Performance and emissions of a small scale generator powered by a spark ignition engine with adaptive fuel injection control

Applied Energy 121 (2014) 196–206

Contents lists available at ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

Performance and emissions of a small scale generator powered by a spark ignition engine with adaptive fuel injection control Adrian Irimescu ⇑, Gabriel Vasiu, Gavrila˘ Trif Tordai ‘‘Politehnica’’ University of Timisoara, Pta Victoriei No 2, 300006 Timisoara, Romania

h i g h l i g h t s  Analysis of a cost competitive injection system with adaptive control strategy for stationary spark ignition engines.  Good running characteristics and power quality for the proposed control system.  Efficiency and emissions analysis during multi-fuel operation of a small scale generation unit.  Brief financial analysis of payback periods for the proposed injection system compared to carburetor fueling.

a r t i c l e

i n f o

Article history: Received 16 July 2013 Received in revised form 29 October 2013 Accepted 25 January 2014 Available online 28 February 2014 Keywords: Distributed generation Spark ignition engines Alternative fuels Iso-butanol Natural gas Regulated emissions

a b s t r a c t Distributed generation of electricity is more and more viewed as a solution for reducing transmission losses and provide better catering for the needs of end users. Small-scale generation is therefore likely to increase its share in the energy sector, as it ensures high degree of flexibility, quick start-up and good performance in combination with intermittent power sources such as solar or wind. One drawback of small scale generators driven by internal combustion engines is, however, low fuel conversion efficiency and high specific emissions compared to medium or high scale power units. A new control strategy for fuel injection and emissions reduction is proposed to mitigate both aspects, while ensuring flexibility in the choice of fuels for a spark ignition engine powered generator. Performance and emissions are compared for carburetor and fuel injection combined with the use of a three way catalytic converter, with the latter solution proving to be more efficient and environmentally friendly. Significant improvements in fuel conversion efficiency and reductions of carbon monoxide and unburned hydrocarbons emissions were obtained by employing the proposed setup and control strategy. Flexibility in the use of different fuel types was evaluated by performing measurements with gasoline, iso-butanol and combined use of alcohol and natural gas. Financial aspects are also covered through a brief analysis of initial capital costs and payback time in order to offer a more detailed view of both fuel systems. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Electric power production units are changing their characteristics, as more and more renewable sources are added to the grid. Within this context, when aging capacity needs to be replaced on an extensive scale in order to ensure stability in combination with intermittent wind and solar [1], distributed generation could provide a solution that is easier to implement from the financial perspective, and provide an optimum choice for more environmentally conscious, sustainable and renewable energy based cities [2]. Even in extremely harsh conditions, such systems can provide the ⇑ Corresponding author. Address: Faculty of Mechanical Engineering, Bld Mihai Viteazul 1, 300222 Timisoara, Romania. Tel.: +40 722823416; fax: +40 356818783. E-mail address: [email protected] (A. Irimescu). http://dx.doi.org/10.1016/j.apenergy.2014.01.078 0306-2619/Ó 2014 Elsevier Ltd. All rights reserved.

required electric energy, as well as heating [3]. Small scale generators can be brought online with reduced capital cost and much quicker than large utilities. One drawback of such power units driven by internal combustion engines is that they usually feature relatively low conversion efficiency and longer downtime. Another issue that needs to be considered is fuel availability. Natural gas is set to increase its share in energy production as more and more shale gas projects develop, providing a cost competitive and low carbon dioxide emissions choice. Therefore, extensive research into distributed generation, grid stability, multi-fuel operation and emissions reduction needs to be performed. In order to provide a somewhat comprehensive picture of the problems associated with the operation of small scale generators driven by internal combustion engines, a literature review pertaining to fundamental aspects, multi-fuel operation and

A. Irimescu et al. / Applied Energy 121 (2014) 196–206

197

Nomenclature k K t U P

g _ m

constant (–) smoke opacity (1/m) time (s) voltage (V) power (W) efficiency (–) mass flow (kg/s)

Subscripts inj injection e electric

environmental impact was performed. While the addition of hydrogen to biogas was found to increase efficiency [4] and combined with the use of exhaust gas recirculation, to significantly reduce NOx emissions [5], large variations in the Wobbe index and stoichiometric air–fuel ratio are not beneficial for most gas fired generators [6]. As the majority of these power units do not feature closed loop control, significant mixture leaning can occur and thus result in very different performance and emissions when the fuel changes its composition. Even severe damage to the engine can occur if abnormal combustion phenomena are present [7]. Running a SI engine powered unit on pure hydrogen was found to deliver improved conversion efficiency [8], mainly due to the ability of running with lean mixtures. This strategy is however limited by increasing NOx emissions when operating below an excess air ratio of 2 [9]. Lean operation is possible even with fuels that feature more narrow flammability ranges than hydrogen, by using two stage combustion [10]. One drawback is that such concepts usually feature increased heat losses and require complex combustion chamber geometry. Other combustion development processes such as so called homogenous charge compression ignition can provide increased efficiency and low raw emissions, but are notoriously difficult to control, even with spark assistance [11]. Replacing gasoline with alcohols (ethanol and more recently methanol and butanol) is a fueling strategy intensely researched, as it reduces dependency on finite energy sources and can offer a reduction of green house gas emissions. The use of methanol presents significant advantages compared to gasoline in terms of efficiency and emissions due to specific properties of this fuel [12]. Several studies have investigated the influence of blending butanol with gasoline on raw emissions [13] and experimental combined with simulation investigations revealed that the alcohol containing mix resulted in improved full load performance even at high engine speed [14]. One important issue that needs to be addressed when using alcohol fuels is the control strategy employed for compensating large variations in relative air–fuel ratio such as is the case for so called ‘flex-fuel’ engines [15]. Essentially, fuel metering is controlled through two main mechanisms, a feed forward component that relies on estimation of air flow into the engine and a feedback loop that is required for precise air–fuel ratio control in order to ensure optimum operation of the exhaust gas treatment system [16]. Lean burn SI engines feature additional complications due to much wider range of air–fuel ratios during normal operation, thus making closed loop control for lean NOx trap regeneration a good opportunity for ethanol content estimation in the fuel blend [17]. Dual fuel operation is usually employed for diesel power units and is relatively rare for SI engines. Studies pertaining to injection of gasoline and butanol separately, obtained reduced knock tendency and unburned hydrocarbon emissions as compared to using

