A water well optimization method based on numerical simulation results

Abstract

The present invention discloses a water well tuning method based on digital modeling results, relates to the technical field of digital modeling results, and is used to solve the problem of water well productivity decline caused by the wrong guidance analysis of the extraction degree on the water well productivity. The method comprises the following steps: using the MAD algorithm to screen data and then calculating the corresponding threshold to analyze the degree of water well productivity decline; using the support vector machine to construct an extraction impact assessment model through extraction data; analyzing the degree of influence of extraction on water well productivity according to the results of the extraction impact assessment model; using fuzzy reasoning to evaluate the adjustment level based on the degree of water well productivity decline and the degree of influence of extraction on water well productivity; implementing different adjustment operations for different adjustment levels; collecting extraction adjustment feedback data for the water wells adjusted by extraction; using the logistic regression algorithm to analyze the degree of water well productivity recovery according to the extraction adjustment feedback data; and performing fracturing operations on the water wells with low recovery degrees. Thus, the analysis error caused by the extraction degree is reduced, and the water well productivity guarantee is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of digital modeling results, and more specifically, to a water well tuning method based on digital modeling results. Background Art

[0002] Digital modeling technology mainly analyzes and solves practical problems by establishing mathematical models. Various optimization algorithms can be used to optimize different systems. The application of digital modeling technology in water well tuning systems can improve water well production capacity and reduce human resources.

[0003] The prior art has the following deficiencies:

[0004] In the past, before adjusting the production capacity of a water well, it was necessary to conduct extraction to determine the production capacity of the well. When the production capacity of the well decreased, fracturing operations would be performed to increase the production capacity. The impact of the degree of extraction on the production capacity of the well was not taken into consideration, resulting in excessive fracturing causing damage to the formation and exacerbating the decline in water well production capacity.

[0005] In view of the above problems, the present invention proposes a solution. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a water well optimization method based on digital modeling results, which solves the problems raised in the above-mentioned background technology by comprehensively analyzing the reasons for the decline in water well productivity.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] Step S1, collecting water well data, setting a threshold according to the water well data, screening out water wells with reduced production capacity and marking them;

[0009] Step S2, collecting abnormal data and extraction data of marked water wells, using MAD algorithm to filter data according to abnormal data, and then calculating corresponding thresholds to analyze the degree of decline in water well productivity, using support vector machine to build an extraction impact assessment model based on extraction data, and analyzing the degree of impact of extraction on water well productivity according to the results of extraction impact assessment model;

[0010] Step S3, comprehensively considering the degree of water well productivity decline and the degree of impact of extraction on water well productivity, using fuzzy reasoning to evaluate the adjustment level, performing fracturing operation if the adjustment level is high, and performing extraction adjustment if the adjustment level is moderate;

[0011] Step S4, collecting the extraction adjustment feedback data, analyzing the degree of water well production capacity recovery using a logistic regression algorithm based on the extraction adjustment feedback data, and performing fracturing operations on water wells with low recovery degrees.

[0012] In a preferred embodiment, in step S1, the water well data is the water level difference of the water wells. The water level differences of multiple wells are obtained by calculating the average value and standard deviation, and a threshold is set for the water level difference of the wells to screen the wells. A sample time point is set, and the historical water levels of multiple wells are obtained at the sample time point. Then, the current water levels of the corresponding wells are measured, and the water level differences of multiple wells are obtained by subtracting them in turn, and merged into a water level difference data set.

[0013] In a preferred embodiment, the specific steps of screening water wells are as follows:

[0014] Calculate the mean and standard deviation of the water level difference data set and mark them as a and b, and set the water level difference threshold to a+b; obtain the historical water level of the current well at the sample time point, obtain the water level measurement data of the current well, calculate the water level difference of the current well, and determine whether to mark the current well by comparing the water level difference of the current well with the water level difference threshold;

[0015] If the current water level difference of the well exceeds the water level difference threshold, the current well is marked; if the current water level difference of the well is lower than the water level difference threshold, it is not marked.

