Neural Network Model to Forecast of Part Measuring Uncertainties

The labor intensity of control in technological processes of part manufacture is 30 % of the total labor intensity. The goal of the article is to create a numerical model that allows forecasting timely measuring uncertainties of free form surfaces of parts during their inspection on coordinate measuring machines (CMM). Forecasting is performed with the help of the neural network. A training set is created for the neural network by generating actual surfaces of parts containing data on production deviations and modelling the process of actual surface measurement. All the parts of the model have been implemented in the MATLAB system. The forecasts of the measuring uncertainty for blade body edges of the compressor have been made. 97 % of the obtained results do not exceed 10 % of the maximum measuring uncertainty value.