Document Type : Research Articles
Author
University of Birjand
Abstract
Railway traction vehicles transfer forces between rails and wheels through an adhesion coefficient. In order to prevent wheel locking and shorten stopping distances, estimating the adhesion conditions between rails and wheels is an essential task in railway operations. Since the adhesion condition is influenced thru many factors, its estimation technique is complex. This paper presents an intelligent square root cubature Kalman filter (ISRCKF) to estimate adhesion force. The proposed method has the advantage that it does not require to know the noise statistics. This method integrates the differential evolution (DE) algorithm to tune the SRCKF by solving the optimal values of the covariance matrix Q and measurement noise matrix R. It can also decrease the error because of unknown noise, and increase the accuracy. Furthermore, it exhibits a consistent enhancement in numerical stability due to the assurance that all resultant covariance matrices remain positive semi-definite. This innovative approach plays an active role in optimizing the utilization of the current adhesion while reducing wheel wear by mitigating high creep values. The outcomes demonstrate that the suggested approach yields superior estimation accuracy and exhibits a swifter convergence rate in comparison to alternative methods.
Keywords
Main Subjects
)