A Learning and Estimating Method of Indicator Diagram of Pumping Unit

Abstract

The invention discloses an oil pumping unit indicator diagram learning estimating method comprising the following steps: S1, pre-processing original data records, so the data records can be complete and usable; S2, space-regularizing pre-processed data records, so each data record can be mapped to an uniform displacement sequence, wherein the displacement refers to the oil pumping unit displacement; S3, using regularized data records to compute the characteristic coefficient alpha (s), wherein the characteristic coefficient represents an influence factor on the motor power changes by oil pumping unit load changes, s represents oil pumping unit polished rod displacement; S4, using the learned characteristic coefficient alpha (s) to estimate to-be-estimated oil pumping unit indicator diagram data. The oil pumping unit indicator diagram learning estimating method is less in learning parameters, small in processing space-time cost, and high in robustness and accuracy.

Description

technical field [0001] The invention relates to the field of data mining, in particular to a method for learning and estimating an indicator diagram of a pumping unit. Background technique [0002] During the work and management of the oil well, the staff need to keep abreast of the dynamic parameters of the pumping unit, and analyze various information such as the working condition of the pumping unit and the state of downhole oil. These measures are crucial to confirm whether the oil well is working normally and to troubleshoot and resolve abnormal conditions, and are necessary means to ensure oil pumping efficiency and maintain oil well safety. [0003] The dynamometer diagram of the pumping unit is the relationship curve between the load and displacement of the polished rod of the pumping unit. It has become an important and main method of oil well condition diagnosis. Therefore, in actual work, we need to obtain the dynamometer diagram of the pumping unit in real time...

AI Technical Summary

Problems solved by technology

[0004] However, it is not an effective method to directly use the load cell to measure the load. This technology has the following defects: On the one hand, since the load cell has been in the process of stretching under alternating loads, it needs to withstand a huge tensile force for a long time , its elastic coefficient is easy to change, which affects the measurement accuracy of the sensor and accelerates the failure, resulting in a short service life of the sensor and making it difficult to achieve long-term effective monitoring; on the other hand, the sensor needs to be installed on the frame and polished rod of the pumping unit It is connected to the main control instrument through a long-distance lead wire, which is complicated and inconvenient to disassemble and install. The maintenance workload is heavy, the cost is high, and it is prone to man-made damage and theft, resulting in a decrease in system stability and reliability.

[0007] Second, some steps are computationally complex

On the one hand, it leads to a large space-time overhead, and on the other hand, it is also prone to singularities

[0008] Third, the accuracy of the results needs to be improved

The dynamometer diagram obtained by indirect calculation does not coincide well with the actual measured value, which limits the prospect and significance of its practical application

Images