Fuel Ratio Calculation For Natural Gas Fuelled Spark Engineering Essay

Air-fuel ratio (AFR) is a crucial parameter for combustion controls in internal combustion engines. An incorrect AFR metering for reciprocating internal combustion engine causes high toxic gases emissions formulation, serious fuel consumption problems and unbearable combustion noise and combustion deterioration. Traditionally, the AFR is obtained by direct measurement of intake air and the fuel either injected into the combustion chamber or pre-mixed at the carburetor. However, the accurate AFR obtained from direct measurement is difficult due to measuring equipments resolution prone to errors.

This paper describes a method for accurate determination of air-fuel ratio based on exhaust emission gas analysis as an additional tool used to be validated the conventional direct air fuel flow rates measurement. This method explains all the possible parameters that may affect the accuracy of air-fuel ratio is conducted which includes the instrument error, ambient conditions, the assumed water-gas shift reaction constant, the humidity of the atmospheric air and the inclusion of nitrogen in the air-fuel ratio model.

Results show that four essential exhaust gas emission concentrations, namely carbon monoxide, carbon dioxide, oxygen and hydrocarbons are adequate for obtaining an accurate air-fuel ratio. The fuel type and the range of parameters that may affect the accuracy of air-fuel ratio are properly defined. This paper will also present experimental results of a bi-fuel natural gas spark ignition engine to be compared with computational results. The results of this investigation will be used to develop a new dedicated natural gas engine.

Keywords: AFR, Bifuel engine, Natural gas, Emissions, Water-gas reaction.

Notation:

A* mole number of atmospheric air

W mole number of water vapour present in the atmospheric air

Pg saturated water vapour pressure at ambient temperature

carbon percentage in fuel

Hydrogen percentage in fuel

Oxygen percentage in fuel

nitrogen percentage in fuel

sulfer percentage in fuel

Ei mole fraction of emission species for ith component

Np is the total number of moles of exhaust products

INTRODUCTION

Today, vehicle emission is one of the major sources of environmental pollution. Petroleum fossil fuels used in vehicle are the key resources of pollutants in the environment. Worldwide much effort is going on vehicle and engine system to reduce exhaust emissions. Recently, the research [1,2] is going on with high accelerating effort to develop dedicated higher efficiency natural gas fueled engine. This is due to good ignition and physicochemical properties together with simplicity of the natural gas such as compressed natural gas (CNG) fuel’s molecule that gives soot-free complete combustion. The stringent emission requirements together with the increasing demand of the natural gas vehicle have led to active researches on engine system for further reduction of pollutants emission. However, due to the lack of CNG refueling station, manufactures are converting gasoline engine into bi-fuel engine in order to satisfy customer’s demand rather than producing new-dedicated vehicles as currently happening in Malaysia. The number of registered vehicles in Malaysia is 12 millions with 51% of them are using gasoline engines [3]. The government of Malaysia is trying to retrofit gasoline engine to natural gas engine. Moreover, European Union (EU) and the association of emission control by catalyst (AECC) have suggested retrofitting engines associated, fuel management system and catalytic converters on the original gasoline engines. This is due to reduce exhaust emissions as well as to meets the US and EURO emissions requirement (http://www.aecc.be…). However, the reduction of harmful exhaust emissions and greenhouse gases will continue to direct technical development of fuels, engine after treatment systems and complete natural gas vehicles. In the past [4], most of the light-duty vehicles were actually adopted for bi-fuel solutions. It was not properly retrofitted for natural gas such as optimization of spark advance and plug gap.

This paper is an attempt to develop AFR model be used to calculate AFR from exhaust gas emissions. It is found in the literature [5,6] that many simplified models which employ AFR calculation based upon exhaust gas emissions analysis that do not mention the effect of parameters such as the assumed water-gas equilibrium constant, wetness in air, the inclusion of oxides of nitrogen, etc. on the computed AFR. Although their predicted results are acceptable, it is necessary to study these effects in order that general formulation can be used with confidence for various types of engine and fuel.

It is also found [7] that none of models includes sulfer (S) in their fuel formulae which may affect on AFR. None of their models [5,6,7] derive the AFR of natural gas. It can be said that natural gas is not simply the methane whose chemical formulae is CH4. Natural gas contains some oxygen and nitrogen in the molecule that must be included and to be counted during calculation from natural gas emissions.

This paper briefly derive a complete model which includes all the parameters that may have a direct effect on AFR calculations, such as the accuracy of the emission analyzers, the assumed water-gas shift reaction constant, the humidity and temperature of the atmospheric air and the inclusion of oxides of nitrogen in the AFR model. Other results such as performance and exhaust emissions from gasoline and CNG fueled have been presented with discussions.

