new methodologies to measure in real time fuel

0 downloads 0 Views 459KB Size Report
Keywords: vehicle, fuel consumption, flow sensor, air-mass sensor, Lambda ..... Karl-Heinz Dietsche, Maria Klingebiel, Automotive Handbook, Robert Bosch ...
15th International Conference on Experimental Mechanics

PAPER REF: 2862

NEW METHODOLOGIES TO MEASURE IN REAL TIME FUEL CONSUMPTION OF INTERNAL COMBUSTION ENGINES. Henrique Fonseca(*), Carlos Ferreira, Telmo Fernandes Escola Superior de Tecnologia e Gestão (ESTG), Instituto Politécnico de Leiria, Leiria, Portugal (*) Email: [email protected]

ABSTRACT The precise assessment of the instant fuel consumption in vehicles is an important feature. Besides making the driver feedback more accurate it also enables the development of tools and driver support systems aiming a more fuel-efficient driving. To this extent, this work presents two new methodologies to measure the fuel consumption in both gasoline and Diesel combustion engines. The first one uses a Mass Air Flow (MAF) sensor together with a wide band Lambda sensor. The second one applies a volumetric flow sensor developed based on Bernoulli’s equation. Both proposed methodologies where tested on a Diesel vehicle running in a chassis dynamometer. Results of both methods showed an error reduction of 79% and 81%, respectively, when compared with the vehicle on-board data analysis. Keywords: vehicle, fuel consumption, flow sensor, air-mass sensor, Lambda sensor, combustion engines INTRODUCTION Fuel consumption is one of the main concerns in modern vehicles. Actually, the instant fuel consumption of internal combustion engines is determined as a function of the fuel pressure (at the injection rail) and of the injection time. The computed value is subsequently used by the on-board computer to assist the driver. Although, due to factors such as variations in fuel pressure, mechanical variations in fuel injectors (wearing, obstruction, etc..) this methodology lacks precision. Moreover, a precise assessment of the fuel instant consumption will allow new possibilities in the optimization of engines and in the development of driver assist systems with the aim of reducing the vehicle fuel consumption. As example, innovative cruise control systems; “Eco-Cruise Control” (Sangjun Park, Hesham Rakha, Kyoungho Ahn, Kevin Moran 2011) are being developed with the aim to reduce the vehicle fuel consumption by knowing the topographical information. Such a system could be further optimized with the feedback of the vehicle precise fuel consumption. For heavy-duty vehicles, the SAE J1321 standard (SAE International 1986) specifies a rigorous fuel consumption test procedure. This standard based on gravimetric measurements of fuel tanks has been applied in the optimization of engines (L. Joseph Bachman, Anthony Erb, Cheryl Bynum 2006). Still, such a standard requires equipment (extra fuel tank) and instrumentation not practicable in light vehicles. On behalf of the above, this work evaluates the precision of the actual on-board methodology (engine management system) and presents two new methodologies to measure fuel consumption in real time. The first proposed methodology uses a wide band lambda sensor, to measure the vehicle Air Fuel Ratio (AFR) in the vehicle exhaust system, together with an MAF sensor at the engine admission. The second one uses a developed flow sensor, which ICEM15

1

Porto/Portugal, 22-27 July 2012

measures the fuel volumetric flow. This sensor must be installed in the engine fuel supply line and in the fuel return line. PROPOSED METHODOLOGIES The first methodology to measure in real time the fuel consumption of an internal combustion engine is based upon the knowledge of the air-mass, admitted by the engine, and the air-fuel ratio present on the exhaust system. If the intake volumetric air flow is measured with an MAF sensor, already available in most vehicles, Eq.1 (JC Dixon, 2007) can be used to convert the volumetric air flow (litres per second) into air-mass weight (grams per second).

