Prediction of spraying process parameters based on BP neural network

Abstract

In order to solve the problem that there are many process parameters in spraying process, which have complex influence on coating quality and have uncertain process parameters, a prediction method of spraying process parameters was proposed based on BP neural network algorithm. In the process of spraying, there are many factors that affect the quality of coating, and their interaction is complex, so it is difficult to find an accurate mathematical model in accordance with its rules. In this paper, BP neural network is used to establish the nonlinear mapping relationship between spraying process parameters and coating quality indexes, which is used as the prediction model of spraying process parameters. Taking ambient temperature, humidity, paint viscosity and film thickness as the input of neural network, and flow proportional valve pressure, atomization pressure and fan control pressure as the output, BP neural network model was established, and the prediction ability of the model was improved by optimizing the structure of the model. After optimizing the structure of BP neural network model, the prediction results show that the average relative error of flow proportional valve pressure is 0.4%. The average relative error of atomization pressure is 0.43%. The average relative error of sector control pressure is 0.61%. The established BP neural network model can meet the prediction requirements, and can clearly describe the relationship between environmental parameters, paint parameters, pressure parameters and coating quality, so as to better control the spraying process.