基于改进神经动态规划算法的CFB锅炉床温控制器

A New Improved Controller Based on Improved Neural Dynamic Programming Algorithm for CFB Boiler Bed Temperature

摘要:

循环流化床锅炉燃烧系统是一个复杂的、大时滞、强耦合、非线性系统,维持燃烧室内正常的床温,是硫化床锅炉稳定运行的关键；针对传统的CFB锅炉床温控制器存在抗干扰能力弱、稳定时间长、控制精度不高等问题,提出了一种新型的基于改进神经动态规划(NDP)算法的锅炉床温控制器,该控制器的模型模块采用Elman神经网络；在某工控点下,将其用于CFB锅炉燃烧控制系统,并与传统的PID控制器进行比较；仿真结果表明:基于改进NDP算法的控制器,可以更好地拟合CFB锅炉燃烧系统的实际数学模型,具有动态性能好、收敛速度快、控制精度高等优点。

Abstract:

Circulating fluidized bed boiler combustion system is a complex, large time delay, strong coupling and non-linearity system. Keeping the regular bed temperature is the key point of making the circulating fluidized bed boiler work stably. Aimed at the weak anti-interference ability, long stable time and low control precision in the traditional CFB boiler bed temperature controller, an new CFB boiler bed temperature controller for improved Neural Dynamic Programming(NDP) methods which the model network is Elman neural network was designed. Finally, the simulation was finished on the circulating fluidized bed boiler combustion system. Comparing with the traditional PID controller, the results show that this new controller not only can better fitting practical mathematical model of CFB boiler combustion system, but also has many advantages such as good dynamics, fast convergence rate, high controlled resolution.