A beam pumping unit well condition diagnosis method based on evaluation function

Abstract

The invention discloses a method for diagnosing well conditions of a beam pumping unit based on an evaluation function. The method uses the accuracy of five aspects, namely, lower left angle difference, center distance difference, relative distance, module length distance, and center angle difference, to individually judge the well conditions to calculate a weight coefficient group of the evaluation function. The actual power diagram is analyzed by the weight coefficient group to match the power diagram in a sample library. The calculation amount is small and the time consumption is short. In addition, the method uses multiple weight coefficients to analyze the actual power diagram, and has high analysis accuracy, accurate data, and high feasibility.

Description

Technical Field

[0001] The invention relates to the field of oil production engineering, and in particular to a beam pumping unit well condition diagnosis method based on an evaluation function. Background Art

[0002] The beam pumping unit has always been the first choice of oil well equipment due to its simple structure and high reliability. As the main downhole equipment in the oil production system, the oil pump may be affected by sand, water, and gas due to the complex environment around the well, and it may fail to work normally. The condition of the oil pump will directly affect the output of crude oil in the oil well. Therefore, it is particularly important to diagnose the downhole working conditions. The power diagram analysis is currently the best way to diagnose the working conditions of the pumping well. In actual work, in order to accurately analyze the power diagram, it is often necessary to compare the actual power diagram with the sample power diagram. The current power diagram recognition mainly relies on artificial neural networks. However, although neural networks have many advantages, they also have some defects that are difficult to overcome. First, in order to achieve appropriate judgment accuracy, the training of neural networks requires a lot of computing time, which is large in calculation and takes too long to train. Second, the performance of neural networks is too dependent on the choice of their structure and the choice of initial points. Neural networks with different structures show great differences in performance for the same data. To avoid the above problems, the present invention proposes a beam pumping unit well condition diagnosis method based on an evaluation function. The method analyzes the actual power diagram by designing a weight coefficient to match the power diagram in the sample library. Compared with the neural network, the method has a small amount of calculation and is time-saving. In addition, the method designs multiple weight coefficients, and the analysis of the actual power diagram is accurate and has high feasibility. Summary of the invention

[0003] To solve the above problems, the present invention proposes a beam pumping unit well condition diagnosis method based on an evaluation function. The method analyzes the actual power diagram by designing a weight coefficient to match the power diagram in the sample library. Compared with the neural network, the method has a small amount of calculation and a short time consumption. In addition, the method designs multiple weight coefficients, and the analysis of the actual power diagram is accurate and has high feasibility.

[0004] A beam pumping unit well condition diagnosis method based on an evaluation function comprises the following steps:

[0005] Step S1, assuming that there are n sample power diagrams in the sample library, the actual power diagram and the sample power diagram are normalized;

[0006] Step S2, calculate the lower left angle difference of the i-th sample work diagram , Center distance difference , relative distance , module length , Center Angle Difference ;

[0007] Step S3, after simulation, it can be obtained by The accuracy of judging the well condition alone is: , , , , ,use Accuracy calculation weight coefficient when judging well condition alone , five weight coefficients Composition weight coefficient group ;

[0008] Step S4, based on the i-th sample work graph Composition of evaluation index group , in the sample library n Evaluation index group Constructing the sample evaluation index matrix , n is a natural number greater than 1, T is the transpose symbol, for The transpose of

[0009] Step S5: construct an evaluation matrix , its evaluation matrix B is a 1xn matrix, n is a natural number greater than 1;

[0010] Step S6, select the smallest value in the evaluation matrix B , at this time, the well condition of the i-th sample in the sample library is considered to be the actual well condition. i 1~ n A number in n is a natural number greater than 1.

[0011] Furthermore, in step S1, the normalization method is:

[0012] ,

[0013] in, y is any number in a set of data, is the minimum value of a set of data. The maximum value of a set of data.

[0014] Further, in step S2, the lower left angle difference of the i-th sample work diagram The calculation method is as follows:

[0015] Calculate the angle between vector AB and vector AD of the actual work diagram , calculate the vector of the i-th work graph in the sample library With vector Angle , the lower left angle difference The calculation formula is as follows:

[0016] ,

[0017] in, is the lower left angle difference, is the angle between vector AB and vector AD in the actual work diagram, is the vector of the i-th work graph in the sample library With vector The angle of is the absolute value symbol, i 1~ n A number in n is a natural number greater than 1.

