Study On Robust Control For Vehicle Active Electromagnetic Suspension

Vehicle suspension is an important system which affects vehicle ride comfort andhandling stability significantly. The complicated and changable operation conditionsdemand higher requirements for suspension systems. Thus, it is difficult for traditionalpassive suspensions to satisfy the desired vehicle performances. In recent years, as the fastdevelopment of electromagnetic materials and electronic control technologies, the researchof electromagnetic active suspension has attracted more attentions from academia andindustry.In order to explore the key problems of electromagnetic suspension characteristicsand control system, e.g., the effects of the electrical motor characteristics along with itsservo control system on the suspension performances, and the approaches to deal withmodel internal uncertainty and suppress external disturbance, the study on dynamiccharacteristics and robust control for a novel electromagnetic suspension actuator based onmixed uncertainty modeling is presented in the dissertation.A prototype of electromagnetic suspension actuator is designed and developed for anintermediate-class car. Firstly, the design process and development of the actuatorprototype are described in detail. Then, the tests and measurements for electricalresponsing characteristics, damping characteristics, back electromagnetic forcecharacteristics are carried out and the test results are analyzed. Based on the experimentallyidentified parameters, a mathematic model for the electromagnetic suspension actuator isestablished.The proposed control system for the electromagnetic suspension consists of amain-loop control and a servo-loop control. The former is used to calculate desiredsuspension active force according to the control law of suspension system, while the lattertakes charge of the force traking between actuator output force and the desired suspensionactive force. A hysteresis current control approach is adopted for the servo-loop control soas to track the desired electromagnetic force effectively. According to the control rules ofcurrent controller, the complicated motor control system is simplified. Then, the control current constraint for the actuator and the operating states of the motor and battery areobtained by analyzing the simplified control circuit, which provides essentials for thefollowing study on robust controller design.After the servo-loop control is designed and analyzed, and comprehensivelyconsidering the uncertainties of actual suspension system, two robust controllers usingmixed μ synthesis and gain-scheduled H_∞control approaches are designed respectively forthe main-loop.While in the controller design for the electromagnetic suspension, the uncertaintiesare comprehensively considered both in actual suspension system itself and thedisturbances in different vehicle operation conditions, thus a mixed μ synthesis controlleris designed. After the internal model uncertainties and external disturbances are analysed,linear fractional transformation theory is applied to seperate the known information ofuncertainties from system feedback interconnection. Then, a mixed-uncertainties model forthe electromagnetic active suspension, including parameter uncertainties and unmodelleddynamic characteristics, has been built. In order to reduce the conservativeness, a mixed μsynthesis approach is adopted to design the mixed μ synthesis controller and D-G-Kiteration process is used to solve the proposed controller. The simulation results of themain-loop μ synthesis controller show that the robust performances using mixed μsynthesis are better than those by using H_∞controller in the case of existing the mixeduncertainties within limited bounds. Moreover, mixed μ analysis approach can obtainbetter robust performance with the requirement of robust stability meanwhile.Since the effects of each uncertainty factor on suspension performances are different,the relatively sensitive uncertainties to suspension performances are selected by usingfrequency analysis for single uncertainty and a gain-scheduled H_∞controller is designedparticularly for the electromagnetic suspension. Firstly, the suspension system istransformed into a parameter-dependent system, i.e., a linear parameter varying system.Secondly, LMI approach is used to solve the gain-scheduled controller satisfying thequadratic H_∞performance, which can ensure both robust stability and robust performanceof the LPV system. Finally，by using a Kalman Filter state observer and ARX onlineparamenter identification, varying parameters are online obtained for the main-loop controland hence to online determine the real-time gain-scheduled H_∞controller. Combining themain-loop with servo-loop control, simulations are carried out showing that the proposedcontroller can adopt the large range of uncertainty and meanwhile achieve significant improvement of vehicle ride comfort.