Design of an Environmental Monitoring System for Additive Manufacturing Workshops based on Two-Level Data Fusion

Abstract

As an important element in the framework of intelligent manufacturing, additive manufacture works in the complex and changing workshop environment. These environments will bring great changes and instability to the data measured by every single sensor. And the traditional evaluation ways by one parameter only is limited by the one-dimensional view of it, which is not enough to get the precision that we need for the environment grade in the real time control application. In this paper, we solve these problems by giving an environmental monitoring system for the additive manufacture workshop, which is based on the dual level data fusion architecture of Kalman filtering and Bayesian networks. The system implementation is using an STM32 micro controller, a USR-EG828 edge gateway, embedded system program, Python development environment and the Modbus communication protocol. The results of experiments show that after the first fusion stage, the temperature data has greatly decreased the fluctuation, and the data of integration is more like the actual conditions. The estimated environmental parameters are very close to the real measured ones. In particular, the mean absolute error (MAE), root mean square error (RMSE) and maximum absolute error (MAXAE) for the temperature are decreased from 0.855 °C, 0.9776 °C, and 1.7543 °C (with one sensor only) to 0.373 °C, 0.2985 °C and 0.6127 °C, respectively. Similar improvements on the noise suppression and the measure precision can also be seen for the humidity, TVOC, and PM2.5 data. The second level data fusion gets perfect accuracy on the classification of the environmental grades for all ten validation samples. In summary, the introduced framework can help to make very accurate monitoring of many parameters and a complete assessment of the environmental grades in the manufacturing environment, which can largely improve the reliability of evaluating the ambient conditions. This research provides a practical technical way for the management of environmental factors in the additive manufacturing facilities.