Knock Reduction Measures in the Gas Fuelled Internal Combustion Engine

Studies on the influence of applying various technologies for combustion knock reduction have been presented in the paper. Among others, investigation concerning the following: over-expanded cycle, variable valve timing, internal and exhaust gas recirculation, leaning the combustible mixture and cooling the in-cylinder charge were of the interest. The research works were focused on impact of these technologies on both knock intensity reduction, and engine performance and toxic emissions. Results presented in the paper were coming from experimental investigation based on in-cylinder combustion pressure data acquisition. Additionally, knock intensity calculation methods were discussed. They are based on in-cylinder combustion pressure pulsations. Combustion knock intensity expressed by the maximum peak of the incylinder pressure pulsations shows a strong negative correlation with both the EGR ratio and relative equivalence ratio – lambda. With respect to a catalytic converter installed on the exhaust pipe line, applying EGR appears as better solution for knock reduction then leaning the combustible mixture because the catalytic converter needs stoichiometric mixture for effective NOx reduction. Furthermore, application of the over-expanded cycle to the hydrogen or coke gas fueled IC engine significantly reduces intensity of potential knock by 50 % in comparison to Otto cycle for all loads. Additionally, over-expanded cycle contributes to increase in engine thermal efficiency. Summing up, all the presented measures and technologies can be successfully implemented into practice in stationary engines as well as in traction engines, both of them working on either natural gas or gaseous renewable fuels.

Strong environment protection regulations on carbon dioxide emission from vehicles force scientists to work on several liquid and gaseous substances which can substitute fossil fuels. Thus, investigation is concentrated on applying hydrogen or biomass based fuels as alcohols, bio-oils and methane rich gases for this purpose. It was found that alternative fuels can be burnt in both spark ignition and compression ignition engines as it was discussed in. However, burning gaseous renewable fuels in compression ignition engines needs a pilot diesel fuel for combustion initiation. On the other hand, the spark ignited engine features with knock effect occurring at unfavorable operating conditions, particularly in case hydrogen is applied as a fuel. As far as knock comes from auto-ignition of premixed combustible mixture, such phenomena might also be observed in a dual-fuel compression ignition engine fueled with premixed gaseous fuels.
Among other things, combustion knock is one of the most important constraints, which significantly reduces both the performance, torque output and efficiency of the spark ignited (SI) internal combustion (IC) engine. On the other hand, hydrogen as a gaseous fuel for the IC engine, gives it a significant advantage over other commonly used fuels including gasoline, diesel, and ethanol [1]. Thus, mixture of hydrogen and natural gas can be considered as valuable fuel to the SI engine. 20 % H 2 in natural gas mixture is known as hythane and was successfully implemented in test buses. These advantages include potential of near zero emissions of the regulated emissions of carbon monoxide (CO), nitric oxides (NO x ), and hydrocarbons (HC) while simultaneously eliminating carbon dioxide (CO 2 ) emissions, and greatly improved cold start capability. Hydrogen's combustion properties enable the development of an engine that would meet all current and future emissions standards. It is reported [1,2] that hydrogen as engine fuel has great potential for improving engine thermal efficiency due to its rapid combustion, which approaches the ideal constant-volume heat release combustion process.