[0016] In a preferred embodiment, in step S2, the water well abnormal data and extraction data include bottom hole pressure and water level difference coefficient; the extraction data include heavy metal content in the water well and average power of extraction equipment; the MAD algorithm is used to analyze the degree of decline in water well productivity using the bottom hole pressure and water level difference coefficient, and the extraction impact assessment model is constructed by support vector machine using the heavy metal content in the water well and the average power of extraction equipment to analyze the degree of influence of extraction on water well productivity;

[0017] By setting the sampling time point, the bottom hole pressure, heavy metal content in the water well and the average power of the extraction equipment are collected at the sampling time point, and the water level difference coefficient is obtained by processing the water level difference calculated when screening the water wells.

[0018] In a preferred embodiment, the specific steps of calculating the corresponding threshold after the MAD algorithm screens the data are as follows:

[0019] The bottom hole pressure and water level difference coefficient of sufficient water wells are collected and merged into a pressure data set and a water level difference coefficient data set, marked as F and S respectively;

[0020] Calculate the median: Calculate the median of the pressure data set and the median of the water level difference coefficient data set and record them as f and s;

[0021] Calculate the absolute deviation: for the bottom hole pressure and water level difference coefficient of each water well, calculate the absolute difference between the bottom hole pressure of each water well and the median of the pressure data set, and the absolute difference between the water level difference coefficient of each water well and the median of the water level difference coefficient data set, that is, take the absolute value of the difference between the bottom hole pressure of each water well and the median of the pressure data set, and the water level difference coefficient and the median of the water level difference coefficient data set;

[0022] Calculate the median absolute deviation: Take the median of all absolute differences to get the median absolute deviation MAD. The calculation formula of MAD is as follows:

[0023] , ;

[0024] The bottom hole pressure threshold can be set based on the calculated median absolute deviation (MAD). , Water level difference coefficient threshold ;

[0025] Collect the current water well abnormal data. If the bottom hole pressure of the current water well exceeds the bottom hole pressure threshold And the water level difference coefficient of the current well is lower than the water level difference coefficient threshold , then it is judged that the current water well production capacity decline is low; otherwise, it is judged that the current water well production capacity decline is high.

[0026] In a preferred embodiment, the support vector machine constructs an extraction impact assessment model to analyze the impact of extraction on water well productivity. The specific steps are as follows:

[0027] The heavy metal content of N wells and the average power of the corresponding well extraction equipment are collected and marked as , ;Will and by The two-dimensional vector is constructed by setting N heavy metal content thresholds and average power thresholds respectively. and , construct a two-dimensional vector in the same way , set the binary classification label y as the output result and y is 0 or 1;

[0028] Will The linear kernel function is used for calculation, and the penalty parameter c is set. If the calculation result exceeds the sum of the result calculated by the linear kernel function for N heavy metal content thresholds and the average power threshold and the penalty function, the output result is 1, and the extraction has a high impact on the current water well production capacity; otherwise, the output result is 0, and the extraction has a low impact on the current water well production capacity.

[0029] In a preferred embodiment, the specific calculation process of the linear kernel function is as follows:

[0030] Use the linear kernel function to calculate a linear hyperplane to separate the categories of input data. The formula is:

[0031] ,in are two vectors, Represents the vector inner product, which is obtained by taking the average value of the vector inner product calculated by N heavy metal content thresholds and the average power threshold. , set the penalty parameter c, when The linear kernel function is used to calculate the result of more than When the penalty parameter c is added to the calculated result, the output result y=0, otherwise the output result y=1.

[0032] In a preferred embodiment, in step S3, the degree of decline in water well productivity and the degree of influence of extraction on water well productivity are defined as input variables, which are divided into different fuzzy sets respectively;

[0033] The well adjustment level is defined as the output variable and divided into fuzzy sets;

[0034] Formulate a set of fuzzy rules to describe the impact of different input variables on the output variables;

[0035] Fuzzy reasoning is performed based on fuzzy rules to determine the water well adjustment level.

