基于QP改进模型的离心泵性能预测方法  

Performance Prediction Method of Centrifugal Pump Based on Improved QP Model

作　　者：吴跃忠 张婷 费铭昊 吴登昊 任芸[3] WU Yuezhong;ZHANG Ting;FEI Minghao;WU Denghao;REN Yun(College of Metrology&Measurement Engineering,China Jiliang University,Hangzhou 310018;Zhejiang Leo Group Co.,Ltd.,Taizhou 317511;Zhijiang College,Zhejiang University of Technology,Shaoxing 312030)

机构地区：[1]中国计量大学计量测试工程学院,杭州310018 [2]利欧集团股份有限公司,台州317511 [3]浙江工业大学之江学院,绍兴312030

出　　处：《机械工程学报》2023年第13期280-289,共10页Journal of Mechanical Engineering

基　　金：浙江省自然科学基金(LGG21E090002,LY21E060004);中国博士后科学基金(2021M691383)资助项目。

摘　　要：离心泵广泛的应用于工业生产的各个领域,其用电量占工业电动机用电总量的22%。为了实现离心泵机组能耗的有效监测,以及解决传感器测量技术存在的成本高、故障率高等问题,需要发展一种新的无传感在线监测技术。在标准流量Q-轴功率P(QP)预测模型的基础上,以一台异步电机驱动的多级离心泵为研究对象,通过水力性能测试获取了不同工况下的性能数据。借助离心泵相似定理,修正当前转速下的离心泵特性曲线,建立流量与轴功率和转速以及扬程与流量和转速的数学预测模型。基于不同转速下预测模型的流量和扬程误差分析结果,提出一种基于转速差加权算法的QP改进模型。开展4种不同转速差权重函数的流量和扬程预测模型的误差分析,结果表明:带有权重函数Ⅲ的QP改进模型平均预测误差最小,其流量平均误差为3.66%和扬程平均误差为1.51%,与标准QP模型相比,其预测精度得到大幅提升。该QP改进模型进一步提升了离心泵运行状态的预测精度,为离心泵性能无传感在线监测技术的发展提供了一定理论参考。Centrifugal pumps are widely used in various fields of industrial production,and they account for 22% of the industrial electricity consumption.A new sensorless online monitoring technology is proposed to achieve effective monitoring of energy consumption of centrifugal pump units by replacing the sensor monitoring methods due to high cost and high failure rate.On the basis of a standard flow rate Q and shaft power P(QP) prediction model,a multistage centrifugal pump driven by an asynchronous motor is tested in lab to obtain performance data under different operating conditions.The centrifugal pump characteristic curves at the different speeds are modified to regular speeds by using the affinity law.A predictive mathematical model of flow rate versus power and speed,and head versus flow rate and speed is developed.An improved QP model with speed difference weighting algorithm is proposed based on the results of flow and head error analysis at different speeds,and carried out error analysis on the flow rate and head estimation models with four typical weighting functions.Results showed that the improved QP model with weight function Ⅲ has the smallest average prediction errors,its flow rate relative error is 3.66% and head relative error is 1.51%.The prediction accuracy is significantly improved compared with the standard QP model.The improved QP estimation model not only improves the prediction accuracy of centrifugal pump operation status,but also provides a theoretical basis for the development of senseless online monitoring technology of centrifugal pump.

关 键 词：离心泵 性能预测 无传感测量 QP模型 

分 类 号：TG156[金属学及工艺—热处理]