ρG =

P RG ⋅ TK

Eq.1 Equation for density of gases (J.C. Dixon, 2007)

Where ρ G is the density of the air in kg/m3, P is the pressure of the admitted air in Pa, RG is the dry gas constant (equal to 287,05 J for dry air) and TK is the temperature of the air in kelvin. In turn, the pressure and the temperature of the admitted air could be measured with the Manifold Absolute Pressure (MAP) sensor and the Intake Air Temperature (IAT) sensor, also available in vehicles. In the exhaust system a wide band Lambda sensor will be necessary to assess the air-fuel ratio. The use of a wide band sensor has the advantage of also enabling the application of this methodology to Diesel engines, where a Lambda value, λ, in the range 1.2 to 6 is verified. Given the measured value of Lambda, air-fuel ratio for particular fuel can be calculated by multiplying stoichiometric air-fuel ratio (AFR(stoich)) of that fuel by Lambda (RRSaraf, Dr.PKSaxena, 2009), in the case of diesel the stoichiometric air-fuel ratio is respectively 14.5 shares of air for one of fuel, in order to find the instantaneous fuel consumption (Eq. 2).

Fuel (grams) =

Air Mass (grams) λ × AFR(stoich )

Eq.2 Fuel consumption form air-mass and Lambda

The second proposed methodology employs a prototype sensor developed based in the Bernoulli’s equation (J.C. Dixon, 2007) (J. Fraden, 2004) (Karl-Heinz Dietsche, Maria Klingebiel, 2007). The developed sensor measures the fuel volumetric flow in a given fuel line. And when installed in the engine fuel supply line and in the fuel return line, can be used to assess the engine instant fuel consumption. The volumetric flow rate Q (m3/s) of liquid, through an orifice, is generally described as in Eq.3 (J.C. Dixon, 2007):

Q = Cd ⋅ A ⋅ u T Eq.3 Volumetric flow rate in an orifice (J.C. Dixon, 2007)

2

15th International Conference on Experimental Mechanics

where Cd is the discharge coefficient, A is a reference area (m2), generally the passage minimum area, and u T is the theoretical exit speed (or velocity), in turn given by Eq.4 (J.C. Dixon, 2007):

uT =

2( P1 − P2 )

ρ

Eq.4 Theoretical exit speed u T in an orifice (J.C. Dixon, 2007)

where ρ is the density of the liquid (kg/m3) and (P1 - P2) is the pressure difference across the orifice. Thus, the vehicle fuel consumption will be given by the difference between the fuel flow rate from the reservoir (main fuel line) less the fuel returning to the reservoir through the return line. The measured volumetric flow (in m3/s) is then multiplied by the density of the fuel used and corrected by the fuel temperature, to obtain the fuel-mass consumption (in grams). PRACTICAL ESSAYS To validate the proposed methodologies and to determine their effectiveness/precision on measuring the instant fuel consumption, practical essays were performed using a 2004 Volkswagen® Sharan 1.9 TDI (turbocharged Diesel engine). Fig 1 shows the test vehicle on the chassis dynamometer during the essays.

Fig.1 Test vehicle during the practical essays

The vehicle was tested for 55 sets of varying engine load and speed combinations in a calibrated chassis dynamometer (Brace, C. J., Burke, R., Moffa, J., 2009). For each set one minute for stabilization of the vehicle conditions and two minutes for acquisition of data were taken. The vehicle sensors Bosch® 0280 218 019 and Bosch® 0 281 002 399 (pressure and temperature) were used to acquire the air-mass flow and the manifold absolute pressure (and the intake air temperature), respectively. A Bosch® wide band Lambda sensor, Bosch® LSU4, was installed in the vehicle exhaust system according to the manufacturer instructions. A ICEM15

3

Porto/Portugal, 22-27 July 2012

Motorola® MPX4250D pressure sensor was utilised to measure the differential pressure in the prototype fuel flow sensor. The Fig. 2 shows the prototype flow sensor, installed in the main fuel supply line (after de fuel filter). The discharge coefficient, Cd, of the prototype sensor was empirically determined with a set of experimental essays.

Fig.2 Prototype flow sensor installed in the vehicle; 1-Fuel supply line, 2- Differential Pressure Sensor and 3- Orifice

The fuel lines between the vehicle reservoir and the engine were modified in a way that the fuel used during the essays was kept on a recipient sited over a digital scale (shown in Fig 3). The digital scale, which outputs the digital value of the instant fuel weight, was used for the precise measurement of the consumed fuel.

