Design and application of a fault diagnosis system for machine tool angle heads based on Boolean matrix filtering and an optimized BP neural network

Abstract

In this paper, a fault diagnosis method based on Boolean matrix filtering and optimizing backpropagation (BP) neural networks is proposed for angle heads of computer numerical control (CNC) machine tools. The matrix filtering is first carried out with the fault case database and the fault cause symptom Boolean matrix according to the fault types and characteristics of the machine tool angle head. On the basis of the combination of multiple fault causes obtained from the initial filtering, the Euclidean distance method is used to narrow the results of fault cause filtering. The BP neural network model with weight vector is established and optimized to perform an accurate diagnosis. Finally, the fault diagnosis and management system of the angle head of a CNC machine tool, integrating with the Boolean matrix filtering method, Euclidean distance method, and BP neural network model, is developed and implemented with the Python language and QT development framework.