Combustion knock results from the spontaneous ignition of the end gas mixture inside the engine cylinder ahead of the propagating flame [3]. When this abnormal combustion occurs, there is a rapid release of the chemical energy in the remaining unburned mixture, causing high local pressures and generating propagating pressure waves in amplitude of several hundred (kPa) across the cylinder space. These pressure waves are transmitted through the engine body and results in the combustion ping heard by observers. The large amplitude pressure waves of the hot combusted gases in the cylinder are the primary reason for mechanical engine failure through increased thermal and mechanical stress (Fig. 1). The pressure waves that result from combustion knock occur at frequencies that are acoustic vibration modes of the cylinder geometry. Hence, intensity of combustion knock can be easily calculated on the basis of high frequency oscillations on pressure inside the engine cylinder during combustion. The amplitude of these oscillations can be as high as peak of combustion pressure, so they could increase the incylinder pressure twice [4][5][6]. This combustion causes faster wear of the engine, thus from this point of view is considered as unwelcome combustion phenomenon. Detection of combustion knock is the necessary measure in the modern engine control system to prevent the engine from working with abnormal combustion as combustion knock is. The detection is usually realized by a vibration sensor installed on the engine body. There is almost linear correlation between knock intensity and engine body vibration, thus such monitoring of combustion knock can be easily conducted. Although, running the engine under knock is undesirable due to shortening the engine durability, but working just at the edge of combustion knock can provide favorable conditions for maximum engine performance. It is particularly observed when bad quality fuels are applied. The knock detecting system provides real-time monitoring of the engine and in case of Наука техника. Т. 19, № 4 (2020) Science and Technique. V. 19, No 4 (2020) и the knock the control system retards spark timing to eliminate this improper combustion. The best for the engine is to work without the knocking. Thus, prediction of potential combustion knock is the main target of research in the field. It particularly stimulates development in knowledge of knock genesis, knock precursors, fuel properties, cylinder design, etc. Theory and modeling the combustion knock can be valuable tools for computational simulation of engine working cycles and their analysis. Although, several, experimentally determined, correlations between knock intensity and quantities as pre-ignition temperature and ignition lag are satisfactory recognized, but origin of this phenomenon still needs to be theoretically examined in details. Thus, it makes the sub-model of combustion knock more universal and provides opportunity of knock prediction in various engine types as well as experimental models do it for specified engines with limits in working conditions.
Combustion knock occurring in the IC spark ignited engine is a phenomenon, which is sensitive to several quantities. They can influence on knock intensity at various impacts. Several quantities are identified and their impact on the knock is moreless recognized. They can be distinguished as follows:  temperature of pre-ignition gases -it has significant impact on an induction time (pre-ignition reactions) and strongly impacts on combustion knock [1,4]. Compression ratio is one of the parameters influencing on the pre-ignition temperature;  O-H (oxygen to hydrogen) ratio, which might be expressed by the lambda -relative equivalence ratio, which is a factor showing hydrogen active radicals dilution in fresh combustible mixture;  oxygen dilution in the intake charge -provided by exhaust gas recirculation (EGR);  presence of hot spots, exhaust residualsthey can work as extra sources for uncontrolled airfuel ignition;  initial swirls, intake manifold shape, combustion chamber shape;  engine load -fuel dose -higher engine load requires higher amounts of fuel delivered to the engine cylinder, hence it provides energy released during combustion and also influences the knock intensity.
Majority of these above mentioned quantities are not independent from each other, for example fuel dose and lambda, that could change the knock. As known through literature survey [1,2,7,8], research on knock reduction in the SI engine by diluting air-fuel combustible mixture is rich of credible results. It is confirmed, that engines working on air-gas lean mixtures are more resistant to knock [7]. Similar impact is observed, while applying EGR. It contributes to remarkable reduction in knock intensity [8]. These both measures can be used to reduce knock intensity in stationary gas fueled engines for heat and power generation. The paper presents measures to effectively reduce combustion knock intensity even to completely eliminate it. Therefore, there are two major effective methods to reduce combustion knock in the engine. Knock can be reduced by either applying EGR or fueling the engine with lean air-fuel mixture. Investigation presented concerns impact of both these measures on reducing combustion knock intensity in the SI reciprocating engine fueled with hydrogen rich gases. Investigation involves hydrogen and natural gas as fuels to the IC engine. Hydrogen addition to methane contributes to increase in laminar flame speed (LFS) of this combustible mixture, that shortens entire combustion duration in the engine cylinder and finally leads to higher engine overall efficiency. Natural gas with a calorific value of 9.7 kW⋅h/nm 3 constitutes a high-value fuel for effective power generation.