Fig.3 Alternative fuel reservoir system;1- Recipient, 2- Digital scale and 3-Fuel supplies lines

The transfer function of the used MAF sensor was precisely determined in a calibrated flow meter (Super Flow reference SF-120E). For all the other sensors were used the technical transfer functions provided by the manufactures. The data values of all sensors used in practical essays, plus the digital scale, were acquired using a NI® PCI-6225 data acquisition board installed in a PC running NI® LabVIEW dedicated routines. Moreover, a vehicle diagnostic system, brand TEXA®, model NAVIGATOR TXT, was used to acquire the vehicle calculated fuel consumption (litres/hour) and fuel injected (milligrams/engine stroke) via its On-Board Diagnostic (OBD) port. The values acquired gave the fuel consumption in grams for all the essays, allowing the comparison with the digital scale and with the new methodologies proposed and implemented. 4

15th International Conference on Experimental Mechanics

Note that the Exhaust Gas Recirculation (EGR) valve of the test vehicle was turn OFF during the essays. This procedure was necessary to allow the correct measurement of the MAF value,  since it not easy to quantify how much quantity of exhaust gas recirculated gone into the combustion chamber during the engine operation. RESULTS For the 55 sets of varying engine load and speed tested, Fig. 4 shows a comparison between the vehicle fuel consumption determine by the engine management system (obtained via the vehicle OBD port) and the one measured with the digital scale. In turn, the results for the vehicle fuel consumption obtained using the first proposed methodology (MAF & Lambda sensor) are compared to the ones obtained with the digital scale in Fig. 5. Using the value measured with the digital scale as reference, a summary of the error (given all the test points) for the first proposed methodology and the engine management system calculation is presented in Table 1. 300

Fuel consumption (g/min)

250 200 150 100 50

Digital Scale (g/min) Vehicle OBD Port (g/min)

15 00 15 /815 00 /1 17 710 50 /1 17 340 50 /2 20 400 00 /1 61 0 22 50 22 /900 50 /1 25 913 00 /1 25 210 00 /2 27 200 50 /1 4 30 74 00 / 30 760 00 /1 6 32 15 50 32 /963 50 /1 35 750 00 /1 1 37 55 50 / 37 602 50 /1 2 40 75 00 40 /770 00 /1 40 0

0

Engine speed (rpm )/Load (N)

Fig.4 Comparison of instant fuel consumption, vehicle OBD port fuel consumption versus fuel weight (digital scale)

Fuel consumption (g/min)

300 250 200 150 100

Digital Scale (g/min)

50

Method MAF and Lambda sensors (g/min)

15 00 / 15 815 00 /1 17 710 50 /1 17 340 50 /2 20 400 00 /1 61 0 22 50 /9 0 22 50 0 /1 25 913 00 /1 25 210 00 /2 27 200 50 /1 47 4 30 00 /7 30 6 00 0 /1 61 5 32 50 /9 6 32 50 3 /1 35 750 00 /1 15 5 37 50 / 37 602 50 /1 27 5 40 00 /7 40 7 00 0 /1 40 0

0

Engine speed (rpm )/Load (N)

Fig.5 Comparison of instant fuel consumption, first proposed methodology (MAF and Lambda sensor) versus fuel weight (digital scale)

ICEM15

5

Porto/Portugal, 22-27 July 2012

Table 1 – Comparison between the first proposed methodology and vehicle OBD port, error compared to the digital scale. Fuel Weight (digital scale) (g) OBD Port (g) Lambda Sensor + Air Mass sensor (g) Total

17223

19582

17710

Error (%)

0%

13.6%

2.8%

Experimental tests were also made with the developed fuel flow sensor, allowing the comparison between its results, the consumption determine by the engine management (OBD port) and the digital scale. The prototype sensor was installed both in the engine fuel supply line and in the fuel return line. The results obtained, up to the date, with the second proposed assessing methodology versus the digital scale are shown in Fig. 6. Table 2 summarises the measuring error of the prototype fuel flow sensor compared to the consumed fuel weight given by the vehicle OBD port and the digital scale.

