An electric submersible pump fault warning method and device based on current change analysis

Abstract

The present invention relates to the technical field of equipment fault diagnosis, and particularly relates to a method and device for electric submersible pump fault early warning based on current change analysis. The method constructs a current normalization formula, a time normalization formula, and a current eigenvalue calculation formula, and constructs a fault diagnosis model based on this. After training the fault diagnosis model with several typical current eigenvalues, several measured current eigenvalues are calculated according to the current operation data. Moreover, several measured current eigenvalues are compared with several typical current eigenvalues to determine several different current characteristic types, and the fault types of the electric submersible pump are determined by combining several different current characteristic types, so as to improve the efficiency of electric submersible pump fault troubleshooting, and improve the refined and intelligent management level of the electric submersible pump, and ensure the production efficiency of the oil well platform.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment fault diagnosis, and particularly to an electric submersible pump fault warning method and device based on current change analysis. Background Art

[0002] The electric submersible centrifugal pump, hereinafter simply referred to as the electric submersible pump, is widely used in oilfield production. Especially in deep wells with a large depth reduction of one kilometer, due to its characteristics of high lift and large flow rate, the electric submersible pump is the preferred equipment. Exactly because the working place of the electric submersible pump is generally in deep wells with a depth reduction of one kilometer, when a fault occurs in the electric submersible pump, it will bring relatively complex maintenance problems. One is that when a fault occurs in the electric submersible pump, it often takes a long time for the staff to rush to the location of the electric submersible pump, which will affect the normal working time of the electric submersible pump. The other is that the fault type can only be determined according to the actual situation after the staff arrives at the scene, which will delay the maintenance progress. Both situations will delay the production progress. Therefore, it is necessary to monitor the running state of the electric submersible pump in real time and be able to warn of the faults of the electric submersible pump, so as to be able to solve the fault problems in advance and ensure the production efficiency of the oil well platform.

[0003] The Chinese patent publication number in the prior art: CN111461551B, discloses an electric submersible pump fault warning method based on deep learning and SPC criteria. The method of this technical solution is to construct a CNN-LSTM model and perform standardized training, and then estimate the monitoring parameter values of the electric submersible pump to be evaluated. Based on the gap between the monitoring parameters and the normal state, calculate the health degree of the electric submersible pump, and determine whether the operation of the electric submersible pump is abnormal according to the health degree, so as to trigger the alarm mechanism, and reduce the process of manual detection, saving labor costs. However, this technical solution does not determine the fault type when the electric submersible pump is abnormal based on the obtained monitoring parameter values, and thus cannot enable the staff to start corresponding treatment measures in time, which will affect the production efficiency of the oil well platform. Summary of the Invention

[0004] Therefore, the present invention provides an electric submersible pump fault warning method and device based on current change analysis to solve the problem in the prior art that a current characteristic value calculation formula is not constructed according to the current change situation and the fault type existing in the electric submersible pump is not analyzed by comparing a number of measured current characteristic values and a number of typical current characteristic values obtained.

[0005] To achieve the above object, the present invention provides an electric submersible pump fault warning method based on current change analysis, including:

[0006] Construct a current standardization formula, a time standardization formula, and a current characteristic value calculation formula, and construct a fault diagnosis model based on the current standardization formula, the time standardization formula, and the current characteristic value calculation formula;

[0007] Set up a training dataset, obtain a number of typical current characteristic values corresponding to the fault points from the training dataset, and train the fault diagnosis model based on the number of typical current characteristic values;

[0008] Periodically collect the current operation data in the ESP, and import the current operation data into the trained fault diagnosis model to obtain a number of measured current characteristic values;

[0009] Compare the number of measured current characteristic values with the number of typical current characteristic values to determine a number of different current characteristic types;

[0010] Combine the number of different current characteristic types to determine the fault type of the ESP.

[0011] Furthermore, the process of constructing the current normalization formula, the time normalization formula, and the current characteristic value calculation formula includes:

[0012] The current normalization formula is as follows:

[0013] ,

[0014] where, is the measured current value at the i-th point, with the unit of A; is the rated working current value, with the unit of A; is the first sensitivity factor; is the current value after normalization at the i-th point;

[0015] The time normalization formula is as follows:

[0016] ,

[0017] where, is the measured time value corresponding to the measured current value at the i-th point, with the unit of h; is the longest time among several measured current curves, with the unit of h; is the second sensitivity factor; is the time value after normalization at the i-th point;

[0018] The current characteristic value calculation formula is as follows:

[0019] ,

[0020] where, is the measured current characteristic value, with the unit of A / h; is the characteristic value correction coefficient.

[0021] Further, the process of determining a number of the different current characteristic types includes:

[0022] The different current characteristic types include a stable current characteristic type, a fluctuating current characteristic type, a shutdown characteristic type, a slowly decreasing characteristic type, a rapidly decreasing characteristic type, and an increasing characteristic type;

[0023] If the absolute value of the measured current characteristic value is less than or equal to the stable current characteristic value, it is determined as the stable current characteristic type;

[0024] If the absolute value of the measured current characteristic value is greater than the first current characteristic value and less than or equal to the second current characteristic value, and the proportion of the measured characteristic value is less than the proportion of the second characteristic value and greater than the proportion of the first characteristic value, it is determined as the fluctuating current characteristic type;

[0025] If the measured current characteristic value is less than or equal to -1000 and the number of shutdowns is greater than or equal to 1, it is determined as the shutdown characteristic type;

[0026] If the measured current characteristic value is less than the negative of the first current characteristic value and greater than or equal to the negative of the third current characteristic value, and the proportion of the measured characteristic value is greater than or equal to the proportion of the third characteristic value, it is determined as the slowly decreasing characteristic type;

[0027] If the measured current characteristic value is less than the negative of the third current characteristic value and greater than or equal to -1000, and the proportion of the measured characteristic value is greater than or equal to the proportion of the third characteristic value, it is determined as the rapidly decreasing characteristic type;

[0028] If the measured current characteristic value is less than the fourth current characteristic value and greater than or equal to the first current characteristic value, and the proportion of the measured characteristic value is greater than or equal to the proportion of the fourth characteristic value, it is determined as the increasing characteristic type.

