基于GA-SVM的燃气轮机NO_(x)排放预测研究

GA-SVM-based NO_(x)Emission Prediction Study for Gas Turbines

为解决燃气轮机发电站存在NO_(x)超标排放问题,根据燃气轮机运行过程的环境变量和工艺过程参数,提出一种基于GA-SVM模型的NO_(x)排放预测方法。以UCI中的燃气轮机运行数据集进行试验。本文提出RMSE和MAE评估模型。结果表明:该模型输出均方根误差R MSE为8.79 mg/m^(3),平均绝对误差(MAE)为6.23 mg/m^(3),并通过与KNN模型和RF模型进行比较,验证了所提方法能够对燃气轮机NO_(x)排放浓度进行准确的预测。

To solve the problem of excessive NO_(x)emissions in gas turbine power stations,a NO_(x)emission prediction method based on GA-SVM model is proposed according to the environmental variables and process parameters of gas turbine operation process.Experiments are conducted with the gas turbine operation dataset in UCI.The RMSE and MAE evaluation models are proposed in this paper.The results show that the model outputs a root mean square error RMSE of 8.79 mg/m^(3)and a mean absolute error(MAE)of 6.23 mg/m^(3).By comparing with the KNN and RF models,the proposed method is verified to provide accurate prediction of gas turbine NO_(x)emission concentrations.