1. Introduction

The performance improvement of present internal combustion engines is an important task for engine manufactures, this being motivated by different reasons, like increased human concern on the environment impact of the ongoing use of such engines or the development of alternative electrically-powered vehicles. A high-voltage spark plug is the common device to initiate combustion at each cycle in a reciprocating engine. Usually having a “J” shape, the electric spark plug (ESP) is placed near the relatively cold wall of the cylinder, at a fixed position; this influences the formation and the development of the flame kernel during the initial stage of combustion. In addition, diluted air-fuel mixtures are more difficult to be ignited and operation at high-pressure inside the engine combustion chamber is limited when an ESP is used. Therefore, enhanced systems are investigated to initiate faster and more robust combustion and among these a) high energy spark plug ignition (like capacitive discharge, continuous discharge or high frequency multi-charge ignition); b) pulse power ignition (such as repetitive pulse spark or transient plasma ignition); c) radio-frequency ignition (spark, corona or micro-wave plasma ignition) or d) laser ignition (LI) can be mentioned.

LI offers several advantages compared to the ignition by ESP [1,2]. Thus, the laser beam is delivered into the engine cylinder through an optical window that is positioned on the cylinder wall; in this way the ESP protruding electrode is removed, and the flame kernel can develop freely, with no quenching. In addition, the ability to manipulate a laser beam offers the possibility to access any internal part of the cylinder; consequently, attempts can be made to increase the engine efficiency by igniting the fuel at various points, different of that where the ignition is done by the ESP. Furthermore, one laser beam can be manipulated, or several beams can be used, to target simultaneously more points of ignition inside the engine cylinder (i.e. to achieve a spatial control of the ignition). This approach can be accompanied by the temporal control of the laser beam, in which not only one but more laser pulses are delivered in a very short time for producing a single-ignition event (this is the so called ‘burst-mode operation’). Diluted air-fuels mixtures can be ignited or operation at high pressure can be achieved in these ways with LI.

For the first time the LI was successfully done by Dale et al. in a one-cylinder ASTM-CFR engine running at 1000 rpm speed [3,4]. A CO₂ laser at 10.6 μm wavelength was used in those experiments. The recorded data showed improved engine performances, i.e. increased peak pressure in the engine, higher brake power and lower brake specific fuel consumption for ignition by laser in comparison with ignition by ESP. Differences were not significant for carbon monoxide CO and total unburned hydrocarbons THC emissions, but more nitric oxide NO was produced by LI. This behavior was attributed to a faster combustion and thus to higher cylinder temperature during LI. Exhaust gas recirculation (EGR) was used to reduce NO. Worth mentioning is that LI allowed an extension of lean limit of operation, from air-fuel ratio A/F = 12:1 for normal operation up to A/F = 22.5:1; at this point the engine ran steady with LI, but barely could be operated by the ESP. It was also observed that moving the point of LI away of the cylinder wall produces much faster combustion and thus a more rapid rise of the cylinder pressure is obtained compared with classical ignition.

A four-cylinder engine was fully operated, for the first time only by LI, at the University of Liverpool [5]. The engine, a Ford Mondeo 1.6 litre Zetec motor, was ran on various 800 rpm to 2400 rpm speeds and on a wide 50 N⋅m to 75 N⋅m load torques, as well as at various ignition timings. Electro-optically Q-switched Nd:YAG laser with emission at 1.06 μm were employed for LI. The measurements concluded that LI improves the engine combustion stability, which was characterized in terms of coefficient of variance in indicated mean effective pressure (COV_IMEP) and of peak cylinder pressure position.

One could observe that all these experiments were performed with tabletop lasers of large size and heavy weight, and low output efficiency. Therefore, the need for a compact laser with dimension close to that of an ESP and that in addition must be robust enough to perform in adverse conditions of vibrations, temperatures and pressure, for example as those found in an automobile engine, triggered further research in this area. The solution to this technical challenge was proposed by Koefler et al. [6], consisting of a Nd:YAG laser passively Q-switched by Cr⁴⁺:YAG saturable absorber (SA) and that is pumped by a fiber-coupled diode laser operated in quasi-continuous wave (quasi-cw) mode. Based on this scheme, laser spark-plug prototypes were built by several research groups, like Kroupa et al. [7], Tsunekane et al. [8], Pavel et al. [9], Tauer et al. [10] or Schwarz et al. [11]. The world’s first gasoline engine vehicle that was ignited only by laser spark plugs was developed in Japan, being reported by Taira et al. in 2013 [12]. A few years later, LI of an automobile was also done in Romania [13–15]. Detailed aspects regarding the development and application of LI in different fields can be found in the review works published by Tauer et al. [10], Morsy [16], O’Briant et al. [17] about LI for aerospace propulsion, or Dearden and Shenton [18] and Pavel et al. [19] regarding LI in reciprocating engines.

