The use of artificial intelligence in the tasks of optimizing cutting modes and predicting processing accuracy for unstable processing conditions of a batch of parts on CNC machines

The article considers the possibility of using artificial intelligence (AI) in the technological preparation of production (TPP) at the stage of calculating optimal cutting modes, considering unstable processing conditions of a batch of parts on the example of flat grinding performed on CNC machines. The problems of predicting processing accuracy and optimization of cutting modes for CNC machines have not yet been fully solved due to the complexity of their solution associated with: too high dimensionality of the optimized parameters, need to solve a large number of complex interrelated mathematical models of the processing process, optimization of numerous parameters of cutting modes and considering the limitations of the objective function in a multidimensional states space of the processing process, complexity of considering the multidirectional influence of various unstable technological factors on the process of the allowance removal, complexity of the models of formation of the technological size and quality parameters of the processed surface. The high dimensionality of tasks requires huge computing power of supercomputer technology, which no manufacturing enterprise has. The use of AI allows overcoming the «curse of dimensionality». It is proposed to solve the problems of predicting processing accuracy and optimizing cutting modes in the production conditions by using a trained convolutional neural network (NN) used for pattern recognition, which allows calculating optimal cutting modes for CNC machines and predicting processing accuracy. NN training is performed on a multiple sample (one hundred thousand or more CNC operations), with ready optimal cutting modes. Preparation of a sample of operations with ready optimal solutions is performed in advance on a supercomputer, using software created on the basis of the developed technique of complex structural and parametric optimization of cutting modes for CNC machines, considering the influence of various variable technological factors on the processing of a batch of parts.