AIR FUEL RATIO FORMULATION

A general formula for the composition of fuel can be represented as For conventional petroleum-based fuels, oxygen will be absent; for fuels containing alcohols, oxygen will be present. The overall combustion reaction can be written as – fuel + oxidizer →products. The fuel is while the oxidizer, which is the atmospheric air is composed of 78.09% N2, 20.95%O2, 0.93% Ar, and 0.03% CO2. The products are CO, CO2, O2, H2O, CHx, NO, N2, NO2, Ar, SO2 and soot particles (which are mainly solid carbon). The amount of solid carbon present is usually sufficiently small ( 0.5 percent of fuel mass) for it to be omitted from the analysis as suggested in [8]. Hence, the overall combustion reaction can be written explicitly as

The above equation can be written as-

Mole fraction adds upto 1:

Normally the atmospheric air contains water-wetness that has been included in the above combustion equation. The wetness in air is related to the unknown A* by

…………………………………………………………(3)

where is the atmospheric pressure and is the partial pressure of water vapour in the atmospheric which can be expressed as a function of relative humidity ().

…………………………………………………………(4)

Hence the AFR is

…………………………..(5)

Now AFR can be calculated with the assessment of the value A*. This can be resolved by considering the conservation of atoms for carbon (C), hydrogen (H) and oxygen (O) in the reaction.

C- atom balance:

H-atom balance:

O-atom balance:

N – atom balance:

Generally one does not use an instrument to measure water vapour concentration. Furthermore, complications arise in most instruments as they do not function properly if water condenses in them. A common way to handle this problem is to remove the condensate from the sample prior to delivery to the instruments. The concentration is then referred to as ‘dry’ as opposed to ‘wet’. Denoting a dry concentration with a superscript zero, the dry concentration of any species i to the wet concentration can be related by

……………………………………………………………….(9)

Now equation 6 to 8 can be rewritten as

In consideration of carbon, hydrogen and oxygen atom balance equations and equation 4, there are five unknown available, that is A*, W, and Np, since , and are all measurable. However, there are only four equations which are inadequate to solve for the five unknowns. A common practice to deal with this program is to introduce a water-gas shift reaction to govern the CO, CO2, H2 and H2O, that is

…………………………………..(13)

which possesses an equilibrium constant K

or

Where K is a function of the equilibrium temperature,. Values of 3.5 and 3.8, obtained from the JANAF Table, are commonly used which correspond to equilibrium temperature at 1738K and 1814K, respectively.

More generally,

; for 400 K<Teq<3200 K ……(15)

Combining equations (4), (10) to (12) and (14), unknown A* is determined and presented as the function of emission concentrations, that is

………………………….(16)

Hence …………………………(17)

EXPERIMENTAL SETUP & PROCEDURES

A schematic diagram of the experimental apparatus is given in Fig. 1. A proton magma -12 valves 4 cylinder engine was set on a test bed. The engine has a bore 75.5 mm, stroke 82 mm, capacity 1,468 cc and a compression ratio 9.2. An eddy current ‘Heenan’ dynamic dynamometer model Froude MK1 was connected to the engine. The load imposed on the engine by the dynamometer is governed by the amount of excitation current passing through the field coil. The method by which the coil current is controlled gives the dynamometer varying types of power/speed characteristics, such as constant torque, propeller law or constant speed. All the electronic equipment, together with its manipulative controls and indicators, etc is mounted on ‘CP Cadet6’ control unit.

Eight thermocouples (type k) were installed to measure the temperature of engine cooling water, oil, inlet air, fuel and exhaust gas. The airflow rate was measured with an air flow

Gas conversion kit

CNG

Gas main supply

Gasoline tank

ECU

Exhaust Gas

Gas Analyzer

Air Intake

Air Flow

Meter

Control Room

Data Acquisition System

Switch Box

Dynamo-

meter

Connecting Shaft

Gasoline Engine

4-cylinder

Figure 1. Schematic diagram of the experimental set-up

meter. The fuel system was so designed that the engine can run on gasoline and natural gas. The gasoline consumption was measured with fuel flow meter (load cell arrangement) and natural gas was measured with vortex gas flow transmitter. All the flow meters were incorporated with engine control system through interfacing cards.

RETROFITTING

The CNG stored under a maximum pressure of 200 bar flowed to the engine through a gas conversion kit. The kit supplied CNG fuel to the engine carburetor at approximately atmospheric pressure (0.8 bar) so that carburetor can effectively use it. The conversion kit is a three stages pressure regulator based ‘high pressure conversion kit’ model Tartarini -RP/76M. Specification of test engine and gas conversion kit is shown in Table 1. A shut off solenoid is included to present gas flow when the engine is not operating or when operating on gasoline.

