摘要: 本课题的主要设计目的是利用函数逼近的方法实现对于在碳阳极生产工艺过程中几个流程中的不同变量进行模型的构建,从而能达到控制碳阳极质量的生产。本设计从阳极生产的石油焦煅烧、阳极焙烧这两个过程中选取温度、压力等变量进行函数拟合,利用最小二乘法、BP神经网络方法建立模型实现控制。对于阳极焙烧过程中,烟气压力这一变量模型,本文采用的是小波神经网络方法进行建模,并加以PSO方法优化;而在焙烧温度控制模型方面,采用的是基于遗传算法的优化神经网络进行模型的建立。24104 毕业论文关键词: 碳阳极;Matlab;函数逼近;阳极焙烧;建模; 神经网络
Matlab function approximation based on modeling of the carbon anode quality control applications
Abstract: The main purpose of this project is to design the use of function approximation methods to achieve carbon anode for different variables in the production process of the process to build several models, which can achieve the quality control of the production of carbon anodes. The design of the anode production from calcined petroleum coke, anode baking process to select both the temperature, pressure and other variables function fitting, using the least squares method, BP neural network modeling to achieve control.For the anode baking process,the flue gas pressure of the variable model,this paper uses a wavelet neural network modeling and optimization methods to PSO;while roasting temperature control model,the use of a genetic algorithm optimization neural network based for modeling building.
Keywords: carbon anode; Matlab; function approximation; anode ; neural networks