f

fuel

Abbreviations COV coefficient of variation LHV lower heating value MAP manifold absolute pressure PRP prime power UEGO unheated exhaust gas oxygen (sensor) SI spark ignition

the two fuel types blended [18]. Injection phasing was also found to have an influence on emissions during stoichiometric operation and to a lesser degree when using lean mixtures [19]. Similar results were obtained for the use of gasoline with ethanol and dimethylfuran [20,21]. Given that, as mentioned before, natural gas is likely to increase its share in the energy mix, dual fueling by using combined gaseous and liquid combustibles may be a strategy that can offer great flexibility, with increased performance and reduced emissions [22]. In addition to the reduction in carbon dioxide emissions when using natural gas instead of gasoline, methane enriched biogas can offer close to ‘carbon neutral’ operation (given that it is obtained from biomass) without major differences in performance as compared to the fossil fuel [23]. Following the literature review pertaining to fundamental aspects of internal combustion engines operation, the use of alcohols and dual fuelling of SI power units, this study investigated the implementation of a cost competitive solution for ensuring increased efficiency and low emissions for a small scale generator powered by a spark ignition (SI) engine. The ability of the proposed injection system to operate within the same power quality range as the original equipment that features carburetor fueling was studied for gasoline, iso-butanol, as well as dual fueling, by using natural gas combined with the four carbon atoms alcohol. Electric efficiency and emissions for all investigated conditions were compared and a brief economic assessment was also performed to provide a more detailed view of issues associated with small scale distributed generation using multi-fuel SI engines.

2. Experimental setup and injection control The generator used in the experimental trials was powered by a single cylinder SI engine (Fig. 1). Table 1 shows the main engine and generator characteristics. No modifications were made to the original ignition system (that featured fixed spark advance) and mechanical speed governor; the carburetor was kept as well. Given that cost is an important issue for small scale generators, a simplified injection system was developed in order to reduce the number of sensors and help mitigate this financial aspect. The proposed control system relies on the readings from an absolute manifold pressure sensor for feed-forward control and the use of a narrow band exhaust gas oxygen sensor for closed loop operation. In this way, close to stoichiometric air–fuel mixtures can be delivered to the engine so that high efficiency can be ensured for a three way catalytic converter. One Hall type sensor mounted near the engine’s flywheel was used for triggering the injection control and synchronize it to the rotational velocity. Rather than relying on so called ‘injection maps’ that contain predetermined injection time values for specific engine speed and load settings, the

198

A. Irimescu et al. / Applied Energy 121 (2014) 196–206

Fig. 1. Experimental setup and schematic of the fuel systems.

Table 1 Engine and generator characteristics [24,25]. SI engine

Generator

Number of cylinders Displacement Compression ratio Bore x Stroke Rated power Operating principle Specific fuel consumption at rated power Control system Fuel system Ignition system

1 389 cm3 8.2:1 88  64 mm 8.4 kW@3000 rev/min four stroke 313 g/kWh mechanical governor connected to the throttle valve Carburetor, liquid fuel at 3 bar and gas injection at 0.03 bar Transistorized magneto

proposed fuel system employs a simple strategy that uses only the measurement of absolute pressure in the intake manifold to adjust fuel flow when rapid changes in engine load occur; readings from a narrow band oxygen sensor are used for performing fine adjustments in order to maintain the air–fuel ratio as close as possible to stoichiometric values. As a direct result of using the simplified injection system, the number of sensors required for proper operation was significantly reduced, and the possibility of using a much wider range of fuels was ensured. A schematic of the logic tree used for the control algorithm is given in Fig. 2. The same controller and software were used for metering liquid as well as gaseous fuel types. Fig. 3 shows traces recorded for the three input

Type

AC current threephase, asynchronous self-excited, selfregulated, brushless insulation class H

Three-phase generation PRP Single-phase generation PRP Frequency Power factor (cos u)

6 kVA (4.8 kW)/400 V/8.7 A 5 kVA/230 V/21.7 A 50 Hz 0.8

parameters, as well as the voltage used for opening the fuel metering valve during gasoline operation. Near stoichiometric fueling was obtained for all investigated conditions when using the injection system. Electric signals within the ±10 V range were recorded at a frequency of 30 kHz for every monitored parameter, with an accuracy of ±3.1 mV using a high speed data acquisition system capable of up to 250  103 samples/s. Fuel consumption was measured using the gravimetric method through measurements at 1 min intervals of stable operation when employing carbureted fueling, while during the trials with the injection system, flow rates were calculated based on recorded injection time values. Prior to the measurements performed for this study, the injection system’s volumetric flow was characterized for the entire range of injection pulse durations at a frequency of 50 Hz, also using the gravimetric method. Gaseous fuel consumption was recorded using a volumetric flow meter. Electric power was measured with an accuracy better than ±2% and fuel consumption error was within the same ±2% range. Therefore, fuel conversion efficiency was determined with an accuracy better than ±4%. This latter parameter was calculated using Eq. (1):

gf ¼ P e =ðm_ f  LHVÞ;

Fig. 2. Schematic of the fuel injection control algorithm.

