A method for predicting the efficiency of oil pumping system

Abstract

The invention discloses a method for predicting the efficiency of an oil pumping unit system, and relates to the technical field of oil pumping units. The method comprises: firstly, dividing and classifying all historical data accumulated by the oil pumping unit; then, using a method of calculating correlation, completing the screening of component sub-data corresponding to four types of influencing factor data, and setting influence coefficients for the four types of influencing factor data; by constructing a prediction model, inputting a set first formula and a method for obtaining relevant data in the formula into the prediction model, and building a mapping relationship between a characteristic value of the efficiency of the oil pumping unit system and the efficiency of the oil pumping unit system, and training the prediction model; finally, obtaining the influencing factor data of the oil pumping unit and inputting it into the trained prediction model, and the prediction model outputting the predicted system efficiency of the oil pumping unit, thereby improving the timeliness and accuracy of the existing prediction of the efficiency of the oil pumping unit system.

Description

Technical Field

[0001] The invention relates to the technical field of oil pumping units, and in particular to a method for predicting the efficiency of an oil pumping unit system. Background Art

[0002] The efficiency of the pumping system refers to the ratio of the effective working energy of lifting the liquid to the ground by transferring the ground electric energy to the downhole liquid to the system input energy. The efficiency of the pumping system consists of two parts: the ground efficiency and the downhole efficiency. The bare rod rope hanger is the boundary. The efficiency above the rope hanger is the ground efficiency, and the efficiency below the rope hanger is the downhole efficiency. The ground efficiency includes the motor efficiency, the belt-reduction box efficiency and the four-bar efficiency. The downhole efficiency includes the packing box efficiency, the sucker rod efficiency, the pump efficiency and the tubing efficiency.

[0003] At present, the mechanism model is mostly used to predict the system efficiency of the pumping unit. However, due to the uncertainty, complexity and timeliness of the pumping unit system efficiency, and because the change of the pumping unit system efficiency is restricted by multiple influencing factors, and the influencing factors maintain a complex relationship, if all influencing factors are used to predict the pumping unit system efficiency, the calculation of the mechanism model will be extremely complicated, and the real-time prediction cannot be guaranteed. If only the machine learning model is used to screen the main control influencing factors affecting the pumping unit system efficiency, firstly, all the main control influencing factors cannot be unified (different types and specifications), and secondly, comprehensiveness cannot be taken into account (such as the main control influencing factors are all working parameter data of the equipment), and the accuracy of the prediction cannot be guaranteed, which ultimately leads to the overall prediction effect is not ideal. Accurate prediction of the pumping unit system efficiency can take preventive measures in advance, control the efficiency of the pumping unit production system, and achieve the effect of energy saving, consumption reduction, cost reduction and efficiency improvement. Therefore, a method for predicting the efficiency of the pumping unit system is urgently needed to solve the problem that the existing prediction effect of the efficiency of the pumping unit system is not ideal. Summary of the invention

[0004] In order to solve the problems existing in the above-mentioned prior art, the present invention provides a method for predicting the efficiency of an oil pumping unit system, the method comprising:

[0005] Obtain all historical data accumulated by the pumping unit within a preset period;

[0006] Based on the data type, all historical data are divided into system efficiency data and influencing factor data, wherein the influencing factor data includes the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit;

[0007] Establishing a first curve graph of the system efficiency data with respect to time and a second curve graph of a plurality of influencing factor data with respect to time;

[0008] The correlation between the first curve graph and all the second curve graphs is calculated in sequence to obtain a first calculation result, and based on the first calculation result, the sub-data corresponding to the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit are screened;

[0009] After pre-processing the sub-data constituting the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit, a third curve graph of the four types of influencing factor data with respect to time is established;

[0010] Calculating the correlation between the first curve graph and each third curve graph to obtain a second calculation result;

[0011] Based on the second calculation result, an influence coefficient is set for the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit;

[0012] The first formula is used to calculate the system efficiency characteristic value of the oil pumping unit:

[0013]

[0014] is the system efficiency characteristic value of the pumping unit, A is the system efficiency constant, , , and are the influence coefficients corresponding to the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit, respectively. , , and They are respectively the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit;

[0015] Constructing a prediction model, inputting the first formula into the prediction model, building a mapping relationship between the system efficiency characteristic value of the oil pumping unit and the system efficiency of the oil pumping unit, and training the prediction model;

[0016] The influencing factor data of the oil pumping unit is obtained and input into the trained prediction model to obtain the predicted system efficiency of the oil pumping unit.

