Introduction The harmful impact of the motor vehicles on the environment and the growing concern regarding the pollution of the environment resulted with the change in development requirements of Internal Combustion (IC) engines that are now focused on the prevention of future global pollution. Greenhouse gases are proven to be one of the main reasons for global warming [1]. Therefore, there is a number of regulations that aim to reduce carbon dioxide (CO2) emissions from motor vehicles [2]. Reduction of CO2 emissions can be achieved by using the energy sources without carbon (electrical energy generated from renewable sources [3,4], hydrogen [5,6]), by using fuels with lower carbon ratios (e.g. methane) or by reducing fuel consumption [7]. On a large scale the first and the second solution could be a long-term solution, while for the short to medium term solution the best way to reduce CO2 is by reducing fuel consumption. Compression ignition (CI) engines have shown significant reduction of fuel consumption over the last decade, while spark ignition (SI) engines still have not reached efficiencies and CO2 emission of CI engines. On the other hand, harmful emissions (nitrogen oxides (NOX), particles) of CI present a significant concern which triggered a number of notifications regarding restrictions of use of certain CI engines in some city centres (e.g. Stuttgart, München) [8–10]. Therefore, further reduction of fuel consumption of SI engines while simultaneously keeping the low emissions of harmful exhaust gasses (total hydrocarbons (THC), carbon monoxide (CO), NOX) is required. For achieving higher efficiency in contemporary SI engines, several technologies are usually applied: turbocharging [11], optimization of combustion chamber design and in-cylinder flow, variable valve timing [12], direct injection [13], etc. Turbocharging, as one of the main technologies for achieving higher engine efficiency, reduces the ratio of friction and pumping losses by the increase of engine load. At the same time, the increase of engine load increases the tendency of the engine to knock [14]. Engine knock, which is one of the abnormal types of combustion, can cause permanent engine damage [15]. Over the years it has been accepted that knock is a consequence of auto-ignition of the gas in front of the turbulent premixed flame, called “end-gas” [16]. The end-gas is compressed not only by the piston movement but also by the expansion of the burned gases. Therefore many factors can increase the probability of knock occurrence, such as increased inlet temperature and pressure as well as elevated compression ratio. Therefore, one of the main obstacles in furtherer development of the SI engines and a in further increase of the engine load is the occurrence of knock [17]. For that reason a development of different knock suppression strategies is being performed [15], e.g. retarding of the spark timing, enriching of the mixture, using of fuels with higher octane numbers, using of cooled exhaust gas recirculation (EGR), cooling of the intake air and enhancing the turbulence as a measure to increase the speed of normal combustion. Literature review shows that exhaust gas recirculation (EGR), which was first implemented for reduction of NOX emissions, might be a useful method for suppression of knock [18]. In [19] with the increase of EGR ratio to 20%, the brake specific fuel consumption (BSFC), emission of NOX and of particle number were reduced by 7%, 87% and 36%, respectively. One of the recently developed engines described in [20] achieved an increase of efficiency with respect to its previous version by employing cooled EGR. Furthermore, [21] showed that the pumping loss gradually decreases with the increase of exhaust gas recirculation rate, while at the same time the efficiency of the high-pressure cycle increases due to the decrease of the heat transfer and exhaust gas energy loss. In [22] simple knock model coupled with a comprehensive cycle simulation of the engine showed that in addition to the suppression of knock, the use of EGR resulted with a slight increase of the brake thermal efficiency. The thermodynamic reasons for the above mentioned increase included slightly lower heat transfer and increase of the ratio of specific heat. The levels of EGR in SI engines are limited since it was shown that the increase of EGR influences combustion stability presented by the coefficient of variation of IMEP (CoVIMEP). In [24] it was shown that CoVIMEP increases with higher EGR ratio and decreases with higher compression ratio and higher intake air pressure. On the other hand, in [13] it was shown that retarding of the spark timing results in an increase of CoVIMEP, while [12] showed that EGR enables advanced spark timing, so the effect of increased CoVIMEP with the increase of EGR is partially compensated by the advanced spark timings. However, there is a limit on the level of EGR. In [25] with 10% of the EGR and advanced spark timing the low level of CoVIMEP was obtained, but further