The System Design And Control Methodology Of The Waste Heat Recovery System For The Internal Combustion Engine Under Transient Driving Cycle

Due to the increasingly serious energy and environmental problems,the energysaving and emission reduction of ICE(ICE)has become more and more urgent.At the same time,the Waste Heat Recovery(WHR)technology for vehicles has attracted more attention.However,under the actual driving cycles,the operating conditions of the ICE change frequently and violently,which has posed great challenges to the safety and efficiency of the WHR system.Therefore,this study has focused on design and controller of the WHR system under driving cycles,by experiments and simulations.As the preliminary study for driving cycles,the tests under varying engine condition can be used to explore the operating characteristics of the WHR system under the whole condition range and the transient conditions of the ICE,which further guides the design of the controller.In order to obtain the characteristics of the waste heat source and the ORC system under the varying engine condition,a test bench was built for the WHR system of the heavy duty diesel engine.Firstly,the heat balance test is used to analyze the energy flow of the target diesel engine.The distribution of each energy flow under the whole condition range is obtained,and the variation of the exhaust gas properties under the whole condition range is analyzed.Then the performance of working fluids(R245fa and R123)was compared under the varying engine condition.The results show that the R245 fa working fluid has relatively better performance under most conditions and can be used for subsequent simulation studies.Later,by the transient response test of the ORC system,it can be seen that the condensing pressure can remain steady under the severely changing condition,which supports the simplification of the model in the future controller.Finally,by manipulating the control parameter,it is concluded that the superheat degree has little effect on the system performance,which set a lower requirement of the controller design,and the evaporating pressure should be as large as possible.Before the simulation work begins,this study proposed a single-stage recuperative ORC to improve the system efficiency and reduce the heat ejection demand,based on the characteristics of the experimental results and the limited cooling capacity on the vehicle.Then based on GT-Suite software,the modelling methods for the transient performance of diesel engine and recuperative ORC are explored.The experimental results are used to verify the model,and the accuracy meets the requirements of the following research.In order to optimize the performance of the WHR system under the target driving cycles,it is necessary to optimize the size of the key components.Facing the challenges of the varying exhaust gas under driving cycles,the nonlinearity of the system and the coupling of the constraints,the design space exploration method is adopted to systematically solve the problems during the ORC design process and proved to be effective and universally applicable.By setting three target spaces(Design Inputs space,System Inputs space and Control Inputs space)and applying simplification,the computational cost is greatly reduced.Once the target driving cycle and the target ICE are set,the optimal system is found in the Design Inputs space,so that the maximum net output power can be obtained within all the System Inputs space under possible optimal control.In the meantime,the corresponding control parameters are required to be in the Control Inputs space.During the optimization,the GT-Suite model is called by Matlab and parallel computing is applied,which can greatly reduce the computational time.Based on the optimal system above,a Map-based feedback-in-loop control method is proposed for the driving cycle.Firstly,the Particle Swarm Optimization(PSO)algorithm is used to obtain the optimal control parameter map of the ORC system and the corresponding state reference map.Then,two pairs of parameters with weak coupling are identified for PID feedback control.Finally the entire closed-loop system is constructed.In this chapter,the control effects of the open loop system based on the control map and the closed loop system with PID feedback are compared.The results show that the map-based control method ignores the transient characteristics of the system,and it is impossible to adjust the controller parameters based on the current state.Therefore,for the system with large thermal inertia like ORC,both systems are not satisfactory enough.At the same time,because there is no real-time optimization in the controller and the constraints cannot be applied,these also prevent the ORC cycle system from achieving the optimal performance or meeting the constraints.Then a nonlinear Model Predictive Control(MPC)systems is designed for the same situation.Firstly,the full-order ORC model is simplified by model order reduction to the reduced-order model that can be used for control,without losing much accuracy.Then the receding-horizon control and particle swarm optimization are combined to build a nonlinear MPC controller based on the simplified model.Finally,the nonlinear state estimator closes the feedback loop.The results show that the MPC controller can make full use of the actuators to meet the nonlinear constraints,as well as apply the real-time optimization control.The control effect is greatly improved.However,due to the inaccuracy of the model and the estimator,it is still possible to violate the constraint under certain conditions.Since there is a trade-off between the computational cost and the optimization,it’s hard to guarantee the operating states of the system always close to the optimal results.