基于单神经元PSD算法的闪蒸罐压力自适应控制

Adaptive Control of Flash Tank Pressure Based on Single Neuron PSD Algorithm

传统PID无法自学习整定参数,适应时变、耦合、非线性系统的能力较弱。结合过控系统中闪蒸罐压力的特点,设计了具备增益自学习、自整定的单神经元PSD控制器。通过SCL编译成控制块,基于CFC完成单神经元自适应PSD-PID复合串级控制的组态,利用SMPT-1000实验平台验证,结果表明,闪蒸罐压力在SCL编译的单神经元PSD控制器作用下,实现了自适应控制,调节速度优于传统PID算法,整体进程加快1.3%,系统产量和回收料均有增加。在结合工艺流程的基础上,开发了WinCC人机界面,实时监控、调试闪蒸罐压力PSD控制器,同时直观展示系统的浓度、产量、回收工艺流量等重要指标。

Traditional PID can not self-learn to set parameters,and the ability to adapt to time-varying,coupling and nonlinear systems is weak.Combined with the characteristics of flash tank pressure in the over-control system,a single-neuron PSD controller with gain self-learning and self-tuning was designed.By SCL compiled into a control block,and based on single neuron adaptive PSD accomplished CFC-compound PID cascade control configuration.By using SMPT-1000 experimental platform verification,the results showed that the flash tank pressure in SCL compiled under the action of single neuron PSD controller,the adaptive control,adjust speed is superior to the traditional PID algorithm,the whole process to speed up 1.3%,the system output and recycling material are increased.On the basis of combining the process flow,the WinCC man-machine interface was developed to monitor and debug the flash tank pressure PSD controller in real time,and at the same time,the concentration,output,recovery process flow and other important indicators of the system were visually displayed.