[0018] Further, in step S2, the center distance difference of the i-th sample work map The calculation method is as follows:

[0019] According to the center point O of the actual power diagram and the center point of the i-th power diagram in the sample library Position, calculate the modulus length of vector OA, vector OB, vector OC and vector OD respectively , , and , take the average value of the sum of the modulus lengths of vector OA, vector OB, vector OC and vector OD as the center distance of the actual work diagram, denoted by L; similarly, calculate the vectors ,vector ,vector and vector Length of mold , , and , take the vector ,vector ,vector and vector The average value of the sum of the modulus lengths is the center distance of the sample power diagram, recorded as , center distance difference The calculation formula is as follows:

[0020] ,

[0021] ,

[0022] ,

[0023] in, is the modulus of vector OA, is the modulus of vector OB, is the modulus of the vector OC, is the modulus of the vector OD, For vector The module length, For vector The module length, For vector The module length, For vector The module length, is the center distance difference, L is the center distance of the actual power diagram, is the center distance of the sample power diagram, is the absolute value symbol, i 1~ n A number in n is a natural number greater than 1.

[0024] Further, in step S2, the relative distance of the i-th sample work map The calculation method is as follows:

[0025] According to the distance between the actual work map and the i-th work map of the sample, that is, the vector Length of mold ,vector Length of mold ,vector Length of mold and vector Length of mold , take the vector ,vector ,vector and vector The average value of the sum of the modulus lengths is taken as the relative distance between the actual power diagram and the power diagram in the sample library. , and its calculation formula is as follows:

[0026] ,

[0027] in, is the relative distance, For vector The module length, For vector The mold length, For vector The module length, For vector The module length, i 1~ n A number in n is a natural number greater than 1.

[0028] Further, in step S2, the modulus distance of the i-th sample work diagram is The calculation formula is as follows:

[0029] ,

[0030] in, is the module length, is the modulus length of the actual work diagram AB vector, is the i-th work graph in the sample library The magnitude of the vector, is the modulus length of the actual work diagram AD vector, is the i-th work graph in the sample library The magnitude of the vector, is the modulus length of the actual work diagram BC vector, is the i-th work graph in the sample library The magnitude of the vector, is the modulus length of the actual work diagram CD vector, is the i-th work graph in the sample library The magnitude of the vector, is the absolute value symbol, i 1~ n A number in n is a natural number greater than 1.

[0031] Further, in step S2, the central angle difference of the i-th sample power diagram The calculation formula is as follows:

[0032] ,

[0033] in, is the center angle difference, is the angle between the AC vector and the BD vector of the actual power diagram, is the i-th work graph in the sample library Vector and The angle between the vectors, is the absolute value symbol, i 1~ n A number in n is a natural number greater than 1.

[0034] Furthermore, in step S3, the weight coefficient The calculation formula is as follows:

[0035] ,

[0036] in, is the weight coefficient, for The accuracy of judging well conditions alone, for The accuracy of judging well conditions alone, for The accuracy of judging well conditions alone, for The accuracy of judging well conditions alone, for The accuracy when judging the well condition alone.

[0037] The technical effects of the present invention are:

[0038] 1. Avoid high-computation judgment methods such as neural networks, which have low actual computational complexity and short time consumption;

[0039] 2. The present invention designs multiple weight coefficients, and the analysis of the actual power diagram is accurate and highly feasible. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A flow chart of the steps of a beam pumping unit well condition diagnosis method based on an evaluation function provided by the present invention;

[0041] Figure 2 A power diagram situation in the actual power diagram provided by the present invention;

[0042] Figure 3 This is a power diagram situation of the sample power diagram provided by the present invention. DETAILED DESCRIPTION

[0043] The following is only a preferred embodiment of the present invention. The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments, so that those skilled in the art can understand the present invention. It should be pointed out that for those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention defined and determined by the attached claims without departing from the principle of the present invention, all inventions and creations using the concept of the present invention are protected.

[0044] Figure 1 A flow chart of the steps of a beam pumping unit well condition diagnosis method based on an evaluation function provided by the present invention is as follows: Figure 1 As shown, a beam pumping unit well condition diagnosis method based on an evaluation function comprises the following steps:

[0045] Step S1, assuming that there is a sample power diagram in the sample library, the actual power diagram and the sample power diagram are normalized; the normalization method is:

[0046] ,

[0047] in, y is any number in a set of data, is the minimum value of a set of data. The maximum value of a set of data.

[0048] Step S2, calculate the lower left angle difference of the i-th sample work diagram , Center distance difference , relative distance , module length , Center Angle Difference .

