CMAC复合PID在柴油发电机组转速控制中的应用  

Application of CMAC Neural Network Coupled with PID Controller on Speed Control of Diesel Generating Set

作　　者：石勇[1] 齐自达[1] 张连瑜[1] 张红光[1] 

机构地区：[1]哈尔滨工程大学动力与能源工程学院,黑龙江哈尔滨150001

出　　处：《内燃机学报》2012年第6期563-568,共6页Transactions of Csice

基　　金：国家自然科学基金资助项目(50909024);中央高校基本科研业务费专项资金资助项目(GK20306022)

摘　　要：船用柴油发电机组调速系统是复杂的非线性系统,需要采用先进的控制技术以提高其转速响应特性与稳定性.提出了一种小脑模型神经网络(CMAC)与PID复合控制的柴油机调速策略,利用传统PID来实现反馈控制,保证系统的稳定性并抑制扰动,CMAC神经网络对控制器进行前馈控制,确保系统的控制响应速度,提高控制精度.在搭建dSPACE半实物仿真平台并对算法进行了仿真验证之后,使用MicroAutoBox进行了D6114发电机组配机试验.试验结果表明,机组具有良好的动静态特性,转速波动率为0.15%,标定负载突变超调6.7%,稳定时间为2.5s,完全满足二级电站对柴油机调速系统的要求,通过与D6114发电机组原控制器下的发电频率变化曲线对比,证明了其具有与产品级控制器相近的控制性能.Speed governing system of marine diesel generating set is a complicated non-linear and time- variable system. Advanced controlling technology is required to increase its speed dynamic response and stability. A controller combined with cerebellar model articulation controller （CMAC）neural network and PID is proposed and implemented to the speed control of marine diesel generating set. The PID feedback controller is employed to keep system stability. CMAC feedforward controller has the advantage of fast- convergence rate and ability of local learning and can maintain dynamic response and precision. Simula- tion was made in hardware-in-loop system built by dSPACE and MicroAutoBox, the controller was ex- perimentally evaluated on the diesel generating set of D6114. The overshot with load changing from full load and zero load is within 6.7% and speed fluctuation is within 0.15%. Comparison on power generation frequency shows that the CMAC-PID controller has the similar performance with that of the production controller.

关 键 词：船用柴油发电机组 小脑模型关联控制器 试验验证 

分 类 号：TK427[动力工程及工程热物理—动力机械及工程]