Abstract:Aiming at the problem of spindle accuracy compensation for mechanical manufacturing equipment, a spindle rotation accuracy prediction method combined the accuracy of machine tools and machining parameters was studied.In order to obtain the index of spindle rotation precision, the method of measuring the actual rotation position of the spindle based on multipoint measurement was studied.The least squares method was used to solve the maximum circle, the minimum circle, and the optimal circle of the spindle face runout, and the maximum error and minimum error of the end surface roundness were further obtained.A BP neural network model was established to integrate the machine tool accuracy factors and machining accuracy factors to obtain input indicators. The maximum error, minimum error and other measured indicators were used as output indicators to train neural networks and to obtain network weights.The accuracy of the spindle revolution of the machine tool was verified.The experimental result shows that the prediction deviation of the training network is 05%.The error prediction results considering machine tool body accuracy and machining parameters can be used to instruct machine tool selection,engineering technicians can select the optimal machining method according to different machining requirements, to improve the machining quality.