Font Size: a A A

Research On Error Compensation Method Of Ultrasonic Flowmeter Based On Convolutional Neural Network

Posted on:2022-08-10Degree:MasterType:Thesis
Country:ChinaCandidate:X R KangFull Text:PDF
GTID:2492306602955499Subject:Control Engineering
Abstract/Summary:PDF Full Text Request
Ultrasonic flow meters have been widely used in petrochemical,aerospace,construction,medical and other fields due to their advantages of no pressure loss,low cost,easy installation,wide range coverage,and high accuracy.With the continuous expansion of the scope of application,the accuracy requirements of various applications for ultrasonic flowmeters are also increasing.In this paper,the compensation method for the measurement error of the ultrasonic flowmeter installed behind the gate valve is studied.The research content is as follows:(1)Analyze the measurement error of ultrasonic flowmeter caused by factors such as temperature,flow fluctuation and unstable flow field from the principle,improve the asymmetry coefficient to quantitatively evaluate the flow field state,and in-depth study the gate valve opening,the gate valve rotation angle,and the gate valve Measurement error caused by unstable flow field caused by distance.(2)Aiming at the unstable flow field caused by the gate valve in the pipeline,the influence of factors such as the opening of the gate valve,the distance between the ultrasonic flowmeter and the gate valve and the installation angle of the gate valve on the measurement error of the ultrasonic flowmeter was studied through CFD simulation experiments,and the ultrasonic flowmeter test was designed and processed.For the sample,a real-flow experiment platform was built,and the CFD simulation results were verified through the real-flow experiment.(3)Based on the analysis result of the measurement error,an error compensation model of the ultrasonic flowmeter based on the convolutional neural network is established.The gate valve opening,the installation angle of the transducer and the installation position are taken as input,and the error is taken as the output for model calculation.training.After the model training is completed,the prediction error is compensated back to the original measurement value,and the flow measurement error of the ultrasonic flowmeter after the gate valve is better compensated.
Keywords/Search Tags:flow measurement, multi-channel ultrasonic flowmeter, error compensation, convolutional neural network
PDF Full Text Request
Related items