ð1Þ

where gf is the fuel conversion efficiency, Pe electric power measured in W was calculated as Umeas/UratedPrated, fuel consumption _ f in kg/s and LHV in J/kg. m

199

MAP [V]

Trigger [V]

A. Irimescu et al. / Applied Energy 121 (2014) 196–206 Table 2 Accuracy and resolution for measured parameters.

6 5 4 3 2 1 0 -1

3.5 kW

6 5 4 3 2 1 0 -1

Parameter

Resolution

Accuracya (%)

Range

Mass (kg) Liquid fuel flow (dm3/min) Gas fuel flow (m3/min) Density (kg/m3) Voltage (V) Electric power (W) Engine speed (rev/min) O2 (%)

0.0001 0.001

±0.0003 ±1 ±1 ±1 ±1 ±2 ±1 ±1 ±3

0–10.1 – – – 0–500 0–10,000 0–4500 0–4 4–22 0–10 0–20 0–20,000 0–4000 0–99.99

CO (%) CO2 (%) HC (ppm) NOx (ppm) Opacity (1/m)

3.5 kW

a

1

0.01 0.1 0.01 0.1 1 1 0.01

Relative value of reading.

UUEGO [V]

1 3.5 kW

0.8 0.6 0.4 0.2

Injection [V]

0 6 5 4 3 2 1 0 -1

3.5 kW

0

0.02

0.04

0.06

0.08

0.1

Time [s] Fig. 3. Voltage readings from the sensors and control signal for the fuel metering valve during gasoline fueling.

For each operational point, five measurements were performed in order to test for repeatability, which was found to be within the ±2% accuracy range for fuel consumption and power measurements. All tests were executed at an ambient pressure of 1 atm, temperatures in the 295–298 K range and relative humidity of 50%. Only steady state operation was investigated, with the engine fully warmed-up. Operational load range was chosen so that it is as representative as possible for actual use, namely up to 70% of PRP, given that generators are usually sized by employing an appropriate safety coefficient of 1.3 [26]. Four load settings were investigated, for which resistive power consumers rated at 1, 2, 3 and 4 kW (at 230 V AC, 50 Hz single phase power), were connected to the generator; measurements at idle were also performed, equivalent to 0 kW load. All measurements were performed around the rated rotational speed of 3000 rev/min, (equivalent to a single phase electric power frequency of 50 Hz, given that engine and generator shafts are directly linked) resulting in a generator frequency range of 48–51 Hz during stable operation. Lower frequencies of 45 Hz and as high as 52 Hz were recorded at the times when resistive loads were connected and disconnected. These conditions were however brief and stabilization was achieved within a few seconds. Exhaust gas concentrations of carbon monoxide (CO), carbon dioxide (CO2) and unburned hydrocarbons (HC) were recorded using the non-dispersive infrared measurement principle, while for nitrogen oxides (NOx) and oxygen (O2) electrochemical methods were used. Opacity measurements were performed using the

light absorption principle. Accuracy and resolution values for all recorded parameters are compiled in Table 2. As previously stated, given that the control algorithm features no injection maps, operation with a wide range of fuels should be possible without performing any changes to the engine or control software. In order to test this ability, measurements were performed with gasoline, iso-butanol and combined operation of alcohol with natural gas. This latter condition featured a 10% substitution rate (mass flow based) of iso-butanol with natural gas from the main, and was designed to simulate the operation of the generator unit in combination with small scale biomethane production installations, for which gas quality and quantity is known to vary within wide ranges [27], thus requiring the use of a support fuel. Investigations on the cold start performance of the control strategy with each fuel type were beyond the scope of this research. Nonetheless, it should be mentioned that compensation for the change in relative air–fuel ratio was done by the proposed control system without any intervention on the software when switching from one fuel type to another. Average valve opening times increased by 30% for iso-butanol compared to gasoline and by 15% when the alcohol was partially substituted with natural gas. Properties of the three fuel types used in the trials and listed in Table 3, were evaluated using a combined literature review and specific determinations; gasoline density was measured according to ISO 3675, research and motor octane numbers were evaluated using the specifications of ISO 5164 and 5163 respectively, while natural gas composition was determined through gas chromatography according to ISO 6975.

