遗传算法和神经网络的DFB激光器温控系统

摘要

To solve the problems of nonlinearity and delay property existing in DFB laser temperature control system,the composite control structure based on genetic algorithm and neural network is proposed. In the system,the microprocessor as the system core processor is used to design the temperature control system,and the Pt resistance,TEC semiconductor refrigerator, temperature sensor and temperature control actuator are used as the control units. The neural network positive model was con⁃structed to analyze the physical characteristics of the controlled object. The neural network control is used to map the control laws,and the fast searching ability of genetic algorithm is used to train the weight coefficient of neural network. The designed system was verified with the experiment. The results show that the temperature control accuracy of the system is ±0.002 ℃,the range of temperature control is 5~70 ℃,the overshoot is less than 8%,the designed system can realize the control effect of high precision and wide range,and has better engineering application value.%针对DFB激光器温度控制系统普遍存在的非线性和延迟性等问题，提出了基于遗传算法和神经网络的复合控制结构，采用微处理器作为系统的处理器核心设计了温度控制系统，并利用铂电阻、TEC半导体制冷器、温度敏感器和温控执行器作为控制单元，再通过构造神经网络正模型分析被控对象的物理特性，用神经网络控制实现控制律的映射，同时利用遗传算法的快速搜索能力训练神经网络的权系数。最后，对设计的系统进行了实验验证，结果表明，该系统的温度控制精度为±0.002℃，控制范围为5~70℃，超调量低于8%，能够实现高精度和宽范围的控制效果，具有较好的工程应用价值。