Knock intensity determination
As mentioned, knock intensity is determined on the basis of pressure pulsation intensity. Fig. 2 and 3 presents combustion pressure and high-pass filtered pressure pulsations for hydrogen combustion in the SI engine under two various compression ratios (CR) of 10 and 12. As observed, pressure pulsations from Fig. 3 (CR = 12) are several times higher than these from Fig. 2 (CR = 10). Hydrogen as a fuel was not selected by coincidence. It is observed, that pressure pulsations start immediately after combustion development. They are not formed at the end combustion phase as is typical for gasoline knock. As seen in Fig. 4 combustion knock by end-gas self-ignition is ac-    Knock metrics. There are several metrics for in-cylinder pressure pulsations. Here the peak pressure (PP) is proposed as the credible metrics which are easy to be determined. Thus, PP metrics from an individual combustion event as defined as follows max HFPO ( ) .
And PP for test series can be calculated where PP i -peak pressure (PP) of the HFPO i taken from the single combustion event (Fig. 5a); PP -mean of the PP i (test series usually contains 120 combustion events) (Fig. 5b).  Based on the PP metrics correlation between knock intensity and compression ratio was determined for the hydrogen SI fuelled engine (Fig. 6). As seen, combustion knock to specific limit features with so-called combustion instabilities, which come from unstable combustion. Typical knock as end gas self-ignition defined by Heywood for the gasoline fueled engine starts with its intensity several times higher, as seen for CR over 12 in Fig. 6 [9].

In-cylinder exhaust residuals impact
In-cylinder exhaust gases residuals (ExR) influence on engine in-cylinder combustion progress as diluent. However, exhaust residuals can differ from external EGR with temperature and moisture content. As known, diluent simply lengthens combustion duration, so reduces combustion knock intensity. In this case, higher exhaust residuals might lead to opposite correlation due to increase in temperature of fresh charge by higher ExR content [10,11].
Experimental set-up. The tests were conducted on the single cylinder research engine based on Ricardo's Hydra platform. The engine specifications are provided in Tab. 1 [10]. The engine setup is shown in Fig. 7a. Variable valve overlap was achieved through phasing the intake valve timing relative to a fixed exhaust valve timing. The intake valve cam center line location (ICCL) varies from 90 to 150 crank angle (CA) deg after TDC (top dead center), while the exhaust valve cam centerline location (ECCL) was fixed at 95 CA deg before TDC. The valve profiles are shown in Fig. 7b. Detailed test specifications are provided in Tab. 2.   Results and discussion. As observed in Fig. 8 ExR plots are in line with engine overlap. Moreover, higher engine rotational speed (2000 rpm) forces higher exhaust residuals percentage ratio.
Results presented in Fig. 9 confirm that higher ExR contributes to shorter combustion duration, both for these two specific phases: CA0-10 and CA10-90. Hence, temperature impact from higher и ExR on combustion duration is more significant rather than its dilution.
As expected, shorter combustion duration is strictly correlated with higher heat release rate (HRR) as depicted in Fig. 10.
Overlap, CA deg  Knock intensity expressed by metrics PP goes down with higher overlap, as is seen in Fig. 11. The correlation is strongly exponential, hence one can conclude that it is influenced by chemical reac-tion rate that is affected by both temperature and exhaust residuals content in the fresh charge.
Overlap, CA deg Fig. 11. Knock intensity PP vs overlap

External exhaust gases impact
As mentioned, diluent can reduce knock intensity. Exhaust gases consisted of two major species CO 2 and H 2 O can play crucial role in diluting combustible mixture without changing its equivalence ratio (relative equivalence ratio λ). As observed in Fig. 12a, b both higher percentage of EGR and higher ratio of leaning the combustible mixture leads to increase combustion duration [7,8].