3. Results and discussion 3.1. Fuel system effects An important requirement for electric power generators is that rotational speed is maintained within a tight range in order to supply the required voltage and frequency. Variations of ±1% of rated rotational velocity are acceptable [30] during stable operation; to test the ability of both carburetor and proposed injection control to comply with this requirement, an analysis of engine speed was performed at the four chosen load settings. Fuel conversion efficiency and emissions were evaluated as well, along with a brief economic analysis of payback periods for the additional costs associated with the implementation of the more complex control system. It should be mentioned that all results presented in this subsection were obtained with gasoline fuel. A drop in engine speed was recorded as load increased (Fig. 4) and supplied voltage decreased from 225 V to as low as 200 V

A. Irimescu et al. / Applied Energy 121 (2014) 196–206

Gasoline [28]

Isobutanol [29]

Natural gas

LHV (MJ/kg) Density (kg/m3) at 15 °C Octane number [RON/MON] Stoichiometric air–fuel ratio (–)

42.9 720–775 95/85 14.7

33.1 802 113/94 11.2

49.6 0.691 130/120–130 17.1

Engine speed [rev/min]

3100

Engine speed [rev/min]

3100

Engine speed [rev/min]

3100

Engine speed [rev/min]

3100

Engine speed [rev/min]

throughout the studied load interval. This was due to the limited capacity of the mechanical governor to cover the entire load range. Therefore, the inability of the generator to comply with voltage requirements of ±10% for grid power [31] is not related to the injection control strategy or the fuel system. It should be noted, however, that the generator did comply (for the studied load range) with the requirement of voltage fluctuations between + 10% and 15% specified for users and networks not connected to transmission lines. For both fuel systems the engine speed (results presented in Fig. 4 were recorded during 5 s of steady state operation for each operating point) showed comparable values at the same power level. The difference in stability is more evident at low load, while at higher electric power both injection and carburetor featured comparable variations of engine speed. This can be attributed to the response time of the oxygen sensor in the first case and different flow phenomena for the simpler fuel system, coupled with changes in the indicated mean effective pressure. More detailed

3100

3000 Average and ±1% limits

2900

0 kW

Carburetor

2800

3000 2900

1 kW

2800

3000 2900

1.9 kW

2.7 kW

3000 2900

3.5 kW

3000 2900 2800 1

2

3

Time [s]

4

3100

3100

Engine speed [rev/min]

2800

0

3100

Engine speed [rev/min]

2800

Engine speed [rev/min]

Characteristic

investigations with an analysis of in-cylinder pressure would provide improved insight into such phenomena, but were beyond the scope of this research. Nonetheless, both fuel systems showed similar control capability with regard to voltage and frequency of supplied electric power. Stability, quantified through the coefficient of variation (COV) for engine speed, was also found to be roughly the same when switching from the carbureted to the fuel injected setup (Fig. 5). A slight improvement was recorded for the latter system, but the overall trend was the same and differences in the COV (also calculated for 5 s for each load level during stable operation) were not significant. It should be noted that from this point on, the lines presented in correlation with individual points in all illustrations are second order polynomial fit values that were found to be most representative for the variation trend of each series. Supplying excess fuel in carbureted engines is not an unusual strategy, given that slightly rich mixtures are preferred to lean ones at high load, thus greatly reducing the tendency to knock; improved stability at idle is obtained through the same procedure. Without any changes made to carburetor settings from the time when the equipment was purchased, the air–fuel mixture delivered to the engine was rich throughout the load range, and therefore fuel consumption was higher for the simple feed system compared to the one using the solenoid metering valve (Fig. 6). This is the main reason for the improvement in fuel conversion efficiency when employing the closed loop injection system. An average increase of 15% in efficiency is in line with

Engine speed [rev/min]

Table 3 Fuel properties.

Engine speed [rev/min]

200

5

3000 Average and ±1% limits

2900

0 kW

Injection

2800

3000 2900

1 kW

2800

3000 2900 1.9 kW

2800 3100 2.7 kW

3000 2900 2800 3100 3.5 kW

3000 2900 2800 0

1

2

3

Time [s]

Fig. 4. Engine speed at different electric loads for carburetor (left) and fuel injection (right).

4

5

201

A. Irimescu et al. / Applied Energy 121 (2014) 196–206

0.9 Carburetor Injection

3000

COV engine speed [%]

Engine speed [rev/min]

3050

2950 2900 2850

Carburetor Injection

0.8 0.7 0.6 0.5 0.4

2800 0

1

1.9

2.7

3.5

0

1

Electric power [kW]

1.9

2.7

3.5

Electric power [kW]

Fig. 5. Average engine speed (left) and coefficient of variation (right) at different loads.

38 35 32 29 26 Carburetor Injection

23 20 0

1

2

3

Electric power [kW]

4

employed, the combination of engine-exhaust gas treatment system needs to be considered. In fact, catalytic conversion efficiency can be considered as very high for NOx, given that raw emissions during stoichiometric fueling feature concentrations in the order of thousands ppm for this pollutant, with a very steep drop as mixtures become rich [33]. Smoke opacity was significantly reduced when employing the electronically controlled fuel metering method, with the same trend of an increase as load was higher for both systems. The variation of opacity seems to be decoupled from that of the relative air–fuel ratio, when analyzing emissions recorded for carburetor fueling. Given that rich mixtures result in significantly higher particulate emissions, it would have been expected that smoke opacity be lowered as the air–fuel mixture was slightly leaned out when load increased. This observation is important for the evaluation of fuel effects, considering the fact that the method used for such determinations is sensitive to the concentration of hydrocarbons in the exhaust gas stream [34]. Therefore, the discussion related to smoke opacity in the next section is more pertinent, given that a somewhat clear decoupling of HC and opacity measurements can be identified. In order to evaluate the effects of using the proposed fuel system in view of emissions standards, specific CO and combined HC with NOx emissions were calculated and compared to prescribed limits (Fig. 9). Requirements of CO limits imposed for non-road SI power units [35] within the power rating of the engine used in the study are fulfilled only in the high load range when using the carburetor, while for unburned and nitrogen oxides prescribed values are exceeded only in the light load region. It should be mentioned that for the standard considered, only CO limits are mandatory for all fuel types, and combined HC and NOx values apply only to gas fuelled engines. Nonetheless, the comparison gives a good idea of the proposed system’s capability to comply with current and future emissions standards.