[0029] Further, the process of determining the fault type of the electric submersible pump includes:

[0030] Combining a number of the different current characteristic types to obtain an analysis characteristic type group;

[0031] If all in the analysis characteristic type group are the fluctuating current characteristic types, it is determined that the fault type is that the well fluid in the environment where the electric submersible pump is located contains gas;

[0032] If in the analysis characteristic type group are successively the stable current characteristic type and the shutdown characteristic type, it is determined that the fault type is that the electric submersible pump suddenly shuts down;

[0033] If in the analysis characteristic type group are successively the rapidly decreasing characteristic type and more than two shutdown characteristic types, it is determined that the fault type is that the electric submersible pump frequently restarts;

[0034] If the stable current feature type, the slowly decreasing feature type, and the shutdown feature type are in sequence within the analysis feature type group, it is determined that the fault type is insufficient well fluid supply in the environment where the electric submersible pump is located;

[0035] If the stable current feature type, the increasing feature type, and the shutdown feature type are in sequence within the analysis feature type group, it is determined that the fault type is the electric submersible pump overloading and shutting down;

[0036] If the fluctuating current feature type, the increasing feature type, and the shutdown feature type are in sequence within the analysis feature type group, it is determined that the fault type is that the well fluid in the environment where the electric submersible pump is located contains impurities;

[0037] If the stable current feature type, the slowly decreasing feature type, the stable current feature type, and the shutdown feature type are in sequence within the analysis feature type group, it is determined that the fault type is the malfunction of the underload protection of the electric submersible pump;

[0038] If the stable current feature type, the slowly decreasing feature type, the fluctuating current feature type, and the shutdown feature type are in sequence within the analysis feature type group, it is determined that the fault type is the pump gas lock of the electric submersible pump.

[0039] Further, the process of obtaining several of the measured current feature values includes:

[0040] Obtain the eigenvalue variance according to several of the measured current feature values within a preset period;

[0041] Based on the comparison result between the eigenvalue variance and the preset eigenvalue variance, or, based on the comparison result between the average value of several historical eigenvalue variances obtained within several historical periods and the eigenvalue variance, determine whether the calculation process of the measured current feature values is qualified;

[0042] When it is determined that the calculation result of the measured current feature values is unqualified, determine the reason according to the comparison result between the difference between the maximum value and the minimum value among several of the measured current feature values and the preset difference.

[0043] Further, the process of determining the reason according to the comparison result between the difference between the maximum value and the minimum value among several of the measured current feature values and the preset difference includes:

[0044] Calculate the difference between the maximum value and the minimum value among several of the measured current feature values, and denote it as the eigenvalue amplitude difference, and denote the preset difference as the preset feature amplitude difference;

[0045] Determine the reason for the unqualified calculation of the measured current eigenvalue according to the comparison result between the eigenvalue amplitude difference and the preset eigenvalue amplitude difference;

[0046] Determine the corresponding processing based on the reason, including: increasing the first sensitivity factor and the second sensitivity factor or reducing the eigenvalue correction coefficient.

[0047] Further, the process of increasing the first sensitivity factor and the second sensitivity factor includes:

[0048] Determine to increase the first sensitivity factor and the second sensitivity factor based on the comparison result between the pump setting depth and the preset pump setting depth, and the increase amplitudes of the first sensitivity factor and the second sensitivity factor are in a direct proportion relationship with the pump setting depth.

[0049] Further, the process of reducing the eigenvalue correction coefficient includes:

[0050] Determine to reduce the eigenvalue correction coefficient based on the comparison result between the amplitude difference value and the preset amplitude difference value, and the reduction amplitude of the eigenvalue correction coefficient is in a direct proportion relationship with the amplitude difference value;

[0051] Wherein, the amplitude difference value is the difference between the eigenvalue amplitude difference and the preset eigenvalue amplitude difference.

[0052] Further, after completing the increase adjustment of the first sensitivity factor and the second sensitivity factor, determine to reduce the eigenvalue correction coefficient based on the comparison result between the motor temperature and the preset motor temperature, and the reduction amplitude of the eigenvalue correction coefficient is in a direct proportion relationship with the motor temperature.

[0053] The present invention also provides an electrical submersible pump fault warning device based on current change analysis, including:

[0054] A model construction module for constructing a current normalization formula, a time normalization formula, and a current eigenvalue calculation formula, and constructing a fault diagnosis model based on the current normalization formula, the time normalization formula, and the current eigenvalue calculation formula;

[0055] A training module, connected to the model construction module, for setting a training data set, obtaining a plurality of typical current eigenvalues corresponding to the fault points from the training data set, and training the fault diagnosis model based on the plurality of typical current eigenvalues;

[0056] A data acquisition module, connected to the model construction module, for periodically acquiring the current operation data in the electrical submersible pump, and importing the current operation data into the trained fault diagnosis model to obtain a plurality of measured current eigenvalues;

[0057] A comparison module, which is respectively connected to the model construction module and the training module, is used to compare a number of the measured current characteristic values with a number of the typical current characteristic values to determine a number of different current characteristic types, and to combine a number of different current characteristic types to determine the fault type of the ESP;

[0058] An analysis module, which is connected to the model construction module, is used to determine whether the calculation process of the measured current characteristic values is qualified based on the comparison result between the variance of a number of the measured current characteristic values and a preset variance, and to generate a corresponding instruction when it is unqualified;

[0059] An optimization module, which is respectively connected to the analysis module and the model construction module, is used to determine a first sensitivity factor and a second sensitivity factor or to determine an eigenvalue correction coefficient based on the instruction.