[0017] The present invention is implemented through the following technical scheme: since a large amount of data will be accumulated during the operation of the oil pump, the first thing to do is to classify all the historical data accumulated by the oil pump within a preset period, and based on the data type, the accumulated historical data are divided into system efficiency data, influencing factor data and other data. The system efficiency data is the historical data of the system efficiency of the oil pump, and the influencing factor data is the historical data that affects the system efficiency of the oil pump; and the influencing factor data is generally divided into four categories, namely, the operating condition data of the oil pump (such as service life, loss degree, etc.), the working parameter data of the oil pump (such as stroke, stroke frequency, pump diameter, etc.), the working environment data of the oil pump (such as viscosity, water content, gas-oil ratio, etc.) and the production data of the oil pump (such as oil pressure, casing pressure, etc.), and the above steps complete the division and classification of historical data.

[0018] Since there are many data on the factors affecting the efficiency of the oil pumping system, if only the machine learning model is used to screen the main controlling factors affecting the efficiency of the oil pumping system, it is impossible to guarantee comprehensiveness (all four types of influencing factor data are included). Therefore, this scheme first establishes a first curve graph of the oil pumping system efficiency data with respect to time, and then establishes a second curve graph of each influencing factor with respect to time, and then calculates the correlation between the first curve graph and each second curve graph in turn (such as using MATLAB software to implement) to obtain a first calculation result. Through the first calculation result, the sub-data corresponding to the oil pumping unit operating condition data, working parameter data, working environment data and production data are screened. The above steps complete the screening of the sub-data corresponding to the four types of influencing factor data.

[0019] After pre-processing the operating condition data, working parameter data, working environment data and production data of the oil pumping unit, a third curve graph of the four types of influencing factor data with respect to time is established, and the correlation between the first curve graph and each third curve graph is calculated to obtain a second calculation result. According to the second calculation result, different influence coefficients are set for the four types of influencing factor data. The above steps complete the setting of the influence coefficients of the four types of influencing factor data, and the influence coefficients are used to reflect the degree of influence of different types of influencing factor data on the efficiency of the oil pumping unit system.

[0020] This scheme calculates the system efficiency characteristic value of the oil pumping unit by setting the first formula, and introduces the influence coefficients of four types of influencing factor data and four types of influencing factor data after preset processing in the first formula, and sets the system efficiency constant (so that the system efficiency characteristic value is reflected in the form of percentage), and finally constructs a prediction model, inputs the first formula into the prediction model, which also includes the acquisition method of various types of data in the first formula and inputs it into the prediction model, and builds a mapping relationship between the system efficiency characteristic value of the oil pumping unit and the system efficiency of the oil pumping unit, and then trains the prediction model. The influencing factor data of the oil pumping unit system efficiency to be predicted is obtained and input into the trained prediction model. After the prediction model completes the screening and preset processing of the input influencing factor data, the corresponding oil pumping unit system efficiency characteristic value is calculated by the first formula, and then the predicted system efficiency of the oil pumping unit is finally output through the mapping relationship between the oil pumping unit system efficiency characteristic value and the oil pumping unit system efficiency. The above steps complete the prediction of the oil pumping unit system efficiency.

[0021] As an optional technical solution, screening the sub-data corresponding to the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit includes:

[0022] Calculating the correlation between the first curve graph and all second curve graphs and placing them into a preset set;

[0023] Filter all the correlations in the preset set that are greater than the set value and name them as first correlations;

[0024] Determine whether all the influencing factor data corresponding to the first correlation degree fully include the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit;

[0025] If yes, respectively count the sub-data corresponding to the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit;

[0026] If not, sort all the correlations in the preset set, traverse all the correlations in the sorted preset set, obtain the last first correlation corresponding to a value greater than the set value as the second correlation, and traverse all subsequent correlations in sequence starting from the second correlation until the influencing factor data corresponding to the screened correlations all contain the operating data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit, and respectively count the sub-data corresponding to the operating data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit.

