An Intelligent Calibration Testing of Electricity Meter using XGBoost for Manufacturing 4.0

The manufacturing company with a core business of manufacturing electricity meters has a testing system for the metrology industry, namely calibration testing. The part of calibration testing on an electricity meter is the verification test, which uses the method of comparing it to the standard meter test bench to calculate the accuracy error of the measurement. However, the problem is that carrying out this test requires a long cycle time, and it is difficult to increase production capacity without adding a standard meter calibration test bench which has an expensive investment. The smart factory concept that supports industry 4.0 can open up opportunities for research on information technology using artificial intelligence with machine learning models to solve this problem with the idea of creating intelligent testing. This research requires data collection on a calibration test bench machine, then processed to find model predictions so that they can be implemented into an intelligent test using the XGBoost Regression with Hyperparameter Tuning and Optimization methods as the Goal of this Research. In the results of this research, the evaluation of the XGBoost using the Hyperparameter Tuning and Optimization method, which is implemented in this case, could improve the accuracy and RMSE data testing modelling comparing other scenario models as defined before in the literature review. So, this can be an excellent solution to be applied in metrology manufacturing, especially verification tests in Manufacturing Calibration Testing on Electricity Meter, which is faster and low-investment testing with the implementation of an Intelligent Manufacturing Calibration Test.