[0036] In a preferred embodiment, in step S4, the extraction adjustment feedback data includes a bottom hole pressure change amount and a water level difference coefficient change amount, the bottom hole pressure change amount is the difference between the bottom hole pressure of the water well after the extraction adjustment and the bottom hole pressure of the water well before the extraction adjustment, and the water level difference coefficient change amount is the difference between the water level difference coefficient of the water well after the extraction adjustment and the water level difference coefficient of the water well before the extraction adjustment;

[0037] The specific steps of using the logistic regression algorithm to analyze the recovery degree of water well productivity are as follows:

[0038] The calculated bottom hole pressure change and water level difference coefficient change are de-unitized and marked as p and q respectively. , where z is the result of the logistic regression formula, x is the weighted sum of the bottom hole pressure change and the water level difference coefficient change, , and are the weight values ​​of p and q respectively, and the calculated result z is used as the recovery coefficient;

[0039] The bottom hole pressure changes and water level difference coefficient changes of the water well samples are collected respectively, and multiple recovery coefficients are calculated through the logistic regression formula, and the average value of the calculated recovery coefficients is used as the recovery threshold; the bottom hole pressure changes and water level difference coefficient changes of the current water well are collected, and the recovery coefficient of the current water well is calculated by the logistic regression formula and compared with the recovery threshold to determine the recovery degree of the current water well.

[0040] The technical effects and advantages of the water well tuning method based on digital modeling results of the present invention are as follows:

[0041] The present invention collects water well data, analyzes the water well data, selects water wells with reduced production capacity and marks them, collects water well abnormal data and extraction data for the marked water wells, analyzes the degree of reduction in water well production capacity according to the water well abnormal data, analyzes the degree of influence of extraction on the production capacity of the marked water wells according to the extraction data, evaluates the adjustment level by comprehensively considering the degree of reduction in water well production capacity and the degree of extraction influence, performs fracturing operation when the adjustment level is high, does not perform adjustment when the adjustment level is low, performs extraction adjustment on the water well when the adjustment level is moderate, collects extraction adjustment feedback data, analyzes the degree of recovery of water well production capacity according to the extraction adjustment feedback data, and performs fracturing operation on the water well with a low recovery degree. Thus, the analysis error caused by the extraction degree is reduced, and the water well production capacity guarantee is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 The present invention is a schematic diagram of the structure of a water well optimization method based on digital modeling results. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] The present invention collects water well data, analyzes the water well data, selects water wells with reduced production capacity and marks them, collects water well abnormal data and extraction data for the marked water wells, analyzes the degree of reduction in water well production capacity according to the water well abnormal data, analyzes the degree of influence of extraction on the production capacity of the marked water wells according to the extraction data, evaluates the adjustment level by comprehensively considering the degree of reduction in water well production capacity and the degree of extraction influence, performs fracturing operation when the adjustment level is high, does not perform adjustment when the adjustment level is low, performs extraction adjustment on the water well when the adjustment level is moderate, collects extraction adjustment feedback data, analyzes the degree of recovery of water well production capacity according to the extraction adjustment feedback data, and performs fracturing operation on the water well with a low recovery degree. Thus, the analysis error caused by the extraction degree is reduced, and the water well production capacity guarantee is improved. Example

[0045] Step S1, collect water well data, set thresholds based on the water well data, filter out water wells with reduced production capacity and mark them.

[0046] Step S2, collect the abnormal data and extraction data of the marked water wells, use the MAD algorithm to filter the data according to the abnormal data, and then calculate the corresponding threshold to analyze the degree of decline in the water well production capacity, use the support vector machine to build an extraction impact assessment model based on the extraction data, and analyze the impact of extraction on the water well production capacity according to the results of the extraction impact assessment model.