Fuel conversion efficiency [%]

Fuel consumption [g/min]

measurements of exhaust gas concentrations for five chemical species, values that provided the basis for calculating relative air–fuel ratio (Fig. 7). While the formula used for calculating the latter parameter is accurate only around stoichiometric air–fuel ratios [32], the fact that the relative differences of around 15% are in line with the ones of conversion efficiency seems to make them reliable enough in order to give a broad idea of the main phenomena behind the modification of efficiency and emissions when comparing the two fuel systems. Carbon monoxide concentration in the exhaust gas stream is directly related to the relative air–fuel ratio and was much higher when using the carburetor compared to the injection system (Fig. 8). Unburned hydrocarbons (HC) emissions were significantly lower for the more complex fuel system for the same reason. A decreasing trend can be observed for both CO and HC concentrations as load increased, when investigating the carburetor setup; this variation can also be attributed to the change in relative air– fuel ratio (and therefore variation in combustion efficiency), given that mixtures were leaner for higher load. An interesting observation is that nitrogen oxide emissions are roughly the same for both fuel systems. This can be explained by the relatively rich mixtures (and thus reduced NOx formation rates) during carburetor operation and close to stoichiometric fueling when using the injection system. While the contribution of nitrogen oxides formation mechanisms (namely prompt, which is more likely to be prominent for rich mixtures, and thermal, which is more likely to occur for stoichiometric and lean operation due to higher temperature levels and oxygen concentrations in the burned gas region) is most likely very different due to the variation in air–fuel ratio, no evaluation could be performed solely on the data recorded in the present study. Even if more complex investigation, such as in-cylinder pressure analysis or spectroscopic measurements specific for optically accessible engines were to be

18 15 12 9 6 Carburetor Injection

3 0 0

1

2

3

Electric power [kW]

Fig. 6. Fuel consumption (left) and conversion efficiency (right).

4

202

A. Irimescu et al. / Applied Energy 121 (2014) 196–206

Volumetric efficiency [%]

Relative air-fuel ratio [-]

1.1 1 0.9 0.8 0.7

Carburetor Injection

80

70

60

50

Carburetor Injection

40

0.6 0

1

2

3

0

4

1

Electric power [kW]

2

3

4

Electric power [kW]

Fig. 7. Measured relative air–fuel ratio (left) and calculated volumetric efficiency (right).

18

18 Carburetor Injection

15

12

CO2 [%]

CO [%]

15

9 6

12 9 6 Carburetor Injection

3

3

0

0 0

1

2

3

0

4

1

2

3

4

Electric power [kW]

Electric power [kW] 300

80 Carburetor Injection

Carburetor Injection

250

NOx [ppm]

HC [ppm]

60 200 150 100

40

20 50 0

0 0

1

2

3

4

0

1

2

3

4

Electric power [kW]

Electric power [kW] 0.5

Carburetor Injection

K [1/m]

0.4 0.3 0.2 0.1 0 0

1

2

3

4

Electric power [kW] Fig. 8. Measured exhaust gas concentrations for pollutant emissions.

After performing the analysis of power quality, conversion efficiency and pollutant emissions, it was evident that the fuel injection system, in combination with a three way catalytic converter, ensured greatly improved performance and reduced environmental impact. Economy and reliability is crucial for any equipment, and the first parameter is decisive in consumer choice when comparing different small scale electric power units. While sensors used in the automotive field have proven extremely

reliable, therefore making the simple carburetor completely comparable with fuel injection from this point of view, costs increase significantly when employing the complex system. Therefore, the economics of choosing one system instead of the other need to be evaluated. An initial assessment of costs associated with fitting an integrated injection and emissions control system such as the one used for the experiments, revealed that compared to the carburetor

203

A. Irimescu et al. / Applied Energy 121 (2014) 196–206

recuperated within 1000–4600 h of operation, depending on the settings of the carburetor (Table 4); calculations were performed considering an average load factor of 50% and assuming a fuel price of 1.5 Euros/liter or equivalent to 0.17 Euros/kWh. Recommended oil change interval is 100 h or 6 months, while the emissions durability period given by the manufacturer is 500 h; this should give an idea of the expected lifetime of such equipment. At first glance, calculated payback periods seem extremely long, especially if carburetor settings are properly adjusted and checked. Nonetheless, given that fuel prices feature an increasing trend, the benefits of using fuel injection systems could be greater. Further improvements can be obtained by optimizing ignition timing, given that the injection control unit can be upgraded for adjusting spark advance as well.

40

2000 Carburetor

30

1000

20 EPA CO limit

HC+NO x [g/kWh]

CO [g/kWh]

Injection

1500

10

500 EPA HC+NO x limit

0 3.5

0 1

1.9

2.7

Electric power [kW]

3.2. Effects of fuel properties

Fig. 9. Comparison of specific emissions with limits imposed through EPA standards [35] for the appropriate engine category.

