Performance of ML methods in estimating net energy produced in a combined cycle power plant

One of the most fundamental problems in the generation of electricity is the supply of energy demand, and the inability to store excess energy produced. This nature of generation increases the cost of the generated energy and brings with it the need for intelligent energy management. When the need of energy in demand side and the output of energy produced in a generator is known, right amount of electrical energy can be supplied to the consumer with lower cost. The accurate estimation of the net energy produced in a power plant is an important task to supply low cost energy. Recently, machine learning (ML) approaches are proposed to predict the net hourly electrical energy produced from power plants. In this paper, the performance of ML methods in estimating net energy produced in a combined cycle power plant is presented. Performance of k-NN, Simple Linear, Linear, Decision Tree, Bayesian Linear, Gaussian Naive Bayes, and RANdom SAmple Consensus (RANSAC) Regression methods were evaluated. k-NN, Linear Regression and RANSAC regressions achieved the best performance among the other tested ML methods.