increase of the EGR resulted in an unallowable increase of CoVIMEP. In [23] this limit was slightly higher. Although, at 15% EGR level the CoVIMEP was slightly increased (CoVIMEP = 2.5%) it was still far below the usual limits. The further increase of the EGR above 15% resulted in an unallowable increase of the CoVIMEP. In order to effectively use the EGR as a method for suppression of knock, engine designers have to understand the sources of EGR influence on knock. The EGR suppresses knock by three main factors: influence of EGR on flame speed, chemical influence on auto-ignition tendency (ignition delay) and thermal influence on the end-gas temperature [26]. The influence on the flame speed is shown as a decrease of flame speed with the increase of the EGR amount. Decrease of flame speed results with the decrease of mass burning rate and therefore pressure and temperature profiles of the end-gas change. On the one hand, slower combustion results in lower in-cylinder pressures and temperatures, therefore, reducing the tendency towards knock, while on the other hand the available time for auto-ignition increases. Chemical influence on auto-ignition is defined as the influence of species that come from the EGR on the chemical kinetic behaviour of the mixture and therefore on the tendency towards auto-ignition. On one hand, it was shown that the CO2 and water vapour (H2O) that come from the EGR could reduce chemical ignition delay for the same pressure, temperature and excess air ratio conditions [27]. It was also shown that the NOX (formed during combustion), which is recirculated back into the engine cylinder, could increase engine tendency to knock [28]. In [29] it was shown that the influence of NOX is different for different in-cylinder pressure vs end-gas temperature (p-Tend-gas) history due to the nitrogen oxide (NO) reactivity (oxidation) at different temperatures. Reactivity of NOX depends on whether the p-Tend-gas history is in high, in NTC (Negative Temperature Coefficient) or in low-temperature regime. If p-Tend-gas history is in low or NTC temperature regime, the addition of NOX causes knock suppression, while if it is in the high-temperature regime the NOX can promote knock. Finally, thermal influence is defined as the influence of the EGR on the end-gas temperature level. The influence of EGR is two-folded. First, the EGR has different temperature compared to the fresh intake air, therefore the mixture of the EGR and intake air has a different temperature than fresh intake air, while the size of this difference depends on whether the EGR was cooled or uncooled. The second influence is through the thermal properties of the EGR mixture, where EGR has higher specific heat, and therefore the increase of the temperature of the end-gas caused by the compression is lower. In both thermal cases, the increase of the end-gas temperature will result in more pronounced tendency towards knock and vice versa. As it can be noticed, the use of EGR introduces various effects that are in some cases similar and in other cases are opposite. Even though the use of EGR in SI engines has been researched over the years [30] and there are publications regarding its use, literature review showed that there is no in-depth experimental analysis of the sources of EGR influence on SI combustion with the emphasis on end-gas temperature. This study aims to bridge that gap by evaluation of the influence of the end-gas temperature, which changes with the addition of EGR, on the tendency of the engine towards knock. In the study experimentally acquired results are shown and discussed in three parts. First part of the study shows the benefits and limits of the EGR dilution on the combustion, second part emphasize peak end-gas temperature and knock indicators, and third part shows the analysis at different engine speed and engine load provided by boosted intake pressure showing the influence on the engine efficiency. Experimental setup The experimental work is performed on the experimental setup at Laboratory for IC Engines and Motor Vehicles of the Faculty of Mechanical Engineering and Naval Architecture in Zagreb. The experimental setup consists of alternating current (AC) Dynamometer, SI engine upgraded with EGR, intake air heater, indicating equipment, emission sampling, air flow and fuel flow measurement. Additionally, temperature and pressure of the intake, exhaust, engine head and oil are measured. Acquisition of data from SI engine in various operating conditions was performed by the use of in-house built acquisition system which is prepared so that it stores the data of the boundary conditions (pressure, temperature, throttle position, etc.) and emissions. The IC engine used in this experimental setup is heavily modified HATZ 1D81Z, originally a single cylinder diesel engine. It has two valves per cylinder and combustion chamber of a toroidal type. In order to run as SI engine, the modifications were made to the piston, ignition and fuel injection systems. Also, for the