Automatic Calibration System Of Pipe Prover Based On Image Recognition

In the transportation process of the petroleum in the industry,the pipe prover is a common flowmeter verification device.The accuracy of the pipe prover cycle calibration directly affects the metering performance of the flowmeter,relating to the economic benefits and reputation of the oilfield.Aiming at the problems of the current pipe prover verification system,such as low level of automation,low accuracy and inefficiency and high labor intensity of the verification operator,this paper studies and designs an automatic pipe prover verification system based on image recognition.The main work and achievements are as follows:(1)According to the verification principle of pipe prover,the requirements of system function and technical parameter are determined,at the same time,the overall design of system is completed.Then the software and hardware of the pipe prover verification system are completed,which are divided into three parts: PLC measurement and control system,metering neck liquid level control and measurement system and upper computer control system.(2)A metering neck liquid level control system adopting the segmentation control idea is designed.The liquid level position is estimated by the cumulative flow of the flowmeter's detection inlet.Then stepped control the inlet flow rate of verification system,so as to control the metering liquid level in the best visible range of the sight tube;Besides,a kind of cameraliquid level following control system is designed,which can judge the liquid level position by acquiring liquid level image information in real time.At the same time,controlling the stepper motor to move the camera to a level relative to the liquid level,which can realize automatic tracking of the liquid level and improve the speed and accuracy of liquid level image acquisition.(3)A metering neck liquid level identification method based on image recognition is proposed,which realizes the automatic identification of the metering neck liquid level and lays a foundation for the system's fully automated verification.The liquid level image segmentation algorithm is designed by using filtering algorithm,highlight suppression algorithm,edge detection technology and K-Means clustering algorithm.It solves the segmentation problem between sight tube image and scale image in liquid level image.According to the reading line and shape characteristics of the concave liquid surface,a concave liquid surface reconstruction and recognition algorithm is proposed.By adapting the steps of graying,fitting and reconstruction,the liquid level pixel coordinate values are obtained.Then based on the equalized brightness algorithm and edge feature search algorithm,SVM algorithm and prior knowledge,this paper designs a scale parameter recognition algorithm,which realizes the identification of scale numbers and graduation line.Finally,the liquid level measurement value is determined by combining the relationship between the scale value and the concave liquid surface.(4)The pipe prover verification system is tested for liquid level control,liquid level identification and system function.The results show that the liquid level control system is accurate,the liquid level recognition and vernier caliper reading error is less than ±0.5mm,the system function is perfect and it meets the design requirements,which improves the verification accuracy and automation level.