LI of a real automobile engine [12] was obtained for lean air-fuel mixtures up to A/F~23, showing, in comparison with ignition by ESP, similar or improved engine combustion stability in terms of COV_IMEP; the data were recorded for the engine conditions of 1200 rpm speed and 73 N⋅m torque. Engine operation on a speed range from 1500 rpm to 2000 rpm and high 770 mbar up to 920 mbar loads (taken as the absolute pressure in the intake manifold) was investigated by Pavel et al. [13] at stoichiometric λ~1 air-fuel ratio. Improved combustion stability in terms of COV_IMEP and COV_Pmax (the coefficient of variability of maximum pressure) was observed. In addition, specific emissions like hydrocarbons HC, CO, nitrogen oxides NO_x and carbon dioxide CO₂ were determined. Measurements showed decreases of CO and HC, but also concluded that a slight increase of CO₂ and NO_x emissions is obtained from the engine charge ignited by LI. A similar behaviour, i.e. reduced CO and HC emissions and increased of CO₂ and NO_x emissions, was observed for the same 2000 rpm speed and high 920 mbar load. Some other results regarding ignition by laser can be mentioned, being obtained on different engines. A research performed by Saito et al. [20] on single-cylinder engine that was operated at different EGR rates showed better COV_IMEP, lower THC and increased NO_x emissions in comparison with ignition by ESP. Reductions of NO_x emissions, but also increases of CO and unburned hydrocarbons UHC were reported by Gupta et al. [21] for a 6-cylinder natural gas engine ignited by lasers. The investigations made by Grzeszik on a single-cylinder engine showed decreased fuel consumption and lower NO_x emissions for multi-point LI [22]. One could observe that some data are contradictory, especially regarding the NO_x emissions, most probably due to the specific engine operating conditions tracked in the experiments.

In this work new results obtained on a conventional multipoint fuel injection passenger automobile engine with four-cylinders that was operated by a compact LI system replacing the original spark ignition device are presented. Based on previous basic researches [13–15], more experiments were carried in order to study the effect of LI on the engine performance in terms of power, stability and pollutant emissions. The engine was operated at selected conditions considered as representative for the city traffic driving, of 2000 rpm speed and load equivalent of 2 bar brake mean effective pressure (BMEP). Furthermore, stoichiometric λ~1 air-fuel mixture, as well as a lean λ~1.25 air-fuel mixture, were considered. In general, the engine brake power and combustion stability improved under LI, with less influence at λ~1, but with significant rates for operation with the lean air-fuel mixture. Thus, when the engine was operated at λ~1.25, at optimum ignition timing, an increase in power by nearly 29% was evaluated for the engine ignited by laser in comparison with ignition by ESP. For optimum timing condition, significant reductions of brake specific fuel consumption (BSFC) and CO, with relatively same level of THC emissions at both λ values were measured for LI, while increases of NO_x emissions have been registered when this new LI system was used instead of the classical one. The faster combustion twisted by LI was considered as possible explanation for these results.

2. Experimental conditions

2.1 The laser spark-plug system

A block diagram of the experimental LI set-up is shown in Fig. 1(a). The LI system was realized with four laser spark plugs (LSP), each LSP resembling a classic ESP. In order to build such a compact laser, a composite Nd:YAG/Cr⁴⁺:YAG ceramic medium (Baikowski Co., Japan) was used in a monolithic-scheme resonator. The Nd:YAG active ceramic medium with 1.1-at.% Nd doping and about 8.5 mm length was diffusion bonded to a Cr⁴⁺:YAG SA ceramic with initial transmission T_i = 0.40. The total length of Nd:YAG/Cr⁴⁺:YAG was around 11 mm. The monolithic oscillator was obtained by coating the resonator high reflectivity mirror, of reflectivity R> 99.9% at the lasing wavelength of 1.06 μm, on the free surface of Nd:YAG. Because the optical pumping was made through this side, it was also coated with high transmission (T> 98%) at the pump wavelength of 807 nm. The resonator out-coupling mirror, with reflectivity R~50% (± 5%) at 1.06 μm, was coated on the exit side of Cr⁴⁺:YAG SA.

