Abstract

Today, due to the rapidly increasing environmental pollution, the importance of energy obtained from renewable energy sources is increasing. One of these important renewable energy sources is Hydroelectric Power Plants. Using hybrid-machine learning algorithms, non-hybrid machine learning algorithms and deep learning models, the electrical energy produced by Hydroelectric Power Plant was estimated based on seventeen different input parameters such as flow rate and generator current. In this study, hyperparameters were tuned and feature selection was made with a Genetic Algorithm in order to increase the accuracy rate in artificial intelligence models. In addition, all artificial intelligence models were evaluated according to the processing time, Coefficient of determination, Mean Squared Error, Mean Absolute Error and Root Mean Square Error performance evaluation criteria. It was seen that the electrical energy production value was determined with the Random Forest Algorithm, which is one of the non-hybrid machine learning algorithms, with an accuracy rate of 99.641%, with Genetic Algorithm + Random Forest Algorithm, which is a hybrid machine learning algorithm, with an accuracy rate of 99.672% and with Deep Neural Network, which is one of the deep learning models, with a 99.99% accuracy rate. According to the results, it was determined that the Deep Neural Networks model determined the electrical energy production value with a higher accuracy rate compared to hybrid and non-hybrid machine learning algorithms.

摘要

今天，由于环境污染的迅速增加，从可再生能源中获得能源的重要性正在增加。这些重要的可再生能源之一是水力发电厂。使用混合机器学习算法、非混合机器学习算法和深度学习模型，根据 17 个不同的输入参数（如流量和发电机电流）估算水力发电厂产生的电能。在这项研究中，调整了超参数并使用遗传算法进行了特征选择，以提高人工智能模型的准确率。此外，所有人工智能模型都根据处理时间、判定系数、均方误差、平均绝对误差和均方根误差性能评价标准。可见电能产值是用随机森林算法确定的，它是一种非混合机器学习算法，准确率为99.641%，用遗传算法+随机森林算法，这是一种混合机器学习算法，准确率为99.672%，深度神经网络是深度学习模型之一，准确率为99.99%。根据结果​​，确定与混合和非混合机器学习算法相比，深度神经网络模型以更高的准确率确定电能产值。可见电能产值是用随机森林算法确定的，它是一种非混合机器学习算法，准确率为99.641%，用遗传算法+随机森林算法，这是一种混合机器学习算法，准确率为99.672%，深度神经网络是深度学习模型之一，准确率为99.99%。根据结果​​，确定与混合和非混合机器学习算法相比，深度神经网络模型以更高的准确率确定电能产值。可见电能产值是用随机森林算法确定的，它是一种非混合机器学习算法，准确率为99.641%，用遗传算法+随机森林算法，这是一种混合机器学习算法，准确率为99.672%，深度神经网络是深度学习模型之一，准确率为99.99%。根据结果​​，确定与混合和非混合机器学习算法相比，深度神经网络模型以更高的准确率确定电能产值。准确率为 99.672%，深度神经网络是深度学习模型之一，准确率为 99.99%。根据结果​​，确定与混合和非混合机器学习算法相比，深度神经网络模型以更高的准确率确定电能产值。准确率为 99.672%，深度神经网络是深度学习模型之一，准确率为 99.99%。根据结果​​，确定与混合和非混合机器学习算法相比，深度神经网络模型以更高的准确率确定电能产值。