基于改进ANFIS的铝生产中铁含量时间序列预测

Time Series Prediction of Iron Content in Aluminum Production Based on Improved ANFIS

在铝生产工业中对杂质铁含量的预测有利于提高铝液的生产水平。针对传统预测模型需要大量数据进行训练,耗时长且效率低下等问题,提出了一种改进ANFIS的时序预测模型(GNG-AMDEA-ANFIS)。首先利用GNG模型对原始数据集动态跟踪找到奇异点,然后采用自适应变异差分进化算法和梯度下降法对ANFIS网络所需参数进行优化,同时构建ANFIS网络模型。最后,结合贵阳某铝厂铝生产过程中铁含量数据,完成对该模型的性能验证。结果表明文章建立的模型能够对铁含量进行精确预测与监督,节省计算成本的同时提高了预测效率。

The prediction of impurity iron content in aluminum production industry is beneficial to improve the production level of liquid aluminum.To solve the problems of traditional forecasting model,such as large amount of data,long time consuming and low efficiency,a time series forecasting model with improved fuzzy neural network(GNG-AMDEA-ANFIS)was proposed.First,the GNG model is used to find the singular points for the dynamic tracking of the original data set.Then,adaptive mutation differential evolutionary algorithm and gradient descent method are used to optimize the required parameters of the ANFIS network,and the ANFIS network model is constructed at the same time.Finally,combined with the data of iron content in the production process of an aluminum plant in Guiyang,the performance verification of the model is completed.The results show that the model established in this paper can accurately predict and supervise the iron content,save the calculation cost and improve the prediction efficiency.