Combustion Characterization in Diesel Engine via Block Vibration Analysis

<div class="section abstract"><div class="htmlview paragraph">This study aims at building efficient and robust artificial neural networks (ANN) able to reconstruct the in-cylinder pressure of Diesel engines and to identify engine conditions starting from the signal of a low-cost accelerometer placed on the engine block. The accelerometer is a perfect non-intrusive replacement for expensive probes and is prospectively suitable for production vehicles. In this view, the artificial neural network is meant to be efficient in terms of response time, i.e. fast enough for on-line use. In addition, robustness is sought in order to provide flexibility in terms of operation parameters. Here we consider a feed-forward neural network based on radial basis functions (RBF) for signal reconstruction, and a feed-forward multi-layer perceptron network with tan-sigmoid transfer function for signal classification. The networks are trained using measurements from a three-cylinder real engine for various operating conditions. The RBF neural network is trained with time series from in-cylinder pressure signals and vibration signals measured on a cylinder which is distant from the one in which the pressure signal is measured. The accuracy of the predicted pressure signals is analyzed in terms of mean square error and in terms of a number of pressure-derived parameters. The location of the accelerometer has little influence on the accuracy of the reconstruction. This is confirmed also by the fact that the perceptron network, constructed in the second part of the work, is able to distinguish, from the accelerometer signal, among motored and fired conditions for any of the cylinders. Here, training data are again composed of time series obtained from the accelerometer, plus the corresponding target classes (fired/non-fired). Despite of the noisy character of the vibration signal and the distance from the cylinders, the perceptron network classifies correctly almost 100% of the signals.</div></div>

<div class="section abstract"><div class="htmlview paragraph">Water Injection (WI) has become a key technology for increasing combustion efficiency in modern GDI turbocharged engines. In fact, the addition of water mitigates significantly the occurrence of knock, reduces exhaust gas temperatures, and opens the possibility to reach optimum heat release phasing even at high load.</div><div class="htmlview paragraph">This work presents the latest development of a model-based WI controller, and its experimental validation on a GDI TC engine. The controller is based on a novel approach that involves an analytic combustion model to define the spark advance (SA) required to reach a combustion phase target, considering injected water mass effects. The calibration and experimental validation of the proposed controller is shown in detail in the paper. At first, the focus is on the open-loop branch, to evaluate the performance of the combustion model and its ability to manage Spark Advance (SA) taking in account the phasing implications of WI, maintaining a pre-defined combustion phase target. Then the closed-loop (CL) chain is introduced, defining a structure that allows reaching the target while keeping knock intensity (KI) levels under an established threshold using two different control levers (SA and WI). At the same time, the controller has been designed to minimize water consumption. The proposed controller has then been implemented in a Rapid Control Prototyping environment, to validate the strategy on a real engine. As shown in the paper, this approach allows to manage both combustion phasing and knock level in an engine equipped with WI.</div><div class="htmlview paragraph">The closed-loop controller was initially based on in-cylinder pressure signals both for knock intensity and combustion phasing measurements. To allow the proposed controller to be ready for an on-board implementation, there is the need to replace the pressure signal used from the in-cylinder pressure sensor, not available on commercial vehicles. For this, a specific signal processing algorithm has been developed to extract the angular combustion phase from the accelerometric signals available on-board, and it has been validated with experimental data.</div></div>

<div class="section abstract"><div class="htmlview paragraph">In the present work, an Auto Regressive Moving Average (ARMA) model and a Discrete Wavelet Transform (DWT) are applied on vibrational signals, acquired by an accelerometer placed on the cylinder block of a Spark Ignition (SI) engine, for knock detection purposes. To the aim of tuning such procedures, the same analysis has been carried out by using the traditional MAPO (Maximum Amplitude of Pressure Oscillations) index and an Inverse Kinetic Model (IKM), both applied on the in-cylinder pressure signals. Vibrational and in-cylinder pressure signals have been collected on a four cylinder, four stroke engine, for different engine speeds, load conditions and spark advances. The results of the two vibrational based methods are compared and in depth discussed to the aim of highlighting the pros and cons of each methodology. The presented outcomes show the capability of the vibration based detection algorithms in accurately monitor the presence of knock phenomena, and to determine its intensity according to the IKM and MAPO based methods. The comparison with a standard processing of the knock sensor output reveals the higher sensitivity of the proposed methodologies. Therefore, the possibility of implementation in modern on-board control units is foreseen, as well.</div></div>