Table 1. Test engine specifications

Characteristic

Proton Magma -12 Valves

Displacement, cc

1,468

Compression Ratio

9.2

Bore, mm

75.5

Stroke, mm

82

Max Output (DIN) PS/rpm net

(kW/rpm)

87/6000

(64/6000)

Max Torque (DIN) kg.m/rpm net

(Nm/rpm)

12.5/3500

(122/3500)

Carburetor

Down-draft 2-barrel

Specification of NGV Carburetion System Tested:

Proton 12-Valves 1.5S

Stage

1st Stage – 3.4 Bar

2nd Stage – 0.8 Bar

3rd Stage – negative pressure

Regulator

Tartarini Model RP/76-M

Mixer

Remote Extractor

Control

Time Advance Processor

Model 529

A downdraft 2-barrel carburetor is used for gas-air mixture. In fact only the top-half body with the venturi mixer is used and fixed upon the gasoline carburetor, so that it is easy to switch the engine to run on gasoline or gas alternatively.

Previous studies [9,10] indicate that the ignition delay of natural gas is twice that of gasoline, and that by increasing the spark advance to 28 (BTDC) degrees crank angle, a similar burning rate and combustion duration to gasoline can be achieved. For both the fuels the maximum pressure was achieved at 10.5 ATDC degrees crank angle. An electronic control unit (ECU), model MEMS -spi is equipped with engine for its ignition and fuel injection. Engine operating computer communicates with ECU through a specially designed interface and controls the engine operation either as automatic or manual. The proper plug gap for natural gas is larger than that for petrol. It was observed that the shortest combustion duration and maximum torque achieved with 1.00 mm plug gap compared with 0.85 mm for gasoline. In this experiment, plug gap was fixed to 1.20 mm for natural gas.

The engine was operated at steady state conditions with wide open throttle (WOT) and half throttle (50%) position at the speed range from 1500 rpm to 5500 rpm with 500-rpm interval. At this test conditions, AFR was calculated and verified by measured. These results will be used in near future to further development CNG fuel ignition and injection system to meet the requirement of lambda sensor for a particular engine speed. The circulating coolant water temperature was set to 85 oC to reduce the effect of cooling cylinder on combustion. This would reduce HC emissions as suggested by Boam et al. [11] and Hotchmuch et al. [12]

An Auto-check (Model 974/5) and a Bacharach model CA300NSX analyzers (standard version, k-type probe) were used to measure the concentration of carbon monoxide (CO), unburn hydrocarbon (HC) and oxides of nitrogen (NOx). The emission results were taken three (3) times repeatedly (with reproducibility of 96%) and the average results were taken for analysis.

COMPRESSED NATURAL GAS AND GASOLINE

The composition of a natural gas fuel varies with location, climate and other factors. It is anticipated that such changes in fuel properties affect emission characteristics and performance of CNG fuel in engines as shown by Byung et al. [13].

Physicochemical properties of CNG used in this experiment are shown in Table 2. From this table a chemical formulae for CNG was derived as, where and

Table 2. Typical natural gas Compositions

Component

Mole(%)

Methane

83.44

Ethane

10.55

Propane

1.31

Isobutane

0.13

Normal-butane

0.07

isopentane

0.01

Hexane

0.01

Carbon dioxide

4.17

Nitrogen

0.31

Density (kg/cm3)

0.81

Gross Calorific value (MJ/kg)

49.00

Molecular weight

19.19

Specific Gravity

0.64

The gross calorific value and specific gravity of the used gasoline fuel (C8H18) are 45 kg/m3 and 0.692 respectively.

RESULTS AND DISCUSSION

Engine performance:

The brake power output versus engine speed over the range from 1500 to 5500 rpm for both gasoline and CNG fuels is shown in Fig. 2. It can be seen that the maximum brake power obtained by gasoline fuel is 47 kW followed by CNG fuel (39 kW) which occurs at 5000 rpm. The CNG fuel produces an average (over all the speed range) of 15% less

Figure 2. Brake power vs engine speed

brake power due to the effect of reducing charge energy density per injection into the engine cylinder as compared to gasoline fuel [14]. This is mainly due to the gaseous nature of CNG fuel that reduces both the air volumetric efficiency and charge energy density per injection into the engine cylinder, hence power reduces from CNG fuel.

