船舶柴油发电机组双回路系统神经网络控制研究

针对船舶柴油发电机组转速控制与励磁控制，设计了CMAC神经网络与PID控制器相结合的并行控制系统，并分别运用到船舶柴油发电机组系统的2个控制回路中。CMAC神经网络控制算法具有在线学习速度快和局部泛化能力强的特点；在控制中，CMAC神经网络通过对发电机控制的学习，获得了系统广义被控对象的逆动态逼近模型，以此实现对系统的控制。在某大型船舶电力仿真系统的运用中，发电机负载特性测试的结果表明控制系统的2回路之间的协调性好，系统控制质量满足有关船舶规范的要求。

Research of neural networks control of two loops marine diesel engine generator set system

Aimed at producing a control system of rotational speed and excitation for a marine diesel engine generator set,a cerebellar model articulation controller(CMAC) neural networks and PID parallel controller was designed and applied in two control loops of a marine diesel engine generator set system.The CMAC neural networks control algorithm possesses fast learning speed with an online method and strong ability of local generation.The system generalized inverse approach dynamic model of control object system was obtained through learning from a generator control by CMAC neural networks in control process.Then the control was realized.Applied to a large marine power simulation system,the test result of the load characteristic of the generator indicated good quality of coordinate between two control loops.The quality of system control satisfies the requirements of related marine criterion.