基于BP神经网络的燃煤机组NOx排放浓度预测模型

为了准确预测燃煤电厂NOx排放量,笔者利用BP神经网络对NOx排放浓度建立了预测模型,对某电厂660 MW机组实时的不同机组负荷、脱销进口烟温、进口NOx浓度、进口O2浓度、出口NOx浓度的样本进行训练,得出训练模型.实践证明,通过训练后的BP神经网络模型对未知的NOx排放浓度进行预测,预测精度达到93.48％以上,完全满足实际中的预测需求。

Forecasting model of NOx emission concentration for coal-fired unit based on BP neural network

In order to Aaccurately forecast the quantity of NOx emission, the author establishes the forecasting model based on BP neural network, and acquires the training mode by training the sample of real time olad of different units, denitrated flue temperature at the entrance, NOx concentration at the entrance, O2 concentration at the entrance and NOx concentration at the exit from a 660 MW unit. The practice proves that to forecast the unknown NOx emission concentration by trained BP neural network model can increase the forecasting accuracy to 93.78%, which meets the requirement in practical forecasting.