Table 4 Estimated payback periods for the fuel injection and emissions control system compared to carburetor fueling. Relative air–fuel ratio carburetor setting

Hours of operation to payback

0.80

0.85

0.90

0.95

970

1380

2190

4610

version, the new generator unit would feature a price increase of 60–70%, comparable with more complex versions equipped with an inverter and control module designed for improved economy. When considering fuel savings, the additional cost would be

After evaluating the benefits of using the proposed injection system as compared to the simpler carburetor, the effects of using different fuels was investigated. Gasoline was replaced with isobutanol and combined alcohol with natural gas dual fueling was employed to test multi-fuel operation. All these conditions were studied using the injection system; following the first step that featured gasoline operation, for which the natural gas feed was closed, the fuel tank was emptied and iso-butanol was added, after which in the third stage tests were performed with the gaseous fuel line opened. An interesting observation is that average engine speed was consistently higher (even though differences were reduced, lower than 1%) when switching from gasoline to iso-butanol (Fig. 10). This could be explained by a slight improvement of volumetric efficiency when using the alcohol. The drop recorded when combining iso-butanol with natural gas can be explained by different combustion development, given that methane features lower

1

2950

engine speed

3050

[%]

Gasoline Isobutanol Isobutanol & Natural gas

2900

COV

Engine speed [rev/min]

3100

3000

2850

0.9

Gasoline Isobutanol Isobutanol & Natural gas

0.8 0.7 0.6 0.5

0

1

1.9

2.7

0

3.5

1

1.9

2.7

3.5

Electric power [kW]

Electric power [kW]

4

18

Efficiency variation [%]

Fuel conversion efficiency [%]

Fig. 10. Average engine speed (left) and coefficient of variation (right) at different loads and for three fuel types.

15 12 9 6 Gasoline Isobutanol Isobutanol & Natural gas

3

Gasoline Isobutanol Isobutanol & Natural gas

2 0 -2 -4 -6

0 0

1

2

3

Electric power [kW]

4

0

1

2

3

Electric power [kW]

Fig. 11. Efficiency (left) and the effects of fuel properties compared to gasoline (right).

4

204

A. Irimescu et al. / Applied Energy 121 (2014) 196–206

Volumetric efficiency [%]

Relative air-fuel ratio [-]

1.1 Gasoline Isobutanol Isobutanol & Natural gas

1.05

1

0.95

0.9

80

70

60 Gasoline Isobutanol Isobutanol & Natural gas

50

40 0

1

2

3

0

4

1

2

3

4

Electric power [kW]

Electric power [kW]

Fig. 12. Measured relative air–fuel ratio (left) and calculated volumetric efficiency (right) for different fuel types.

18

0.5 Gasoline Isobutanol Isobutanol & Natural gas

16

CO2 [%]

CO [%]

0.4 0.3 0.2 0.1

14 12 10 Gasoline Isobutanol Isobutanol & Natural gas

8

0

6 0

1

2

3

4

0

1

Electric power [kW]

3

4

80

30 Gasoline Isobutanol Isobutanol & Natural gas

Gasoline Isobutanol Isobutanol & Natural gas

60

NOx [ppm]

25

HC [ppm]

2

Electric power [kW]

20 15 10

40

20 5 0

0 0

1

2

3

0

4

1

2

3

4

Electric power [kW]

Electric power [kW] 0.5

Gasoline Isobutanol Isobutanol & Natural gas

K [1/m]

0.4 0.3 0.2 0.1 0 0

1

2

3

4

Electric power [kW] Fig. 13. Effect of fuel properties on measured exhaust gas concentrations for pollutant emissions.

burning speed and requires more advanced ignition settings [36]. Nonetheless, all three fueling strategies featured comparable average engine speed and variations of this parameter. The coefficient of variation was also comparable to carburetor fueling, even though higher values were recorded when changing the fuel from gasoline to iso-butanol and then combining the alcohol with natural gas. While the first phenomena can be attributed to changes in

vaporization behavior (given that iso-butanol features very low saturation point), the latter increase when adding the gaseous fuel is more likely to be caused by changes in combustion development. A slight drop in efficiency was recorded for iso-butanol fueling compared to gasoline (Fig. 11). This is in line with findings of previous work developed on an automotive power unit that found a decrease of up to 9% [37]. The difference between the two fuel

A. Irimescu et al. / Applied Energy 121 (2014) 196–206

types was up to 5% relative values for this study, a phenomenon that can be attributed to the fact that for this setup the fuel metering valve was fitted upstream of the throttle, as compared to the one used in the previous work [37] that was port injected. Therefore, fuel evaporation was most likely more complete for the present study. This explanation seems to be confirmed by the slight improvement when substituting 10% of the liquid fuel with natural gas. Measured relative air–fuel ratio values ranged from slightly rich to lean, from 0.98 to 1.01 (Fig. 12). This is comparable to the range achieved by automotive injection systems that employ narrow band oxygen sensors. Slight improvements in volumetric efficiency when switching from gasoline to iso-butanol are in line with the observed trend in average engine speed. Reduced differences in calculated volumetric efficiency between the three fuel types, at the same electric power level, can be attributed to the aforementioned increased accuracy of measured relative air–fuel ratio around the stoichiometric point, compared to rich mixtures, as was the case for carburetor fueling. A slight reduction of carbon dioxide emissions was to be expected when changing the fuel from gasoline to iso-butanol and then combining liquid and gaseous energy sources, mainly due to lower carbon content in the fuel; differences were, however reduced (Fig. 13). Comparable values of CO emissions were recorded for all three fuel types, showing that the catalytic converter operated at optimum efficiency for all investigated conditions. A slight increase of HC concentrations in the exhaust gas stream can be identified especially at high load for the alcohol and iso-butanol combined with natural gas, as compared to gasoline operation; differences were however reduced for this parameter as well. Comparable values of NOx concentrations were recorded for all three fuel types, with a trend of rising exponentially as load was increased. This was to be expected, as temperatures are higher throughout the working cycle when engine load increases; catalytic conversion efficiency can be estimated as very high when comparing recorded nitrogen oxides concentrations with values usually found for raw emissions. Consistently high conversion efficiency of carbon monoxide, unburned hydrocarbons and nitrogen oxides in the exhaust gas stream serves as further validation of the control algorithm. Given that a clear decoupling of HC and smoke opacity measurements was established in the previous sub-section of the article, the effects of reduced particulate emissions when switching from gasoline to iso-butanol and then combined alcohol with natural gas fueling can be attributed to fuel effects rather than sensitivity of the measuring equipment to certain characteristics of the unburned fuel. While reduced as overall effect, the decrease in smoke opacity is in line with findings of other work that investigated the use of butanol in port injection engines [13,38]. Even lower particulate emissions were recorded when substituting part of the alcohol with natural gas, a variation that was to be expected, given that fuel droplets that enter the cylinder are the main reason for high values in this emissions category.

