Pumping well multi-well working fluid level depth prediction method based on dynamic and static information feature fusion neural network

Abstract

The invention discloses a pumping well multi-well working fluid level depth prediction method based on a dynamic and static information feature fusion neural network, and belongs to the field of soft measurement, the method comprising the steps of collecting historical data of a plurality of sucker-rod pumping wells on site, and obtaining factors highly related to the depth of the underground working fluid level; constructing a prediction model structure, and dividing oil well historical operation data into a training set, a verification set and a test set in proportion; and taking Huber loss as a loss function of the neural network, and optimizing parameters of the dynamic and static information feature fusion neural network by adopting a gradient descent method to obtain an optimal value. According to the invention, the multi-well working fluid level depth prediction of the sucker-rod pumping well in different underground environments is realized, the prediction precision is high, and the stability is high.

Description

technical field [0001] The invention belongs to the field of soft measurement, in particular to a method for predicting the dynamic liquid level of multiple pumping wells. Background technique [0002] According to statistics, there are about 920,000 oil wells in the world, of which more than 90% are artificial lift mechanical oil production wells, and rod pumping systems are generally used. Moreover, the power consumption of rod pumped oil wells accounts for more than 30% of the total power consumption of the oil field, which is one of the main factors affecting the cost of oil production, so the potential for energy saving is huge. Accurate prediction of dynamic fluid level and fluid production is the key to self-tuning of pumping wells. [0003] The current method of manually measuring the dynamic liquid level has the problems of complex testing equipment, high cost, single test function, heavy workload and high cost of on-site installation, use and maintenance, inabilit...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the dynamic measurement of multi-well dynamic fluid levels of rod pump oil wells in the prior art, and the problem of abnormal data in the data set during the modeling process, the present invention proposes a multi-well dynamic fluid measurement method based on dynamic and static information feature fusion neural network. surface prediction method

Images