垃圾焚烧过程选择性催化还原（SCR）脱硝系统出口NO _x 排放浓度的准确预测对提高数据质量和喷氨控制水平具有重要意义。垃圾焚烧过程存在显著的非线性、多变量耦合和时间序列特性，给NO _x 排放浓度的精准预测带来了巨大挑战。针对此问题，该文将最大信息系数（MIC）、主成分分析（PCA）和长短期记忆（LSTM）神经网络相结合，提出了一种SCR脱硝系统出口NO _x 排放浓度预测模型。首先，采用MIC方法计算各变量间的最大归一化互信息值，选择和NO _x 排放浓度相关性较大的特征变量，再结合最大冗余原则剔除冗余变量。随后，基于PCA方法获得各主成分方差的累计贡献率，提取主成分特征，得到最优输入特征变量集。最后，利用LSTM神经网络建立SCR出口NO _x 排放浓度预测模型。结果表明，相比反向传播神经网络模型和支持向量机模型，该文提出的模型具有最优的预测精度和泛化能力，其测试集平均绝对百分比误差为6.33%，均方根误差为4.71 mg/m³，决定系数为0.90。研究结果为实现垃圾焚烧过程SCR脱硝系统的喷氨智能控制提供了理论基础。

Accurately predicting the emission concentration of NO _x at the outlet of the selective catalytic reduction (SCR) denitrification system in the waste incineration process is of great significance for enhancing data quality and optimizing ammonia injection. However, the waste incineration process exhibits significant nonlinearity, multivariate coupling, and time-series characteristics. These factors pose substantial challenges to achieving accurate prediction of NO _x emissions. To solve this problem, this paper presents a prediction model for the emission concentration of NO _x at the outlet of SCR denitrification system by integrating maximum information coefficient (MIC), principal component analysis (PCA) and long short-term memory (LSTM) neural networks. First, MIC is employed to assess the maximum normalized mutual information values among variables, and the input variables that exhibit the strong-est correlation with NO _x emission concentration are selected while the redundant variables are eliminated based on the principle of maximum redundancy. Then, PCA is utilized to obtain the cumulative contribution rate of the va-riance of each principal component, extract the principal component features, and obtain the optimal input feature variable set. Finally, an emission prediction model of NO _x at the outlet of SCR denitrification system is established based on the LSTM neural network. The results indicate that, as compared with the back propagation neural network model and the support vector machine model, the proposed model exhibits higher accuracy and generalization ability, achieving a mean absolute percentage error of 6.33%, a root mean squared error of 4.71 mg/m³ and a determination coefficient of 0.90. This research lays a theoretical foundation for achieving the intelligent control of SCR denitrification system in the waste incineration process.