[0027] Among them, when the sub-data composed of the four categories of influencing factor data are screened again, the correlation between all the influencing factor data and the pumping system efficiency has been obtained through the first calculation result, and all the calculated correlations are put into a preset set, and all the correlations in the preset set with correlations greater than the set value are screened and named as first correlations, and it is judged whether the influencing factor data corresponding to all the first correlations contain the four categories of influencing factor data. If so, then in the influencing factor data corresponding to all the first correlations, the sub-data of the four categories of influencing factor data are counted respectively; if not, the correlations in the preset set are sorted, and then the preset set is traversed to obtain the last first correlation greater than the set value as the second correlation, and the subsequent correlations are traversed starting from the second correlation until the influencing factor data corresponding to the screened correlations contain the four categories of influencing factor data, and then the sub-data of the four categories of influencing factor data are counted respectively.

[0028] As an optional technical solution, pre-processing the sub-data consisting of the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit includes:

[0029] Acquire all sub-data consisting of the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit and the corresponding data types;

[0030] Based on the data type, for a single sub-data, multiple data intervals are set, and the number of intervals of any two data intervals is equal;

[0031] Using a first preset method to preset a specific value for each of the data intervals;

[0032] When it is determined that the sub-data is located in the first data interval, a specific value corresponding to the first data interval is obtained as the value assigned to the sub-data.

[0033] Among them, since all the influencing factor data are divided into four categories, and whether it is the subsequent establishment of the third curve graph or the application of the four categories of influencing factor data in the first formula, each category of influencing factor data needs to be unified (the same measurement standard), so this solution adopts the method of converting the specific influencing factor data into specific values, and the specific value range corresponding to each influencing factor data is the same. The specific assignment method is to set several data intervals for each influencing factor data, and use the first preset method to set a specific value for each data interval. When it is determined that the specific influencing factor data falls within a certain data interval, the specific value corresponding to the data interval is directly obtained, which is the assignment of the influencing factor data.

[0034] As an optional technical solution, the first preset method includes:

[0035] Based on the historical data, all system efficiency data corresponding to each of the data intervals are obtained;

[0036] Calculate the average value of the system efficiency data corresponding to each of the data intervals to obtain a third calculation result;

[0037] Based on the third calculation result, drawing a thermal value diagram of the data interval corresponding to the system efficiency data;

[0038] Based on the heat value map, a specific value is preset for each of the data intervals.

[0039] Specifically, when it comes to the data of a single influencing factor, the impact of the data on the system efficiency is not positively correlated. Therefore, this solution obtains the corresponding mean value of the system efficiency for each data interval, and uses the mean value of the system efficiency to draw a heat value map of the system efficiency corresponding to the data interval. Through the heat value map, specific values ​​are preset for each data interval.

[0040] As an optional technical solution, the method further includes obtaining the system efficiency constant:

[0041] Based on the range of the specific value, a value of the system efficiency constant is obtained.

[0042] The system efficiency constant is used to express the system efficiency characteristic value in the form of a percentage, so the value of the system efficiency constant is determined based on a specific value range.

[0043] As an optional technical solution, pre-processing the sub-data constituting the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit further includes:

[0044] After assigning values ​​to all sub-data corresponding to the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit, a second formula group is used to calculate the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit;

[0045] The second formula group is specifically:

[0046]

[0047] , , and The sub-data respectively constitute the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit after being assigned; , n, p and They are the number of sub-data constituting the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit.

[0048] Among them, the four types of impact factor data are further unified through the second formula group. The core idea is that after all sub-data in each type of impact factor data are assigned values, the mean value of each type of impact factor data is calculated through the second formula group.

[0049] As an optional technical solution, the method further includes:

[0050] constructing a risk model and training the risk model;

[0051] Inputting the predicted system efficiency of the oil pump obtained by the prediction model into the trained risk model;

[0052] The risk model determines whether the prediction system efficiency is within a first preset range, and if so, outputs an instruction to execute the oil pump to work, and if not, uses a second preset method to adjust the oil pump.

[0053] Among them, most of the existing mechanism models for predicting the efficiency of the oil pumping system only predict the efficiency of the oil pumping system, but lack follow-up measures. Therefore, this scheme also provides a risk model, which mainly processes the prediction results of the prediction model, and inputs the oil pumping system efficiency predicted by the prediction model into the risk model. The risk model determines whether the predicted system efficiency is within a first preset range. If so, it outputs an instruction to execute the oil pump to work. If not, it is necessary to use the second preset method to adjust the oil pump. Because it is not within the first preset range, there are two situations. One is that the predicted oil pumping system efficiency exceeds the first preset range, and the other is that the predicted oil pumping system efficiency does not exceed the first preset range, so they need to be processed separately.