<div class="htmlview paragraph">Making use of spark-plug-washer type cylinder pressure sensors and a high-performance 16-bit microprocessor, the authors have developed a new control system (Nissan ECCS) of ignition timing for gasoline engine.</div> <div class="htmlview paragraph">Use of this system results in effective control, enabling each engine to deliver maximum torque and minimum fuel consumption at all conditions, regardless of changes in environmental conditions, etc. Features of this system are summarized as follows: <ol class="list nostyle"> <li class="list-item"> <span class="li-label">(1)</span> <div class="htmlview paragraph">It is able to be realized that the ignition is kept near the MBT (Minimum advance for the Best Torque) by a feedback control, detecting accurately the peak timing (crank angle) of cylinder pressure from low frequency component signal.</div></li> <li class="list-item"> <span class="li-label">(2)</span> <div class="htmlview paragraph">The best performance of acceleration is obtained at all conditions, regardless of changes in fuel characteristics and environmental conditions.</div></li> <li class="list-item"> <span class="li-label">(3)</span> <div class="htmlview paragraph">This system is able to absorb variability in the dimensions and characteristics of components in the same specification engines, and then the ignition timing for the best fuel consumption is performed.</div></li></ol></div>

<div class="section abstract"><div class="htmlview paragraph">Ambient temperature conditions, engine design, fuel, lubricant and fuel injection strategies influence the cold start performance of gasoline engines. Despite the cold start period is only a very small portion in the legislative emission driving cycle, but it accounts for a major portion of the overall driving cycle emissions. The start ability tests were carried out in the weather controlled transient dynamometer - engine test cell at different ambient conditions for investigating the cold start behavior of a modern generation multi-point fuel injection system spark ignition engine. The combustion data were analyzed for the first 200 cycles and the engine performance and emissions were analyzed for 300 s from key-on. It is observed that cumulative fuel consumption of the engine during the first 60 s of engine cold starting at 10 °C was 60% higher than at 25 °C and resulted in 8% increase in the value of peak speed of the engine. The rate of pressure rise was significantly higher and prolonged for a number of cycles at 10°C compared to 25 °C and 45 °C. The cycle-to-cycle variation in the cylinder pressure at 10 °C was three times higher than at 25 °C. The first 60 sec of the cold start cumulative CO emission at 10 °C was approximately 3 times higher; cumulative HC emission was 3.5 times higher; cumulative NOx emission was 50% lower than that of at 25 °C. The particles in the size range of 50-200 nm are accounted for 60% at 10 °C and 6-8% at 25 °C & 45 °C ambient temperature conditions. The exhaust particles at the low ambient temperatures increased the exhaust particulate mass by 30 times at 10 °C in comparison with that of at 25 °C. The accumulated unburned fuel during the cold start period combusted abruptly and caused for the higher peak speed and exhaust emissions.</div></div>

