基于神经网络的间隔调速模型研究

Study on Neural Network Based Space-interval Speed-control Model

车组溜放速度控制是驼峰自动化的重点和核心内容。由于间隔调速位位于驼峰咽喉 ,溜放坡度大 ,车组速度快 ,车辆密集 ,所以间隔调速是速度控制的难点。传统的间隔调速思想是 :首先根据车组溜放的物理数学模型 ,建立车组溜放方程 ,确定出口定速 ,然后调节车组的溜放速度 ,使之达到出口定速 ,即以“出口定速”为控制目标的静态间隔调速。这种控制方法由于没有实时考虑车组间的间隔 ,所以容易导致溜放事故或解体作业效率的降低。随着溜放作业自动化的发展 ,传统的静态间隔调速模型开始受到挑战 ,建立根据前后车组间的距离—间隔动态控制出口速度的间隔控制模型 ,应当成为当前驼峰自动化的研究重点。本文是利用智能控制和神经网络原理 ,建立动态控制出口速度的间隔制动位速度控制模型。该模型以前后车组间的实时间隔作为控制参数 。

The core of automated hump is to control the speed of a rolling car. Tools for interval adjustment are located at the throat of the hump, where the slope is great and most of the cars go through. So it is difficult to adjust the space interval between two neighboring rolling cars. The main idea of the traditional speed control system is to do in two steps. First, an equation is created according to the physical mathematical model. The purpose of the equation to define the speed of the rolling car when it comes out of the retarder. Secondly, adjust the rolling car′s speed until it reaches the ″defined speed″ . The traditional speed control system is a static system to make the speed of rolling car equal to the ″defined speed″ when it comes out of the retarder. But that system is subject to accident or results in low efficiency. With the automation of the control of rolling cars, the traditional speed control system is being challenged. Emphasis should be put on the system that dynamically changes the rolling car′s speed according to space interval between two neighboring rolling cars. This paper establishes such a speed control system. This system is based on Intelligent Cybernetics and theories of Neural Network. In the system, space interval is a real time input parameter that can change the rolling car′s speed dynamically.