 

Fig. 1 (a) Block diagram of the experimental set-up. DL: Diode laser; OF: Optical fiber; LSP: Laser spark plug; CYL: Engine cylinder; ECU: Electronic control unit. (b) The engine equipped with the LI system is presented.

Download Full Size | PDF

The optical pump was done at 807 nm, employing fiber coupled (fiber with diameter of 600 μm and numerical aperture NA = 0.22) diode lasers (JOLD-120-QPXF-2P, Jenoptik, Germany) that were operated in quasi-cw mode. The pump-pulse duration was fixed at 250 μs whereas the repetition rate could be varied from a few Hz up to 100 Hz. For compactness, the pump light delivered from an optical fiber was coupled into Nd:YAG with a single achromatic lens of 6 mm focal length. By fixing the space between OF and this lens, to about 5 mm, laser pulses at 1.06 μm with energy between 1.7 mJ and up to 5.9 mJ could be obtained from the Nd:YAG/Cr⁴⁺:YAG by positioning the laser medium at different distances from the lens. In the final set-up each laser spark plug yielded pulses with 4 mJ energy and about 0.8 ns duration, which corresponds to a pulse peak power of nearly 5 MW; the pump pulse energy at 807 nm was around 40 mJ. The laser beam was then collimated and finally it was focused into the engine combustion chamber; worthwhile to mention, the focusing position was chosen to match the point where ignition is made by the corresponding classical spark plug. A sapphire window was used as the interface between an LSP and the corresponding engine cylinder. All optical elements (i.e. the Nd:YAG/Cr⁴⁺:YAG medium, lenses and the sapphire plate) were embedded in the laser spark metallic body with an epoxy adhesive resistant to temperature variation on a large −70°C to 170°C range and that possess high shear and peel strength. More information about the design and laser performances of such LSP could be found in [13].

A Renault multipoint fuel injection passenger car engine with four cylinders in line and 1598 cm³ total displacement volume, rated power of 64 kW at 5500 rpm speed, was used in the experiments; the engine was mounted on a test bench provided with the appropriate equipment for control and for collecting the relevant parameters of engine’s operation. The four LSP system was installed on the test engine, as shown in Fig. 1(b). In order to assure stable laser operation each LSP was positioned in a tubular-shape copper block that was cooled by re-circulating water. A personal computer was used to set, through an USB port, the temperature and the characteristics of the current pulses for each diode laser. It is important to mention that ignition triggered by LI was done with only one laser pulse per each cycle (i.e. not in a burst mode) in order to have similar conditions and thus to make a fair comparison with the ignition induced by ESP. The energy of electrical discharge for the ESP system was around 30 mJ [23].

2.2 The engine test bed

A schematic of the test bench layout is shown in Fig. 2. For ignition command the engine had an open ECU (Continental EMS 3132) that was related to an ECU control system (ETAS-INCA type). Ignition timings and air-fuel ratios were regulated by the means of INCA V7 2.1 software modifying the reference calibration of the engine. For loading the engine and measurement of its brake torque (this is the reactive torque developed by the dynamometer, which in steady-state condition equals the torque developed by the engine under test) and speed a Froude Consine AG 250-UK eddy current dynamometer (DYNO) was coupled to the engine by a dedicated drive shaft. The control of the test bed system (i.e., the control of the engine-dyno unit and variation of the low speed parameters), was done with a TEXCEL v6 instrument having installed the software Froude; data acquisition was made on 16 channels. A dynamic fuel meter AVL 733S in tandem with a fuel temperature conditioner AVL 753C were used to measure the engine fuel consumption. The in-cylinder pressures, the electric discharge parameters and the lift of the injector needle were acquired with an AVL Indimodul 621 system that was also coupled with a micro-IFEM AVL device. Cylinder 1 and 4 were equipped with an AVL ZF43 pressure detector; in addition, an AVL GU21D sensor was mounted on cylinder 1 by crossing the cylinder head.