[0054] As an optional technical solution, the second preset method includes:

[0055] Obtain the two endpoint values ​​of the first preset range, and determine whether the value of the prediction system efficiency exceeds the two endpoint values. If so, save the input influencing factor data corresponding to the prediction system efficiency to a preset data set; if not, calculate the difference between the prediction system efficiency and the minimum value of the two endpoint values, and process the difference using a third preset method.

[0056] Among them, two endpoint values ​​of the first preset range are obtained, and when it is judged that the value of the predicted system efficiency exceeds the two endpoint values, it means that the system efficiency predicted at this time is higher than the conventional system efficiency of the oil pumping unit, and the influencing factor data corresponding to the input of the predicted system efficiency is saved in the preset data set for subsequent adjustment of the oil pumping unit; when it is judged that the value of the predicted system efficiency does not exceed the two endpoint values ​​and is not within the first preset range, it means that the system efficiency predicted at this time is lower than the conventional system efficiency of the oil pumping unit, and then the difference between the predicted system efficiency and the minimum value of the two endpoint values ​​is calculated, and then the difference is processed by the third preset method.

[0057] As an optional technical solution, the third preset method includes:

[0058] It is determined whether the difference is within a second preset range. If so, all the pumping units are adjusted based on the data stored in the preset data set; if not, a warning instruction is output.

[0059] Among them, if it is judged that the difference is within the second preset range, it means that the difference between the predicted system efficiency and the conventional system efficiency is not large at this time, so all the influencing factor data accumulated in the preset data set can be used to make corresponding adjustments to the oil pump; if it is judged that the difference is not within the second preset range, it means that the difference between the predicted system efficiency and the conventional system efficiency is too large at this time. If the oil pump is executed at this time, the loss caused is large, so an early warning instruction is output to allow relevant personnel to inspect or maintain the entire oil pump system.

[0060] As an optional technical solution, the method further includes optimizing the prediction model;

[0061] Obtaining actual system efficiencies of multiple groups of the pumping units within a set period and placing them into set A;

[0062] Obtain predicted system efficiencies of multiple groups of the pumping units within the set period through the prediction model and put them into set B;

[0063] Based on the principle of permutation and combination, the differences between all actual system efficiencies in the set A and all predicted system efficiencies in the set B are calculated in sequence to obtain a fourth calculation result;

[0064] Based on the fourth calculation result, the prediction model is optimized.

[0065] Among them, this solution also provides a method for optimizing the prediction model, by setting set A to collect multiple groups of actual system efficiencies of oil pumps within a set period, by setting set B to collect multiple groups of predicted system efficiencies within a set period, and then based on the principle of permutations and combinations, the differences between all actual system efficiencies in set A and all predicted system efficiencies in set B are calculated in sequence to obtain a fourth calculation result, and the prediction model is optimized through the fourth calculation result.

[0066] One or more technical solutions provided by the present invention have at least the following technical effects or advantages:

[0067] The present invention discloses a method for predicting the efficiency of an oil pumping unit system. First, all historical data accumulated by the oil pumping unit are divided and classified; then, the corresponding component sub-data of four types of influencing factor data are screened by calculating the correlation, and the influencing coefficients are set for the four types of influencing factor data; by constructing a prediction model, a set first formula and a method for obtaining relevant data in the formula are input into the prediction model, and a mapping relationship between a characteristic value of the oil pumping unit system efficiency and the oil pumping unit system efficiency is established, and the prediction model is trained; finally, the influencing factor data of the oil pumping unit is obtained and input into the trained prediction model, and the prediction model outputs the predicted system efficiency of the oil pumping unit, thereby improving the timeliness and accuracy of the existing prediction of the efficiency of the oil pumping unit system.

[0068] The present invention also discloses a risk model, which processes the pumping unit system efficiency predicted by the prediction model, and distinguishes the pumping unit prediction system efficiency by setting a first preset range and a second preset range. When the prediction system efficiency is within the first preset range, an instruction to execute the pumping unit operation is output; when the prediction system efficiency exceeds the first preset range, a preset data set is set to save the corresponding input influencing factor data; the difference between the prediction system efficiency and the smaller endpoint value of the first preset range is calculated, and when the difference is within the second preset range, the entire pumping unit system is adjusted through the data accumulated in the preset data set, and when the difference is not within the second preset range, an early warning instruction is output.

