METHOD FOR GENERATING A DATA SET FOR TRAINING A NEURAL NETWORK IN A TRANSPORT CONVEYOR MODEL

Abstract

The object of research is a stochastic input flow of material coming in the input of a conveyor-type transport system. Subject of research is the development of a method for generating values of the stochastic input material flow of transport conveyor to form a training data set for neural network models of the transport conveyor. The goal of the research is to develop a method for generating random values to construct implementations of the input material flow of a transport conveyor that have specified statistical characteristics calculated based on the results of previously performed experimental measurements. The article proposes a method for generating a data set for training a neural network for a model of a branched, extended transport conveyor. A method has been developed for constructing implementations of the stochastic input flow of material of a transport conveyor. Dimensionless parameters are introduced to determine similarity criteria for input material flows. The stochastic input material flow is presented as a series expansion in coordinate functions. To form statistical characteristics, a material flow implementation based on the results of experimental measurements is used. As a zero approximation for expansion coefficients, that are random variables, the normal distribution law of a random variable is used. Conclusion. It is shown that with an increase in the time interval for the implementation of the input material flow, the correlation function of the generated implementation steadily tends to the theoretically determined correlation function. The length of the time interval for the generated implementation of the input material flow was estimated.