Linear Regression-Based Machine Learning Model for Dynamic Load Flow Analysis

The nonlinear nature of the power system load introduces variability in the operation of the power system. These loads need to be modelled and handled accurately to get a close insight into the power system’s operation. The load flow algorithm helps to determine various informative indexes. The L-index, or fast voltage stability index, further shows the degree of stability of the system. Most of the literature shows that power system loads follow normal distribution. With the help of a dynamic load flow program that takes normally distributed loads as input and supplies unknown quantities in the slack bus and generator bus as output, we can apply advanced machine learning algorithms to these datasets to get the hidden model between input and output data. In the current decade, we are witnessing the application of artificial intelligence (AI) and machine learning (ML) algorithms in all fields of engineering. These advanced algorithms mostly rely on the datasets and algorithms to predict hidden patterns inside the datasets. In this paper, the dynamic load models are used as a set of input data, and the load flow results are used as a set of output data for the supervised machine learning algorithm. A linear regression-based supervised machine learning algorithm based on regression analysis is used to find the input-output relationship model in this paper. The obtained model is then tested and validated against test data. The model received from the ML algorithm shows an accurate relationship between input and output data sets. The proposed scheme is tested on the IEEE-6 bus system.