[0069] The present invention also discloses a method for optimizing the prediction model, by setting a set A to collect multiple groups of actual system efficiencies of oil pumps within a set period, by setting a set B to collect multiple groups of predicted system efficiencies within a set period, and then based on the principle of permutations and combinations, sequentially calculating the differences between all actual system efficiencies in set A and all predicted system efficiencies in set B to obtain a fourth calculation result, and optimizing the prediction model through the fourth calculation result. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of the present invention, and do not constitute a limitation on the embodiments of the present invention;

[0071] Figure 1 It is a schematic flow chart of a method for predicting the efficiency of an oil pumping unit system in the present invention. DETAILED DESCRIPTION

[0072] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict.

[0073] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those within the scope of this description. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0074] Example

[0075] Please refer to Figure 1 , Figure 1 The figure is a flow chart of a method for predicting the efficiency of a pumping unit system in the present invention, the method comprising:

[0076] Obtain all historical data accumulated by the pumping unit within a preset period;

[0077] Based on the data type, all historical data are divided into system efficiency data and influencing factor data, wherein the influencing factor data includes the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit;

[0078] Establishing a first curve graph of the system efficiency data with respect to time and a second curve graph of a plurality of influencing factor data with respect to time;

[0079] The correlation between the first curve graph and all the second curve graphs is calculated in sequence to obtain a first calculation result, and based on the first calculation result, the sub-data corresponding to the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit are screened;

[0080] After pre-processing the sub-data constituting the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit, a third curve graph of the four types of influencing factor data with respect to time is established;

[0081] Calculating the correlation between the first curve graph and each third curve graph to obtain a second calculation result;

[0082] Based on the second calculation result, an influence coefficient is set for the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit;

[0083] The first formula is used to calculate the system efficiency characteristic value of the oil pumping unit:

[0084]

[0085] is the system efficiency characteristic value of the pumping unit, A is the system efficiency constant, , , and are the influence coefficients corresponding to the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit, respectively. , , and They are respectively the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit;

[0086] Constructing a prediction model, inputting the first formula into the prediction model, building a mapping relationship between the system efficiency characteristic value of the oil pumping unit and the system efficiency of the oil pumping unit, and training the prediction model;

[0087] The influencing factor data of the oil pumping unit is obtained and input into the trained prediction model to obtain the predicted system efficiency of the oil pumping unit.

[0088] The specific embodiments of the present invention are as follows:

[0089] Divide and classify all historical data accumulated by the oil pumping unit: obtain all historical data accumulated by the oil pumping unit within a preset period, and divide all historical data into system efficiency data and influencing factor data based on the data type. The influencing factor data is divided into four categories, namely the operating condition data of the oil pumping unit (such as age, loss degree, etc.), the working parameter data of the oil pumping unit (such as stroke, stroke frequency, pump diameter, pump depth, rod column combination, motor power, etc.), the working environment data of the oil pumping unit (such as viscosity, water content, gasoline ratio, wellbore structure, etc.) and the production data of the oil pumping unit (such as output, oil pressure, casing pressure, etc.). Historical data also includes other data, which are all data that need to be filtered except the system efficiency data and influencing factor data.

[0090] Screen the sub-data corresponding to the four types of influencing factor data: establish the first curve graph of system efficiency data on time and the second curve graphs of several influencing factor data on time, calculate the correlation between the first curve graph and all the second curve graphs in turn to obtain the first calculation result. There are relatively mature technologies for calculating the correlation between curves, such as the corrcoef function of maltab software. Then, based on the first calculation result, screen the sub-data corresponding to the four types of influencing factor data, and the specific implementation method is as follows:

[0091] Calculate the correlation between the first curve graph and all second curve graphs and put them into a preset set,

[0092] Filter all the correlations in the preset set that are greater than the set value and name them as first correlations;

[0093] Determine whether the influencing factor data corresponding to all the first correlations fully include the operating data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit; if so, respectively count the sub-data corresponding to the operating data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit; if not, sort all the correlations in the preset set, traverse all the correlations in the sorted preset set, obtain the last first correlation corresponding to a value greater than the set value as the second correlation, and traverse all subsequent correlations in sequence starting from the second correlation until the influencing factor data corresponding to the screened correlations fully include the operating data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit, and respectively count the sub-data corresponding to the operating data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit.

[0094] Among them, all the calculated relevances are put into a preset set, and all the relevances in the preset set whose relevances are greater than the set value are screened and named as the first relevances, and it is determined whether the influencing factor data corresponding to all the first relevances all contain the four categories of influencing factor data. If so, then in the influencing factor data corresponding to all the first relevances, the sub-data of the four categories of influencing factor data are counted respectively; if not, the relevances in the preset set are sorted, and then the preset set is traversed to obtain the last first relevance greater than the set value as the second relevance, and the subsequent relevances are traversed with the second relevance as the starting point, until the influencing factor data corresponding to the screened relevances all contain the four categories of influencing factor data, and then the sub-data of the four categories of influencing factor data are counted respectively.