[0047] Step S3, comprehensively considering the degree of water well productivity decline and the impact of extraction on water well productivity, using fuzzy reasoning to evaluate the adjustment level, and performing fracturing operation if the adjustment level is high, and performing extraction adjustment if the adjustment level is moderate.

[0048] Step S4, collecting the extraction adjustment feedback data, analyzing the degree of water well production capacity recovery using a logistic regression algorithm based on the extraction adjustment feedback data, and performing fracturing operations on water wells with low recovery degrees.

[0049] The specific implementation is as follows:

[0050] In step S1, the well data is the well water level difference. The well water level can be measured and obtained by using an electronic water level meter. The well historical data is accessed to obtain the well historical water level. The well water level difference can be obtained by subtracting the well historical water level from the well water level.

[0051] The water level difference of a well is a direct reflection of the well productivity. The larger the water level difference of a well, the higher the relative decline in the well productivity. The water level difference of multiple wells can be calculated to obtain the average value and standard deviation, and the threshold value of the water level difference can be set to screen the wells. The specific steps are as follows:

[0052] Select a historical time point as the sample time point, access the historical data of the well, obtain the historical water levels of multiple wells at the sample time point, use the electronic water level meter to measure the corresponding well water level, make subtractions in turn to obtain the water level differences of multiple wells and merge them into a water level difference data set, calculate the average and standard deviation of the water level difference data set and mark them as a and b, and the water level difference threshold can be set to a+b. Obtain the historical water level of the current well at the sample time point, obtain the measurement data of the current well water level, calculate the current well water level difference, and determine whether to mark the current well by comparing the current well water level difference with the water level difference threshold.

[0053] If the current water level difference of the well exceeds the water level difference threshold, that is, the current water level difference of the well is large, the current water well production capacity decreases relatively, and the current well is marked; if the current water level difference of the well is lower than the water level difference threshold, that is, the current water level difference of the well is small, the current water well production capacity decreases relatively, and it is not marked.

[0054] It should be noted that an electronic water level meter is a common water level detection device. It can obtain water level information through electronic sensors and convert it into electronic signals for display or transmission to other devices. Electronic water level meters can be used to measure the water level of a well. Well historical data usually refers to the long-term water level and water volume change data of a specific well. These data are obtained by regularly sampling and monitoring the well. Well historical data can be accessed to obtain well historical water level information. There is no unique method for setting the water level difference threshold, and the above method is only an example.

[0055] In step S2, the abnormal data of the water well are marked as the bottom hole pressure and the water level difference coefficient, and the extraction data includes the heavy metal content of the water well and the average power of the extraction equipment. The bottom hole pressure can be obtained by using a pressure sensor, and the water level difference coefficient can be calculated based on the current water level difference of the water well; the heavy metal content of the water well can be captured by an inductively coupled plasma emission spectrometer and the heavy metal content data of the water well can be sent to the data receiving device, and the average power of the extraction equipment can be obtained by accessing the industrial PLC control system to obtain the average value of the power of each device.

[0056] The greater the bottom hole pressure, that is, the greater the reserves in the well, the lower the decline in the well productivity; the water level difference coefficient is the result of normalizing the water level difference of the well. The smaller the water level difference coefficient, that is, the smaller the change before and after the well, the lower the decline in the well productivity. There are many methods for normalization processing, which will not be analyzed here.

[0057] The heavy metal content in a water well reflects the water quality of the well. When the water well production capacity decreases, the more frequent the extraction, the higher the heavy metal content, that is, the greater the impact of extraction on the water well production capacity; the average power of the extraction equipment is the electrical power of the extraction equipment when extracting the water well. When the water well production capacity decreases, the greater the average power of the extraction equipment and the more frequent the extraction, that is, the greater the impact of extraction on the water well production capacity.

[0058] For abnormal water well data, the MAD algorithm can be used to filter the data and then calculate the corresponding threshold. This example uses the bottom hole pressure as an example:

[0059] The bottom hole pressures of sufficient water wells are collected and combined into a pressure data set labeled F.