 

Fig. 2 Schematic of the engine test bench. The four LSP devices are shown in the inset.

Download Full Size | PDF

The signals needed for data recording were triggered by a crank angle encoder AVL 365 that was connected to the AVL Indimodul 621. Pressure traces were recorded in groups of 500 consecutive cycles at each point of operation. Exhaust emissions have been measured by a gas analyser Horiba EXSA 1500L before the catalytic reactor (CATA) of the engine and registered in the TEXCEL computer. The accuracy for CO and NO_x measurements is ± 1% full scale, for a gas flow rate of 3 l/min and response time of maximum 23 s. The duration of measurements in steady-state condition was one to 3 minutes. A separated automation system, dedicated to control the engine coolant temperature by means of a water-water heat exchanger, was used to maintain its value in the normal range of 90°C (± 5°C). The oil temperature was also kept under 95°C. The engine was operated at constant 2000 rpm speed and constant load of BMEP = 2 bar, considered as specific for the use of an automobile in normal city traffic, and different ignition timings, which were investigated for the optimum one in terms of maximum brake torque. Inset of Fig. 2 presents the four LSP before being installed on the engine.

3. Results and discussion

The optimization of ignition timing has the objective of maximising the engine brake torque; in constant speed - constant load operation mode this means finding of maximum engine brake power, i.e. of the power developed by the engine at the output shaft. This is usually associated with the position of 50% from mass fraction burned around 8°CA after top dead center. When an engine operates with different air-fuel ratios, the optimum ignition timing is not in the same position and usually lean mixture operation leads to advanced ignition timings. In our experiments, for stoichiometric λ∼1 air-fuel ratio and at the optimum ignition timing, found as −40°CA for both ignition systems, higher power by 7.9% was obtained with LI (engine brake power of 5.86 kW) compared with the conventional ESP ignition (engine brake power of 5.43 kW), as shown in Fig. 3(a). For the lean λ∼1.25 air-fuel mixture operation the optimum timing was determined to be −46°CA when the engine was ignited by laser; at this point the engine brake power reached 5.91 kW. On the other hand, the engine brake power was 4.58 kW with the classical ignition system at the optimum timing of −44°CA. Thus, the LI system improved the engine brake power by 29% relative to the conventional ESP ignition. This behaviour could be explained through a more stable engine operation ensured by faster ignition and more robust combustion initiation induced by higher efficiency of LI in the breakdown stage of flame kernel formation with a substantial reduction of the initial stage of normal combustion.

 

Fig. 3 (a) The engine brake power and (b) The engine BSFC versus ignition timing for engine speed of 2000 rpm and BMEP = 2 bar at λ~1 and λ~1.25. Lines represent fittings with polynomial functions.

Download Full Size | PDF

The BSFC is one of the most significant parameters characterising the engine efficiency. It relates the engine fuel consumption by the brake power and because the engine was operated in constant speed-constant load condition its variation is normally to be like that of brake power but reversed. Figure 3(b) shows the change of BSFC relative to the ignition timing for the before mentioned testing conditions. At the optimum timings previously stated for the stoichiometric λ~1 and for the lean λ~1.25 air-fuel mixtures the decreasing of BSFC was of 7.4%, respectively of 21% when the engine ignition system was changed from classical ESP system to the new LI device.

One of the main important causes by which performance, efficiency and emissions of the petrol engines are negatively affected is the cycle-by cycle variation in pressure. The occurrence of this undesirable phenomenon is related to multiple causes, among which the local fluctuations in velocity flow field, the local variations of the mixture quality in the vicinity of the ignition source, as well as the fluctuation of the plasma channel between the spark plug electrodes contribute massively. It is well recognized that any improvement in the ignitability of the cylinder charge can reduce the amplitude of this phenomenon with beneficial effects on petrol engines operation. The cyclic variability is much more accentuated in lean or diluted mixtures in which also the ignition conditions are much more severe. To characterize this harmful phenomenon several parameters are usually used, such as: COV_IMEP, the coefficient of variation in indicated mean effective pressure (COV_IMEPH), as well as the coefficient of variation in peak fire pressure (COV_PFP) or in angular position of the peak fire pressure (COV_APFP). The most commonly is COV_IMEPH that defines the cyclic variability in indicated work per cycle; a maximum acceptable value of 10% to avoid drivability vehicle problems is considered as a limit for this parameter [23]. The variation of COV_IMEPH for our experiments is presented in Fig. 4(a). It shows that in stoichiometric λ~1 air-fuel mixture operation the COV_IMEPH is well enough under the requested limit, having quite the same value of ~1.3% for both ignition systems, while in lean mixture condition only the LI system could keep it under the 10% threshold. Its value of 15.9% for ESP is reduced by 50%, to 7.9%, when the LI system was used in optimum ignition timing conditions.

