Rod-pumped well pump efficiency prediction system and method based on time series data

A technology for pumping wells and forecasting systems, applied in forecasting, nuclear methods, neural learning methods, etc., can solve problems such as low accuracy, difficulty in data collection, and inability to return data in time, so as to achieve accurate and reliable prediction results and avoid The effect of reducing production

Pending Publication Date: 2020-11-24
CHINA UNIV OF PETROLEUM (BEIJING) +1
View PDF1 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] In the process of oil well production, the main problems of oil well pump efficiency prediction include: 1. The pump efficiency of oil wells is constantly changing during the life cycle of oil wells. The pump efficiency model of the oil well is constantly changing; 2. It is difficult to collect data. Since the oil fields are mostly located in the Gobi, desert and other fields, limited by communication conditions, many collected data cannot be sent back to the oil field command center in time, resulting in accurate predictions not high

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Rod-pumped well pump efficiency prediction system and method based on time series data
  • Rod-pumped well pump efficiency prediction system and method based on time series data
  • Rod-pumped well pump efficiency prediction system and method based on time series data

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0024] This embodiment discloses a pump efficiency prediction system for pumping wells based on time series data. Pumping wells generally use pump equipment to recover crude oil in the formation to the surface or near-surface gathering and transportation systems, such as oil storage tanks or gathering and transportation pipelines. The prediction system includes an input module 100, a elimination module 200 and a prediction module 300.

[0025]The input module 100 is mainly used to obtain the second working parameter and its corresponding existing pump efficiency, to obtain the first working parameter associated with the pump efficiency of the oil well, and a verification set composed of historical production data. The input module 100 is data-connected with the prediction module 300 through the elimination module 200, and produces a long-term and short-term neural network prediction model based on the second working parameter training data. The input module 100 and the predic...

Embodiment 2

[0042] This embodiment may be a further improvement and / or supplement to Embodiment 1, and repeated content will not be repeated here. In the case of no conflict or contradiction, the whole and / or part of the content of the preferred implementations of other embodiments may serve as supplements to this embodiment.

[0043] There are many factors affecting the pumping efficiency of pumping wells, including: static data of oil wells (reservoir petrophysical properties, wellbore trajectory, etc.); historical dynamic daily data (time, oil pressure, casing pressure, pump depth, theoretical pump displacement, water cut, etc.) well fluid viscosity, dynamic liquid level, etc.); equipment operating condition data (time, current, voltage, power, power factor, instantaneous power consumption, service time, pump efficiency, system efficiency), etc. Accurate selection of the main characteristic parameters affecting oil well pump efficiency is of great significance for analyzing the main co...

Embodiment 3

[0062] This embodiment may be a further improvement and / or supplement to Embodiment 1, 2 or their combination, and repeated content will not be repeated. This embodiment discloses that the whole and / or part of the preferred implementation manners of other embodiments may be used as a supplement to this embodiment under the condition that no conflict or contradiction is caused.

[0063] This embodiment discloses a test method for a long-term and short-term neural network model, which is used for the prediction module 300 to test its trained prediction model. Evaluate the generalization ability of the prediction model on the test set, that is, the pump efficiency prediction effect of the model. The prediction effect evaluation indicators adopted mainly include: determination coefficient R 2 ; mean absolute deviation MAD; mean relative error MAPE; mean square error RMSE; Hill inequality coefficient TIC.

[0064] (1) Determination coefficient R 2 : To characterize the extent to...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a rod-pumped well pump efficiency prediction system and method based on time series data, a rod-pumped well adopts an oil pump to extract and send crude oil in a stratum to agathering and transportation system. The prediction system comprises an input module for obtaining a first working parameter associated with the oil well pump efficiency; and a prediction module usedfor predicting the subsequent pump efficiency according to the first working parameter. Before that the prediction module predicts the subsequent pump efficiency, the input module receives a second working parameter in the working process of the rod-pumped well and the corresponding pump efficiency of the existing oil well to form a training set, an elimination module in data connection with the input module acquires a principal component parameter associated with the oil well pump efficiency intensity from the second working parameter; an elimination module inputs principal component parameters and the existing oil well pump efficiency into the prediction module in a time sequence mode according to preset time steps, and the prediction module generates a long-short-term neural network prediction model based on the principal component parameters and corresponding existing oil well pump efficiency training to predict the subsequent pump efficiency of the long-short-term neural network prediction model.

Description

technical field [0001] The invention relates to the technical field of petroleum engineering intelligent mining, in particular to a system and method for pumping efficiency prediction of pumping wells based on time series data. Background technique [0002] Pump efficiency refers to the ratio of the actual fluid production of the pumping well to the theoretical displacement of the pump. The pump efficiency of pumping wells is the basis for understanding oil reservoirs, improving oil well work systems, and formulating scientific and reasonable development adjustment plans. Oil well pump efficiency can be used to reflect the current production capacity of the oil well, the dynamic change of the production capacity of the oil well, the working condition of the oil well pumping equipment and the effect of the reaction measures. Improving the pumping efficiency of pumping wells is one of the effective measures to slow down the natural decline rate in oilfields. The factors affec...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06N3/04G06N3/08G06N20/10G06Q10/04G06Q50/02E21B47/008
CPCG06N3/049G06N3/08G06Q10/04G06Q50/02G06N20/10
Inventor 檀朝东曹晟李玉泽魏方方宋健李小民檀晨
Owner CHINA UNIV OF PETROLEUM (BEIJING)
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products