[0060] Calculate the median: Calculate the median of the pressure data set and record it as f.

[0061] Calculate the absolute deviation: For the bottom hole pressure of each water well, calculate the absolute difference between the bottom hole pressure of each water well and the median of the pressure data set, that is, take the absolute value of the difference between the bottom hole pressure of each water well and the median of the pressure data set.

[0062] Calculate the median absolute deviation: Take the median of the absolute difference between all bottom hole pressures and the median of the pressure data set to obtain the median absolute deviation MAD. The calculation formula of MAD is as follows:

[0063] ,in The pressure data set data, You can choose 1, 2, 3, etc. is the median of the pressure data set.

[0064] The bottom hole pressure threshold can be set based on the calculated median absolute deviation MAD, which is calculated according to the formula: .in is the bottom hole pressure threshold, and f is the median of the pressure data set.

[0065] Similarly, the water level difference coefficient threshold can be calculated as By collecting the abnormal data of the current well and comparing it with the calculated threshold, the current well productivity decline is analyzed. If the bottom hole pressure of the current well exceeds the bottom hole pressure threshold, And the water level difference coefficient of the current well is lower than the water level difference coefficient threshold , then it is judged that the current water well production capacity decline is low; otherwise, it is judged that the current water well production capacity decline is high.

[0066] When analyzing the impact of extraction on water well productivity, we first collected the heavy metal contents of N water wells and marked them as , and then collect the average power of the corresponding water well extraction equipment and mark them as follows: , the extraction impact assessment model can be constructed by the support vector machine method. The specific steps are as follows:

[0067] Will and by ... The two-dimensional vector is constructed by setting N heavy metal content thresholds and average power thresholds respectively. and , construct a two-dimensional vector in the same way , set the binary classification label y as the output result and y is 0 or 1. The binary classification label is used as the output result, where y=1 represents a positive example, that is, when the output result is 1, it is judged that the impact of extraction on the water well productivity is high; y=0 represents a negative example, when the output result is 0, it is judged that the impact of extraction on the water well productivity is low.

[0068] The fit algorithm can be used to train the model according to the input vector Set the value of the output result y. Use the linear kernel function to calculate a linear hyperplane to separate the categories of the input data. The formula can be:

[0069] ,in and are two vectors, Represents the vector inner product, which is obtained by taking the average value of the vector inner product calculated by N heavy metal content thresholds and the average power threshold. , set the penalty parameter c, when The linear kernel function is used to calculate the result of more than When the calculation result of is added with the penalty parameter c, the output result y=0, otherwise the output result y=1. The penalty parameter c can be set according to the actual situation and will not be described here.

[0070] The current heavy metal content in the water well and the average power of the extraction equipment are collected through the extraction impact assessment model. If the output result y=0, the extraction has a low impact on the current water well production capacity; if the output result y=1, the extraction has a high impact on the current water well production capacity.

[0071] It should be noted that the pressure sensor is an electronic device used to measure pressure. It can measure pressure by converting pressure changes into electrical signals. The pressure sensor can be used to measure the bottom pressure of a water well. The inductively coupled plasma emission spectrometer is a device for detecting a variety of heavy metals. It can detect the specific content of various heavy metals in the water well and calculate the comprehensive index of heavy metals in the form of a percentage. The industrial PLC control system integrates a power measurement module, which can automatically connect to equipment such as water pumps, automatically collect power data, and display the measurement results on the human-machine interface of the PLC. The MAD algorithm is a statistical method for detecting outliers. This method uses the median and absolute median deviation to identify outliers in a data set, and can also be used to set thresholds. Support vector machine is a machine learning algorithm that can build relevant models for classification kernel regression analysis. The threshold setting method in the above steps is not unique and can be adjusted according to actual conditions.

[0072] In step S3, the water well adjustment level is evaluated by fuzzy reasoning based on the degree of water well productivity decline and the impact of extraction on water well productivity. The specific steps are as follows:

[0073] The degree of water well productivity decline and the impact of extraction on water well productivity are defined as input variables, which are divided into different fuzzy sets.

