Neural network with fuzzy set-based classification for short-term load forecasting

Abstract

Electric power utilities require forecast of system some drawbacks such as inaccurate prediction, difficulty in demand or electrical load for one to seven days ahead. This modeling processes, numerical instability, requirement of paper studies a short-term electric load forecasting technique large historical database, and demand of high human using a multi-layered feedforward Artificial Neural Network (ANN) and a fuzzy set-based classification algorithm. The Recently. the application of the artificial neural network hourly data is subdivided into various class of weather (ANN) to short-term load forecasting has gained a great deal conditions using the fuzzy set representation of weather of interest and several researchers have reported the variables and then the ANN'S are trained and used to perform effectiveness of the ANN approach [ 6, 7, 8 1. Unlike the the load forecasting up to 120 hours ahead with a remarkable previous techniques, the ANN leams the patterns from the accuracy. inputs and outputs of the utilities' system and then, it creates expertise.

Short-term load forecasting in non-residential Buildings

“…The best record to our knowledge in short-term load forecasting presented a MAPE of the 1.53 % [17]. According to [8], simply the reduction of the 1% in the average forecast error may save hundreds of thousands or even millions of dollars.…”

Short-term load forecasting in air-conditioned non-residential Buildings

“…Finally, the AR model shows the best overall performance, with very acceptable marks, specially with normal data. The best record to our knowledge in short-term load forecasting presented a MAPE of the 1.53 % [37]. According to [1], simply the reduction of the 1% in the average forecast error may save hundreds of thousands or even millions of dollars.…”

“…In our problem domain, non-residential buildings, this possibility is not plausible since an 1% error may be a deviation of some kW. Hence, [20] accomplished a 1,945% (two points above the AR model in dataset 2, only a 1,31% difference if we take just weekdays) using a neural network (as in the case of [37]), with all the problems aforementioned that NN present (for instance, [37] uses a fuzzy set-based classification algorithm to improve the classification ability of their NN). Still, Neural networks (also Bayesian networks) offer a worse trade-off between the difficulty of design, parametrisation, etc., and the performance, in comparison, for instance, with time series.…”

Short-Term Load Forecasting for Microgrids Based on Artificial Neural Networks

“…While the solutions studied in the literature [35][36][37][38][39][40][41][42][43][44][45][46][47][48][49][50][51][52][53][55][56][57][58][59] present sometimes good prediction efficiency figures (normally their MAPEs are around 2%), they deal almost exclusively with big areas, and mainly entire countries, and they are never applied to smaller environments of the size of small cities or microgrids. Therefore, they do not give any evidence of how will they behave when applied to highly variable load curves.…”

“…In the work of Kim et al [47], an ANN produces a provisional load forecast, then, a fuzzy expert system is used to modify the provisional load forecast on the basis of temperature data and day type (workday/holiday). Daneshdoost et al [48] classify data into 48 fuzzy subsets by temperature and humidity, then each subset is modeled by its own ANN. Senjyu et al [49] present a hybrid correction method where fuzzy logic, based on -similar days‖, corrects the neural network output to obtain next-day load forecast.…”