Modeling And Optimal Control Of Automotive Engine Thermal Management System For Fuel Saving

With the development of automotive technology,automotive engine has gradually evolved into a highly complex system with mechinery-electricity-heat coupled.The current engine control system mainly focuses on the control of dynamic-electical chain,but lacks the research on the optimal control of thermal chain.The optimal control of the thermal chain is an effective way to avoid engine knock and fuel gelation,and a necessary technology to realize the efficient operation of the engine.Research shows that the thermal management system based on the optimal control of the thermal chain is expected to have 10-15% fuel saving potential.However,the modeling and control of the engine thermal management system still faces many challenges:(1)Most modeling methods are oriented to the design and thermodynamic analysis of the cooling system structure.Although the models have high accuracy,they cannot be used for the model-based controller design.(2)The nonlinear characteristics of the heating/cooling/convective heat transfer process,the transmission delay of the coolant and the external disturbances such as vehicle speed bring challenges to the high-precision tracking control of coolant temperature.(3)The contradiction between the slow changing heat transfer and the fast changing power transfer limits the selection of the optimal control strategy and the implementation in the embedded system.(4)A serious coupling between the electricity storage unit(battery)and the heat storage unit(engine)exists in the power system with multi energy source,which challenges to the real-time energy-heat optimal control.Focusing on the above problems,this dissertation conducts research on the thermodynamic modeling and real-time optimal control of thermal management system.The specific contents are as follows:(1)For the high order of the cooling system model that based on the components temperature,a third-order thermodynamic model of engine cooling system is proposed based on the medium temperature,and the heating/cooling/convective heat transfer process are modelled based on mechanism.The thermodynamic model has the advantages of high accuracy and low order.Furthermore,for the real-time implementation in the embedded systems,the thermodynamic model is simplified as a first-order model by ignoring the temperature difference among the medium.The reduced-order model slightly reduces the transient accuracy,but ensures the steady-state accuracy,which balances the requirements for the model complexity and accuracy.(2)For the requirement of coolant temperature tracking control,a temperature tracking control architecture based on a Smith predictor and an extended state observer is proposed for the coolant transmission delay and unmeasurable disturbance.The control architecture separates the compensation for the delay and unmeasurable disturbance from controller design.For the nonlinear characteristics of heating/cooling/convective heat transfer process in SISO system,which is equipped with electronic fans,a nonlinear triple-step-method tracking controller for coolant temperature is proposed.Furthermore,for the temperature control of the over-actuated MISO system,the power consumption of cooling system is introduced as the optimization objective to limit the freedom degree of the control law,and the receding horizon optimization control method of coolant temperature is proposed.Compared with the traditional switch control and PI control,the control precision and response time are improved significantly.(3)For the control problem of integrated mechinery-electricity-heat control of engine,a minimum fuel consumption control strategy based on energy-heat integrated optimal control is proposed.For the contradiction between the slow change of heat transfer process and the fast change of power transfer process,the maximum principle is applied to realize the real-time implementation in the embedded system.For the condition that the future vehicle speed is unknown,a stochastic dynamic programming controller based on the Markov chain vehicle speed predictor is proposed,which realizes the real-time suboptimal control by using the empirical speed information.(4)For the control problem of energy-heat coupling control of hybrid electric vehicle,a single/multi time scale hierarchical energy-heat optimal control method is proposed by combining the maximum principle and forward dynamic programming.The control strategy divides the contradiction of the predictive horizon,control interval and model's order into different levels of the control system.The simulation results show that the hierarchical control method significantly improves the computing efficiency compared to the centralized control method,the proposed integrated energy and thermal management control strategy can reduce fuel consumption by about 8% compared to the traditional energy management strategy.The paper makes detailed derivation and verification of the mentioned engine thermal management control strategy,and realizes systematic research from theory to applications.Based on the current researches,the future work will focus on the integrated energy-thermal control of multi-source power system under the intelligent transportation system,and fast solution algorithm for dynamic optimization with state constraints.