A collaborative fuzzy-neural approach for long-term load forecasting in Taiwan

Forecasting the long-term load in a country is a critical task for the government. In addition, establishing a precise upper bound for the long-term load avoids unnecessary power plant investment. For these purposes, a collaborative fuzzy-neural approach is proposed in this study. In the proposed approach, multiple experts construct their own fuzzy back propagation networks from various viewpoints to forecast the long-term load in a country. To aggregate these long term load forecasts, fuzzy intersection is applied. After that, a radial basis function network is constructed to defuzzify the aggregation result and to generate a representative/crisp value. The practical case of Taiwan is used to evaluate the effectiveness of the proposed methodology. According to the experimental results, the proposed methodology improved both the precision and accuracy of long term load forecasting by 40% and 99%, respectively. In addition, the proposed methodology made it possible to accurately forecast the average and peak values of the annual energy consumption at the same time.

► We estimate the annual energy consumption in Taiwan.
► The average value, peak load, and the lowest possible value of the annual energy consumption are estimated successively.
► Both the precision and accuracy of long-term load forecasting are improved.
► Through collaboration, the forecasting performance of an existing approach can be enhanced.