基于人工神经网络的锅炉受热面积灰预测研究现状及展望

Status quo and prospect of boiler heating surface based on artificial neural network

人工神经网络(Artificial Neural Network,ANN)可对锅炉受热面积灰数据间的非线性关系进行很好地描述,较好地预测锅炉受热面积灰程度。对ANN积灰预测模型的构建、向量数据的选取和处理、预测评价指标和网络优化方面的相关研究现状进行综述。在预测模型的构建中,输出向量选取锅炉实际运行中与受热面积灰程度呈单向变化的参数,输入向量采用基于热力学平衡的影响因素法,考虑锅炉设计及运行工况、受热面状况进行选取,并进行奇异值剔除、数据去噪等处理,采用平均影响值等算法进行最终确定。在网络优化上开展多种先进算法与网络模型相结合,并借鉴其他预测领域的ANN模型,开展基于深度神经网络的受热面积灰预测的研发是未来重要的研究方向。

Artificial Neural Network(ANN)can be well described with nonlinear relationship between heat-resistant area of boiler,which is better to predict the degree of heating area ash of the boiler.The paper summarized the relevant research status from the construction of the ANN ash prediction model,the selection and processing of vector data,predicting evaluation indicators,and network optimization.It is believed that in the construction of the prediction model,the output vector selects the parameters of the actual operation of the boiler in a one-way change in the degree of grayness.The input vector used the influencing factor method based on thermodynamic balance,considering the boiler design and operating conditions,the heat surface condition was selected,and the strange value was removed,data denoising and other algorithms,the algorithm,etc.The network optimization was carried out in combination with network models,and drew on the ANN model in other predictive fields to the study,and the research and development based on deep neural network was the future of an important research direction.