[0049] Figure 2 A power diagram situation of the actual power diagram provided by the present invention, such as Figure 2 As shown, the actual power diagram is ABCD The parallelogram formed by the two sides has O as its center point; Figure 3 A power diagram situation of the sample power diagram provided by the present invention, such as Figure 3 As shown, the sample power diagram is based on They form an approximate parallelogram. as its center point.

[0050] Among them, the lower left angle difference of the i-th sample work diagram is The calculation method is as follows:

[0051] Calculate the angle between vector AB and vector AD of the actual work diagram , calculate the vector of the i-th work graph in the sample library With vector Angle , the lower left angle difference The calculation formula is as follows:

[0052] ,

[0053] in, is the lower left angle difference, is the angle between vector AB and vector AD in the actual work diagram, is the vector of the i-th work graph in the sample library With vector The angle of is the absolute value symbol, i 1~ n A number in n is a natural number greater than 1.

[0054] Among them, the center distance difference of the i-th sample work map is The calculation method is as follows:

[0055] According to the center point O of the actual power diagram and the center point of the i-th power diagram in the sample library Position, calculate the modulus length of vector OA, vector OB, vector OC and vector OD respectively , , and , take the average value of the sum of the modulus lengths of vector OA, vector OB, vector OC and vector OD as the center distance of the actual work diagram, denoted by L; similarly, calculate the vectors ,vector ,vector and vector Length of mold , , and , take the vector ,vector ,vector and vector The average value of the sum of the modulus lengths is the center distance of the sample power diagram, recorded as , center distance difference The calculation formula is as follows:

[0056] ,

[0057] ,

[0058] ,

[0059] in, is the modulus of vector OA, is the modulus of vector OB, is the modulus of the vector OC, is the modulus of the vector OD, For vector The module length, For vector The module length, For vector The module length, For vector The module length, is the center distance difference, L is the center distance of the actual power diagram, is the center distance of the sample power diagram, is the absolute value symbol, i 1~ n A number in n is a natural number greater than 1.

[0060] Among them, the relative distance of the i-th sample work map is The calculation method is as follows:

[0061] According to the distance between the actual work map and the i-th work map of the sample, that is, the vector Length of mold ,vector Length of mold ,vector Length of mold and vector Length of mold , take the vector ,vector ,vector and vector The average value of the sum of the modulus lengths is taken as the relative distance between the actual power diagram and the power diagram in the sample library. , and its calculation formula is as follows:

[0062] ,

[0063] in, is the relative distance, For vector The module length, For vector The mold length, For vector The module length, For vector The module length, i 1~ n A number in n is a natural number greater than 1.

[0064] Among them, the modulus distance of the i-th sample work diagram is The calculation formula is as follows: ,

[0065] in, is the module length, is the modulus length of the actual work diagram AB vector, is the i-th work graph in the sample library The magnitude of the vector, is the modulus length of the actual work diagram AD vector, is the i-th work graph in the sample library The magnitude of the vector, is the modulus length of the actual work diagram BC vector, is the i-th work graph in the sample library The magnitude of the vector, is the modulus length of the actual work diagram CD vector, is the i-th work graph in the sample library The magnitude of the vector, is the absolute value symbol, i 1~ n A number in n is a natural number greater than 1.

[0066] Among them, the central angle difference of the i-th sample work diagram is The calculation formula is as follows:

[0067] ,

[0068] in, is the center angle difference, is the angle between the AC vector and the BD vector of the actual power diagram, is the i-th work graph in the sample library Vector and The angle between the vectors, is the absolute value symbol, i 1~ n A number in n is a natural number greater than 1.

[0069] Step S3, after simulation, it can be obtained by The accuracy of judging the well condition alone is: , , , , ,use Accuracy calculation weight coefficient when judging well condition alone , five weight coefficients Composition weight coefficient group ;

[0070] Weight coefficient The calculation formula is as follows:

[0071] ,

[0072] in, is the weight coefficient, for The accuracy of judging well conditions alone, for The accuracy of judging well conditions alone, for The accuracy of judging well conditions alone, for The accuracy of judging well conditions alone, for The accuracy when judging the well condition alone.

[0073] Step S4, based on the i-th sample work graph Composition of evaluation index group , in the sample library n Evaluation index group Constructing the sample evaluation index matrix , n is a natural number greater than 1, T is the transpose symbol, for The transpose of .

[0074] Step S5: construct an evaluation matrix , its evaluation matrix B is a 1xn matrix, n is a natural number greater than 1.