[0060] Compared with the prior art, the beneficial effect of an ESP fault warning method based on current change analysis of the present invention is that the method constructs a current normalization formula, a time normalization formula and a current characteristic value calculation formula, and constructs a fault diagnosis model based on this. After training the fault diagnosis model with a number of typical current characteristic values, a number of measured current characteristic values are calculated according to the current operation data, and a number of measured current characteristic values are compared with a number of typical current characteristic values to determine a number of different current characteristic types. By combining a number of different current characteristic types, the fault type of the ESP is determined, so as to improve the efficiency of troubleshooting the ESP, and improve the refined and intelligent management level of the ESP, and ensure the production efficiency of the oil well platform.

[0061] Further, by combining a number of different current characteristic types obtained in chronological order, the present invention can accurately determine the fault type of the corresponding ESP, and then realize the determination of the corresponding fault without manual on-site detection, and play a warning function, so that the equipment maintenance personnel can carry out targeted fault repair, thereby improving the repair efficiency of the equipment and ensuring the production progress of the oil well platform.

[0062] Further, the present invention further determines the calculation result of the measured current characteristic values based on the comparison result between the eigenvalue variance and the preset eigenvalue variance, determines whether the calculation of a number of the measured current characteristic values in the current period is qualified, and when it is determined that the calculation result of a number of the measured current characteristic values in the current period is unqualified, the specific reason can be processed accordingly, so as to improve the calculation qualification rate of the measured current characteristic values, thereby ensuring the accuracy of the subsequent fault diagnosis result.

[0063] Further, when the present invention compares the eigenvalue variance with the preset eigenvalue variance, it can further compare the average value of several historical eigenvalue variances in several historical cycles before the current preset cycle with the eigenvalue variance to re-determine whether the calculation result of the measured current eigenvalue is qualified, thereby improving the accuracy of the determination result.

[0064] Further, when the present invention determines that the calculation result of the measured current eigenvalue is unqualified, it can calculate the eigenvalue amplitude difference through the maximum value and the minimum value among several measured current eigenvalues, determine the reason based on the comparison result between the eigenvalue amplitude difference and the preset eigenvalue amplitude difference, and generate a corresponding processing method based on the reason, thereby improving the calculation qualification rate of the measured current eigenvalue.

[0065] Further, when the present invention determines that the problem lies in the first sensitive factor in the constructed current normalization formula and the second sensitive factor in the time normalization formula, it can also determine to increase the first sensitive factor and the second sensitive factor based on the comparison result between the pump setting depth and the preset pump setting depth, thereby reducing the influence on the normalization formula caused by the increase in the current value due to the excessive pumping resistance of the electric submersible pump at a deeper depth.

[0066] Further, when the present invention determines that the problem lies in the eigenvalue correction coefficient in the constructed current eigenvalue calculation formula, it can also determine to reduce the eigenvalue correction coefficient based on the comparison result between the amplitude difference value and the preset amplitude difference value, thereby reducing the influence caused by the large variation amplitude among several measured current eigenvalues. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 It is a module schematic diagram of an electric submersible pump fault warning device based on current change analysis in the present invention;

[0068] Figure 2 It is a flowchart of an electric submersible pump fault warning method based on current change analysis in the present invention;

[0069] Figure 3 It is a logical decision diagram for determining whether the calculation result of the measured current eigenvalue is qualified based on the eigenvalue variance in the present invention;

[0070] Figure 4 It is a logical decision diagram for determining the reason for the unqualified calculation of the measured current eigenvalue based on the eigenvalue amplitude difference and the corresponding processing method in the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0071] To make the objectives and advantages of the present invention more clearly understood, the present invention will be further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0072] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present invention and do not limit the protection scope of the present invention.

[0073] It should be noted that in the description of the present invention, unless otherwise clearly defined and limited, the term "connection" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0074] Please refer to Figure 1As shown in the figure, it is a schematic diagram of the modules of a device for a submersible pump fault warning method using current change analysis in this embodiment. The device is used to execute the submersible pump fault warning method based on current change analysis. The submersible pump in this embodiment is used for oil and gas field exploration and development. The device includes a model construction module, a training module, a data acquisition module, a comparison module, an analysis module, and an optimization module. The model construction module is used to construct a current normalization formula, a time normalization formula, and a current eigenvalue calculation formula, and construct a fault diagnosis model based on the current normalization formula, the time normalization formula, and the current eigenvalue calculation formula; the training module is connected to the model construction module, and is used to set a training data set, obtain a number of typical current eigenvalues corresponding to the fault points from the training data set, and train the fault diagnosis model based on the number of typical current eigenvalues; the data acquisition module is connected to the model construction module, and is used to periodically collect the current operation data in the submersible pump, and import the current operation data into the trained fault diagnosis model to obtain a number of measured current eigenvalues; the comparison module is respectively connected to the model construction module and the training module, and is used to compare a number of measured current eigenvalues with a number of typical current eigenvalues to determine a number of different current feature types, and combine a number of different current feature types to determine the fault type of the submersible pump; the analysis module is connected to the model construction module, and is used to determine whether the calculation process of the measured current eigenvalues is qualified based on the comparison result between the variance of a number of measured current eigenvalues and a preset variance, and generate a corresponding instruction when it is unqualified; the optimization module is respectively connected to the analysis module and the model construction module, and is used to determine a first sensitivity factor and a second sensitivity factor or determine an eigenvalue correction coefficient based on the instruction. By collecting the current values generated during the real-time operation of the submersible pump, and then calculating a number of measured current eigenvalues through the fault diagnosis model according to the collected current values, and comparing a number of measured current eigenvalues with typical current eigenvalues, it is possible to determine the current feature type in the submersible pump in the current cycle, and based on the current feature type, it is possible to determine the fault type that has occurred or may occur in the submersible pump, so that maintenance personnel can take corresponding maintenance measures in advance, thereby improving the efficiency of submersible pump fault detection, improving the refinement and intelligent level of the submersible pump, and ensuring the production efficiency of the oil well platform.

