Fabrizio Ponti, Vittorio Ravaglioli, Gabriele Serra, Federico Stola
Evaluation of MFB50 is very useful for combustion control, since it gives an evaluation of the combustion process effectiveness. Real-time monitoring its value enables to detect for example the kind of combustion that is taking place (useful for example for HCCI applications), or could provide important information to improve real-time combustion control. While it is possible to determine the position where the 50% of mass burned inside the cylinder is reached using an in-cylinder pressure sensor, this work proposes to obtain this information from the engine speed fluctuation measurement. In-cylinder pressure sensors in fact are still not so common for on-board applications, since their cost will constitute an important portion of the whole engine control system cost. Engine speed measurement is instead already performed in modern engine control systems and therefore being capable of obtaining MFB50 related information from this signal means obtaining it at approximately no additional costs.
The MFB50 estimation procedure presented in this paper is based on the measurement of the engine speed fluctuations and it mainly consists in two separated steps. As a first step, a torsional behavior model of the powertrain configuration is developed. The engine-driveline torsional model enables to estimate the indicated torque frequency components, from the corresponding components of the instantaneous engine speed fluctuation. This estimation can be performed cycle by cycle and cylinder by cylinder. As a second step, the analysis of the relationship between MFB50 and the phase of the frequency components over an engine cycle allowed defining the final estimation algorithm that reconstructs the MFB50 starting from the instantaneous engine speed fluctuation analysis. The developed approach has been applied with success to a diesel engine mounted on-board a vehicle.
Instantaneous Engine Speed Measurement and Processing for MFB50 Evaluation
Author: Fabrizio Ponti, Vittorio Ravaglioli, Gabriele Serra, Federico Stola
Combination of In-Cylinder Pressure Signal Analysis and CFD Simulation for Knock Detection Purposes
Author: Enrico Corti, Claudio Forte
A detailed analysis of knocking events can help improving engine performance and diagnosis strategies. The paper aim is a better understanding of the phenomena involved in knocking combustions through the combination of CFD and signals analysis tools.
CFD simulations have been used in order to reproduce knock effect on the in-cylinder pressure trace. In fact, the in-cylinder pressure signal holds information about waves propagation and heat losses: for the sake of the diagnosis it is important to relate knock severity to knock indexes values. For this purpose, a CFD model has been implemented, able to predict the combustion evolution with respect to Spark Advance, from non-knocking up to heavy knocking conditions.
The CFD model validation phase is crucial for a correct representation of both regular and knocking combustions: the operation has been carried out by means of an accurate statistical analysis of experimental in-cylinder pressure data.
The simulation results allow relating the combustion characteristics to their effect on the in-cylinder pressure signal. It is then possible to highlight critical issues regarding typical knock detection methodologies, while disclosing novel approaches. One of the main results is the validation of knock detection strategies based on the low-frequency content of the pressure signal. These strategies can be used together with standard high-frequency based techniques in order to improve the detection robustness.
Statistical Analysis of Indicating Parameters for Knock Detection Purposes
Author: Enrico Corti, Claudio Forte
Specific power and efficiency of gasoline engines are influenced by factors such as compression ratio and Spark Advance (SA) regulation. These factors influence the combustion development over the crank angle: the trade-off between performance and the risk of irreversible damages is still a key element in the design of both high-performance (racing) and low-consumption engines.
This paper presents a novel approach to the problem, with the objective of defining a damage-related and operating conditions-independent index. The methodology is based on the combined analysis of indicating parameters, such as Cumulated Heat Release (CHR), Indicated Mean Effective Pressure (IMEP) and 50% Mass Fraction Burned (MFB50), and typical knock detection parameters, estimated by means of the in-cylinder pressure sensor signal.
