基于人工神经网络建模软件的铝电解应用研究

Application of Aluminium Electrolysis Based on Artificial Neural Network

人工神经网络具有自学习、自组织、自适应以及很强的非线性函数逼近能力,可针对具体情况,建立起相应的神经网络。针对铝电解而建立起的可视化人工神经网络建模系统,对铝的电解率进行预测,神经网络的预测值与实验值相比较,误差小于1%,证明利用人工神经网络法研究铝电解是可行的。

The Artificial Neural Network(ANN) is self-learning, self-organizing and self-adapting, so it can be applied widely. Visual Artificial Neural Network Build Software can be used to build some applied ANN for special research. An Artificial Neural Network for the Aluminum electrolysis can forecast the ration of Aluminum electrolysis, with the forecast result being equal to the experiments. Applying the Artificial Neural Network to investigate the Aluminum electrolysis is effective.