Energy Management in Electrical Power System Employing Machine Learning

Fossil fuel depletion has led to the increase in usage of solar energy. Many plants have been setup in various parts of the India to harness this energy. One of the major contributors is Karnataka. The electricity demand here is met using both energy from non-renewable energy and solar energy. The technology used to meet the demand is by injecting all the power from all the sources to a single transmission line. This is then distributed to different feeders according to its needs. Solar energy is not used to its fullest since it is only injected to the transmission line and extra energy is only stored for reserve. A framework is been introduced in order to make this solar energy used to its full extent and usage of other power is only at times when solar energy is not able to meet the demand. This framework proposes a unique switching strategy by maximizing the usage of solar employing various machine learning algorithms. The data of all the powers and demand are taken from Karnataka power transmission corporation limited (KPTCL) official website for prediction using machine learning techniques. Out of many algorithms used such as linear regression, logistic regression, decision tree, random forest and support vector machines (SVM), it is found that random forest is most efficient and gives the best switching configuration for maximum usage of solar energy.