 

Fig. 4 Variation of (a) COV_IMEPH and (b) COV_α50% versus ignition timing for engine speed of 2000 rpm and BMEP = 2 bar at λ~1 and λ~1.25.

Download Full Size | PDF

Another parameter that offers information about combustion stability and its optimization is the angular position of 50% for the mass fraction burned MFB (α50%). Usually at optimum ignition timings, for stoichiometric λ~1 air-fuel mixture the regular values of α50% are in the range of 8°CA to 10°CA after top dead center, but with a relatively high coefficient of variation which approaches of 20%. For the lean λ~1.25 air-fuel mixture condition, the α50% of the MFB is delayed by 4°CA to 8°CA, having higher values for ESP, and the coefficient of variation of this parameter is approximately doubled. The evolution of this parameter, COV_α50% versus ignition timing is presented in Fig. 4(b). It shows that by advancing ignition timings the values of this coefficient grow significantly and the new range of COV_α50% for optimum timings becomes 35% to 45%.

Figure 5 presents the variation of combustion duration for the initial stage [Fig. 5(a)] and for the main stage [Fig. 5(b)]. These durations have been considered from the ignition event up to 5% mass fraction burned for the initial stage, and respectively from 5% to 90% of mass fraction burned for the main stage. Both ignition systems present normal trends of increased duration of the initial stage and reduced duration of the main stage for advanced ignition timings. It is worth to notice that in lean λ~1.25 air-fuel mixture operation there is a clear difference between the classical ESP ignition system and the LI one. For all ignition timings the difference in shortening combustion duration is evident at optimum timings, the reduction being of 3.2% in initial stage and of 18.8% in the main stage. For the stoichiometric λ~1 air-fuel mixture condition these differences are not so apparent. This result clearly emphasizes that LI has a better effect in lean mixtures, where ignitability conditions are severe, being thus able to reduce the length of initial stage and so the total combustion duration.

 

Fig. 5 (a) Initial stage of combustion and (b) Main stage of combustion versus ignition timing for engine speed of 2000 rpm and BMEP = 2 bar at λ~1 and λ~1.25.

Download Full Size | PDF

The variation of the carbon monoxide CO emission is shown in Fig. 6. By advancing the ignition timing the CO emission usually goes up due to the stronger dissociation reaction of CO₂, which is accentuated by the increasing of maximum temperature in cycle. This trend was observed in our experiments only for the stoichiometric λ~1 air-fuel mixture operation, when the total combustion duration was increased by 4°CA for ESP ignition and reduced by 2°CA for LI (Fig. 5). That is probably why the CO increasing effect is not so apparent for LI in stoichiometric condition. However, the LI system was able to reduce the CO emission over the whole ignition timing investigated by an average value of 20%, and by 22% at optimum timing of −40°CA. For the lean λ~1.25 air-fuel mixture operation, the general trend is reversed, and the CO emission is decreased by the advancement of ignition timings. This behaviour could be associated with a reduction of total combustion duration by 3°CA for ESP ignition and by 1°CA for LI (Fig. 5). Although, it can be emphasized that LI ensured a reduction of CO emission for all ignition values tested and by 37% for the optimum timings.

 

Fig. 6 Variation of CO emissions versus ignition timing for engine speed of 2000 rpm and BMEP = 2 bar at λ~1 and λ~1.25.

