Machine Learning-Based Predictive Techno-Economic Analysis of Power System

This paper proposes a predictive techno-economic analysis in terms of voltage stability and cost using regression-based machine learning (ML) models and effectiveness of the analysis is validated. Predictive analysis of a power system is proposed to address the need for faster and accurate analyses that would aid in the operation and control of modern power system. Several methods of analyses including metaheuristic optimization algorithms, artificial intelligence techniques and machine learning algorithms are being developed and used. Predictive ML models for two modified IEEE 14-bus and IEEE-30 bus systems, integrated with renewable energy sources (solar and wind) and reactive power compensative device (STATCOM) are proposed and developed with features that include hour of the day, solar irradiation, wind velocity, dynamic grid price and system load. An hour-wise input database for the model development is generated from monthly average data and hour-wise daily curves with normally distributed standard deviations. The data feasibility tests and output database generation is performed using MATLAB. Linear and higher order polynomial regression models are developed for the 8760hr database using Python 3.0 in JupyterLab and a best-fit predictive ML model is identified by analysing the coefficients of determination. The voltage stability and cost predictive ML models were tested for a 24hr input profile. The results obtained and the comparison with the expected values are furnished. Prediction of the outputs for the test data validate the accuracy of the developed model.

Abstract licence: CC BY-NC-ND

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.