模型预测方法在城市固体废弃物热转化过程中的应用  

Model prediction for thermal conversion process of municipal solid waste

作　　者：袁浩然[1,2] 赵丹丹[1,2] 陈勇[3] 

机构地区：[1]中国科学院广州能源研究所,广东广州510640 [2]中国科学院可再生能源与天然气水合物重点实验室,广东广州510640 [3]中国科学院广州分院,广东广州510070

出　　处：《化工学报》2013年第6期1893-1899,共7页CIESC Journal

基　　金：国家重点基础研究发展计划项目(2011CB201501);中国科学院知识创新工程重要方向项目(KSCX2-EW-G-1-5);国家自然科学基金国际合作项目(51161140330)~~

摘　　要：城市固体废弃物(MSW)热转化过程预测模型对研究其实验过程和热处理装置的设计和运行具有重要的指导意义。目前,应用于MSW热转化过程的模型可归纳为动力学模型、热力学平衡模型、多元回归模型和神经网络模型四类。模型研究主要集中于进行MSW热转化特性、产物热值、污染物排放、灰渣熔点等方面的预测。动力学模型适用于较窄的温度区间,对于较大的温度区间需要分段建模。热力学平衡模型局限于少数明确定义机理的反应,对于稍微复杂的系统,通常需要先对反应机理进行进一步研究。多元回归模型也需要先确定函数类型。MSW热转化过程的不确定性和非线性特点,使得神经网络这种非线性建模和预测方法以其良好的自适应、自学习能力成为模型研究的主要发展趋势。Prediction model of municipal solid waste （MSW） thermal conversion process can be applied for study of experimental as well as design and operation of heat treatment device. At present, the prediction models can be classified as kinetic model, thermodynamic equilibrium model, multiple regression model and neural network model. The study on prediction models mainly focuses on predicting MSW thermal conversion characteristics, heat value of products, pollutant emissions, ash melting point and so on. The kinetic model is appropriate for narrow temperature interval and segment models should be established for wide interval. The thermodynamic equilibrium model is restricted to few reactions comprising specific mechanism. For some complex systems, the reaction mechanism should be further investigated generally. The function type should be also confirmed in the multiple regression model. The nondeterminacy and nonlinearity of MSW thermal conversion process makes the nonlinear neural network prediction model become main development tendency of model study for its predominant self-adaption and self-learning.

关 键 词：城市固体废弃物(MSW) 热转化过程 非线性 预测模型 

分 类 号：X705[环境科学与工程—环境工程]