[0075] Specifically, in this embodiment, by selecting several groups of dynamic (such as current, voltage, motor temperature, pump inlet pressure, pump outlet pressure, pump suction inlet temperature, etc.) data sources corresponding to the real-time operation of the electrical submersible pump and static (such as production start time, pump setting depth, middle depth of oil layer, inner diameter of tubing, outer diameter of tubing, inner diameter of casing, outer diameter of casing, model, crude oil density, density of water, density of natural gas, saturation pressure, dissolved gas-oil ratio, geothermal gradient, motor model, motor power, number of stages, head, etc.) data sources corresponding to the oil well and wellbore, and using them as the training data set, the corresponding current characteristic value when a fault occurs is selected as the typical current characteristic value, and then the fault diagnosis model is trained using the typical current characteristic value.

[0076] Please refer to Figure 2 shown, which is a schematic flow chart of a fault warning method for an electrical submersible pump based on current change analysis in this embodiment. The process includes the following steps:

[0077] S1: Construct a current normalization formula, a time normalization formula, and a current characteristic value calculation formula, and construct a fault diagnosis model based on the current normalization formula, the time normalization formula, and the current characteristic value calculation formula.

[0078] S2: Set the training data set, obtain several typical current characteristic values corresponding to the fault points from the training data set, and train the fault diagnosis model based on the several typical current characteristic values.

[0079] S3: Periodically collect the current operation data in the electrical submersible pump, and import the current operation data into the trained fault diagnosis model to obtain several measured current characteristic values.

[0080] S4: Compare the several measured current characteristic values with the several typical current characteristic values to determine several different current characteristic types.

[0081] S5: Combine the several different current characteristic types to determine the fault type of the electrical submersible pump.

[0082] Among them, S3 includes the following steps:

[0083] S31: Determine whether the calculation process of the measured current characteristic value is qualified based on the comparison result between the variance of the several measured current characteristic values and the preset variance, and generate a corresponding instruction when it is unqualified.

[0084] S32: Determine the first sensitivity factor and the second sensitivity factor or determine the eigenvalue correction coefficient based on the instruction.

[0085] Furthermore, the process of constructing the current normalization formula, the time normalization formula, and the current characteristic value calculation formula includes:

[0086] The current normalization formula is as follows:

[0087] ,

[0088] where, is the measured current value at the i-th point, with the unit of A; is the rated working current value, with the unit of A; is the first sensitivity factor; is the current value after normalization at the i-th point.

[0089] The time normalization formula is as follows:

[0090] ,

[0091] where, is the measured time value corresponding to the measured current value at the i-th point, with the unit of h; is the longest time among several measured current curves, with the unit of h; is the second sensitivity factor; is the time value after normalization at the i-th point.

[0092] The calculation formula for the current characteristic value is as follows:

[0093] ,

[0094] where, is the measured current characteristic value, with the unit of A / h; is the characteristic value correction coefficient.

[0095] where, and are also correction factors for normalizing the data to a reasonable range. The purpose of performing normalization calculations on time and current is to enable the time value and current value to meet the requirements of subsequent calculation of the current characteristic value.

[0096] Furthermore, the process of determining several different types of current characteristics includes:

[0097] The different types of current characteristics include a stable current characteristic type, a fluctuating current characteristic type, a shutdown characteristic type, a slow decrease characteristic type, a rapid decrease characteristic type, and an increase characteristic type.

[0098] If the absolute value of the measured current eigenvalue is less than or equal to the stable current eigenvalue, it is determined as the stable current characteristic type; if the absolute value of the measured current eigenvalue is greater than the first current eigenvalue and less than or equal to the second current eigenvalue, and the proportion of the measured eigenvalue is less than the proportion of the second eigenvalue and greater than the proportion of the first eigenvalue, it is determined as the fluctuating current characteristic type; if the measured current eigenvalue is less than or equal to -1000 and the number of shutdowns is greater than or equal to 1, it is determined as the shutdown characteristic type; if the measured current eigenvalue is less than negative of the first current eigenvalue and greater than or equal to negative of the third current eigenvalue, and the proportion of the measured eigenvalue is greater than or equal to the proportion of the third eigenvalue, it is determined as the slow decrease characteristic type; if the measured current eigenvalue is less than negative of the third current eigenvalue and greater than or equal to -1000, and the proportion of the measured eigenvalue is greater than or equal to the proportion of the third eigenvalue, it is determined as the rapid decrease characteristic type; if the measured current eigenvalue is less than the fourth current eigenvalue and greater than or equal to the first current eigenvalue, and the proportion of the measured eigenvalue is greater than or equal to the proportion of the fourth eigenvalue, it is determined as the increase characteristic type.

[0099] Among them, the form of the stable current characteristic type is as follows:

[0100] ,

[0101] Among them, is the stable current eigenvalue, with the unit of A / h; is basically stable around 0, with the fluctuation not exceeding ±0.05.

[0102] Among them, the form of the fluctuating current characteristic type is as follows:

[0103] and ,

[0104] Among them, is the first current eigenvalue, with the unit of A / h; is the second current eigenvalue, with the unit of A / h; is the proportion of the measured eigenvalue, is the proportion of the first eigenvalue, is the proportion of the second eigenvalue; A / h, A / h, and the numerical fluctuation error between the two does not exceed 10%; here represents the proportion of the number of positive or negative measured current eigenvalues in the total number of measured current eigenvalues within the preset period, , .

[0105] Among them, the form of the shutdown characteristic type is as follows:

[0106] ,

[0107] wherein, is the number of shutdowns.

[0108] Wherein, the form of the slow decrease feature type is as follows:

[0109] and ,

[0110] wherein, is the third current eigenvalue, with the unit of A / h; is the proportion of the third eigenvalue; A / h, and the fluctuation error of this value does not exceed 10%; here represents the proportion of the number of measured current eigenvalues that are negative in the total number of measured current eigenvalues within the preset period, .

[0111] Wherein, the form of the rapid decrease feature type is as follows:

[0112] and ;

[0113] Wherein, the form of the increase feature type is as follows:

[0114] and ,

[0115] wherein, is the fourth current eigenvalue, with the unit of A / h; is the proportion of the fourth eigenvalue; A / h, and the fluctuation error of this value does not exceed 10%; here represents the proportion of the number of measured current eigenvalues that are positive in the total number of measured current eigenvalues within the preset period, . It should be noted that according to the actual situation and empirical comparison, and can be set to other values, and no specific limitation is made here.