<div class="section abstract"><div class="htmlview paragraph">Some SI (spark-ignition) engines fueled with gasoline for industrial machineries are designed based on the conventional diesel engine in consideration of the compatibility with installation. Such diesel engine-based SI engines secure a combustion chamber by a piston bowl instead of a pent-roof combustion chamber widely applied for SI engines for automobiles.</div><div class="htmlview paragraph">In the development of SI engines, because knocking deteriorates the power output and the thermal efficiency, it is essential to clarify causes of knocking and predict knocking events. However, there has been little research on knocking in diesel engine-based SI engines. The purpose of this study is to elucidate knocking phenomena in a gasoline engine with a re-entrant piston bowl and swirl flow numerically and experimentally. In-cylinder visualization and pressure analysis of knock onset cycles have been experimentally performed. Locations of autoignition have been predicted by 3D-CFD analysis with detailed chemical reactions. The prediction accuracy of the location of autoignition has been examined by comparing with experimental results. Initial and boundary conditions for 3D-CFD are obtained by TPA (Three Pressure Analysis), which is a 1D engine cycle simulation using measured intake/exhaust and in-cylinder pressures as the input data. Locations and timing of autoignition have been observed through an endoscope attached to the cylinder head. The visualization area through the endscope is determined based on the location of autoignition predicted by 3D-CFD analysis. The in-cylinder pressure is also measured simultaneously, and the maximum amplitude of the pressure oscillations after applying a high-pass filter to the in-cylinder pressure is used as knock intensity (KI) to determine knock onset cycles. From results of the in-cylinder pressure measurements and analysis by TPA, the relationships between in-cylinder conditions and KI are analyzed. The comparison between visualized and calculated location of autoignition have been shown good agreement in the case that both the maximum and the low engine speed conditions.</div></div>

<div class="section abstract"><div class="htmlview paragraph">In recent years, emission regulations have been getting increasingly strict. In the development of engines that comply with these regulations, in-cylinder pressure plays a fundamental role, as it is necessary to analyze combustion characteristics and control combustion-related parameters. The analysis of in-cylinder pressure data enables the modelling of exhaust emissions in which characteristic temperature can be derived from the in-cylinder pressure, and the pressure can be used for other investigations, such as optimizing efficiency and emissions through controlling combustion.</div><div class="htmlview paragraph">Therefore, a piezoelectric pressure sensor to measure in-cylinder pressure is an essential element in the engine research field. However, it is difficult to practice the installation of this pressure sensor on all engines and on-road vehicles owing to cost issues. Therefore, there have been several studies aiming to estimate the in-cylinder pressure using only the data available from the Engine Control Unit (ECU) without an additional pressure sensor.</div><div class="htmlview paragraph">An in-cylinder pressure prediction model for conventional light-duty diesel engines had been established by authors and described in this paper. First, the pressure during the compression stroke was estimated using a variable polytropic index. Next, the Wiebe function was applied to describe pressure during pilot and main combustion. This prediction of the in-cylinder pressure was established for real-time estimation.</div><div class="htmlview paragraph">The complete pressure prediction model consists of subordinate models, where physical phenomena are simplified for the real-time application of the model. The accuracy of the submodels has a direct influence on the error in the parent model the pressure prediction. Therefore, sensitivity analysis was conducted in this study to verify the effect of the submodels, including the polytropic index model, ignition delay model, heat loss model and combustion duration model, on the entire pressure prediction model. This study will contribute to establishing a more accurate pressure prediction through sensitivity analysis and will be useful to many studies that require in-cylinder pressure information.</div></div>

Efficient combustion control has increasingly become a quality requirement for automobile manufacturers because of its impact on pollutant and greenhouse gas emissions. In view of this, the management system development of modern internal combustion engines is mainly aimed at combustion control. The real-time detection of in-cylinder pressure characteristic parameters has a considerable significance on the closed-loop combustion control of the internal combustion engine. This paper presents a detection method in which the start of combustion, peak pressure, maximum pressure rise rate, and phase of maximum pressure rise rate are identified through vibration acceleration signal. In order to analyze the relationship between vibration and in-cylinder pressure signal, experimental data are acquired in a diesel engine by implementing various injection strategies and engine operating conditions (speed and load). The results show that the start of combustion can be detected by analyzing its relationship with the peak position of the filtered vibration signal, and the phase of the maximum pressure rise rate can be identified by examining its relationship with the zero-cross position that is adjacent to the right of the peak position. Moreover, the filtered vibration signals are also truncated in the same length and utilized as inputs for algorithms to detect the peak pressure and the maximum pressure rise rate. The algorithms are mainly performed on data compression (or feature extraction) and target regression. Major algorithms, such as one-dimensional convolutional neural network, compression sensing, wavelet decomposition, multilayer perceptron, and support vector machine, are tested. Various experimental results verify that for the test engine the phase detection accuracy of the start of combustion and maximum pressure rise rate is less than 1.7°CA for a 95% prediction interval width. For the detection of the peak pressure and maximum pressure rise rate, the normalized error threshold is set as 0.05, then the accuracies can be not less than 95%.