4. Conclusion Performance parameters and emissions were investigated for a small scale generator powered by a spark ignition engine. An original control algorithm was validated for an injection system using gasoline, iso-butanol, as well as combined liquid and gaseous fueling. Similar power quality ratings were obtained for both carburetor and proposed fuel control setup. Improved efficiency of up to 15% on average was obtained mainly due to mixture leaning from rich air–fuel ratios when using the carburetor, to stoichiometric ones when employing the closed loop injection system.

205

Significant reductions in exhaust gas CO, HC and NOx concentrations were obtained mainly due to stoichiometric operation and using a three way catalytic converter. The same trend was recorded for smoke opacity as well when substituting gasoline with isobutanol, with additional reductions obtained when part of the liquid fuel flow was replaced with natural gas. A brief economical assessment for the engine used in the study revealed that payback times for the proposed injection system are within the expected lifetime of the equipment, with further benefits identified as possibly achievable through ignition timing control. While the results presented are limited to a single case, they give a good idea of the improvements that can be achieved by using the proposed fuel system in combination with a three way catalytic converter. They also show how the control algorithm performs when using different fuel types, proving that it is capable of multi-fuel operation. Increased fuel conversion efficiency throughout the lifetime of the equipment, low emissions and flexibility in the choice of fuel are the main benefits of the proposed system as compared to the simple and less capital investment intensive carburetor. The significant reduction in CO and HC concentrations in the exhaust gas stream is bound to be an extremely important aspect in light of ever tighter emissions standards. Consistent reduction of particulate emissions could be a major benefit as well, even though this pollutant category is not yet a critical issue for SI power units. Also, fuel flexibility will be a significant advantage when considering fuel availability, an issue that is likely to be very important in the future. Acknowledgement This work was supported in part by CNCSIS-UEFISCSU, project number 22/06.08.2010, PN II-RU code TE 39/2010. References [1] Ross K. Europe’s energy cash crunch bites. Power Eng Int 2013;20:22–6. [2] Manfren M, Caputo P, Costa G. Paradigm shift in urban energy systems through distributed generation: methods and models. Appl Energy 2011;88:1032–48. [3] Obara S, Morizane Y, Morel J. A study of small-scale energy networks of the Japanese Syowa Base in Antarctica by distributed engine generators. Appl Energy 2013;111:113–28. [4] Jeong C, Kim T, Lee K, Song S, Chun KM. Generating efficiency and emissions of a spark-ignition gas engine generator fuelled with biogas–hydrogen blends. Int J Hydrogen Energy 2009;34:9620–7. [5] Lee K, Kim T, Cha H, Song S, Chun KM. Generating efficiency and NOx emissions of a gas engine generator fueled with a biogas–hydrogen blend and using an exhaust gas recirculation system. Int J Hydrogen Energy 2010;35:5723–30. [6] Klimstra J. Natural gas quality: impact on DG. Cogener On-Site Power Prod 2013;13:12–9. [7] Roy MM, Tomita E, Kawahara N, Harada Y, Sakane A. Comparison of performance and emissions of a supercharged dual-fuel engine fueled by hydrogen and hydrogen-containing gaseous fuels. Int J Hydrogen Energy 2011;36:7339–52. [8] Sáinz D, Diéguez PM, Urroz JC, Sopena C, Guelbenzu E, Pérez-Ezcurdia A, et al. Conversion of a gasoline engine-generator set to a bi-fuel (hydrogen/gasoline) electronic fuel-injected power unit. Int J Hydrogen Energy 2011;36:13781–92. [9] Kosmadakis GM, Rakopoulos CD, Demuynck J, De Paepe M, Verhelst S. CFD modeling and experimental study of combustion and nitric oxide emissions in hydrogen-fueled spark-ignition engine operating in a very wide range of EGR rates. Int J Hydrogen Energy 2012;37:10917–34. [10] Szwaja S, Jamrozik A, Tutak W. A two-stage combustion system for burning lean gasoline mixtures in a stationary spark ignited engine. Appl Energy 2013;105:271–81. [11] Olesky LM, Martz JB, Lavoie GA, Vavra J, Assanis DN, Babajimopoulos A. The effects of spark timing, unburned gas temperature, and negative valve overlap on the rates of stoichiometric spark assisted compression ignition combustion. Appl Energy 2013;105:407–17. [12] Vancoillie J, Demuynck J, Sileghem L, Van De Ginste M, Verhelst S, Brabant L, et al. The potential of methanol as a fuel for flex-fuel and dedicated sparkignition engines. Appl Energy 2013;102:140–9. [13] Gu X, Huang Z, Cai J, Gong J, Wu X, Lee CF. Emission characteristics of a sparkignition engine fuelled with gasoline-n-butanol blends in combination with EGR. Fuel 2012;93:611–7.

