Diesel engine airpath system modelling and control

Diesel engines have been used for more than 120 years after their invention by Rudolf Diesel in 1893. Especially along with the increment of number of vehicles on roads day by day, exhaust gases are causing problems for environment and people. Because of this situation, some works have been started firstly in United States of America as since 1960s. Although, it is possible to provide the desired conditions with simple arrangements in early stages, emission standarts have been more strict especially in the last ten years. So, transformation between mechanical to electromechanical systems become necessary to reduce the ratio of emission gases. Operation and performance analysis of diesel engines usually have been made with dynamometer tests on engine test cells in motor companies. In academic and commercial area, researchers are facing serious problems to use engine test cells especially in Turkey. Because there aren't enough facility to provide the testing opportunity. Addition to this, setup of engine test cell facilities need high infrastructure, labor and currency support. Therefore, creating of dynamic model and simulate that on computer systems provide much money and time advantage. Some studies have been carried out in this area for the last twenty years. Early researches include mostly map based modelling and controlling issues; hence, this situation again was in need of much support of engine test cells also, it isn't easy to modify map based models for different engine types. Early in 2000s, some scientists started to create dynamic models and validate these models in engine test cells. This work supplies validated model for engineers in this industry and it isn't hard to modify these dynamic models for different engine types. On the other hand, electronical control unit manufacturing has been controlled by a few companies such as Bosch, Delphi and Continental and almost all automotive companies are supported by them. Otherwise developing countries as Turkey are in need of manufacturing of their own products. Because of these necessity, it is an important step to develop dynamic model and control structure for diesel engines. In this thesis, firstly there are some prior knowledge about historical development of engines. Secondly, some informations are given about operating conditions and performance factors of diesel engines. After that, modelling and control studies of diesel engine airpath system have been carried out to provide lower emission limits, better performance and fuel economy. Airpath system of diesel engine has five sub-systems which are intake manifold and exhaust manifold systems, exhaust gas recirculation system, variable geometry turbocharger system and cylinders. Permanent duty of airpath system is to provide ideal oxygen ratio between compressor and EGR mass flow rate. After the combustion of air-fuel mixture in cylinders, pressured air is discharged by exhaust manifold and energy of pressured air provide power to rotate the turbine vanes. This power is turned over to compressor with support of VGT shaft and operating of compressor discharges air to the intake manifold. If oxygen ratio of air-fuel mixture is high, this causes to high in-cylinder pressure; thus, much come out of NOx gas. NOx gases are constituted because of high temperature in cylinders. Therefore, we should control the oxygen ratio of air-fuel mixture with another subsystem which is called as exhaust gas recirculation system. EGR system works by recirculating a portion of an engine's exhaust gas back to the engine cylinders. This operation reduces the oxygen ratio in the incoming air stream, hence, avoid the creation of more NOx gases in high cylinder temperature and pressures. In spite of that, if much EGR flows to the intake manifold, this will cause to the increase of particle material gases. So, it is important keeping the balance between operation of VGT and EGR systems. This is a complex optimization problem. In the course of creation of system model, complexity and simplicity of this balance has been taken into consideration. Since if the system model is very complex, simulation times will increase and also if the system model is very basic, simulation results will be incorrect. For instance, modelling of in-cylinder pressure is very complex and requires high level mathematical problem so, it is important to reduce the model level while keeping the accuracy. So mean value model of diesel engine airpath with a VGT and EGR system is developed. Aim of modelling work was to setup a model that indicate the gas flow dynamics as manifold pressure dynamics, turbochargers, EGR and actuators with maximum eight states to get shorter simulation times. Number of states can be increased or decreased in order to obtain better results in different operating conditions. Due to privacy policy of engine companies, engine test data can't be provided to validate the model. Therefore, validated engine data are used which are data of Scania heavy truck engine. It is easy to modify model parameters with respect to different type engine data with using curve fitting methods. Airpath model have different level dynamics as temperature, pressure, mass flow rate. Closing of the VGT system increase exhaust manifold pressure, EGR fraction so decrease intake manifold oxygen mass fraction and lambda value. However, this process is going on step by step. Closing of the VGT firstly leads to the increase of exhaust manifold pressure and this leads to the increase in turbocharger speed and compressor mass flow rate. Therefore, lambda value will be higher at the end of the cycle. This operation occurs because of slower dynamics of the turbocharger speed according to faster dynamics of the exhaust manifold pressure. In scientific area, this is called as non-minimum phase behaviour. In addition to this, during this operation if EGR valve position increase, an VGT position decrease, non-minimum phase behaviour for lambda will also increase and cause to the sign reversal and overshoot for lambda. It is similar operation with driving a car during bad traffic conditions. System analysis have been made especially according to these transient conditions to specify the control system requirements. Control structures include PID controllers mostly because of the many operating conditions. High level strategies aren't useful also because of requirement of higher level computer power. Some control goals have been determined to provide the better operation conditions for airpath sytem. These are carrying out the legislated emission levels, to reduce the fuel consumption, carrying out the safety of turbocharger system and obtaining the best performance in these conditions from the engine. System inputs are engine speed, fuel consumption, VGT position and EGR valve position and system outputs are EGR fraction and lambda. Outputs are the most important performance variables because EGR fraction is determining the ratio of PM gases and affecting the NOx ratio, and also lambda is determining the ratio of NOx and affecting the ratio of PM gases. Moreover, fuel consumption can be controlled with using feedforward control approach to consume less fuel with reducing the pumping torque. Pumping torque is reducing the indicated torque because of the pressure difference between the exhaust and intake manifolds. Therefore, suitable control strategy can be used to reduce the pressure difference hence pumping torque. Lambda is controlled by EGR valve position and EGR fraction is controlled by VGT position in this study. This combined control strategy have been designed with respect to handle the non-minimum phase behaviours, sign reversals and overshoots. PID parameters have been determined in three phases. Firstly, relation of lambda – EGR valve position is modelled and simulated in different conditions. After the linearization of this structure, PID controllers are implemented to obtain the required lambda levels. Secondly this operation again is carried out according to the relation of EGR fraction and VGT position. Finally, control structure is created for two input – two output system and optimized results are achieved. It is possible to fulfill all requirements with using PID controllers with using efficient tuning methods.