Figure 3 shows the variation of specific fuel consumption (SFC) versus engine speed. It is found that the specific fuel consumption is lower for CNG fuel. This can be explained that the net weight (in mass basis) of CNG gas per injection into engine cylinder is lower than that of gasoline fuel. The SFC for CNG and gasoline fuels is 300 g/kWh and 380 g/kWh respectively at 5000 rpm. On average, CNG reduces 15% -18% lower SFC than gasoline fuel.

Figure 3. Specific fuel consumption vs engine speed

Exhaust emissions:

Figure 4 and 5 show the emission results obtained from experimental study for engine speed from 1500 to 5500 rpm with WOT position. It can be seen in Fig. 4 that throughout the engine operation, CNG produces much less CO and O2 concentration and higher CO2 concentration in comparison to gasoline fuel. Figure 5 shows CNG reduces HC and increases NOx emissions in comparison to gasoline fuel. As an example at 5000 rpm CNG and gasoline produced HC emissions are 450 and 1000 ppm and; NOx 1500 and

Figure 4. CO, CO2 and O2 concentration vs engine speed at WOT

Figure 5. NOx and HC concentration vs engine speed at WOT

1050 ppm respectively. The reason of producing higher NOx from CNG fuel is mainly due to the simple chemical bond of CNG as mainly CH4 as compared to gasoline C8H18. The high cylinder temperature at the end of compression stroke oxidizes N2 to NOx besides combust CNG fuel such as methane; that might not be happened during gasoline combustion; hence produces lower NOx emission than CNG.

AFR CALCULATION

Data were collected from in-house experiments and other published papers in order to verify the accuracy and usefulness of the AFR formulation. As the used engine is a bi-fuel retrofitted engine for namely gasoline and CNG fuels, hence AFR has been calculated for both the fuels. This attempt will make a discussion on engine test with throttle controlled, comparison between two fuels besides AFR model verification. The chemical formulae of used gasoline was C8H18 and CNG where and obtained from Table 2. The test assumed the water-gas equilibrium constant as 3.73, ambient temperature as 25oC, relative humidity as 0.3 and atmospheric pressure as 1.01325.

Figure 6 shows AFR vs engine speed at WOT position for both the gasoline and CNG fuels. The average AFR of gasoline fuel found for measured, calculated with inclusion and without inclusion NOx is 16.22, 16.33 and 16.10 respectively. For CNG fuel are 15.48, 15.60 and 15.44 respectively. It can be said that calculated value without inclusion NOx is closer to measured value. Hence, it can be noted that only four emissions gaseous such as CO2, CO, O2 and HC are adequate to calculate AFR with confidentially. At 5000 rpm calculated AFR (without inclusion NOx) for gasoline and CNG fuels were 16.45 and 15.47 respectively. It can also be seen that the overall AFR for gasoline is higher than

Figure 6. Air fuel ratio vs engine speed at WOT

CNG fuel. This is due to gaseous nature of CNG fuel that reduces air volumetric efficiency and weight reduction of CNG in mass basis.

Figure 7 shows AFR for gasoline and CNG fuels from 1500 to 3500 rpm at engine half throttle position. The emission results are shown in Table 3. It is found that on average AFR for gasoline and CNG fuels are 19.46 (measured for gasoline), 19.36 (calculated for gasoline), 18.37 (measured for CNG), 18.55 (calculated for CNG). The calculated AFR for both gasoline and CNG fuels showed similar result as their measured value, which indicates the accuracy of calculation as well as AFR model formulation. The maximum AFR found is 19.90 at 2000 rpm with gasoline fuel and lowest 18.37 with CNG fuel. One error reading is found during gasoline fuel as shown at 2000 rpm.

Table 3 Emission data for gasoline (C8H18) and CNG fuels.

Fuels

rpm

CO ppm

CO2%

O2%

NOx ppm

HC, ppm

Gasoline

1500

114

10.4

3.2

1318

150

2000

32

9.25

4.5

950

110

2500

38

9.45

4.4

957

100

3000

56

9.4

4.2

950

120

3500

60

9.34

3.3

970

140

CNG

1500

68

9.6

4.32

2466

86

2000

61

9.8

3.5

2346

84

2500

49

9.5

3.7

2266

76

3000

40

9.4

4.1

1738

70

3500

41

9.5

4.2

1808

89

Error

Figure 7. Air fuel ratio Vs engine speed at half throttle

Figure 8. Air fuel ratio Vs engine speed, using emissions data from ref. **

Figure 8 shows AFR vs engine speed. The results have been obtained from the exhaust emission results from ref. [15]. It can be found that the overall AFR for gasoline fuel is higher than CNG fuel.

