Database of Consumption in CAMEROON (2020)

Many techniques have been used to predict the exact electrical load and reduce losses. Among these techniques, artificial intelligence models (fuzzy logic and ANN) offer greater efficiency compared to conventional techniques (e.g. regression and time series). In this article, a fuzzy logic model is presented for forecasting annual electrical energy consumption in Cameroon. This model is developed according to the evolution of the population, the gross domestic product (GDP) and historical data of the annual consumption of electricity in Cameroon. The development of an effective fuzzy rule base has allowed us to forecast future annual consumption over a ten-year period with a MAPE of 0.011%. A comparison of the results obtained with those of similar models allowed us to conclude that fuzzy logic offers very high accuracy in terms of forecast error.已有诸多技术被用于精准预测电力负荷并降低损耗。其中，人工智能模型（涵盖模糊逻辑与人工神经网络（Artificial Neural Network，ANN））相较传统技术（如回归分析与时间序列方法）具备更高的效率。本文提出了一种面向喀麦隆年度电能消费量预测的模糊逻辑模型。该模型以喀麦隆的人口演化、国内生产总值（Gross Domestic Product，GDP）以及年度电力消费历史数据为依托进行构建。通过搭建高效的模糊规则库，我们得以对未来十年的年度电力消费量进行预测，其平均绝对百分比误差（Mean Absolute Percentage Error，MAPE）仅为0.011%。将本模型的预测结果与同类模型的结果进行对比后，我们可得出结论：模糊逻辑在预测误差维度上具备极高的精度。