Smart Manufacturing with An Artificial Neural Network to Predict Manufacturing Healthiness

In this paper, Smart Manufacturing with an Artificial Neural Network (ANN) system is proposed to perform the prediction of the healthiness of the manufacturing line. There are three main stations in the manufacturing line in Season Malaysia Manufacturing (SMM); Surface Mount Technology, Manual Solder and Functional Testing. With the advancement of big data and artificial intelligence, it is worth to turn the production data into meaningful information to the manufacturer. Since these production data come from different resources, it is necessary to normalize the data so that they are consistent. The data collection was been conducted hourly so that the manufacturer could monitor the production healthiness hourly. The ANN is trained with a range of different configurations. The best ANN model is selected based on two main criteria; misclassification percentage and validation value. With the data available, the best ANN model was selected and incorporated into the design to develop the smart production healthiness monitoring system. The network evaluation shows 3.75% of misclassification and the mean square error (MSE) of 0.0875, in which the network response is satisfactory. The ANN system would certainly help the manufacturer to predict the healthiness of the manufacturing line and take-action on it timely. This will maximize the return on investment with lesser effort and higher productivity.