[0095] Setting influence coefficients for the four types of influencing factor data: After pre-processing the operating data, working parameter data, working environment data and production data of the oil pumping unit, a third curve graph of the four types of influencing factor data with respect to time is established, and the correlation between the first curve graph and each third curve graph is calculated to obtain a second calculation result. According to the second calculation result, different influence coefficients are set for the four types of influencing factor data. The implementation method of pre-processing the four types of influencing factors is as follows:

[0096] Acquire all sub-data consisting of the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit and the corresponding data types;

[0097] Based on the data type, for a single sub-data, multiple data intervals are set, and the number of intervals of any two data intervals is equal;

[0098] A first preset method is used to preset a specific value for each of the data intervals; the first preset method includes: based on the historical data, obtaining all system efficiency data corresponding to each of the data intervals; calculating the average value of the system efficiency data corresponding to each of the data intervals to obtain a third calculation result; based on the third calculation result, drawing a thermal value map of the system efficiency data corresponding to the data interval; based on the thermal value map, preset a specific value for each of the data intervals;

[0099] When it is determined that the sub-data is located in the first data interval, a specific value corresponding to the first data interval is obtained as the value assigned to the sub-data.

[0100] Among them, the sub-data corresponding to the four types of influencing factor data are unified by dividing the data intervals and presetting specific values. This embodiment is implemented by converting the specific influencing factor data into specific values, and the specific value range corresponding to each influencing factor data is the same. The specific assignment method is to set several data intervals for each influencing factor data, and use the first preset method to set a specific value for each data interval. When it is determined that the specific influencing factor data falls within a certain data interval, the specific value corresponding to the data interval is directly obtained, which is the assignment of the influencing factor data.

[0101] Wherein, based on the range of the specific value, the value of the system efficiency constant is obtained. If the range of the specific value is 0-10, then the value of the system efficiency constant is 10, and the value of the system efficiency constant is related to the range of the specific value.

[0102] Among them, the implementation methods of pre-setting the four types of influencing factors also include:

[0103] After assigning values ​​to all sub-data corresponding to the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit, a second formula group is used to calculate the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit;

[0104] The second formula group is specifically:

[0105]

[0106] , , and The sub-data respectively constitute the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit after being assigned; , n, p and They are the number of sub-data constituting the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit.

[0107] Among them, the four types of impact factor data are further unified through the second formula group. The core idea is that after all sub-data in each type of impact factor data are assigned values, the mean value of each type of impact factor data is calculated through the second formula group.

[0108] Predicting the system efficiency of the oil pumping unit: calculating the system efficiency characteristic value of the oil pumping unit by setting the first formula, wherein the four types of influencing factor data and influencing coefficients in the first formula, as well as the system efficiency constant, are specifically obtained and calculated in the embodiment and will not be elaborated on here. The system efficiency characteristic value calculated by the first formula cannot represent the final system efficiency, and it is necessary to establish a mapping relationship between the system efficiency characteristic value of the oil pumping unit and the system efficiency of the oil pumping unit to achieve this. To obtain the mapping relationship between the system efficiency characteristic value of the oil pumping unit and the system efficiency of the oil pumping unit, multiple groups of system efficiency characteristic values ​​can be first obtained by the first formula, and then based on historical data, corresponding multiple groups of system efficiency historical data can be obtained, and the mapping relationship between the system efficiency characteristic value and the system efficiency can be calculated based on the multiple groups of system efficiency characteristic values ​​and the multiple groups of system efficiency historical data. A prediction model is constructed, the first formula and the method for obtaining its data are input into the prediction model, and a mapping relationship between the characteristic value of the pumping unit system efficiency and the pumping unit system efficiency is established to train the prediction model; the influencing factor data of the pumping unit system efficiency to be predicted is obtained and input into the trained prediction model, and after the prediction model completes screening and preset processing of the input influencing factor data, the corresponding characteristic value of the pumping unit system efficiency is calculated through the first formula, and then through the mapping relationship between the characteristic value of the pumping unit system efficiency and the pumping unit system efficiency, the predicted system efficiency of the pumping unit is finally output.