300

Consumption (g/min)

250

200

150

100

Digital Scale (g/min)

50

Developed Flow Sensor (g/min) 0 1500/815

1500/1710

2250/900

2250/1913

3000/760

3000/1615

Engine speed (rpm )/Load (N)

Fig.6 Comparison of instant fuel consumption, second proposed methodology (developed flow sensor) versus fuel weight (digital scale)

Table 2 – Results comparison between the second proposed methodology and the vehicle OBD port, value of the digital scale used as reference. Fuel Weight (digital scale) (g) OBD Port (g) Developed Flow Sensor (g)

6

Total

1490

1652

1463

Error (%)

0%

9,6%

-1.8%

15th International Conference on Experimental Mechanics

DISCUSSION The methodologies proposed and implemented proved to be accurate in the measurement of the instant fuel consumption. The first proposed methodology has the advantage of using sensors that are already present in most gasoline vehicles. Furthermore, for Diesel vehicles is only the installation of a Lambda wide-band sensor is requires. With this methodology, the error, compared to the consumption value given by the digital scale, was only 2.8%. Moreover it reduces the error of the engine management method (OBD port) in almost 79%. Although, results for this methodology will slightly vary if the EGR valve is turn ON, more notorious at low engine velocities. In the essays made up to the date, the second proposed methodology (fuel flow sensor) demonstrated an even better accuracy in the assessment of fuel consumption. Reducing the error obtained with the engine management method (OBD port) in almost 81%. This proposed methodology has the disadvantage of being necessary to install two new sensors in the vehicle fuel lines, which may increase the costs of manufacturing and implementation. CONCLUSIONS This article presents the work carried on to develop sensing methodologies to precise assess the instant fuel consumption in gasoline and Diesel vehicles. Practical essays showed that the proposed methodologies gave very accurate results regarding the instant fuel consumption, reducing in 79% and 81% the error of the vehicle on-board data, respectively. The MAF and Lambda sensor method would be easier and less expensive to implement, since the necessary sensors are already available in gasoline vehicles. In turn, in Diesel vehicles will only be required the installation of a wide band lambda sensor. The developed fuel flow sensor gave even more precise results, although more testing is necessary to prove the accuracy of the methodology over a wide range of engine load/velocity points. To sum up, the proposed assessment methodologies will give the driver an accurate feedback regarding fuel consumption. Moreover, they will also allow for the development of new and improved driving assist/aid systems, aiming to reduce the fuel consumption in vehicles. REFERENCES C.J. Brace, R. Burke, J. Moffa, Increasing accuracy and repeatability of fuel consumption measurement in chassis dynamometer testing. Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, 223 (D9), pp.1163-1177, 2009. J. Fraden, Handbook of Modern Sensors: physics, designs, and applications. 3rd ed., 2004. J.C.Dixon, The Shock Absorber Handbook, 2nd ed., 2007, p.203-207. Karl-Heinz Dietsche, Maria Klingebiel, Automotive Handbook, Robert Bosch GmbH, 7th Edition 2007. L. Joseph Bachman, Anthony Erb, Cheryl Bynum Evaluating, Real-World Fuel Economy on Heavy Duty Vehicles using a Portable Emissions Measurement System, Paper Number 200601-3543 ICEM15

7

Porto/Portugal, 22-27 July 2012

R.R. Saraf, P.K. Saxena, Lambda Characterization of Diesel-CNG Dual Fuel Engine, Second International Conference on Environmental and Computer Science 2009, pp.172. SAE International, Joint TMC/SAE Fuel consumption test procedure – Type II, SAE Surface Vehicle Recommended practice J1321, 1986. Sangjun Park, Hesham Rakha, Kyoungho Ahn, and Kevin Moran, Predictive Eco-Cruise Control: Algorithm and Potential Benefits, 2011 IEEE Forum on Integrated and Sustainable Transportation Systems.

8