[0116] Furthermore, the process of determining the fault type of the ESP includes:

[0117] Combine several of the different current characteristic types to obtain an analysis characteristic type group; if all within the analysis characteristic type group are the fluctuating current characteristic types, determine that the fault type is that the well fluid in the environment where the ESP is located contains gas, that is, long-term gas in the well fluid will cause pump gas lock; if within the analysis characteristic type group are successively the stable current characteristic type and the shutdown characteristic type, determine that the fault type is that the ESP suddenly shuts down; if within the analysis characteristic type group are successively the rapidly decreasing characteristic type and more than two shutdown characteristic types, determine that the fault type is that the ESP frequently restarts; if within the analysis characteristic type group are successively the stable current characteristic type, the slowly decreasing characteristic type, and the shutdown characteristic type, determine that the fault type is insufficient well fluid supply in the environment where the ESP is located; if within the analysis characteristic type group are successively the stable current characteristic type, the increasing characteristic type, and the shutdown characteristic type, determine that the fault type is that the ESP is overloaded and shuts down, that is, during the operation of the ESP, due to excessive load or certain abnormal conditions, the current exceeds its rated value, thus triggering a protection mechanism to stop the pump; if within the analysis characteristic type group are successively the fluctuating current characteristic type, the increasing characteristic type, and the shutdown characteristic type, determine that the fault type is that the well fluid in the environment where the ESP is located contains impurities, that is, the well fluid contains impurities and may cause solid blockage to the ESP; if within the analysis characteristic type group are successively the stable current characteristic type, the slowly decreasing characteristic type, the stable current characteristic type, and the shutdown characteristic type, determine that the fault type is that the underload protection of the ESP fails, that is, the underload protection function of the ESP cannot work properly, and cannot cut off the power supply or give an alarm in time when the ESP is in an underload state, thus possibly causing damage to the ESP; if within the analysis characteristic type group are successively the stable current characteristic type, the slowly decreasing characteristic type, the fluctuating current characteristic type, and the shutdown characteristic type, determine that the fault type is the pump gas lock of the ESP, that is, a large amount of gas accumulates in the pump chamber of the ESP, making the pump unable to work properly.

[0118] Specifically, in this embodiment, by combining several different calculated current characteristic types in chronological order from front to back to determine the analysis characteristic type, and then matching the analysis characteristic type with the typical fault types determined according to past actual situations, it is possible to quickly and accurately determine the fault possibility and fault type of the ESP in the current detection period, thus playing a role in fault warning, providing a basis for improving the refined management and intelligent level of the ESP, and also being able to improve the efficiency of troubleshooting for ESP faults, thus ensuring the production efficiency of the oil well platform where the ESP is located.

[0119] Please refer to Figure 3As shown, it is a logic flowchart for determining whether the calculation result of the measured current eigenvalue is qualified based on the eigenvalue variance in this embodiment. The process of obtaining a plurality of the measured current eigenvalues includes:

[0120] Obtain the eigenvalue variance based on a plurality of the measured current eigenvalues within a preset period; determine whether the calculation process of the measured current eigenvalue is qualified based on the comparison result between the eigenvalue variance and the preset eigenvalue variance, or based on the comparison result between the average value of a plurality of historical eigenvalue variances obtained within a plurality of historical periods and the eigenvalue variance; when it is determined that the calculation result of the measured current eigenvalue is unqualified, determine the reason based on the comparison result between the difference between the maximum value and the minimum value among a plurality of the measured current eigenvalues and the preset difference.

[0121] Specifically, in this embodiment, within the current preset period, by importing the real-time current operation data in the electric submersible pump into the fault diagnosis model, a plurality of measured current eigenvalues corresponding to a plurality of detection nodes within the preset period can be obtained. Then, the eigenvalue variance is calculated based on the plurality of measured current eigenvalues. In this embodiment, after obtaining the current operation data within the preset period for the electric submersible pump, the average value is calculated based on the plurality of measured current eigenvalues corresponding to a plurality of detection nodes. The numerical fluctuations of the measured current eigenvalues within this preset period should be maintained within the vicinity of this average value, and the fluctuation range does not exceed the average value too much. To ensure that the measured current eigenvalues used for determining the fault type are accurate and reasonable, the calculated measured current eigenvalues can be further determined. To make the determination more accurate, the preset eigenvalue variance E0 can be divided into the first preset eigenvalue variance E1 and the second preset eigenvalue variance E2. Set the preset eigenvalue variance standard E3 = 50, E1 = E3 - 10, E2 = E3 + 10. It should be noted that E1, E2, and E3 can also be set to other values according to actual needs. The specific process of comparing the eigenvalue variance E with E1 and E2 is as follows:

[0122] If the eigenvalue variance E is less than or equal to the first preset eigenvalue variance E1, it indicates that the magnitudes of the plurality of measured current eigenvalues within the current preset period are relatively close and the values are relatively concentrated. Therefore, it can be determined that the calculation result of the measured current eigenvalue is qualified. At this time, the measured current eigenvalue meets the requirements for subsequent comparison calculations with a plurality of typical current eigenvalues.

[0123] If the eigenvalue variance E is greater than the first preset eigenvalue variance E1 and less than or equal to the second preset eigenvalue variance E2, it is impossible to accurately determine whether the calculation result of the measured current eigenvalue is qualified. It is necessary to obtain several historical eigenvalue variances in several historical periods before the current preset period, and then calculate the average value based on the several historical eigenvalue variances. Based on the comparison between the variance average value and the eigenvalue variance, it is re-determined whether the calculation result of the measured current eigenvalue is qualified. The several historical measured current eigenvalues ​​in several historical periods here are confirmed data that can be used for subsequent calculations.