<div class="section abstract"><div class="htmlview paragraph">The use of piezoelectric cylinder pressure sensors is very popular during engine testing, but cylinder pressure information is becoming mandatory also in several on-board applications, where Low Temperature Combustion (LTC) approaches require a feedback control of combustion, due to poor combustion stability and the risk of knock or misfire.</div><div class="htmlview paragraph">Several manufacturers showed the capability to develop solutions for cylinder pressure sensing in on-board automotive and aeronautical applications, and some of them have been patented. The most straight-forward approach seems the application of a piezo-electric washer as a replacement of the original part equipping the spark plug; the injector could also be used to transfer the cylinder pressure information to the piezoelectric quartz, in diesel or Gasoline Direct Injections (GDI) engines.</div><div class="htmlview paragraph">The paper describes the features of signals acquired using piezoelectric washers, discussing possible applications, highlighting the factors which impact the sensors accuracy, and proposing algorithms to compensate potential errors in the evaluation of combustion metrics. The sensors have been first tested on a press, then in two different gasoline engines: a naturally aspirated V12 and a turbocharged 2 cylinders with Variable Valve Lift system (VVL). Signals have been compared to those obtained with lab-grade cylinder pressure sensors, with particular attention to peak pressure, combustion phase and knock intensity.</div><div class="htmlview paragraph">The main issue affecting the accuracy of cylinder pressure measurement using the piezoelectric spark plug washer is the effect of temperature variations both on the force transmitted by the thread to the washer and piezoelectricity properties.</div></div>

<div class="htmlview paragraph">This paper presents a study of fuel performance in spark ignition engines in relation to abnormal combustion phenomena such as surface ignition and rumble. This was accomplished by observation of the distribution of flame arrivals at ion gaps located within the engine cylinders, or peak pressure position, in terms of crank angle degrees.</div> <div class="htmlview paragraph">By this means, the surface ignition behavior and mass burning rates of LIB (leaded isooctane-benzene) and commercial fuel components were studied in multi-cylinder engines at 10:1 and 12:1 C.R. Additional information was obtained in two instrumented CFR engines operated at 10:1 and 6.75:1 C.R.</div> <div class="htmlview paragraph">It is concluded that the abnormally high peak cylinder pressures giving rise to rumble result from cycles in which combustion occurs early. Early combustion was caused by deposit-ignition of the charge but was also strongly affected by mass burning rate of the charge. Differences in mass burning rate between fuels also affected rumble differently for this reason. Furthermore, the mass burning rate for any fuel tested was greatly increased by the presence of combustion chamber deposits or by finely divided leaded deposits aspirated into the engine.</div> <div class="htmlview paragraph">A large part of the rumble intensity difference noted between the 0 and 100 LIB reference fuels (benzene + 3 cc TEL and iso-octane + 3 cc TEL) may be accounted for by the difference in their burning rates (10-14%). Whether or not this will be a factor in rating rumble by audible means depends on where the ignition timing is set. However, even when allowance is made for burning rate, the 0 LIB fuel shows more early ignitions than 100 LIB in the presence of either accumulated or aspirated deposits. It is, therefore, concluded that these fuels differ also in their tendency to be ignited by deposits.</div> <div class="htmlview paragraph">This information suggests that when rating the rumble tendencies of fuels and engines, a distinction should be made between rumble caused by fuel burning rate differences and that caused by deposit ignition tendency.</div> <div class="htmlview paragraph">Phosphorus added to the fuel in sufficient concentration was found to change the nature of the deposits and thus minimize rumble by reducing deposit-ignition. However, it did not prevent rumble due to early combustion caused by an increased burning rate of the charge.</div> <div class="htmlview paragraph">In a 10:1 C.R. multicylinder engine, significant differences in rumble tendency were found between cylinders. These differences resulted apparently from variations in fuel-air ratio and other factors that could influence the deposits formed and the mass burning rate of the fuel charge.</div>

