基于样本增量驱动的量子快速学习网络模型与蜂群算法研究及其在循环流化床锅炉燃烧过程优化控制中的应用

For complex characteristics of the circulating fluidized bed boiler (CFBB) combustion process such as large time-delay and strong nonlinearity and time-varying etc., it is difficult to obtain satisfactory control performance using conventional control methods. Therefore, an advanced optimal control theory is used to explore the new optimal control method for the combustion process of CFBB. The project adopts the optimal control system theory with two-layer structure to study optimal control of CFBB combustion process. In two-layer structure, the upper layer is optimal control layer and the lower layer is basic automation control layer. The core task of the upper layer is to provide the optimal set value for the lower layer achieving the combustion optimization control target. Accordingly, the research work is surrounds optimization control layer. First of all, the optimal model of combustion process based on quantum neural network is established. Then, the feedforward algorithm is investigated to achieve online correction for the threshold or incentive function. At the same time, the quantum swarm optimization algorithm based on three-dimensional search of Bloch sphere is studied. In addition, the optimization theory, the artificial intelligence and other related knowledge is used to build intelligent link of the upper system, to design the upper structure and to explore multi-objective optimization algorithm and the control strategy. Finally, the effectiveness of the adopted and proposed theory and method of this project are verifying through the simulation experiment. The project desires to explore a new way for solving the problems of CFBB combustion process optimization control .

针对循环流化床锅炉(CFBB)燃烧过程的大滞后、强非线性、时变性等复杂特性，采用常规的控制方法难以获得满意的控制效果，于是运用先进的优化控制理论探索CFBB燃烧过程优化控制的新方法。本项目采用具有二层结构的优化控制系统理念来研究CFBB燃烧过程的优化控制，即上层为优化控制层，下层为基础自动控制层。上层的主要任务是为下层提供能够实现燃烧过程优化控制目标的最优设定值。依此，围绕优化控制层展开研究工作。首先建立基于量子神经网络的燃烧过程优化模型，然后研究对模型权阈值或激励函数实现在线修正的前馈算法和基于Bloch球面三维搜索的量子蜂群优化算法。之后，利用优化理论和人工智能等相关知识，构建上层系统中的智能环节、设计上层系统结构以及探索多目标优化算法和控制策略。最后，通过实验研究验证项目中所采用和提出的理论与方法的有效性和实用性。本项研究渴望为解决CFBB燃烧过程优化控制中存在的问题探索一条新途径。

针对循环流化床锅炉(CFBB)燃烧过程的大滞后、强非线性、时变性等复杂特性，采用常规的控制方法难以获得满意的控制效果，于是运用先进的优化控制理论实现了CFBB燃烧过程优化控制的新方法。本项目采用具有二层结构的优化控制系统理念来实现CFBB燃烧过程的优化控制。首先建立了基于量子神经网络的燃烧过程优化模型，然后提出了对模型权阈值或激励函数实现在线修正的前馈算法和基于Bloch球面三维搜索的量子蜂群优化算法。之后，利用优化理论和人工智能等相关知识，构建上层系统中的智能环节、设计上层系统结构以及探索多目标优化算法和控制策略。最后，通过实验研究验证项目中所采用和提出的理论与方法的有效性和实用性。该项目的研究不仅具有理论意义，而且对煤炭洁净利用科技的发展具有重大的意义，同时拥有广泛的应用前景与巨大的社会效益。