Trend analysis and evolution of Short Term Load Forecasting Techniques

The study presents a literature based survey on Short Term Load Forecasting Techniques (STLF). Load Forecasts are the elementary factors considered by utilities while planning power generation, capacity building via power infrastructural development, load switching, load shedding, load flows and load sharing to name a few. Hence, rationalizing the importance bestowed to this area of research since the last two decades. Many techniques have been proposed and researched upon inducing the configuration of a research sphere which is still growing. Their fundamentals have evolved from time to time, chronologically ranging from: simple statistical techniques to artificial intelligence based models. The study presents a chronological review of the pertinent literature on popular STLF techniques supported by comparative analysis justifying the trends in technique evolution. Techniques covered under the study are: Single day approach, simulation models and time series models. Under Time series models we have discussed Regression models, Exponential smoothing, Parsimonious stochastic models, Support vector machines, Expert systems, Artificial Intelligence based models, Data mining and Hybrid models. Relevant sub-category models have also been given due coverage.