A novel time-span input neural network for accurate municipal solid waste incineration boiler steam temperature prediction

Abstract

A novel time-span input neural network was developed to accurately predict the trend of the main steam temperature of a 750-t/d waste incineration boiler. Its historical operating data were used to retrieve sensitive parameters for the boiler output steam temperature by correlation analysis. Then, the 15 most sensitive parameters with specified time spans were selected as neural network inputs. An external testing set was introduced to objectively evaluate the neural network prediction capability. The results show that, compared with the traditional prediction method, the time-span input framework model can achieve better prediction performance and has a greater capability for generalization. The maximum average prediction error can be controlled below 0.2 °C and 1.5 °C in the next 60 s and 5 min, respectively. In addition, setting a reasonable terminal training threshold can effectively avoid overfitting. An importance analysis of the parameters indicates that the main steam temperature and the average temperature around the high-temperature superheater are the two most important variables of the input parameters; the former affects the overall prediction and the latter affects the long-term prediction performance.

概要

目的

生活垃圾焚烧炉主蒸汽温度为炉内燃烧调控的重点监控对象. 本文旨在建立一种时域输入的主蒸汽温度神经网络预测模型, 以实现主蒸汽温度未来5 min变化趋势的精准预测, 并且使预测误差控制在1%以内.

创新点

1. 实现了主蒸汽温度的未来趋势预测, 而非当前值预测; 趋势预测的结果能提供操作人员一定的参考价值. 2. 提出了一种时域输入神经网络模型; 该模型能够包含输入输出参数之间的延时特性, 因此能获得更高的预测精度.

方法

1. 通过数据相关性分析与延时性分析, 确定用于预测主蒸汽温度的输入变量, 并减少模型输入层数据维度(表1);2. 提出时域输入算法设计(公式(4)~(5)), 构建时域输入主蒸汽温度神经网络预测模型, 以实现主蒸汽温度未来5 min变化趋势的精准预测(图8);3. 通过调整模型参数, 优化模型结构;4. 通过输入数据敏感度分析, 得出对主蒸汽温度预测影响最大的变量(图14).

结论

1. 本文提出的时域输入神经网络模型比传统神经网络模型的预测精度更高; 2. 时域输入主蒸汽温度神经网络预测模型在未来1 min内可以实现近零预测误差; 3. 根据输入数据敏感度分析可得, 对于本研究的焚烧炉, 主蒸汽温度本身的数据对于其预测的重要性最高; 其次, 高温过热器烟气平均温度对于主蒸汽温度远未来预测的重要性较高.