[0074] For example, "High" and "Low" represent the degree of decline in water well productivity, and "Big" and "Small" represent the degree of impact of extraction on water well productivity.

[0075] Define the water well adjustment level as the output variable and divide it into fuzzy sets, for example, "High", "Medium", "Low" for the water well adjustment level.

[0076] Formulate a set of fuzzy rules to describe the impact of different input variables on output variables. Fuzzy rules can be formulated according to actual conditions, for example:

[0077] The degree of water well productivity decline is marked as Sd, the degree of impact of extraction on water well productivity is Cd, and the water well adjustment level is marked as Su.

[0078] Rule 1: IF (Sd is High) AND (Cd is Small) THEN (Su is High)

[0079] Rule 2: IF (Sd is High) AND (Cd is Big) THEN (Su is Medium)

[0080] Rule 3: IF (Sd is Low) AND (Cd is Small) THEN (Su is Low) ...

[0081] Fuzzy reasoning is performed according to fuzzy rules to determine the water well adjustment level. When the output result is "High", it is judged that the current water well adjustment level is high, and the current water well is subjected to fracturing operation; when the output result is "Medium", it is judged that the current water well adjustment level is moderate, and the current water well is subjected to extraction adjustment; when the output result is "Low", it is judged that the current water well adjustment level is low, and the current water well is not subjected to adjustment.

[0082] It should be noted that the division of fuzzy sets can be adjusted according to actual conditions. The degree of comprehensive well productivity decline and the degree of impact of extraction on well productivity can be set by the judgment result in step 2. For example, when it is judged that the current well productivity decline is high, it is marked as "High", and when it is judged that the degree of impact of extraction on the current well productivity is high, it is marked as "Big". Fracturing operation is a technology to improve the recovery rate of water wells. It improves the fluidity and production capacity of water wells by artificially creating cracks in the formation. Extraction adjustment can be adjusted by setting a downward adjustment ratio to reduce the extraction frequency or reduce the extraction volume, etc., which will not be elaborated here.

[0083] In step S4, if the current water well adjustment level is moderate, the extraction adjustment feedback data is collected after the current water well is extracted and adjusted, and the degree of water well production capacity recovery is analyzed based on the extraction adjustment feedback data. It is determined whether the water well needs fracturing operation based on the degree of water well production capacity recovery.

[0084] The feedback data of the extraction adjustment includes the change in bottom hole pressure and the change in water level difference coefficient. The change in bottom hole pressure is the difference between the bottom hole pressure of the water well after the extraction adjustment and the bottom hole pressure of the water well before the extraction adjustment. The greater the change in bottom hole pressure, that is, the greater the energy storage enhancement of the water well, the higher the recovery degree of the water well.

[0085] The change in water level difference coefficient is the difference between the water level difference coefficient of the well after the extraction adjustment and the water level difference coefficient of the well before the extraction adjustment. The greater the change in water level difference coefficient, the higher the water level change, and the higher the recovery degree of the well. The bottom hole pressure and water level difference coefficient of the well before the extraction adjustment can be obtained by accessing the historical data of the well.

[0086] The calculated bottom hole pressure change and water level difference coefficient change are de-unitized and marked as p and q respectively. , where z is the result of the logistic regression formula, x is the weighted sum of the bottom hole pressure change and the water level difference coefficient change, , and are the weight values ​​of p and q respectively, and their specific values ​​can be set according to actual conditions, such as = = 0.5. The calculated result z is taken as the recovery coefficient.

