Dynamic liquid level prediction method of beam-pumping unit based on particle swarm extreme learning machine

Abstract

The invention provides a dynamic liquid level prediction method of a beam-pumping unit based on a particle swarm extreme learning machine. The dynamic liquid level prediction method comprises the steps that historical dynamic liquid level data are collected through an oil pumping well; the data are preprocessed; test data are input into an extreme learning machine model, and thus a dynamic liquidlevel prediction value of the extreme learning machine is obtained; a root-mean-square error serves as an adaptive value of a particle swarm algorithm; the parameter input layer weight w and the deviation b of the extreme learning machine are optimized through the particle swarm algorithm, and an optimal value is obtained in the global scope; a dynamic liquid level prediction model of the extremelearning machine optimized based on the particle swarm algorithm is obtained; and the processed to-be-tested data are input into the dynamic liquid level prediction model of the extreme learning machine optimized based on the particle swarm algorithm, and a prediction result of the to-be-tested data is obtained. Compared with the mode for measuring the dynamic liquid level through an instrument, the cost of obtaining the dynamic liquid level is lowered, and the risk of accidents caused by the fact that instruments are mounted at a wellhead is reduced. Through experiments, it is shown that thedynamic liquid level prediction method has a good prediction effect.

Description

technical field [0001] The invention belongs to the field of soft measurement, and in particular relates to a dynamic liquid level prediction method of a beam pumping unit based on a particle swarm extreme learning machine. Background technique [0002] The liquid level of the beam pumping motor is a key parameter in the oil extraction process, in order to adapt to the complex oil extraction environment. Efficient, safe, continuous and accurate measurement of dynamic liquid level can grasp the production status of the beam pumping unit and improve the efficiency of oil recovery. [0003] The existing method of measuring the liquid level of the beam pumping motor is the echo method, that is, an empty projectile is fired at the casing position of the wellhead, and the acoustic signal generated subsequently propagates along the pipe wall. When the signal touches the liquid surface, the pulse signal is It will be reflected, and the dynamic liquid level of the pumping well can b...

AI Technical Summary

Problems solved by technology

[0004] The content of the present invention is to provide a dynamic fluid level prediction method of beam pumping unit based on particle swarm extreme learning machine, through which the problems of high intensity of manual measurement and high production cost can be solved

Images