Study On Hump Objective Braking Location Emergency Collision-avoidance Retarder

Shunting hump is the core equipment of the railway marshalling yard and the key equipment to improve the railway freight transportation ability. The core of hump automation is the control of cars' rolling speed, which is realized mainly by retarder braking location. Presently there are two existing problems with retarder braking location: First, the out of control of retarder braking location causing overspeed coupling and cracking up cars, which is ubiquitous and seriously affects railway safety production, and there's no special objective braking location emergency collision-avoidance equipment; Second, the model of cars rolling speed control has defects, analyzing complicated dynamic variables with static formulas and model, the control effect is not good.Aiming at the problems above, this paper is mainly about the study on hump objective braking location emergency collision-avoidance retarder. On basis of brand-new design machinery construction and frictional braking rails, an objective braking location speed control system model is built using Fuzzy Neural Network, and hardware and software of the retarder autocontrol system is designed. The key works of this paper are as follows:Modeling and emulating on objective braking location speed control system. Considering the characteristics of cars' rolling, the Fuzzy Neural Network model which doesn't need a precise model is adopted. The model takes entrance speed at retarder, weight of cars and numbers of cars as input variables, obtains the relations between input and output variables by computing and analyzing sample data, and predicts the speed difference between entrance and release speed, then the release speed is obtained. The forward computing, backward propagation of error signals and parameters modification is taken using BP algorithm improved by adopting additional momentum and variable step size method. The model has self-learning and self-adaptive ability, which can overcome the shortcomings of traditional control model more effectively. By taking emulation and verification using MATLAB, the emulation results show that the model can predict the release speed accurately and effectively.The hardware and software design of the embedded main board and track control board is based on model building, associated with the machinery construction of the retarder. The main function is: receive the overspeed information from the hump host computer, enable the first retarder, then make the enable strategy of the next two retarders by the analysis of Fuzzy Neural Network, and display the brake and release state of the retarder on LCD. The hardware design is taken embedded chip LPC2214 as its core, including RS-485 serial port data reception, brake and release circuit and state expression circuit, and LCD module. The core of software design is analyzing, judging and handling of information in the Fuzzy Neural Network model.