<div class="section abstract"><div class="htmlview paragraph">In the present work, different combustion control strategies have been experimentally tested in a heavy-duty 3.0 L Euro VI diesel engine. In particular, closed-loop pressure-based and open-loop model-based techniques, able to perform a real-time control of the center of combustion (MFB50), have been compared with the standard map-based engine calibration in order to highlight their potentialities. In the pressure-based technique, the instantaneous measurement of in-cylinder pressure signal is performed by a pressure transducer, from which the MFB50 can be directly calculated and the start of the injection of the main pulse (SOI<sub>main</sub>) is set in a closed-loop control to reach the MFB50 target, while the model-based approach exploits a heat release rate predictive model to estimate the MFB50 value and sets the corresponding SOI<sub>main</sub> in an open-loop control.</div><div class="htmlview paragraph">The experimental campaign involved both steady-state and transient tests. The three control techniques were compared in steady-state tests under various conditions, featuring standard as well as PCCI combustion mode, different kinds of fuels, a disturbance added to the pressure signals from in-cylinder transducers (to simulate the effect of an aged or low-cost pressure transducer) and an injector with a reduced mass flow rate mounted on one cylinder. The behavior of the three controls was tested in transient conditions as well, analyzing in particular fast load and speed ramps.</div><div class="htmlview paragraph">The above mentioned testing conditions were performed to evaluate the robustness of the pressure-based and model-based techniques compared to standard calibration map-based control and their outcomes in terms of engine operation stability.</div><div class="htmlview paragraph">The proposed real-time combustion control techniques provided fuel consumption and emissions in line with the conventional map-based control. In addition, they lead to an improvement in combustion stability, which can be an important issue especially when transient operations are considered or when non-conventional combustion modes, such as PCCI, are implemented.</div></div>

<div class="section abstract"><div class="htmlview paragraph">Mainly due to environmental regulation, future Engine Control Unit (ECU) will be equipped with in-cylinder pressure sensors. The introduction of this innovative solution has increased the number of involved variables, requiring an unceasing improvement in the modeling approaches and in the computational capabilities of Engine Control Unit (ECU). Hardware in the Loop (HIL) test system therefore has to provide in-cylinder pressure in real time from an adequate model. This paper describes a synthesis of our study targeted to the development of in-cylinder crank angle combustion model excluding look up tables, dedicated to HIL test bench. The main objective of the present paper is a comprehensive analysis of a reduced combustion model, applied to a direct injection Diesel engine at varying engine operating range, including single injection and multi injection strategies. The developed model has required an important identification step to calibrate the parameters of the combustion model based on single and double Wiebe function. The difficulty to derive one calibrated model for all engine operations is discussed. This study allows to develop two calibrated combustion models one applied to high engine speeds and the other to low and medium engine speeds. The global model is able to run within one crank angle degree required by real time simulations. The precision of the model is analyzed by comparing in-cylinder pressure traces and ignition delay times at varying engine operating conditions.</div></div>