[0075] Step S6, select the smallest value in the evaluation matrix B , at this time, the well condition of the i-th sample in the sample library is considered to be the actual well condition. i 1~ n A number in n is a natural number greater than 1.

[0076] The present invention proposes a beam pumping unit well condition diagnosis method based on an evaluation function. The method uses the accuracy of five data, namely, lower left angle difference X1, center distance difference X2, relative distance X3, module length distance X4, and center angle difference X5, to individually judge the well condition to calculate a weight coefficient group of the evaluation function. The actual power diagram is analyzed by the weight coefficient group to match the power diagram in the sample library. The calculation amount is small and the time consumption is short. In addition, the method uses multiple weight coefficients to analyze the actual power diagram, and the analysis accuracy is high and precise, the data is accurate, and the feasibility is high.

[0077] Although the specific implementation of the invention is described in detail in conjunction with the drawings, it should not be understood as limiting the scope of protection of this patent. Within the scope described in the claims, various modifications and variations that can be made by those skilled in the art without creative work still fall within the scope of protection of this patent.

Claims

1. A beam pumping unit well condition diagnosis method based on evaluation function, characterized in that: The following steps are involved: Step S1, assuming that there are n sample power diagrams in the sample library, the actual power diagram and the sample power diagram are normalized; Step S2, define is the lower left angle difference of the sample work diagram, is the center distance difference of the sample power diagram, is the relative distance of the sample work diagram, is the model length distance of the sample power diagram, is the central angle difference of the sample power diagram, which is calculated from the actual power diagram and the sample power diagram; Calculate the lower left angle difference of the i-th sample work graph , Center distance difference , relative distance , module length , Center Angle Difference ; The lower left angle difference of the i-th sample work diagram The calculation formula is as follows: ， in, is the lower left angle difference, θ is the angle between vector AB and vector AD in the actual power diagram, is the vector of the i-th work graph in the sample library With vector The angle of is the absolute value symbol, i 1~ n A number in n is a natural number greater than 1; The center distance difference of the i-th sample work graph The calculation formula is as follows: ， ， ， in, is the modulus of vector OA, is the modulus of vector OB, is the modulus of the vector OC, is the modulus of the vector OD, For vector The module length, For vector The module length, For vector The module length, For vector The module length, L is the center distance of the actual work diagram, is the center distance of the sample power diagram; The relative distance of the i-th sample work map The calculation formula is as follows: ， in, For vector The module length, For vector The mold length, For vector The module length, For vector The module length; The modulus distance of the i-th sample work diagram The calculation formula is as follows: ， in, is the modulus length of the actual work diagram AB vector, is the i-th work graph in the sample library The magnitude of the vector, is the modulus length of the actual work diagram AD vector, is the i-th work graph in the sample library The magnitude of the vector, is the modulus length of the actual work diagram BC vector, is the i-th work graph in the sample library The magnitude of the vector, is the modulus length of the actual work diagram CD vector, is the i-th work graph in the sample library The magnitude of the vector; The central angle difference of the i-th sample power diagram The calculation formula is as follows: ， Among them, α is the angle between the AC vector and the BD vector of the actual power diagram, is the i-th work graph in the sample library Vector and The angle between the vectors; Step S3, after simulation, it can be obtained by The accuracy of judging the well condition alone is: , , , , ,use Accuracy calculation weight coefficient when judging well condition alone , five weight coefficients Composition weight coefficient group ; Step S4, based on the i-th sample work graph Composition of evaluation index group , in the sample library n The evaluation index groups constitute the sample evaluation index matrix , n is a natural number greater than 1, T is the transpose symbol, for The transpose of Step S5: construct an evaluation matrix , its evaluation matrix B is a 1xn matrix, n is a natural number greater than 1; Step S6, select the smallest value in the evaluation matrix B , at this time, the well condition of the i-th sample in the sample library is considered to be the actual well condition. i 1~ n A number in n is a natural number greater than 1.

2. A beam pumping unit well condition diagnosis method based on evaluation function according to claim 1, characterized in that: In step S1, the normalization method is: ， in, y is any number in a set of data, is the minimum value of a set of data. The maximum value of a set of data.

3. The method for diagnosing well conditions of a beam pumping unit based on an evaluation function according to claim 1, characterized in that: In step S3, the weight coefficient The calculation formula is as follows: ， in, for The accuracy of judging well conditions alone, for The accuracy of judging well conditions alone, for The accuracy of judging well conditions alone, for The accuracy of judging well conditions alone, for The accuracy when judging the well condition alone.

Images