神经网络在线优化PI参数的PLC及组态设计

论文编号:ZD157  包括PLC源程序,开题报告,外文翻译,任务书,及过程材料,S7-200PLC  论文字数:15039,页数:43

摘  要
PID控制是迄今为止在过程控制中应用最为广泛的控制方法，它可以用于补偿系统，以达到大多数特性参数的要求。实践证明PID控制具有结构简单、稳定性能好、可靠性高等优点，尤其适用于可建立精确数学模型的确定性控制系统。但对于时变对象，非线性系统，传统的PID控制显得无能为力。由于神经网络优化PID的控制策略兼有常规PID和神经网络各自的优点，具有自学习、自适应、自组织的能力且能适应被控过程参数的变化。因此本设计采用BP算法进行PI参数的在线调整，能使输出值较好地跟踪给定值，从而达到较好的控制效果。在实际系统中运用S7-200PLC语言实现该算法，并在单容水箱液位控制系统中实现。同时，利用MCGS组态软件对调节曲线进行实时监控。该组态软件能够更直观的监测到控制效果并记录实时曲线的变化，同时又能够对参数进行设置和监控。结果表明，神经网络PI控制器参数调整简单，具有很高的精度和很强的适应性，可以获得满意的控制效果。
 
关键词：PID控制；神经网络；PLC；单容水箱；MCGS
The PLC and Configuration Interface Design of Optimal PI Controller Parameters Based on Neural Network on Line
Abstract
PID control is the most widely used method of control in the process control by now, it can be used for compensation, to reach the majority of the request parameters. It is currently in the process control in the most common controller, and practice has proven PID control with simple structure, stable performance and high reliability, in particular fit with the establishment of precise mathematical model of uncertainty control system. But for the time varying, nonlinear systems, the traditional PID control is powerless. Optimal PID control based on neural network has self-learning, adaptive, self-organization's ability, and can adapt to the changes of process parameters. Because neural network optimal PID control strategy with both conventional PID and neural network with their respective advantages. Therefore the design of BP PI algorithm parameters online adjustments, in order to output value better tracking of a given value, can achieve better control results. Using of S7-200PLC language of the algorithm in actual system, and achieved in a single-capacity tank control system, meanwhile, using of MCGS for real-time monitoring. This configuration software can monitor more intuitive to control results and recording the change of real-time curve, and set up and control the parameters. Results show that the neural network PI parameter adjustment easily with very high precision and strong adaptability, and can get satisfied control results.

Key words：PID control; neural network; PLC; single-capacity tank; MCGS
目  录
1 绪  论 1
1.1 发展前景及历史 1
1.1.1 PID的发展前景及历史 1
1.1.2 神经网络的发展前景及历史 2
1.2 设计的研究方法及内容 3
2 基本理论 4
2.1 常规PID 4
2.1.1 PID控制原理 4
2.1.2 理想PID算法 5
2.1.3 PID算法存在的问题 5
2.2 BP算法 6
3 控制系统的PLC程序设计 10
3.1 单容水箱液位控制系统 10
3.1.1 常规PI控制 10
3.1.2 BP网络PI控制 11
3.2 编程软件 12
3.3 控制系统回路表及地址分配 14
3.4 单容水箱控制系统程序设计 16
3.4.1 总体设计 16
3.4.2 编程思想及主要程序说明 17
4 组态设计 24
4.1 MCGS组态软件概述 24
4.2 软件组态 25
4.3 在线调试 28
4.4 设计结果及分析总结 33
结  论 36
致  谢 38
参考文献 39

相关论文