基于多模型BP神经网络算法的湿化器的湿度预测控制  

Humidity Predictive Control of Humidifier Based on Multi-model BP Neural Network Algorithm

作　　者：苏健 SU Jian(College of Information Engineering,Shenyang University of Chemical Technology,Shenyang 110000)

机构地区：[1]沈阳化工大学信息工程学院,沈阳110000

出　　处：《计算机与数字工程》2024年第5期1587-1591,共5页Computer & Digital Engineering

摘　　要：高流量呼吸湿化器的湿度控制效果,是衡量湿化器品质的重要标准,由于湿化罐出气口没有使用湿度传感器实时采集数据,无法对湿化器出气口的空氧混合气体的湿度做实时监测,所以不能构建一个完整的闭环负反馈的系统,只能构建一个基于预估数据进行控制的开环系统。对于湿化器的湿度预测控制的问题,提出基于多模型切换的BP神经网络算法预测控制,在目标温度平衡点进行线性化,根据流量的变化选择模型。仪器测试实验表明:湿化器的湿度预测控制系统有更高的控制品质,适用于不同温度,不同流量的湿化模式。The humidity control effect of high-flow respiratory humidifier is an important standard to measure the quality of hu-midifier.Since humidity sensor is not used to collect data in real time at the air outlet of humidifier,the humidity of air-oxygen mix-ture at the air outlet of humidifier cannot be monitored in real time,so a complete closed-loop negative feedback system cannot be built.Only an open-loop system can be constructed with control based on estimated data.For humidity predictive control of humidifi-er,a BP neural network algorithm based on multi-model switching is proposed,which is linearized at the target temperature equilib-rium point,and the model is selected according to the change of flow rate.The experimental results show that the humidity predic-tive control system of humidifier has higher control quality and is suitable for different temperature and flow rate.

关 键 词：湿化器 湿度预测 多模型 BP神经网络算法 线性化 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]