Lateral Vibration Detection And Robust Control For EMUs

With the speed increasing on rail vehicles in recent years, the problem of hunting motionof the train is becoming more and more serious, which could bring about hunting motioninstability at some severe circumstances resulting in the deterioration of the quality ofoperation. Meanwhile, strong interaction between wheels and tracks can cause serious damageamong the wheels and rail, and even can lead to derailment. As a result, studying ways ofrestraining train-hunting movements to ensure the security and stationary of EMUs during theoperation is necessary as the representative of high-speed trains.Presently, researchers have revealed the mechanism of hunting movement thatself-structure is the essential reason causing transverse vibration. Meanwhile, a series ofmeasures have been taken by manufacturers to abate train-hunting movement. However,limitations are still existed due to the environmental complexity of the running rail vehicles.Considering the complex environment of the operating EMUs and uncertainty of thevehicle internal structure, it is necessary to provide a real-time monitoring method for thevehicle vibration and adopt algorithm that has strong capacity in anti-disturbance to preventthe hunting phenomenon. Only the assorted use of the line monitoring and real-timecontrolling, can we control it effectively. Formerly, the method of train-vibration testing isoff-line and hardware-used, which is lack of communicating capacity with outsideenvironment. Meanwhile, the hardware has high failure rate and the detection method hasmany shortages. With the rapid development of virtual instrument in recent years, thetendency that software substituting for complex hardware circuits is becoming more and morepopular. Comparing with the traditional hardware machines, the method of software with littlecost in time, financial field and maintenance also can be used in secondary development. Inview of the complexity, hardware-used simulation of train-running has great difficulty. On theother hand, the technology of virtual instrument based on computer simulation can be used indeveloping virtual environment where train-detecting, property-adjusting andperformance-estimating will be included, which provide a method to develop and researchvehicle products with virtual effective, low-cost way. In this paper, virtual instrument(LabVIEW) is used in designing supervision platform of hunting movements, achievingreal-time parameter detecting of transverse vibration in train running, managing and analyzingdata and making real-time estimation of riding quality.Aiming at the factor of external and internal disturbances during process of train running,there are great difficulties in building accurate train-movement model. Most traditionalcontrol algorithms are based on accurate mathematical model and have limitations in thestudying of train running complexity. However, robust control theory has distinguishing advantages in dealing with uncertain problems. In this paper, one single trailer of EMUs ischosen to be as the subject，and H∞mixed sensitivity control is used to analyze train externaluncertain disturbances and internal parameter perturbations, and establish generalizeduncertainty PS/T transverse vibration model. Linear matrix inequality(LMI) is used in bothsupplying the conditions of the existence of the robust H∞controller and solving transversevibration robust controller，which is compared with passive suspension control technology. Bysimulating analysis, the rationality of the robust H∞controller is verified. The closed-looptransverse-controlling system is analyzed and supervised under real-time in the method ofvirtual experiment. The simulation results indicate that in the condition of train-runningvirtual experiment, the robust H∞controller designed not only satisfies the expectedperformance indexes but also ensures that the system has a good robustness in the case ofcontinuous perturbation of parameter of the train model.