[0124] If the eigenvalue variance E is greater than the second preset eigenvalue variance E2, it means that the relative difference between the sizes of several measured current eigenvalues ​​in the current preset period is large, the values ​​are relatively dispersed, and the current values ​​fluctuate too violently in the preset period. Therefore, it can be determined that the calculation result of the measured current eigenvalue is unqualified. At this time, it is necessary to count the maximum and minimum values ​​of several measured current eigenvalues ​​in the current preset period, and then calculate the difference between the two, record it as the eigenvalue amplitude difference, and compare the eigenvalue amplitude difference with the preset eigenvalue amplitude difference to determine the cause of the unqualified. It is mainly determined by analyzing the fluctuation of several measured current eigenvalues ​​in the preset period to determine whether the current measured current eigenvalue can be used as the basis for subsequent calculations.

[0125] Furthermore, an average value is calculated based on several historical eigenvalue variances, which is recorded as the historical variance average value D. The historical variance average value D is compared with the eigenvalue variance E. The specific comparison results are as follows:

[0126] If the historical variance average value D is smaller than the eigenvalue variance E, it means that the fluctuation amplitude of several measured current eigenvalues ​​within the current preset period is larger than the fluctuation amplitude of several historical measured current eigenvalues ​​within the historical period. Therefore, it can be determined that when the eigenvalue variance E is greater than the first preset eigenvalue variance E1 and less than or equal to the second preset eigenvalue variance E2, the calculation result of the measured current eigenvalue is unqualified.

[0127] If the historical variance average value D is greater than or equal to the eigenvalue variance E, it means that the fluctuation amplitude of several measured current eigenvalues ​​in the current preset period is smaller than the fluctuation amplitude of several historical measured current eigenvalues ​​in the historical period. Therefore, it can be determined that when the eigenvalue variance E is greater than the first preset eigenvalue variance E1 and less than or equal to the second preset eigenvalue variance E2, the calculation result for the measured current eigenvalue is qualified.

[0128] See also Figure 4As shown, it is a logic flowchart for determining the reason for the unqualified calculation of the measured current eigenvalue and the corresponding processing method based on the eigenvalue amplitude difference in this embodiment. The process of determining the reason according to the comparison result between the difference between the maximum value and the minimum value among several measured current eigenvalues and the preset difference includes:

[0129] Calculate the difference between the maximum value and the minimum value among several measured current eigenvalues, and denote it as the eigenvalue amplitude difference. Denote the preset difference as the preset eigenvalue amplitude difference; Determine the reason for the unqualified calculation of the measured current eigenvalue according to the comparison result between the eigenvalue amplitude difference and the preset eigenvalue amplitude difference; Determine the corresponding processing based on the reason, including: increasing the first sensitivity factor and the second sensitivity factor or reducing the eigenvalue correction coefficient .

[0130] Specifically, in this embodiment, several measured current eigenvalues calculated within a preset period are compared to determine a maximum value and a minimum value, and then a difference is calculated based on the maximum value and the minimum value. The preset eigenvalue amplitude difference is the preset difference. Set the preset eigenvalue amplitude difference P0 = 50A / h, and P0 can also be set to other values according to needs; The comparison process between the eigenvalue amplitude difference P and the preset eigenvalue amplitude difference P0 is as follows:

[0131] If the eigenvalue amplitude difference P is less than or equal to the preset eigenvalue amplitude difference P0, it means that the change amplitude of the measured current eigenvalue within the current preset period is relatively small and the fluctuation degree is relatively stable. At this time, it can be determined that the reason for the unqualified calculation of the measured current eigenvalue is that there are problems with the constructed current normalization formula and time normalization formula, specifically, there are problems with the first sensitivity factor and the second sensitivity factor. Therefore, it is necessary to re-correct the first sensitivity factor and the second sensitivity factor.

[0132] If the eigenvalue amplitude difference P is greater than the preset eigenvalue amplitude difference P0, it means that the change amplitude of the measured current eigenvalue within the current preset period is relatively large and the fluctuation degree is relatively intense. At this time, it can be determined that the reason for the unqualified calculation of the measured current eigenvalue is that there are problems with the constructed current eigenvalue calculation formula, specifically, there are problems with the eigenvalue correction coefficient, and it is necessary to re-correct the eigenvalue correction coefficient.

[0133] Furthermore, the process of increasing the first sensitivity factor and the second sensitivity factor includes:

[0134] Determine to increase the first sensitivity factor and the second sensitivity factor based on the comparison result between the pump setting depth and the preset pump setting depth, and the increase amplitude of the first sensitivity factor and the second sensitivity factor is in a direct proportional relationship with the pump setting depth.

[0135] Specifically, in this embodiment, the operating environment of the electric submersible pump is a shallow oil well. When the pump setting depth is greater, the pressure of the underground oil fluid will increase, and the resistance of the pump to extract oil will increase. Therefore, the motor needs a greater load to provide sufficient power, which may lead to an increase in the current value. Therefore, in order to standardize more accurately, the adjustment ranges of the first sensitive factor and the second sensitive factor need to be greater; in order to more accurately determine the increase ranges of the first sensitive factor and the second sensitive factor, the preset pump setting depth H0 can be divided into the first preset pump setting depth H1 and the second preset pump setting depth H2. Set the preset pump setting depth H3 = 1200 m, H1 = H3 - 200, and H2 = H3 + 200; the specific process of comparing the pump setting depth H with H1 and H2 is as follows:

[0136] If the pump setting depth H is less than or equal to the first preset pump setting depth H1, the first sensitive factor can be increased to 1.15 times the initial value, and the second sensitive factor can be increased to 1.17 times the initial value using the first factor adjustment coefficient; if the pump setting depth H is greater than the first preset pump setting depth H1 and less than or equal to the second preset pump setting depth H2, the first sensitive factor can be increased to 1.44 times the initial value, and the second sensitive factor can be increased to 1.52 times the initial value using the second factor adjustment coefficient; if the pump setting depth H is greater than the second preset pump setting depth H2, the first sensitive factor can be increased to 2.25 times the initial value, and the second sensitive factor can be increased to 2.33 times the initial value using the third factor adjustment coefficient. Since the greater the pump setting depth H, the greater the influence of the current during acquisition and the time corresponding to the current on the fault diagnosis process, therefore, the magnification factors of the first sensitive factor and the second sensitive factor are greater; the values of the magnification factors can also be set to other values with the goal of reducing the influence brought by the pump setting depth H.