[0087] Take enough water wells as water well samples, collect the bottom hole pressure changes and water level difference coefficient changes of the water well samples, calculate multiple recovery coefficients through the logistic regression formula, and calculate the average value of the recovery coefficient as the recovery threshold. Collect the bottom hole pressure changes and water level difference coefficient changes of the current water well, calculate the current water well recovery coefficient using the logistic regression formula and compare it with the recovery threshold. If the current water well recovery coefficient exceeds the recovery threshold, it is judged that the current water well has a high degree of recovery and the current water well is not adjusted; if the current water well recovery coefficient is lower than the recovery threshold, it is judged that the current water well has a low degree of recovery and the current water well is subjected to fracturing operation.

[0088] It should be noted that there is no unique method for analyzing the degree of water well recovery, and the parameters involved in the formulas of each step are all calculated as denormalized values.

[0089] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0090] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application of the technical solution and the invention constraints. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0091] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0092] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0093] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A water well tuning method based on numerical simulation results, characterized in that: The following steps are involved: Step S1, collecting water well data, setting a threshold according to the water well data, screening out water wells with reduced production capacity and marking them; Step S2, collecting abnormal data and extraction data of marked water wells, using MAD algorithm to filter data according to abnormal data, and then calculating corresponding thresholds to analyze the degree of decline in water well productivity, using support vector machine to build an extraction impact assessment model based on extraction data, and analyzing the degree of impact of extraction on water well productivity according to the results of extraction impact assessment model; Step S3, comprehensively considering the degree of water well productivity decline and the degree of impact of extraction on water well productivity, using fuzzy reasoning to evaluate the adjustment level, performing fracturing operation if the adjustment level is high, and performing extraction adjustment if the adjustment level is moderate; Step S4, collecting the extraction adjustment feedback data, analyzing the degree of water well productivity recovery using a logistic regression algorithm based on the extraction adjustment feedback data, and performing fracturing operations on water wells with low recovery degrees; In step S2, the abnormal data of the water well and the extraction data include the bottom hole pressure and the water level difference coefficient; the extraction data include the heavy metal content of the water well and the average power of the extraction equipment; the MAD algorithm is used to analyze the degree of decline in the water well productivity by using the bottom hole pressure and the water level difference coefficient, and the extraction impact assessment model is constructed by the support vector machine using the heavy metal content of the water well and the average power of the extraction equipment to analyze the degree of influence of extraction on the water well productivity; By setting the sampling time point, the bottom hole pressure, the heavy metal content of the water well and the average power of the extraction equipment are collected at the sampling time point, and the water level difference coefficient is obtained by processing the water level difference calculated when screening the water wells; The specific steps of constructing the extraction impact assessment model using support vector machine to analyze the impact of extraction on water well productivity are as follows: The heavy metal content of N wells and the average power of the corresponding well extraction equipment are collected and marked as , ;Will and by The two-dimensional vector is constructed by setting N heavy metal content thresholds and average power thresholds respectively. and , construct a two-dimensional vector in the same way , set the binary classification label y as the output result and y is 0 or 1; Will The linear kernel function is used for calculation, and the penalty parameter c is set. If the calculation result exceeds the sum of the result calculated by the linear kernel function and the penalty function for N heavy metal content thresholds and the average power threshold, the output result is 1, and the extraction has a high impact on the current water well production capacity; Otherwise, the output result is 0, and the impact of extraction on the current well production capacity is low; In step S3, the degree of water well productivity decline and the degree of impact of extraction on water well productivity are defined as input variables, and they are divided into different fuzzy sets respectively; The well adjustment level is defined as the output variable and divided into fuzzy sets; Formulate a set of fuzzy rules to describe the impact of different input variables on the output variables; Fuzzy reasoning is performed based on fuzzy rules to determine the water well adjustment level.

2. A water well optimization method based on numerical simulation results according to claim 1, characterized in that: In step S1, the water well data is the water level difference of the water wells. The water level differences of multiple wells are obtained by calculating the average value and standard deviation, and a threshold is set for the water level difference of the wells to screen the wells. A sample time point is set, and the historical water levels of multiple wells are obtained at the sample time point. Then, the current water levels of the corresponding wells are measured, and the water level differences of multiple wells are obtained by subtracting them in turn and merged into a water level difference data set.