Knocking combustions have several consequences, therefore they can be detected in many ways. As it is well known, knock excites the combustion chamber resonant frequencies, rising high frequency components in the in-cylinder pressure signal. Additionally, knocking combustions cause a higher heat flux through the combustion chamber walls, increasing the heat losses and lowering the IMEP: this effect can be observed by evaluating IMEP or CHR, that are based on the low-frequency content of the in-cylinder signal spectrum. A knock detection index can then be based on the crossed-observation of two phenomena, one set in the low-frequencies range and the other one on the high frequencies range of the in-cylinder pressure signal.
Knock-related parameters have been analyzed with a statistical approach, in order to define a detection strategy based on one single threshold value, to be used for all the engines, independently of engine speed and load: IMEP and CHR have been related to a typical high-frequency knock index, based on the Maximum Amplitude of Pressure Oscillations (MAPO) evaluated over the combustion angular window. It can be observed that, as knock intensity increases, the relationships between these parameters change. This change can be seized by evaluating the correlation coefficient between the parameters distributions and observing its trend as a function of SA (Spark Advance). The correlation coefficient is intrinsically normalized, therefore the index range is the same for every engine operating condition, and the paper shows that the threshold level, too, is constant. The same approach can be applied to determine SA corresponding to the best combustion phase, simply by evaluating the correlation coefficient for IMEP and MFB50 distributions.
The methodology has been successfully applied to different engines running in different speed and load conditions.
Development of A Control-Oriented Model of Engine, Transmission and Vehicle Systems for Motor Scooter HIL Testing
Author: Enrico Corti, Francesco Migliore, Davide Moro, Paolo Capozzella, Michele Pagano
This paper describes the development of a mathematical model which allows the simulation of the Internal Combustion Engine (ICE), the transmission and the vehicle dynamics of a motor vehicle equipped with a Continuously Variable Transmission (CVT) system. The aim of this work is to realize a simulation tool that is able to evaluate the performance and the operating conditions of the ICE, once it is installed on a given vehicle.
Since the simulation has to be run in real-time for Hardware In the Loop (HIL) applications, a zero-dimensional (filling and emptying) model is used for modeling the cylinder thermodynamics and the intake and exhaust manifolds. The combustion is modeled by means of single zone model, with the fuel burning rate described by Wiebe functions. The gas proprieties depend on temperature and chemical composition of the gas, which are evaluated at each crank-angle. An automatic procedure for identifying the unknown parameters of the model by using experimental data has also been developed. The experimental data needed for model identification are the crank-angle resolved in-cylinder pressure and the measurements performed during typical engine-dynamometer steady state tests: rotational speed, load, fuel consumption, Air Fuel Ratio (AFR)… The transmission and vehicle systems have also been modelled: both clutch and CVT models have been fully described through physically-based equations and geometric characteristics. The CVT model evaluates the gear ratio, once the engine speed and torque are known. The clutch model is able to describe, with a single model, both the situations of engaged and disengaged clutch. The vehicle is modeled as a body subject to the engine tractive force, aerodynamic force and the tires friction force, with an equivalent inertia that takes into account the vehicle mass and the inertia of the engine and transmission elements.
Finally, the model has been validated on the basis of experimental data.
Upgrade of a Turbocharger Speed Measurement Algorithm Based on Acoustic Emission
Author: Davide Moro, Enrico Corti, Matteo De Cesare, Gabriele Serra
The present paper is about the rotational speed measurement of an automotive turbocharger, obtained starting from the analysis of acoustic emission produced by an engine, which have been acquired by a microphone placed under the vehicle hood.
In the first part of the paper several upgrades to increase the overall performance of the speed extraction algorithm are presented and discussed, starting from the basic algorithm that has already demonstrated the methodology capability in a previous paper.
In particular it has been considered a different signal sampling rate in order to extend the applicability of the methodology to a wider range of engines. Also a new processing procedure has been defined to increase the capability of the algorithm to tune on the frequency signal.
A possible calibration procedure has been evaluated in order to set-up a threshold value in the algorithm procedure, based on a off-line analysis of the acoustic signal acquired with the engine running during a typical acceleration phase from idle up to the upper gear shift.