[0109] Wherein, the method further comprises:

[0110] constructing a risk model and training the risk model;

[0111] Inputting the predicted system efficiency of the oil pump obtained by the prediction model into the trained risk model;

[0112] The risk model determines whether the prediction system efficiency is within a first preset range, and if so, outputs an instruction to execute the pumping unit to work, and if not, uses a second preset method to adjust the pumping unit. The second preset method is implemented as follows:

[0113] Obtain the two endpoint values ​​of the first preset range, determine whether the value of the prediction system efficiency exceeds the two endpoint values, if so, save the input impact factor data corresponding to the prediction system efficiency to the preset data set; if not, calculate the difference between the prediction system efficiency and the minimum value of the two endpoint values, and process the difference using the third preset method. The implementation of the third preset method is as follows:

[0114] It is determined whether the difference is within a second preset range. If so, all the pumping units are adjusted based on the data stored in the preset data set; if not, a warning instruction is output.

[0115] Among them, most of the existing mechanism models for predicting the efficiency of the oil pumping system only predict the efficiency of the oil pumping system, but lack follow-up measures. Therefore, the present scheme also provides a risk model, which mainly processes the prediction results of the prediction model, and inputs the oil pumping system efficiency predicted by the prediction model into the risk model. The risk model determines whether the predicted system efficiency is within the first preset range. It should be noted that regarding the setting of the first preset range, the first preset range is different for different types of oil pumps, and can be adjusted according to actual conditions. This embodiment does not make specific limitations. If it is determined that the predicted system efficiency is within the first preset range (such as 25%~30%), the risk model outputs an instruction to execute the oil pump to work; if it is determined that the predicted system efficiency exceeds the first preset range, that is, the predicted system efficiency exceeds 30%, all the influencing factor data input for the predicted system efficiency are saved in the preset data set for subsequent adjustment of the oil pump system; if it is determined that the predicted system efficiency is lower than 25%, the difference between the predicted system efficiency and 25% is calculated to determine whether the difference is within the second preset range (such as 0~5%). If so, it means that the difference between the predicted system efficiency and the conventional system efficiency is not large at this time, so all the influencing factor data accumulated in the preset data set can be used to make corresponding adjustments to the oil pump; if it is determined that the difference is not within the second preset range, it means that the difference between the predicted system efficiency and the conventional system efficiency is too large at this time. If the oil pump is executed to work at this time, the loss caused is large, so an early warning instruction is output to allow relevant personnel to inspect or maintain the entire oil pump system.

[0116] Wherein, the method further comprises optimizing the prediction model;

[0117] Obtaining actual system efficiencies of multiple groups of the pumping units within a set period and placing them into set A;

[0118] Obtain predicted system efficiencies of multiple groups of the pumping units within the set period through the prediction model and put them into set B;

[0119] Based on the principle of permutation and combination, the differences between all actual system efficiencies in the set A and all predicted system efficiencies in the set B are calculated in sequence to obtain a fourth calculation result;

[0120] Based on the fourth calculation result, the prediction model is optimized.

[0121] Among them, this solution also provides a method for optimizing the prediction model, by setting set A to collect multiple groups of actual system efficiencies of oil pumps within a set period, by setting set B to collect multiple groups of predicted system efficiencies within a set period, and then based on the principle of permutations and combinations, the differences between all actual system efficiencies in set A and all predicted system efficiencies in set B are calculated in sequence to obtain a fourth calculation result, and the prediction model is optimized through the fourth calculation result.

