The design and applications for straightening exceeded value of elevator guide rail based on neural network

Because of the shortage for automatic straightening of the elevator guide rail based on database, this paper proposes the design method for that based on neural network. A neural network was built for the automatic straightening process of the elevator guide rail with the initial bending value and the supporting points span as the input parameters, the straightening exceeded value as the output parameter. And the large quantities of successful automatic straightening data shall be as specimens to build a neural network model, which will be trained with the specimens by programing on MFC. Then the neutral network after training shall be called for test by field data. Then the result indicates the method based on the neutral network makes up the shortage based on database effectively, improves the automatic straightening accuracy and efficiency. Meanwhile, this method may provide reference for straightening expert system of the elevator guide rail.