[0137] Further, the process of reducing the eigenvalue correction coefficient includes:

[0138] Determine to reduce the eigenvalue correction coefficient based on the comparison result between the amplitude difference and the preset amplitude difference, and the reduction amplitude of the eigenvalue correction coefficient is proportional to the amplitude difference; where the amplitude difference is the difference between the eigenvalue amplitude difference and the preset eigenvalue amplitude difference.

[0139] Specifically, in this embodiment, the preset amplitude difference B0 can be divided into the first preset amplitude difference B1 and the second preset amplitude difference B2. Set the preset amplitude difference standard B3 = 80 A / h, B1 = B3 - 15, and B2 = B3 + 15. It should be noted that B1, B2, and B3 can all be adjusted according to the actual situation; the specific process of comparing the amplitude difference B with B1 and B2 is as follows:

[0140] If the amplitude difference B is less than or equal to the first preset amplitude difference B1, the eigenvalue correction coefficient can be reduced to 0.95 times the initial value; if the amplitude difference B is greater than the first preset amplitude difference B1 and less than or equal to the second preset amplitude difference B2, the eigenvalue correction coefficient can be reduced to 0.86 times the initial value; if the amplitude difference B is greater than the second preset amplitude difference B2, the eigenvalue correction coefficient can be reduced to 0.75 times the initial value. When the amplitude difference B is large, it is necessary to reduce the eigenvalue correction coefficient, and the reduction factor of the eigenvalue correction coefficient is set to reduce the influence brought by the amplitude difference.

[0141] Further, after completing the increase adjustment of the first sensitive factor and the second sensitive factor, based on the comparison result between the motor temperature and the preset motor temperature, it is determined to reduce the eigenvalue correction coefficient, and the reduction amplitude of the eigenvalue correction coefficient is proportional to the motor temperature.

[0142] Specifically, in this embodiment, the motor temperature here is the temperature generated by the motor during the operation of the electric submersible pump. After completing the adjustment of the first sensitive factor and the second sensitive factor, it is also necessary to consider the influence of the electric submersible pump on the eigenvalue correction coefficient in the current eigenvalue calculation formula at the same time; the general operating temperature range of the electric submersible pump in this embodiment is 40°C - 180°C. The preset motor temperature W0 can be divided into the first preset motor temperature W1 and the second preset motor temperature W2. The preset motor temperature standard is set as W3 = 150°C, W1 = W3 - 20, W2 = W3 + 20. When the motor temperature is in a relatively high range, it will interfere with the collected current, resulting in a larger fluctuation range of the current value. Therefore, it is necessary to determine to reduce the eigenvalue correction coefficient according to the comparison result; the specific process of comparing the motor temperature W with W1 and W2 is as follows:

[0143] If the motor temperature W is greater than 100°C and less than or equal to the first preset motor temperature W1, the eigenvalue correction coefficient can be reduced to 0.98 times the initial value. When the motor temperature is less than or equal to 100°C, it will not affect the current; if the motor temperature W is greater than the first preset motor temperature W1 and less than or equal to the second preset motor temperature W2, the eigenvalue correction coefficient can be reduced to 0.95 times the initial value; if the motor temperature W is greater than the second preset motor temperature W2, the eigenvalue correction coefficient can be reduced to 0.92 times the initial value. The reduction factor of the eigenvalue correction coefficient at this time is different from the eigenvalue correction coefficient reduced according to the amplitude difference; the reduction factor value of the eigenvalue correction coefficient is set to reduce the influence brought by the motor temperature.

[0144] So far, the technical solution of the present invention has been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easily understood by those skilled in the art that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the protection scope of the present invention.

[0145] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various changes and modifications. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for early warning of electric submersible pump failure based on current change analysis, characterized in that: include: Constructing a current standardization formula, a time standardization formula, and a current characteristic value calculation formula, and constructing a fault diagnosis model based on the current standardization formula, the time standardization formula, and the current characteristic value calculation formula; Setting a training data set, obtaining a number of typical current characteristic values ​​corresponding to the fault point from the training data set, and training the fault diagnosis model based on the number of the typical current characteristic values; Periodically collecting current operation data in the electric submersible pump, and importing the current operation data into the trained fault diagnosis model to obtain a number of measured current characteristic values; Comparing the measured current characteristic values ​​with the typical current characteristic values ​​to determine a number of different current characteristic types; combining a plurality of the different current feature types to determine a fault type of the electric submersible pump; The process of obtaining the measured current characteristic values ​​includes: Obtaining a characteristic value variance according to a number of the measured current characteristic values ​​within a preset period; Determining whether the calculation process of the measured current characteristic value is qualified based on a comparison result of the characteristic value variance with a preset characteristic value variance, or based on a comparison result of an average value of several historical characteristic value variances obtained in several historical periods with the characteristic value variance; When it is determined that the calculation result of the measured current characteristic value is unqualified, determining the cause according to a comparison result of a difference between a maximum value and a minimum value among a plurality of the measured current characteristic values ​​and a preset difference; The process of determining the cause according to the comparison result of the difference between the maximum value and the minimum value of the measured current characteristic values ​​and the preset difference value comprises: Calculate the difference between the maximum value and the minimum value of the plurality of measured current characteristic values, record it as the characteristic value amplitude difference, and record the preset difference as the preset characteristic amplitude difference; Determining the reason why the calculation of the measured current characteristic value is unqualified according to the comparison result of the characteristic value amplitude difference and the preset characteristic value amplitude difference; Determining corresponding processing based on the cause includes: increasing the first sensitivity factor and the second sensitivity factor or reducing the characteristic value correction coefficient; The process of increasing the first sensitivity factor and the second sensitivity factor includes: Based on the comparison result of the pump hanging depth and the preset pump hanging depth, it is determined that the first sensitive factor and the second sensitive factor are increased, and the increase range of the first sensitive factor and the second sensitive factor is proportional to the pump hanging depth; After completing the increase adjustment of the first sensitive factor and the second sensitive factor, it is determined to reduce the characteristic value correction coefficient based on the comparison result of the motor temperature and the preset motor temperature, and the reduction range of the characteristic value correction coefficient is proportional to the motor temperature.

