Java Bali Grid Mean Time To Failure Prediction using Machine Learning

The role of Mean Time To Failure (MTTF) is critical to maintain cost-efficient in smart grids, as the predictive maintenance means for ensuring their reliability. Java Bali grid is the most important power network in Indonesia, because it holds 75% energy sales and the foundation of the Indonesian government, business and industrial sectors. One method for prediction in the machine learning is the Linear Regression, as part of the Supervised Learning technique. The regression model extracts the grid overload variable from MV distribution system, where connected load (MVA) divided by the number of MV customers. The model constructs the Linear Regression function, where the slope and intercept values found from the correlation between predicted response and actual response. After confronting the function f(x) and the coefficient of determination (R2), the test results for the MTTF prediction values for Java-Bali’s region shows the higher accuracy when the R2 shows highest value.