电梯零速停靠的RBF神经网络预测算法

针对电梯运行过程中存在爬行距离的问题,提出了基于RBF(Radial Basis Function)神经网络的爬行距离预测模型.将预测的爬行距离增加到电梯速度曲线的匀速段,实现减小或消除爬行距离的目的,从而实现电梯的零速停靠.从电梯运行现场采集大量的原始数据,建立RBF神经网络预测模型,与BP(Back Propagation)预测方法进行仿真比较,结果表明RBF神经网络具有更好的预测效果.给出了应用零速停靠RBF预测算法前后电梯运行的速度曲线,爬行距离减小或消除,电梯的运行时间变短,实现了节能.

RBF neural network prediction algorithm for zero speed parking of elevator

The prediction model of the creeping-in distance based on RBF(Radial Basis Function) neural network is proposed in the paper. The creeping-in distance predicted is added to the uniform motion stage to decrease or eliminate the distance, and the zero speed parking of elevator is realized by this way. A great of original data is collected from the elevator running scene, and the prediction model based on RBF neural network is founded. Compared with the BP(Back Propagation) neural network prediction method, the results indicate that the RBF neural network prediction method possesses better prediction effect. The speed curve of the elevator using RBF neural network prediction algorithm is given, which is compared with the normal speed curve of the elevator. And the creeping-in distance can be decreased or eliminated basically. Meanwhile, the running time of the elevator is shortened, and the energy-saving is achieved.