2. The electric submersible pump fault early warning method based on current change analysis according to claim 1 is characterized in that: The process of constructing the current normalization formula, the time normalization formula and the current characteristic value calculation formula includes: The current normalization formula is as follows: ， in, is the measured current value at the i-th point, in A; is the rated working current value, in A; is the first sensitive factor; is the normalized current value at the i-th point; The time normalization formula is as follows: ， in, is the measured time value corresponding to the measured current value at the i-th point, in units of h; is the longest time among several measured current curves, in h; is the second sensitive factor; is the time value after normalization of the i-th point; The current characteristic value calculation formula is as follows: ， in, is the measured current characteristic value, in A / h; is the eigenvalue correction coefficient.

3. The electric submersible pump fault early warning method based on current change analysis according to claim 2 is characterized in that: The process of determining the different types of current characteristics includes: The different current characteristic types include a stable current characteristic type, a fluctuating current characteristic type, a shutdown characteristic type, a slowly decreasing characteristic type, a rapidly decreasing characteristic type, and an increasing characteristic type; If the absolute value of the measured current characteristic value is less than or equal to the stable current characteristic value, it is determined to be the stable current characteristic type; If the absolute value of the measured current characteristic value is greater than the first current characteristic value and less than or equal to the second current characteristic value, and the proportion of the measured characteristic value is less than the proportion of the second characteristic value and greater than the proportion of the first characteristic value, it is determined to be the fluctuating current characteristic type; If the measured current characteristic value is less than or equal to negative 1000, and the number of shutdowns is greater than or equal to 1, it is determined to be the shutdown characteristic type; If the measured current characteristic value is less than the negative first current characteristic value and greater than or equal to the negative third current characteristic value, and the proportion of the measured characteristic value is greater than or equal to the proportion of the third characteristic value, it is determined to be the slowly decreasing characteristic type; If the measured current characteristic value is less than the negative third current characteristic value and greater than or equal to negative 1000, and the proportion of the measured characteristic value is greater than or equal to the proportion of the third characteristic value, it is determined to be the rapid decrease characteristic type; If the measured current characteristic value is less than the fourth current characteristic value and greater than or equal to the first current characteristic value, and the proportion of the measured characteristic value is greater than or equal to the proportion of the fourth characteristic value, it is determined to be the added characteristic type.

4. The electric submersible pump fault early warning method based on current change analysis according to claim 3 is characterized in that: The process of determining the fault type of the electric submersible pump includes: combining a plurality of said different current feature types to obtain an analysis feature type group; If all of the analyzed feature type groups are of the fluctuating current feature type, it is determined that the fault type is that the well fluid in the environment where the electric submersible pump is located contains gas; If the analysis feature type group includes the stable current feature type and the shutdown feature type in sequence, determining that the fault type is a sudden shutdown of the electric submersible pump; If the analysis feature type group includes the rapid reduction feature type and the shutdown feature type more than twice in sequence, then it is determined that the fault type is the frequent restart of the electric submersible pump; If the analysis feature type group includes the stable current feature type, the slowly decreasing feature type, and the shutdown feature type in sequence, it is determined that the fault type is insufficient well fluid supply in the environment where the electric submersible pump is located; If the analysis feature type group includes the stable current feature type, the increase feature type and the shutdown feature type in sequence, then determining that the fault type is the overload shutdown of the electric submersible pump; If the analysis feature type group includes the fluctuating current feature type, the increasing feature type and the shutdown feature type in sequence, it is determined that the fault type is that the well fluid in the environment where the electric submersible pump is located contains impurities; If the analysis feature type group includes the stable current feature type, the slowly decreasing feature type, the stable current feature type and the shutdown feature type in sequence, it is determined that the fault type is the failure of the underload protection of the electric submersible pump; If the analysis feature type group includes the stable current feature type, the slowly decreasing feature type, the fluctuating current feature type, and the shutdown feature type in sequence, it is determined that the fault type is the pump air lock of the electric submersible pump.

5. The electric submersible pump fault early warning method based on current change analysis according to claim 4 is characterized in that: The process of reducing the characteristic value correction coefficient includes: Based on the comparison result of the amplitude difference and the preset amplitude difference, it is determined to reduce the characteristic value correction coefficient, and the reduction amplitude of the characteristic value correction coefficient is proportional to the amplitude difference; The amplitude difference is the difference between the characteristic value amplitude difference and the preset characteristic value amplitude difference.

6. An electric submersible pump fault warning device based on current change analysis, characterized in that: The method for early warning of electric submersible pump faults based on current change analysis according to any one of claims 1 to 5 comprises: A model building module, used to build a current standardization formula, a time standardization formula and a current characteristic value calculation formula, and build a fault diagnosis model based on the current standardization formula, the time standardization formula and the current characteristic value calculation formula; A training module, connected to the model building module, for setting a training data set, obtaining a number of typical current characteristic values ​​corresponding to the fault point from the training data set, and training the fault diagnosis model based on the typical current characteristic values; A data acquisition module, connected to the model building module, for periodically acquiring current operation data in the electric submersible pump, and importing the current operation data into the trained fault diagnosis model to obtain a number of measured current characteristic values; a comparison module, which is connected to the model building module and the training module respectively, and is used to compare a number of the measured current characteristic values ​​with a number of the typical current characteristic values ​​to determine a number of different current characteristic types, and to combine a number of different current characteristic types to determine a fault type of the electric submersible pump; an analysis module connected to the model building module, for determining whether the calculation process of the measured current characteristic value is qualified based on a comparison result of the variance of the measured current characteristic value and a preset variance, and generating a corresponding instruction when it is unqualified; An optimization module is connected to the analysis module and the model building module respectively, and is used to determine the first sensitive factor and the second sensitive factor or determine the characteristic value correction coefficient based on the instruction.

Images