In the results section, a new graphic representation of the algorithm performance has been introduced in order to obtain a sufficient clear idea of the performance of the rotational speed reconstruction algorithm. The trend of the percent error versus time and also an evaluation of the algorithm performance in every single second of the test are proposed.
The results obtained demonstrate that the methodology is ready to be introduced in the ECU to enhance new strategies for the turbocharger control.
Combined experimental and numerical analysis of knock in spark ignition engines
Author: Claudio Forte, Enrico Corti, Gian Marco Bianchi
A detailed analysis of knocking event can help improving engine performance and diagnosis strategies. The paper aim is a better understanding of the phenomena involved in knocking combustion through the combination of CFD and signals analysis tools. CFD simulations have been used in order to reproduce knock effect on the in-cylinder pressure trace. In fact, the in-cylinder pressure signal holds information about waves propagation and heat losses: for the sake of the diagnosis it is important to relate knock severity to knock indexes values. For this purpose, a CFD model has been implemented, able to predict the combustion evolution with respect to Spark Advance, from non-knocking up to heavy knocking condition. The CFD model validation phase is crucial for a correct representation of both regular and knocking combustions: the operation has been carried out by means of an accurate statistical analysis of experimental in-cylinder pressure data. The simulation results allow relating the combustion characteristics to their effect on the in-cylinder pressure signal. It is then possible to highlight critical issues regarding typical knock detection methodologies, while disclosing novel approaches. One of the main results is the validation of knock detection strategies based on the low-frequency content of the pressure signal. These strategies can be used together with standard high-frequency based techniques in order to improve the detection robustness.
Exhaust Gas Turbocharger Speed Measurement Via Acoustic Emission Analysis
Author: Nicolò Cavina, Davide Moro, Matteo De Cesare, Gabriele Serra
The paper presents a non-intrusive, indirect and low-cost methodology for a real time on-board measurement of an automotive turbocharger rotational speed. In the first part of the paper the feasibility to gather information on the turbocharger speed trend is demonstrated by comparing the time-frequency analysis of the acoustic signal with the direct measurement obtained by an optical sensor facing the compressor blades, mounted in the compressor housing of a spark ignited turbocharged engine. In the second part of the paper, a real time algorithm, to be implemented in the engine control unit, is proposed. The algorithm is able to tune on the turbocharger revolution frequency and to follow it in order to extract the desired speed information. The frequency range containing the turbocharger acoustic frequency can be set utilizing a raw estimation of the compressor speed, derived by its characteristic map. Different tests have been performed on a turbocharged Diesel engine and encouraging results have been achieved, since a satisfactory turbocharger speed reconstruction has been obtained.
Development of Model-Based OBDII-Compliant Evaporative Emissions Leak Detection Systems
Author: Nicolò Cavina, Enrico Corti, Stefano Sgatti, Luca Guidotti, Filippo Cavanna
The paper presents the main results obtained by developing and critically comparing different evaporative emissions leak detection diagnostic systems.
Three different leak detection methods have been analyzed and developed by using a model-based approach: depressurization, air and fuel vapor compression, and natural vacuum pressure evolution. The methods have been developed to comply with the latest OBD II requirement for 0.5 mm leak detection. Detailed grey-box models of both the system (fuel tank, connecting pipes, canister module, engine intake system) and the components needed to perform the diagnostic test (air compressor or vacuum pump) have been used to analyze in a simulation environment the critical aspects of each of the three methods, and to develop “optimal” diagnostic model-based algorithms. Experiments have been initially carried out in a laboratory environment to identify both the main model unknown parameters and the main disturbances, and to acquire data that could be used to validate the simulation results. During a subsequent phase of the project, a prototype vehicle has been setup with various sensors and actuators, and several experiments have been conducted to test the three leak detection model-based methods, while varying environmental and in-vehicle conditions. The critical aspects of each methodology have thus been isolated and compared, also by taking into account actual and possibly future European and North American on-board diagnostic regulations.