基于混合算法优化BP-PID控制的负压吸引器研究

Research on Negative Pressure Aspirator Based on Hybrid Algorithm to Optimize BP-PID Control

摘要: 负压伤口治疗技术(Negative Pressure Wound Therapy, NPWT)是一种创面治疗新技术，负压吸引器应用该技术广泛应用于临床伤口愈合治疗中。真空泵作为负压源是主要的工作部件，针对其存在外部干扰、传感器延迟以及建模不精确等引起的负压强度不稳定的现象，提出一种应用于真空泵的连续负压控制优化方法。首先，提取真空泵及导管等的关键参数，建立负压吸引器创面治疗过程中的数学模型；其次，结合BP神经网络和PID控制，通过识别并优化真空泵伺服动态控制参数，并反馈到控制器输入端；最后，引入灰狼算法(GWO)与粒子群算法(PSO)的混合算法优化BP神经网络，进一步提高神经网络的精度。通过仿真，与传统PID、BP-PID及PSO-BP-PID相比，所提出的混合算法分别提前了1.0s 、0.5 s和0.4 s，稳态误差分别减少了98%、90%和80%，且几乎无振荡和超调，并通过实验验证了应用于负压吸引器上有良好的控制效果。

Abstract: Negative Pressure Wound Therapy (NPWT) is a new wound healing technology, and the negative pressure suction device is widely used in clinical wound healing treatment. The vacuum pump is the main working component as the negative pressure source. Aiming at the unstable negative pressure intensity caused by external interference, sensor delay and inaccurate modeling, a con-tinuous negative pressure control optimization method applied to the vacuum pump is proposed. Firstly, the key parameters of vacuum pump and catheter were extracted, and the mathematical model of vacuum suction device in wound treatment was established; secondly, combining BP neu-ral network and PID control, the dynamic control parameters of vacuum pump servo are identified and optimized, and fed back to the controller input; finally, a hybrid algorithm of Grey Wolf Optimi-zation (GWO) and Particle Swarm Optimization (PSO) is introduced to optimize the BP neural net-work to further improve the accuracy of the neural network. Through simulation, compared with the traditional PID, BP-PID and PSO-BP-PID, the proposed hybrid algorithm is advanced by 1.0s, 0.5s and 0.4s respectively, the steady-state error is reduced by 98%, 90% and 80% respectively, and there is almost no oscillation and overshoot. The experimental results show that the proposed hybrid algorithm has a good control effect on the negative pressure suction.