作　　者： ; ; ; ;

机构地区： 北京交通大学电子信息工程学院全光网络与现代通信网教育部重点实验室

出　　处： 《北京理工大学学报》 2013年第5期523-528,共6页

摘　　要： 为实现射流火焰温度场的控制性能,采用随机分布控制理论实现了射流火焰温度场的建模及其迭代学习控制.采用高斯函数作为输出概率密度函数逼近的基函数,用广义系统状态空间模型建立系统模型,模型的独立状态数量与系统实际的动态阶次一致.采用随机分布预测控制算法实现单批次的控制,在每个批次控制完成后,采用牛顿法优化基函数参数,给出了射流火焰温度场迭代学习控制的仿真结果.通过迭代学习优化了广义随机分布系统模型参数,提高了火焰温度场分布控制性能指标. To improve the performance of controlling jet flame temperature distribution, stochastic distribution control method is adopted to build the model of jet {lame temperature field whose parameters are optimizeed using iterative learning control in this research. Firstly, Gaussian type basis functions were used to approximate output probability density function. A singular state-space model for stochastic distribution system was formulated, where the number of independent states was the same as the actual dynamic order of the plant. Thereafter, predictive control algorithm was used to control each batch. After completing the control of each batch, Newton method was used to optimize the parameters of basis function. Finally, simulation results were given with temperature distribution of jet {lame as its controlled plant. It is indicated that, through optimizing the modeling parameters of singular stochastic system using iterative learning algorithm, the performance index of stochastic distribution control of flame temperature distribution could be improved effectively.

关 键 词： 迭代学习控制 输出概率密度函数 广义系统 温度场分布

领　　域： [自动化与计算机技术] [自动化与计算机技术]