Table 4. Emissions results of a diesel fuel engine (ref. ***)

rpm

CO(%)

CO2(%)

O2(%)

NO

NOx

HC (ppm)

Diesel engine

800

0.0468

7.1

12.1

174

12

1200

0.038

8.8

11.4

248

14.8

1600

0.016

7.3

11.34

192

14.7

2000

0.014

5.8

12.87

158

17

2400

0.016

5.1

12.98

140

18

2800

0.017

4.5

15.39

140

17.5

3200

0.016

4.4

15.96

111

22

Figure 9 shows AFR vs engine speed for diesel fuel. The results were obtained from ref.[16] as is shown in Table 4. It is seen that calculated AFR from exhaust gases is similar as measured value. It is seen that AFR varies from 26 to 50 for diesel fuel.

Figure 9. Air fuel vs engine speed of a diesel engine.

On the basis of discussion from Fig. 6 to 9, it can be said that the AFR model developed in this paper is suitable to calculate air fuel ratio from any internal combustion and for any fuels.

EFFECT OF ASSUMED WATER-GAS EQUILIBRIUM CONSTANT (K) ON AFR CALCULATION.

Figure 10 shows the effect of altering the water-gas equilibrium constant K on AFR calculation. Results show that varying the constant ‘K’ from 1 to 10, AFR changes from 19 to 14 on calculation, when the measured value (AFR) was 16.12. It is interesting that when the value of ‘K’ is selected from between 3.5 to 4, then AFR obtains just similar as measured value. For the value of ‘K’ as 3.5, 3.6, 3.7, 3.8, 3.9 and 4; AFR obtained are 16.29, 16.21, 16.14, 16.07, 16.00 and 15.94 whose average value is 16.10 as shown in Fig. 10. Therefore, the equilibrium constant value from 3.5 to 4 can be assumed reasonably. A slight difference in water-gas constant will not have a large effect on AFR calculation but nevertheless the assumed value of K is important.

Figure 10. AFR vs water-gas shift reaction constant

EFFECT OF RELATIVE HUMIDITY AND TEMPERATURE OF AIR ON AFR CALCULATION

Figure 11 shows the relative humidity and ambient air temperature on the AFR calculation. Results show that varying the relative humidity from 0 to100 per cent or temperature from 25 to 30C gives a variation of maximum 0.5 per cent difference in the computed AFR. Therefore, a variation in humidity or ambient temperature will have little significance in AFR calculation. It can be seen that increasing relative humidity from 0 to 100 percent decreases AFR from 16.08 to 16.02.

Figure 11. AFR vs relative humidity.

7 CONCLUSIONS

EFFECT OF INSTRUMENT ERRORS ON AFR CALCULATION

A through study of the effect of instrument errors on the calculated AFR was conducted and the trend in affecting AFR calculation for all tabulated emissions data was conclusive. This was done by analyzing the sensitivity of the AFR model to changes of each emission concerned. Hence, a sensitivity parameter is defined as

Where or

With this definition, any complete sets of emission concentrations can be used to obtain the sign of the sensitivity parameters, It is seen that

for or [HC]

and

for or [NO2]

These mean that an increase in CO or CO2 or HC concentration will give a decrease in AFR. Whereas, an increase in O2 or NO or NO2 concentration will give an increase in AFR. The sensitivity parameters of AFR due to CO, CO2 and O2 are the same as those obtained by [6], that are within ±0.24. Whereas the sensitivities for HC and NOx (NO+NO2) are within +0.4 and +0.1, respectively. The coverage of the AFR for these sensitivities calculations is between 14 to 20.

The total errors can be obtained by the perturbation technique. Hence,

where is the sensitivity at a particular AFR.

CONCLUSIONS

The following conclusions may be drawn from the present study:

Retrofitted bi-fuel natural gas engine reduces 15% brake power and 15-18% SFC as compared to gasoline fuel.

CNG reduces CO, O2 and HC emissions and increases CO2 and NOx emissions as compared to gasoline fuel.

Only four emission parameters such as carbon monoxide (CO), carbon dioxide (CO2), oxygen (O2) and hydrocarbons (HC) are adequate for determining an accurate AFR.

Other parameters such as the humidity, temperature of ambient air and the assumed water-gas shift reaction constant, are of less importance to AFR calculation provided that a proper range for their values is selected.

From the signs of all sensitivity parameters concerned, it is found that increasing CO or COor HC concentration decreases AFR, whereas increasing O2 or NO or NO2 concentration increases AFR.

ACKNOWLEDGMENTS

The authors would like to thank Mr. Sulaiman bin Ariffin (Lab Assistant) for his technical assistance provided, Ministry of Science, Technology and the Environment for providing financial support through Vote- IRPA No: 33-02-03-3011 and University of Malaya, which made this study possible.