甲醇合成反应中,影响甲醇单程收率的因数多,反应机理十分复杂,难以建立准确的机理模型。本文提出了用非主成分分析方法对输入变量预处理,运用广义回归神经网络的非线性映射能力,建立了甲醇合成单程收率的预测模型,并用此模型对不同时期的甲醇合成催化剂的活性进行评估。实例表明,此模型可对甲醇合成催化剂的活性进行定量评估,对指导甲醇生产具有重要意义。 Methanol synthesis reaction, affecting the single-pass yield of methanol factor, the reaction mechanism is very complex, it is difficult to establish an accurate mechanism model. In this paper, we propose a non-principal component analysis method for the input variables pretreatment, the use of generalized regression neural network nonlinear mapping capabilities to establish a single-pass methanol synthesis yield prediction model, and use this model for different periods of methanol synthesis catalyst activity Evaluation. The example shows that this model can quantitatively evaluate the activity of methanol synthesis catalysts and is of great significance to guide the production of methanol.