3. A water well tuning method based on numerical simulation results according to claim 2, characterized in that: The specific steps for screening water wells are as follows: Calculate the mean and standard deviation of the water level difference data set and mark them as a and b, and set the water level difference threshold to a+b; obtain the historical water level of the current well at the sample time point, obtain the water level measurement data of the current well, calculate the water level difference of the current well, and determine whether to mark the current well by comparing the water level difference of the current well with the water level difference threshold; If the water level difference of the current well exceeds the water level difference threshold, the current well is marked; If the current water level difference of the well is lower than the water level difference threshold, no marking is performed.

4. A water well optimization method based on numerical simulation results according to claim 1, characterized in that ； The specific steps for calculating the corresponding threshold after the MAD algorithm screens the data are as follows: The bottom hole pressure and water level difference coefficient of sufficient water wells are collected and merged into a pressure data set and a water level difference coefficient data set, marked as F and S respectively; Calculate the median: Calculate the median of the pressure data set and the median of the water level difference coefficient data set and record them as f and s; Calculate the absolute deviation: for the bottom hole pressure and water level difference coefficient of each water well, calculate the absolute difference between the bottom hole pressure of each water well and the median of the pressure data set, and the absolute difference between the water level difference coefficient of each water well and the median of the water level difference coefficient data set, that is, take the absolute value of the difference between the bottom hole pressure of each water well and the median of the pressure data set, and the water level difference coefficient and the median of the water level difference coefficient data set; Calculate the median absolute deviation: Take the median of all absolute differences to get the median absolute deviation MAD. The calculation formula of MAD is as follows: 、 ； The bottom hole pressure threshold is set based on the calculated median absolute deviation (MAD). , Water level difference coefficient threshold ; Collect the current water well abnormal data. If the bottom hole pressure of the current water well exceeds the bottom hole pressure threshold And the water level difference coefficient of the current well is lower than the water level difference coefficient threshold , it is judged that the current water well productivity decline is low; Otherwise, it is judged that the current water well productivity decline is high.

5. The water well optimization method based on numerical simulation results according to claim 1 is characterized by: The specific calculation process of the linear kernel function is as follows: Use the linear kernel function to calculate a linear hyperplane to separate the categories of input data. The formula is: ,in are two vectors, Represents the vector inner product, which is obtained by taking the average value of the vector inner product calculated by N heavy metal content thresholds and the average power threshold. , set the penalty parameter c, when The linear kernel function is used to calculate the result of more than When the penalty parameter c is added to the calculated result, the output result y=0, otherwise the output result y=1.

6. The water well optimization method based on numerical simulation results according to claim 1 is characterized by: In step S4, the extraction adjustment feedback data includes a bottom hole pressure change amount and a water level difference coefficient change amount, the bottom hole pressure change amount is the difference between the bottom hole pressure of the water well after the extraction adjustment and the bottom hole pressure of the water well before the extraction adjustment, and the water level difference coefficient change amount is the difference between the water level difference coefficient of the water well after the extraction adjustment and the water level difference coefficient of the water well before the extraction adjustment; The specific steps of using the logistic regression algorithm to analyze the recovery degree of water well productivity are as follows: The calculated bottom hole pressure change and water level difference coefficient change are de-unitized and marked as p and q respectively. , where z is the result of the logistic regression formula, x is the weighted sum of the bottom hole pressure change and the water level difference coefficient change, , and are the weight values ​​of p and q respectively, and the calculated result z is used as the recovery coefficient; The bottom hole pressure change and water level difference coefficient change of the water well samples are collected respectively to calculate multiple recovery coefficients through the logistic regression formula, and the average value of the calculated recovery coefficients is used as the recovery threshold; The bottom hole pressure change and water level difference coefficient change of the current water well are collected, and the current water well recovery coefficient is calculated using the logistic regression formula and compared with the recovery threshold to determine the current water well recovery degree.

Images