Total Cost $ 7,197.17 Figure 7: Items purchased for new fuel measurement system The old fuel measurement system that was used as a comparison is quoted at $77,000. The cheapest alternative found was quoted at $52,000. The cost for the new fuel measurement system costs under $7,200 for materials. The developing (Figure 8) and final non-automated product (Figure 9) can be seen below. Figure 8: Early, outside view of new fuel measurement system Figure 9: Finished, inside non-automated fuel measurement system IV. DISCUSSION The new fuel measurement system was reduced to one tenth the cost of its comparison, while maintaining a lower level of repeatability error at 0.66%. The average fuel consumption for the new fuel measurement system was slightly higher than the fuel consumption rate for the old fuel measurement system. This may have been attributed to varying engine conditions, or differences in filtering gathered data. The standard deviation for the old fuel measurement system in terms of BSFC was slightly lower than the new fuel measurement system. This tells us that there are more outliers in the new fuel measurement system data. However, the repeatability error indicates that these outlier are less extreme than the old fuel measurement system data.
V. CONCLUSION The new fuel measurement system met the goal to be cheaper and match the accuracy of the old fuel measurement system. The recorded list of items purchased allows reconstruction of the new fuel measurement system. Future testing can be done to further measure the accuracy of the fuel measurement system. Measuring fuel consumption using the flow-metric method, volumetric method, or measuring carbon balance are common measuring options that can be compared to. The new fuel measurement system will be used on a 15 Liter Cummins Diesel Engine, commonly found in Class-8 trucks. ACKNOWLEDGMENT Special thanks to Herrick Labs, technician David Meyer, sponsors Eaton and Cummins, SURF program, and Purdue University. REFERENCES [1] A. J. Martyr and M. A. Plint, Engine Testing Theory and Practice, Third. Burlington, MA: Elsevier Ltd., 2007. [2] T. Flauger and S. A. Ifft, Fuel Consumption Measurement With Pulsating Flow, ASME, 2003. [3] R. D. Burke, C. J. Brace, and J. G. Hawley, Critical evaluation of on-engine fuel consumption measurement, Proc. Inst. Mech. Eng. Part D J. Automob. Eng., vol. 225, no. 6, pp. 829 844, 2011. [4] C. Clark, R. Cheesewright, and S. Wang, Response Coriolis Meter Systems, vol. 57, no. 1, pp. 95 99, 2008. [5] E. A. Christie, T. B. Crowie, D. Proctor, and R. R. M. Johnston, The Importance of Fuel Properties in Correlating Carbon Balance and Volumetric Fuel Consumption, SAE Tech. Pap. Ser., 1987. [6] J. L. B. Vilnis Pīrs, Žanis Jesko, Determination methods of fuel consumption in laboratory conditions, vol. 1, pp. 154 159, 2008. [7] C. J. Brace, R. Burke, and J. Moffa, Increasing accuracy and repeatability of fuel consumption measurement in chassis dynamometer testing, Proc. Inst. Mech. Eng. Part D J. Automob. Eng., vol. 223, no. 9, pp. 1163 1177, 2009. [8] Gravimetric Fuel Consumption Meter. [9] M. Dzida and P. Prusakiewicz, The effect of temperature and pressure on the physicochemical properties of petroleum diesel oil and biodiesel fuel, in Fuel, vol. 87, no. 10 11, 2008, pp. 1941 1948. [10] P. Kaub, P. Allen, and C. Lael, Minimizing Errors in Test Cell Fuel Measuring System Calibrations, Pap. Pap. 1990-2002, no. 724, 2001. [11] A. Taylor, Test Cell Set-Up to Enable Drive-Cycle Testing of Variable Valve Actuation Enabled Camless Diesel Engine, Purdue University, 2016.