[0122] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0123] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for predicting the efficiency of a pumping unit system, characterized in that: The method comprises: Obtain all historical data accumulated by the pumping unit within a preset period; Based on the data type, all historical data are divided into system efficiency data and influencing factor data, wherein the influencing factor data includes the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit; Establishing a first curve graph of the system efficiency data with respect to time and a second curve graph of a plurality of influencing factor data with respect to time; The correlation between the first curve graph and all the second curve graphs is calculated in sequence to obtain a first calculation result, and based on the first calculation result, the sub-data corresponding to the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit are screened; After pre-processing the sub-data constituting the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit, a third curve graph of the four types of influencing factor data with respect to time is established; Calculating the correlation between the first curve graph and each third curve graph to obtain a second calculation result; Based on the second calculation result, an influence coefficient is set for the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit; The first formula is used to calculate the system efficiency characteristic value of the oil pumping unit: is the system efficiency characteristic value of the pumping unit, A is the system efficiency constant, , , and are the influence coefficients corresponding to the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit, respectively. , , and They are respectively the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit; Constructing a prediction model, inputting the first formula into the prediction model, building a mapping relationship between the system efficiency characteristic value of the oil pumping unit and the system efficiency of the oil pumping unit, and training the prediction model; Obtaining the influencing factor data of the oil pumping unit and inputting it into the trained prediction model to obtain the predicted system efficiency of the oil pumping unit; The sub-data corresponding to the operating condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit are screened and include: Calculating the correlation between the first curve graph and all second curve graphs and placing them into a preset set; Filter all the correlations in the preset set that are greater than the set value and name them as first correlations; Determine whether all the influencing factor data corresponding to the first correlation degree fully include the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit; If yes, respectively count the sub-data corresponding to the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit; If not, sort all the correlations in the preset set, traverse all the correlations in the sorted preset set, obtain the last first correlation corresponding to a value greater than the set value as the second correlation, and traverse all subsequent correlations in sequence starting from the second correlation until the influencing factor data corresponding to the screened correlations all contain the operating data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit, and respectively count the sub-data corresponding to the operating data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit.

2. A method for predicting the efficiency of an oil pumping unit system according to claim 1, characterized in that: Presetting the sub-data consisting of the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit includes: Acquire all sub-data consisting of the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit and the corresponding data types; Based on the data type, for a single sub-data, multiple data intervals are set, and the number of intervals of any two data intervals is equal; Using a first preset method to preset a specific value for each of the data intervals; When it is determined that the sub-data is located in the first data interval, a specific value corresponding to the first data interval is obtained as the value assigned to the sub-data.

3. A method for predicting the efficiency of an oil pumping unit system according to claim 2, characterized in that: The first preset method includes: Based on the historical data, all system efficiency data corresponding to each of the data intervals are obtained; Calculate the average value of the system efficiency data corresponding to each of the data intervals to obtain a third calculation result; Based on the third calculation result, drawing a thermal value diagram of the data interval corresponding to the system efficiency data; Based on the heat value map, a specific value is preset for each of the data intervals.

4. A method for predicting the efficiency of an oil pumping unit system according to claim 2, characterized in that: The method further includes obtaining the system efficiency constant: Based on the range of the specific value, a value of the system efficiency constant is obtained.

5. A method for predicting the efficiency of an oil pumping unit system according to claim 2, characterized in that: Presetting the sub-data constituting the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit also includes: After assigning values ​​to all sub-data corresponding to the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit, a second formula group is used to calculate the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit; The second formula group is specifically: , , and The sub-data respectively constitute the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit after being assigned; , n, p and They are the number of sub-data constituting the working condition data of the oil pumping unit, the working parameter data of the oil pumping unit, the working environment data of the oil pumping unit and the production data of the oil pumping unit.

6. A method for predicting the efficiency of an oil pumping unit system according to claim 1, characterized in that: The method further comprises: constructing a risk model and training the risk model; inputting the predicted system efficiency of the oil pumping unit obtained by the prediction model into the trained risk model; The risk model determines whether the prediction system efficiency is within a first preset range, and if so, outputs an instruction to execute the oil pump to work, and if not, uses a second preset method to adjust the oil pump.

7. A method for predicting the efficiency of an oil pumping system according to claim 6, characterized in that: The second preset method includes: Obtain the two endpoint values ​​of the first preset range, and determine whether the value of the prediction system efficiency exceeds the two endpoint values. If so, save the input influencing factor data corresponding to the prediction system efficiency to a preset data set; if not, calculate the difference between the prediction system efficiency and the minimum value of the two endpoint values, and process the difference using a third preset method.

8. A method for predicting the efficiency of an oil pumping system according to claim 7, characterized in that: The third preset method includes: It is determined whether the difference is within a second preset range. If so, all pumping units are adjusted based on the data stored in the preset data set; if not, a warning instruction is output.

9. A method for predicting the efficiency of an oil pumping unit system according to claim 1, characterized in that: The method further comprises optimizing the prediction model; Obtaining actual system efficiencies of multiple groups of the pumping units within a set period and placing them into set A; Obtain predicted system efficiencies of multiple groups of the pumping units within the set period through the prediction model and put them into set B; Based on the principle of permutation and combination, the differences between all actual system efficiencies in the set A and all predicted system efficiencies in the set B are calculated in sequence to obtain a fourth calculation result; Based on the fourth calculation result, the prediction model is optimized.

Images