purposes of engine control, the measurement of the crank angle position was added. Modification of the piston included machining of the piston top that resulted in lower engine compression ratio (CR) required for the engine to operate with gasoline in SI combustion mode. CR was lowered from originally 20.5 to 12. Further on, for the measurement of piston position two hall sensors, one on the crankshaft and one on camshaft were implemented. Based on the piston position that was measured, the engine controls regarding fuel injection and ignition timing was operated. For the purpose of control, the signals of fuel injection and ignition were monitored by the engine indicating equipment. The operation of the injection and ignition system was enabled by the in-house solution prepared in LabVIEW software package. BOSCH ZS-K-1X1PME ignition coil with spark plug NGK IRIDIUM CR7EIX was placed centrally on top of the cylinder where diesel injector was removed. Gasoline port fuel injector (BOSCH EV-6-E) was placed on custom made intake manifold with a constant fuel pressure of 3 bar and with fuel mass consumption measured by OHAUS Explorer mass scale. Additionally, intake manifold contains ports for temperature and CO2 emission sample measurement. Besides the equipment for upgrading the engine, additional testbed equipment was used on experimental setup to set up and measure boundary conditions. The intake air mass flow was measured by the TSI 2017L laminar mass flow meter. In order to be able to control the intake and in-cylinder temperature after the measurement of air flow, the intake line was heated by 18 kW OSRAM SYLVANIA SureHeat Air Heater. Further on, for the exhaust gas emissions, two different analyses were used. Acquisition of the emission of CO and CO2 is done by Bosch ETT 8.55 EU gas analyser, while Total Hydrocarbons (THC) are measured by Environnement GRAPHITE 52M heated FID analyzer. EGR flow was controlled by the EGR valve (Valeo 170A9) with a water-cooled heat exchanger. The amount of EGR is controlled by the opening of the electric valve allowing exhaust gas to recirculate from exhaust to the intake manifold. The amount of recirculated EGR is taken as the same proportion as the ratio of the measured CO2 at the intake in relation to the measured CO2 in the exhaust. This is done by using two CO2 analysers. The first one is already mentioned Bosch analyser, placed on the exhaust manifold and second, for the measurement of the intake CO2 Environnement MIR 2M infrared (IR) analyser was used. The amount of the recirculated exhaust gas is calculated by the equation described in [30]. Fast crank-angle based signals typical for combustion engines are acquired and processed by engine indicating equipment. It consists out of the hardware AVL IndiSmart 612 [31] and software package Indicom [32]. Main data that were acquired with this equipment were combustion data from the engine, e.g. in-cylinder pressure, intake pressure and crank position. Further on software calculated and enabled real-time monitoring of the indicated mean effective pressure (IMEP), the coefficient of variation of IMEP (CoVIMEP) and crank angle for 50% mass fraction burned (CA50). Also during measurements monitoring of the knock occurrence expressed via in-cylinder pressure oscillations, spark timing (ST), dwell time, injection timing, fuel spray amount and intake pressure at the intake valve closure (IVC) were possible. For measurement of the intake pressure, the low-pressure AVL LP11DA sensor was used, while for the high in-cylinder pressure AVL GH14DK sensor was used. For the measurement of the high-pressure part of the engine, cycle data was stored with the resolution of 0.1°CA, while the remaining part of the engine cycle was recorded with the resolution of 0.5°CA, all for 300 consecutive cycles per operating point. Measurement procedure Experimental work is performed at 1000 and 1600 rpm, with and without EGR dilution, at compression ratios 10 and 12. In order to have comparable results, some adjustments regarding boundary conditions were made. First, the spark sweeps from no knock to high knock conditions without EGR were performed. The intake pressure was slightly throttled so that there is certain reserve for an increase of intake pressure for compensation of total fuel energy when EGR is introduced. The air to fuel mixture was set to stoichiometric which resulted in the fuel flow that on average gave constant energy of the fuel per cycle. The first part of the study shows external and cooled EGR set to the levels of 0%, 11%, 15% and 20% with the goal of obtaining higher engine efficiency through lowering the tendency to knock and enabling earlier spark timings. During experiment performance and emissions data were measured and evaluated. Through first part of the study it was shown that by using the EGR the abnormal combustion (knock) can be effectively suppressed and that the engine performance in terms of engine efficiency can be improved. It was also shown that spark timing and