206

A. Irimescu et al. / Applied Energy 121 (2014) 196–206

[14] Deng B, Yang J, Zhang D, Feng R, Fu J, Liu J, et al. The challenges and strategies of butanol application in conventional engines: the sensitivity study of ignition and valve timing. Appl Energy 2013;108:248–60. [15] Ahn KH, Stefanopoulou AG, Jankovic M. AFR-based fuel ethanol content estimation in flex-fuel engines tolerant to MAF sensor drifts. IEEE T Contr Syst T 2013;21:590–603. [16] Yildiz Y, Annaswamy AM, Yanakiev D, Kolmanovsky I. Spark ignition engine fuel-to-air ratio control: an adaptive control approach. Control Eng Pract 2010;18:1369–78. [17] Chen X, Wang Y, Haskara I, Zhu G. Optimal air-to-fuel ratio tracking control with adaptive biofuel content estimation for LNT Regeneration. IEEE T Contr Syst T 2014;22:428–39. [18] Venugopal T, Ramesh A. Effective utilisation of butanol along with gasoline in a spark ignition engine through a dual injection system. Appl Therm Eng 2013;59:550–8. [19] Venugopal T, Ramesh A. Experimental studies on the effect of injection timing in a SI engine using dual injection of n-butanol and gasoline in the intake port. Fuel 2014;115:295–305. [20] Wu X, Daniel R, Tian G, Xu H, Huang Z, Richardson D. Dual-injection: The flexible, bi-fuel concept for spark-ignition engines fuelled with various gasoline and biofuel blends. Appl Energy 2011;88:2305–14. [21] Daniel R, Xu H, Wang C, Richardson D, Shuai S. Gaseous and particulate matter emissions of biofuel blends in dual-injection compared to direct-injection and port injection. Appl Energy 2013;105:252–61. [22] Pipitone E, Beccari S. Performances improvement of a S. I. CNG BI-fuel engine by means of double-fuel injection. SAE Technical Paper 2009-24-0058; 2009. doi: http://dx.doi.org/10.4271/2009-24-0058. [23] Subramanian KA, Mathad VC, Vijay VK, Subbarao PMV. Comparative evaluation of emission and fuel economy of an automotive spark ignition vehicle fuelled with methane enriched biogas and CNG using chassis dynamometer. Appl Energy 2013;105:17–29. [24] Honda Engines, GX390 Owner’s manual. ; [accessed March 2013]. [25] MOSA GE 6000 BS Use and Maintenance Manual Spare Parts Catalogue, ; [accessed May 2013].

[26] Kaldellis JK. An integrated model for performance simulation of hybrid wind– diesel systems. Renew Energ 2007;32:1544–64. [27] Irimescu A, Calin L, Jadaneant A. Aspects concerning the efficient use of biogas obtained from sewage water discarded by the beer industry. J Environ Prot Ecol 2009;10:1137–45. [28] Fuel specifications, ; [accessed June 2013]. [29] Wallner T, Miers SA, McConnell S. A comparison of ethanol and butanol as oxygenates using a direct-injection, spark-ignition engine. J Eng Gas Turbines Power 2009;131:032802 [9 pages]. [30] Al-Saedi W, Lachowicz SW, Habibi D, Bass O. Voltage and frequency regulation based DG unit in an autonomous microgrid operation using Particle Swarm Optimization. Int J Electr Power Energy Syst 2013;53:742–51. [31] Standard EN 50160, Voltage characteristics of electricity supplied by public distribution networks. [32] d’Ambrosio S, Finesso R, Spessa E. Calculation of mass emissions, oxygen mass fraction and thermal capacity of the inducted charge in SI and diesel engines from exhaust and intake gas analysis. Fuel 2011;90:152–66. [33] Heywwod JB. Internal combustion engine fundamentals. New York: McGrawHill; 1988. [34] Giechaskiel B, Maricq M, Ntziachristos L, Dardiotis C, Wang X, Axmann H, et al. Review of motor vehicle particulate emissions sampling and measurement: from smoke and filter mass to particle number. J Aerosol Sci 2014;67:48–86. [35] EPA United States Environmental Protection Agency Exhaust Emission Standards for Nonroad Spark-Ignition Engines 19 Kilowatts and Below, Phase 3, Class II, ; [accessed October 2013]. [36] Irimescu A. Comparison of combustion characteristics and heat loss for gasoline and methane fueling of a spark ignition engine. Proc Rom Acad Ser A 2013;14:161–8. [37] Irimescu A. Performance and fuel conversion efficiency of a spark ignition engine fueled with iso-butanol. Appl Energy 2012;96:477–83. [38] Tornatore C, Marchitto L, Valentino G, Corcione FE, Merola SS. Optical diagnostics of the combustion process in a PFI SI boosted engine fueled with butanol gasoline blend. Energy 2012;45:277–87.