<div class="htmlview paragraph">The effects of engine operating conditions on the octane requirement and the resulting knock-limited output were studied on a single cylinder engine using production cylinder heads. A 4-valve cylinder head with port deactivation was used to study the effect of fuel octane, inlet air temperature, coolant temperature, air/fuel ratio, compression ratio and exhaust back pressure. The effect of the thermal environment was studied in more detail using separate cooling systems for the cylinder head and engine block on a 2-valve cylinder head.</div> <div class="htmlview paragraph">The results of this study compared closely with results found in the literature even though the engine and/or operating conditions were quite different in many cases. Guidelines on the effect of various parameters on the octane requirement of an engine can be summarized as follows:</div> <div class="htmlview paragraph"> <ul class="list disc"> <li class="list-item"><div class="htmlview paragraph">Spark Advance: 1 ON/1° SA</div></li> <li class="list-item"><div class="htmlview paragraph">Intake Air Temperature: 1 ON/7 K</div></li> <li class="list-item"><div class="htmlview paragraph">Air/Fuel Ratio: peaks at 5% rich of stoichiometric, decreases by around 2 ON / air/fule ratio around peak</div></li> <li class="list-item"><div class="htmlview paragraph">Intake Pressure: 3-4 ON/10 kPa</div></li> <li class="list-item"><div class="htmlview paragraph">Compression Ratio: 5 ON/CR</div></li> <li class="list-item"><div class="htmlview paragraph">Exhaust Back Pressure: 1 ON/30 kPa</div></li> <li class="list-item"><div class="htmlview paragraph">Coolant Temperature: 1 ON/10 K (Similar effect of head and block temperature., Larger effect at low speed.)</div></li> </ul> </div>

<div class="section abstract"><div class="htmlview paragraph">Forthcoming worldwide emissions regulations will start regulating ammonia emissions from light duty vehicles. At present, most light duty vehicles are powered by gasoline spark ignition engines. Sources of ammonia emission from such engines can be in-cylinder reactions (i.e. combustion) or downstream reactions across aftertreatment devices, particularly three-way catalysts. The latter has been known to be a major source of ammonia emissions from gasoline vehicles and has been extensively investigated. The former (combustion), less so, and thus is the subject of this work. A two-zone thermodynamic spark ignition engine model with a comprehensive chemical kinetics framework (C3MechV3.3 mechanism), after being validated against experimental ammonia emissions data, is used to study ammonia formation during combustion. Reaction pathways responsible for its generation are analysed and the effects of changing the following engine operational and combustion parameters are explored: engine load, start of combustion, combustion duration, fuel-air equivalence ratio, and exhaust gas recirculation fraction.</div><div class="htmlview paragraph">Ammonia production was found to be slower than that of other major pollutant species - starting late during the heat release stage, peaking around the time when the cylinder pressures and temperatures were at their highest, and having a late, prolonged production stage after the end of heat release. Ammonia concentrations did not ‘freeze’ until late into the expansion process. Initial ammonia production was driven by three body elementary reactions involving hydrogen radicals produced from the fuel oxidation/reduction, and the late-stage production was dominated by H<sub>2</sub>O reactions with amino radicals. The net effect of these production pathways on ammonia emissions in response to changes in engine operation was non-monotonic and depended on the dominant pathway at the particular thermal conditions. However, overall trends suggested that emissions increased when engine load increased, combustion duration shortened, combustion timing advanced, fuel-air mixture became richer and exhaust gas recirculation fraction decreased.</div></div>

<div class="htmlview paragraph"> </div> <div class="htmlview paragraph"> </div> <div class="htmlview paragraph">A computer simulation of a high performance two-stroke engine has been validated by a transient test method. The simulation included unsteady gas dynamics, together with detailed mechanical models for the reed valve, etc. The combustion model used Vibe coefficients derived directly from measured cylinder pressures. Cylinder scavenging characteristics which were measured on The Queen's University of Belfast (QUB) single-cycle rig were also incorporated in the simulation.</div> <div class="htmlview paragraph"> </div> <div class="htmlview paragraph">Validation of the model involved use of an inertial dynamometer and data acquisition system which has been developed at QUB. This dynamometer incorporates a flywheel which is directly coupled to the engine. During a test, the engine is accelerated, and the torque and power are calculated from the flywheel speed characteristics. Further, at pre-determined engine speed intervals, the cylinder and exhaust pressures may be recorded. This testing system provided a convenient and rapid experimental method to acquire data.</div> <div class="htmlview paragraph"> </div> <div class="htmlview paragraph">Measurement of power, pressures and temperatures during transient testing are discussed. Also presented are methods which reduce the errors arising from the slow response of thermocouples. The measured results are compared with the computer predictions for power and exhaust pressure. Excellent correlation was achieved.</div> <div class="htmlview paragraph"> </div>