Font Size: a A A

Study Of Intelligence Optimization Control About Locomotives Adhesion

Posted on:2016-01-02Degree:DoctorType:Dissertation
Country:ChinaCandidate:N Z LiFull Text:PDF
GTID:1222330485983275Subject:Power electronics and electric drive
Abstract/Summary:PDF Full Text Request
Optimal control of locomotive adhesion is an effective measure that can fully dig out the potential of wheel-rail adhesion and improve the performance of wheel-rail adhesion, so that the traction/braking capability of locomotive can be maximize. So Optimal control of locomotive adhesion is extremely important measure to improve railway transportation capacity and ensure traffic safety. Obtaining maximum adhesion force is the purpose of the optimal control of locomotive adhesion, so the optimal control of locomotive adhesion is essentially an optimization problem. The primary task of solving the problem is to establish the control performance index function that based on the optimal control theory so that the control quality of the optimal control of locomotive adhesion can be quantitatively assessed, while the existing research about locomotive adhesion control has little mentioned it. In addition, since the locomotive driving system has the characteristics of large inertia and nonlinear, as well as wheel-rail adhesion is complex and changeable uncertainty and strong nonlinearity, it is difficult to accurately get the analytical model of controlled system, it also causes that the optimal control methods based on the accurate mathematical model can’t solve the problem of the optimal control of locomotive adhesion. So the research of locomotive adhesion control strategy which does not rely on the systematic accurate mathematical model is always attracting the attention of researcher.Intelligent control methods, such as neural network control and fuzzy control have advantages that it does not rely on accurate mathematical model and is adaptable to poor modeling of nonlinear dynamic systems. In this thesis, we combined the advantages of intelligent control methods with swarm intelligence optimization algorithm, and researched the theory and method of intelligent optimal control of locomotive adhesion. The main work is as follows:(1) In order to realize the quantitative evaluation of the locomotive adhesion optimization control quality and reduce the dependence to designer’s subjective experiences and knowledges during designing the control system, we studied the construction of the performance index function of intelligence optimization control of locomotive adhesion and intelligent optimization algorithms configuration problems of controller and filter in the design of the locomotive adhesion optimiztion control system. We established the control performance index function of intelligent optimization control of locomotive adhesion, and given specific construction form of the control performance index function in different working conditions, and put forward the dynamic multiple sub-population gravitational search algorithm and grey clustering multi-subpopulation adaptive particle swarm optimization algorithm, and finished the optimization choice of neural network parameters and the confirmation and optimization solution of structure and parameters of nonlinear Volterra filter which can filter the mixed noise (Gaussian noise and impulsive noise) in wheel speed signal.(2) In view of the analytical difficulties to solve the optimal control problem of locomotive adhesion, based on the optimal control theory, we combined swarm intelligence optimization algorithm referred in this thesis with neural network control and fuzzy control technology, put forward the strategy of intelligent optimization control of locomotive adhesion to achieve optimization control of locomotive adhesion in traction/braking conditions. We designed the locomotive adhesion neural network controller and fuzzy controller. The fitness function (that is, the optimization goal) was established based on the control performance index function, and under the guide of optimization goal, the controller parameters was automatic adjusted and eventually converged to the optimal solution by using swarm intelligence optimization algorithm proposed in this thesis, meanwhile the quantitative evaluation of the locomotive adhesion optimization control quality was realized.(3) In order to achieve the optimal match between the wheel/rail adhesion state optimization and traction motor output torque adjustment, through a combination with direct torque control, we researched and proposed the intelligent optimization control strategy of locomotive adhesion. The ideal torque (traction/braking torque) was obtained by intelligent optimization control algorithm of locomotive adhesion and was used as the regulation basis of traction motor output torque. By dynamically regulating the traction motor output torque so that it quickly reached the ideal torque, therefore the current working position of wheel/rail could fast approach the best working point and the optimal adhesive utilization could be obtained.(4) In order to testify the feasibility and effectiveness of the proposed intelligent optimization control strategy of locomotive adhesion and the established performance index function in this thesis, the experimental research of the intelligent optimal control of locomotive adhesion was carried out by using the hardware in the loop simulation. The results showed that the established performance index function of locomotive adhesion optimization control in this thesis was correct and feasible, and it could effectively evaluate the quality of the intelligent optimization control of locomotive adhesion. By using the proposed dynamic multiple sub-population gravitational search algorithm, the gravitational search Gaussian radial basis function (RBF) neural network controller and gravitational search ridgelet neural network controller and the gravitational search fuzzy controller were able to meet the control requirements well, thus, after the shortest regulation time, the locomotive wheel could work steadily in the best adhesive state under various extreme mutation conditions, and the maximum adhesion force was achieved.
Keywords/Search Tags:Optimal control of locomotive adhesion, performance index function, Wheel/rail best working position search, Neural network, Fuzzy control, Swarm intelligence optimization algorithm, Asynchronous traction motor control
PDF Full Text Request
Related items