Long-Term Load Forecasting: A Systematic Review with Focus on Microgrid Application

The long-term load forecasting (LTLF) plays an important role in multiple areas of power distribution system including demand side management and system planning. The LTLF models can be helpful for utility in providing valuable inputs to support grid expansion and electricity purchase agreements. Therefore, the research in the field of load forecasting is ongoing and evolving continuously. This paper presents a systematic review of different approaches that have been reported recently in the literature for LTLF of distribution grid and microgrids. The critical review and comparison among the LTLF models have been carried out based on multiple aspects such as methodology, objective, applicability, forecasting accuracy and error matrix, forecasting horizon and historical data requirements with unique outlook on industry-based acceptability challenges and operational complexities. Many conventional models described in the literature and also being used by the utilities are old and were developed with low resolution data. This paper also highlights the area of improvements in the existing LTLF model as with the current widespread availability of high-resolution data from sensors and smart meters. This review also focusses on the problem addressed by the existing literatures, results obtained by the forecasting models and methods to examine and validate the results. Further, the review work has also been extended to examine the feasibility of real-time deployment of the forecasting models mentioned in the literature. By this approach and comparison of different LTLF models, a few algorithms are highlighted which provide better and improved performance and have the potential of deployment in the distribution level microgrids.