The Research Of Gas Ultrasonic Flowmeter Installation Effect Compensation Based On Neural Network

With the rising share of natural gas in the world's consumption of energy,the natural gas metering instrumentation market is also growing.Compared with the traditional measuring instruments,ultrasonic flowmeter with wide-range ratio,high reliability,high accuracy,no moving parts,etc.,and gradually gain a place in the market.China has abundant natural gas reserves.Since the project of "West-East Gas Transmission" has been carried out,the natural gas market has been developing rapidly.As China's gas ultrasonic flowmeter research started late,the domestic gas ultrasonic flowmeter market has long been monopolized by foreign-funded enterprises.Therefore,the gas ultrasonic flowmeter research and development and its important strategic significance.Gas ultrasonic flowmeter measurement accuracy of many,the main content of this paper is the installation effect of flow field analysis and compensation.In this paper,the BP neural network is introduced in the compensation of gas ultrasonic flowmeter installation effect.By using the strong nonlinear fitting ability of BP neural network,the relationship between the channel and the average flow velocity of the pipeline is obtained through training.The compensation effect of BP neural network on the installation effect is verified by numerical simulation and a simple experimental verification is carried out.The main research contents of this topic are as follows:1.In view of the installation effect of the transducer on the concave and convex mounting and the position shift,a three-dimensional model is established and the flow field is analyzed by means of numerical simulation.The effect of two installation effects on the measurement accuracy is visualized.The data obtained by numerical simulation are processed by the traditional weight coefficient method,and the compensation ability for the installation effect is studied.2.According to the actual problem of gas ultrasonic flowmeter,select the appropriate hidden layer,activation function and cost function and other parameters.For the research group developed three-channel gas ultrasonic flowmeter to build a BP neural network,BP neural network code written in Python.3.Verification of Gas Ultrasonic Flowmeter Installation Effect Compensation Based on Neural Network.The numerical simulation results of different installation effects are divided into training group and testing group.The training group was used to train the neural network,the test group was used to evaluate the fitting of the neural network,and the compensation method of the traditional weight coefficient was compared.Using developed gas ultrasonic flowmeter for a simple experimentalverification.