Modeling And Optimal Control Of Free Piston Engine

The increasingly stringent regulation of energy-saving and environmental protection has propelled the development of novel technologies for new energy vehicle. Before the breakthrough of the battery technology for pure electric vehicle, the hybrid configuration is considered as a reliable solution for the concern of fuel economy. As a novel engine with various advantages, the Free Piston Engine(FPE) is expected to be an alternative auxiliary power unit for application in the next generation of series hybrid vehicle. A free piston engine, if people are not familiar with the concept, is an internal combustion engine without crank. Rather than using a crank to transfer power to a load, a free piston engine needs an alternative mechanism such as electric or hydraulic power takeoff. Compared with a conventional internal combustion engine, by removing the crank, the FPE has some unique advantages. For instance, less moving parts translates to lower friction. More interestingly, the piston motion is no longer constrained so that the displacement is able to change from stroke to stroke which means the compression ratio can be optimized very quickly for a variety of combustion strategies or fuel types leading to a significant improvement on thermal efficiency. However, there are also issues with the added flexibility: the mechanical constraint for the piston motion disappears as the absence of a crank in which case the FPE is very hard to control. This is perhaps the most significant barrier preventing the wide application of FPE technology in production vehicle. Therefore, the reliable control of FPE is a challenging issue and an enable technology to ensure its stable operation.This dissertation focuses on the study of modeling and optimal control of free piston engine. To start with, we introduce a type of dual-piston FPE configuration and its operating principle. As a physics-based model is necessary to evaluate the operating behavior and assist with the control development, we combine the coupled thermal dynamics, electric dynamics, and piston motion dynamics to build a model in the continuous-time domain. According to the validation requirement for controller design, a simulation model is built in AMESim platform and its rationality is preliminarily verified through simulation comparison. The model behavior and dynamic characteristics are analyzed by simulation which indicates the core control variables that directly impact the system operating performance.The control problem is described as the control of FPE piston clearance height at the end of each stroke as well as the constraint management during a load transition. Accordingly, a simplified implicit discrete, control-oriented model is built to describe FPE clearance height dynamics based on energy balance in ideal Otto cycle. Model Predictive Control(MPC) is applied to the multi-variable control task as well as to deal with constraints. Due to the various application scenarios of FPE, two optimization strategies are presented resulting in different control performance. As the measurement of piston position at turnaround point suffers from a single time delay in practice, Newton’s method and an extended state observer are exploited for the state prediction. As demonstrated in simulations, the designed controller achieves good performance in piston position tracking and enforces the constraints during load transitions.With regard of the air-path control in four-stroke, four-cylinder FPE, the controller is augmented by a dual-loop control scheme that not only to track the piston motion set-points and enforce the system constraints, but also maintain the desired airto-fuel ratio. The outer-loop control adjusts the air-charge into cylinder and load change rate by MPC that tracks clearance height set-points. A nonlinear inner-loop air-path controller is designed based on Triple-step method to regulate the throttle to meet airflow demand. Considering the volumetric efficiency drifts as the compression rate changes in FPE, an improved air-path controller is presented that tracks the air-charge directly. An input observer is developed for the air-charge estimation and the estimated error bound is discussed for the parameter tuning. Assisted with the proposed dual-loop controller, the adjustment of ignition position is realized by look-up table which is calibrated based on Extremum Seeking strategy to achieve the optimal cyclic fuel economy.For the purpose of reducing the computational expense of the FPE control system, an iterative Reference Governor is exploited. This study extends the application of Reference Governor to discrete implicit system. In order to further reduce the online iterations of the proposed algorithm, an improved one-step iterative Reference Governor is developed in which the predicting state sequence is obtained only by one-step Newton iteration. The iterative error bound is analyzed and calculated based on which the original constraint set is tightened not only to enforce the actual state constraint but also to boost the computational speed. Finally, the proposed Reference Governor algorithm is applied to the FPE control system. The control performances of iterative Reference Governor and MPC are evaluated and compared by simulation. The comparison study gives the suggestion on the selection of FPE control strategy for the engineering implementation in the future.In this dissertation, the detailed derivation of FPE control is presented. The analysis and validation of the proposed control are verified by the co-simulation of Matlab/Simulink and AMESim. The simulation results are analyzed and discussed which show performances of the proposed FPE control methods are satisfied. This dissertation underscores several opportunities for additional investigation. Due to the current limit of experiment equipment, all the results are based on the simulation. The future work would focus on the Hardware-in-Loop test and the development of test-bench.