Target of the research works was to answer which measure (EGR or lean mixture) for air-fuel mixture dilution with respect to reducing combustion knock is more effective. To solve this problem, the dilution ratio χ was introduced follow the equation where χ -dilution ratio in percentage; n st -number of moles (including fuel and oxidizer) of the stoichiometric air-gas mixture; n dil -number of moles of additional gas working as a diluent. This dilution ratio χ, expressed as diluent mole percentage, makes it possible to compare knock reduction by EGR with knock reduction by lean mixture.
Dilution ratio defined with the equation (3) depends on lambda -relative air-fuel equivalence ratio in case the diluent is air. Hence, following the equation (3) a lean combustible mixture can be treated as the stoichiometric mixture diluted in extra amounts of air. If EGR is applied for diluting the mixture, then the dilution ratio χ dil is a function of the EGR ratio.
where n st -number of moles of exhaust gases recirculating to the engine intake port. Number of moles n i of each gas can be easily calculated where m i -mass of the i agent; MW i -molar weight of the i agent.
Finally, comparison between EGR and leaning impacts were able to be performed. As depicted in Fig. 13, EGR can be considered as more effective measure to reduce combustion knock rather than leaning the combustible mixture. It can be explained by CO 2

Impact of over-expanded cycle on knock intensity
Investigation in applying over-expanded cycle for both knock and exhaust toxic emissions reduction were described in detail in the works [12][13][14][15]. The test bed in the research consists of a singlecylinder spark ignited supercharged engine. The Kistler sensor type 6055 with the Kistler signal charge amplifier type 5011 were used in-cylinder pressure measurements. Pressure in intake manifold was measured by the Motorola MPX 4250 pressure sensor. TDC position and crankshaft positions were measured by the Megatron type MOB 1024/5/BZ/N encoder with resolution of 1024 pulses per revolution. Data measured during the tests were collected with aid of National Instruments data acquisition system NI USB-6251 which was connected to the PC were the data were stored. The over-expanded cycle was achieved by modification of the intake valve timing. In this case the intake valve closure point was located before bottom dead center during intake stroke. It is known as Miller cycle. The test rig specification is presented in Tab. 3.
All measurements were done at wide open throttle conditions in two stages. In the first stage the optimal spark timing was determined for three boost overpressure, in the second stage to find optimal spark timing, the load of the engine was changed by changing a boost overpressure from 0 to 0.6 bar. The excess air ratio λ was maintained constant during the tests and it was close to stoichiometric mixture expressed by λ = 1.
Test matrix is presented in Tab. 4. The optimal ignition timing was determined for three loads of и the engine which were set by changing the boost overpressure. In each set point of the load the ignition timing was changed from 40 to 20 degree bTDC (before top dead center). Body temperature of the engine was maintained at 70 °C and the temperature of the fresh air after supercharger was maintained at 24 °C. Methane was delivered to the engine by a gas mixer which was installed in the intake manifold. The air/fuel ratio was set according to the measurements of air and fuel flows by rotor flow meters.
On the basis of these data the pressure-volume, pressure-time and engine load vs spark timing plots were made. These data allowed to perform analysis on combustion process and knock occurrence. Fig. 14 presents knock intensity PP vs boosting pressure for the engine working with clas-sic Otto cycle and the engine with over-expanded cycle (O-E). As one can notice knock is significantly reduced by applying over-expanded Miller cycle to the engine. It is obviously known that knock is also in strict inline correlation with engine load. As presented in Fig. 15, knock intensity increases with increase in engine load for Otto cycle engine unlike the over-expanded cycle engine, where intensity of knock slightly goes down even the engine load increases.   The results of the research confirm benefits which come from applying the over-expanded cycle to the supercharged SI IC engine. On the basis of the research the following conclusions can be drawn: the optimal ignition timing does not depend on engine cycle type -classic Otto or overexpanded. Moreover, the optimal value is the same for the Otto cycle and the over-expanded cycle engine. In both versions of the engine, increase in boost pressure does not affect on optimal ignition timing, in all cases ignition timing was 28° before TDC.