Download Full Size | PDF

Figure 7 presents the variation of total unburned hydrocarbon emissions THC (equivalent C3) relative to ignition timing. The general trend of increasing THC emissions by advanced ignition timing is in accordance with general influence of reducing oxidation time by shortening combustion duration and with similar results reported by other researchers [20]. LI is maintaining relatively the same level of THC emissions as for the ESP ignition in stoichiometric λ~1 air-fuel mixture operation at optimum timing of −40°CA. As it was already mentioned, in lean λ~1.25 air-fuel mixture operation the optimum timing was changed from −44°CA for ESP ignition to −46°CA for LI. This advancement of optimum ignition timing induced a shortening of combustion duration by 1.4°CA for the initial combustion stage and by 8.8°CA for the main stage of combustion; thus, the total combustion duration was reduced by 10.2°CA in the case of lean λ~1.25 mixture ignited by laser. Such important decreasing of combustion duration was associated with an increasing by 21% of THC emissions expressed in absolute ppm measurement unit. This increase of THC emissions for LI system could be explained by reduction of the total combustion duration when part of the hydrocarbons released form cervices and from the oil film by desorption in expansion cannot be completely burned. However, this apparently significant increasing is though converted into a slight reduction by 5.8% if the THC emissions are expressed in specific measurement unit (g/kW⋅h) [23] due to power improvement by 29% (from 4.58 kW to 5.91 kW) when LI was used.

 

Fig. 7 THC emissions versus ignition timing for engine speed of 2000 rpm and BMEP = 2 bar at λ~1 and λ~1.25.

Download Full Size | PDF

Figure 8 shows the variation of NO_x emissions by ignition timing. The NO_x formation is generally controlled by three parameters: the oxygen level available for reaction, the maximum temperature level and the combustion duration. Sometimes these parameters have contradictory effects. For the same oxygen level, which means the same relative air-fuel ratio, when ignition timing is advanced the general trend is of increasing the NO_x emissions by higher temperatures registered in the first part of combustion even if its duration is reduced. This trend was registered in our experiments for both air-fuel ratios and for the two ignition systems. Thus, comparison between these ignitions systems showed that even a small reduction in the initial stage of combustion occurred when LI was used instead of ESP, an impressive growing of NO_x emissions level, expressed in ppm, by 46% and by 287% was registered in stoichiometric λ~1 air-fuel mixture operation and in lean λ~1.25 air-fuel mixture condition, respectively. As for the THC emissions, these substantial risings are diminished when the NO_x emissions are expressed in the relative (g/kW⋅h) measurement unit. Thus, for the stoichiometric λ~1 mixture operation the increasing is by 35%, while for the lean λ~1.25 air-fuel mixture the notable increase of 287% becomes 203%.

 

Fig. 8 NOx emissions versus ignition timing for engine speed of 2000 rpm and BMEP = 2 bar at λ~1 and λ~1.25.

Download Full Size | PDF

Table 1 summarizes the comparative results obtained on the test engine when the air-fuel mixtures were ignited by classical spark plugs and by laser sparks.

Table 1. Relative Deviations in the Engine Parameters when LI Used Instead of ESP Ignition^a

View Table

It is worth pointing out that improvements in engine brake power and stability, as well as operation in lean air-fuel mixtures, can be achieved by employing LI at two or more locations. In addition, running of the laser device in ‘burst mode’ (i.e. the use of multiple pulses for a single ignition event) can decrease the misfires cases, whose probability could become high in the case of lean air-fuel mixtures. Furthermore, LI at multiple points could improve the reduced probability of ignition that was observed in the case of lean mixtures at high-speed flows [24]. Results reported recently using LI at two locations in a single-cylinder engine showed increased power and better stability in comparison with single-point ignition [25–27]. In these experiments the laser beams were directed into the engine cylinder with a set of reflecting mirrors [25,26] or with a spatial light modulator [27]. Several papers have reported on the development of laser-spark prototypes with multiple-beam output [28–30] and that could be also operated in burst mode. Still, to date there is no such a compact multi-beam laser that could be mounted directly on a four-cylinder engine. This topic could be a challenging subject for further research.

4. Conclusions

The ignition by laser is an attractive topic and the research performed up to now worldwide succeeded in realization of single-beam compact lasers resembling a classical spark plug; some of these devices were applied to passenger gasoline engine vehicles. A lot of researches were dedicated to LI topic because even small enhancements in the cylinder charge ignition can offer an important support for perfections in petrol engines performance, efficiency and emissions. In this sense, the results of the present study can be summarized in the following conclusions:

The research presented here can be continued by extending the investigated engine operation domain towards leaner mixtures, higher loads and superior speeds. New technologies already applied to passenger car petrol engines, as direct injection- spark control compression ignition, will offer probably large opportunities to the LI systems research and development in the effort for improving reliability and becoming cost competitive.