intake pressure need to be optimized to achieve higher efficiency when diluting intake mixture with EGR. In that optimization, generally, the intake pressure has to be increased and spark timing advanced. It was also shown that the increase of the amount of EGR enables advanced spark timings, i.e. advanced combustion phasing (with the same knock limit), through lower end-gas temperature. The advanced combustion phasing with the increased levels of EGR resulted with increased IMEP for the same input of fuel energy [23]. The second part of the study is performed by experimental tests that employ a new approach, where intake temperature is varied by the EGR and the intake air heater (AH) to achieve pre-set end-gas temperatures. The methodology is based on the experimental setup with the SI engine that uses cooled EGR system and air heater placed in the intake which can compensate the temperature change obtained by the application of the cooled EGR. By heating of the intake air the thermal influence of the EGR on the end-gas temperature is annulled in some cases, and in other cases, the temperature is further increased to completely show the influence of temperature on the occurrence of knock. This resulted in a number of operating points with different levels of EGR, different end-gas temperature profiles and knock intensities. Since there is a large number of obtained operating points, a detailed analysis of the results by looking at the pressure profiles and at intake charge mixture compositions were performed only on selected operating points. The selection of operating points was performed with two criteria. First set of operating points had different amounts of EGR, different spark timings and combustion phasing (CA50) positions but had similar peak end-gas temperatures. This set of data aims to look whether there is an influence of combustion phasing if the peak end-gas temperature is the same. The second set, contains data with different levels of EGR, different spark timings and intake temperatures, but have similar CA50 positions. Finally in the third part of the study is expanded to additional cases with higher intake pressure (high load) and lower engine speed of 1000 rpm with compression ratio of 10. Third part of the study investigates EGR influence on the efficiency in order to show weather the improvements of engine performance are result of the improvements in the high pressure cycle or reduced gas exchange losses. Data was acquired with spark sweeps for both naturally aspirated and boosted operating points. Results First part of the study results through two comparisons: one with fixed spark timings and one with optimized spark timings. For the fixed spark timing the comparison is made with different amounts of applied EGR dilution in order to observe the influence of EGR on the combustion with the same spark timing (ignition) conditions. The application of EGR results in reduced IMEP because the combustion temperature is reduced which results in longer combustion duration, lower peak pressures and lower peak of rate of heat release. Further on, in the second comparison, in order to achieve optimal results from the combustion with EGR dilution spark timing was optimized. Optimized spark timing measurement points are selected from the acquired values of IMEP and average MAPO. The optimized conditions for the selected measurement points were highest achieved IMEP without the occurrence of knock. The results showed that the contribution of the diluent effect on IMEP rise was 4.4%, 5.8% and -11.2% for the dilutions of 11%, 15% and 20% respectively. Increase of cylinder pressure with the input of EGR is influenced by the higher intake pressure and by advancement in spark timing. Heat release shows longer combustion duration and lower peak rate of heat release with the application of EGR dilution. In the second part of the study the peak end-gas temperature and knock indicators were observed, the analysis aimed to show whether there is a clearly defined chemical or thermal influence of EGR on knock occurrence. For that purpose, the results of the peak end – gas temperature with respect to knock indicators are observed. For the measure of knock, three different indicators were used. First one is the average Maximum Amplitude Pressure Oscillation (avg. MAPO) value defined according to the procedure described in [33]. MAPO value of each cycle can be considered as the amplitude of high-frequency pressure oscillations caused by knock. MAPO value is obtained for each cycle (300 cycles) and then averaged for each operating point. Besides avg. MAPO value the number of cycles exceeding the MAPO value of 1 bar and of 1.5 bar are taken as additional measures of knock. In the knocking region there is a slight general trend of increase of knock intensity with the increase of end-gas temperature, but also that the spread of temperatures for similar knock intensity is significant. This spread is caused by some other influences, e.g. end-gas pressure and temperature