摘要
针对传统PID控制在单容水箱液位控制系统中存在抗扰动能力弱、参数固化以及非线性适应性差等问题,提出一种融合模糊PID控制、粒子群算法与BP神经网络的液位控制系统优化方法。首先以模糊PID控制为底层控制逻辑,通过粒子群算法对模糊PID的初始参数进行调节,同时结合BP神经网络进行非线性扰动的动态补偿;其次设计仿真实验对优化系统的性能进行验证。实验结果表明,所提方法的液位控制曲线更加接近目标输出曲线,阀门开度陡增20%的扰动测试中,液位波动为0.03 m,稳态恢复时间为0.9 s;阀门开度突降30%的扰动测试中,液位波动为0.04 m,稳态恢复时间为1.0 s,整体性能更加稳定,液位控制精确。
In view of the problems of weak anti-disturbance ability,parameter solidification and poor nonlinear adaptability of traditional PID control in the single-capacity water tank level control system,an optimization method of liquid level control system integrating fuzzy PID control,particle swarm optimization algorithm and BP neural network is proposed,which aims to improve the dynamic performance and robustness of the liquid level control system.Taking fuzzy PID control as the underlying control logic,the imbalance problem of overshoot and steady-state error in the traditional PID control method is solved,the initial parameters of fuzzy PID are adjusted by particle swarm optimization,and the dynamic compensation of nonlinear disturbance is carried out by combining with BP neural network.In order to achieve better control performance of liquid level control system,a simulation experiment is designed to verify the optimal design performance of the PID-based liquid level control system.The experimental results show that the liquid level control curve of the proposed method is closer to the target output curve,the liquid level fluctuation is 0.03 m and the steady-state recovery time is 0.9 s in the disturbance test of the sudden increase of valve opening by 20%,and the liquid level fluctuation is 0.04 m and the steady-state recovery time is 1.0 s in the disturbance test of the sudden drop of valve opening by 30%,so the overall performance is more stable and the liquid level control is accurate.
作者
孙永芳
Sun Yongfang(Department of Intelligent Manufacturing,Shaanxi Institute of Technology,Shaanxi Xi'an,710300,China)
出处
《机械设计与制造工程》
2025年第8期29-33,共5页
Machine Design and Manufacturing Engineering
关键词
模糊PID控制
水箱液位控制
粒子群算法
BP神经网络
动态控制
fuzzy PID control
water tank level control
particle swarm optimization algorithm
BP neural network
dynamic control