profiles, measurement uncertainty, etc. Strong influence of end-gas temperature on transfer from no knock to knock is clearly shown. From acquired data, the selected points are taken in the further in-depth analysis. In order to enable analysis of the results, the selected operating points were classified into 3 main groups (none knock, medium knock and severe knock) where third (severe) group was additionally divided into 3 subgroups in accordance with their knock intensity. It is observed that in general the increase of peak end-gas temperature increases as the conditions move from no knock to severe knock conditions. In order to strengthen the conclusion the detailed analysis of pressure and temperature profiles is made. Results imply that EGR significantly influences ROHR profile, which is caused by the influence of EGR on the flame speed. The increase of EGR results in slower combustion. Because of the slower combustion and of the thermal influence of EGR, the spark timing had to be retarded to obtain similar peak end-gas temperatures. This resulted in different values of CA50. By heating of the intake charge, the in-cylinder temperatures before combustion increased which enabled faster combustion, and required delayed spark timing to obtain similar peak end-gas temperatures. One can notice that heated EGR cases had equal or even faster combustion observed by ROHR profile. Cases with significantly different temperature profiles, pressure histories and different EGR values, but with similar peak end-gas temperatures and p-Tend-gas profiles have similar knock intensities. Results for the third part of the study show the oprating points with and without boost of the engine intake pressure. Results show indicated efficency of the engine with higher efficiency achieved with increase of the amount of the EGR dilution. Presented results imply that positive effects of EGR dilution for the naturaly aspirated operating points are up to the 10% of the dilution and for the boosted operating points that limit was 24%. Limiting value for further dilution was 10% of CoVIMEP, which rises with more freakvent misfires occurrence. When emission of polutants are observed it is shown that emission of HC and NOX show significant decrease for both boosted and naturaly aspirated operating points with application of EGR dilution. Conclusions Conducted study generated from the experimental IC engine research resulted in following conclusions: - The increase of the EGR dilution significantly impacts the increase in CoVIMEP indicating that there is a limit on the possible EGR dilution. - By using the EGR system, detonation combustion can be successfully supressed and thus engine performance improved. - To achieve a higher indicated efficiency with EGR dilution applied adjustment of the spark timing and intake pressure is required. As a rule, with the increase in the EGR dilution, to maintain high efficiency, it is necessary to increase the intake pressure and to shift to earlier spark timing. - By increase of the EGR amount, a higher indicated efficiency is achieved with or without intake boosting, i.e. higher indicated work is achieved with lower fuel consumption with the same level of knock combustion and less emission of pollutants in the engine with EGR relative to the engine with no EGR dilution. - When adding the EGR dilution, the increase in the indicated efficiency is result of the better energy conversion during high-pressure cycle while the negative gas exchange work is approximately constant. - By applying the EGR dilution, nitrogen oxide emissions are significantly reduced as repercussion of the lower in-cylinder temperature. Detailed analysis of the selected operating points revealed the following: - The increase of EGR leads to the slower flame propagation at the similar intake charge temperatures and therefore to lower end-gas temperature. This effect can be compensated by the advance of spark timing or by the increase of intake charge temperature. A possible consequence of the second compensation is that although previously it was observed that fractions of cooled EGR larger than 15% were not possible because of the low flame speed and increased CoVIMEP, the larger fractions of EGR might be possible if the EGR is not cooled. - The results with similar peak end-gas temperature and different CA50 revealed that even though the p-α profiles, Tend-gas-α profiles and EGR levels were significantly different, the intensity of knock was very similar and behaved according to the p-Tend-gas histories, with predominant influence of the peak end-gas temperature. - The results with similar CA50s and with different peak end-gas temperature showed that case with higher peak end-gas temperature have a stronger knock and vice versa, regardless of the small changes in peak pressures or EGR values. This all leads to the overall conclusion that the predominant factor in a tendency towards knock, under the used conditions, is the end-gas temperature profile, i.e. peak end-gas temperature.