A method for automatic parameter fitting of well test interpretation model

Abstract

The present invention provides a method for automatically fitting parameters of a well test interpretation model, which belongs to the technical field of oil and gas field development, and includes: obtaining the main influencing parameters of the fitting of different well test interpretation models; analyzing and calculating the main influencing parameters of the fitting of different well test interpretation models; performing a primary fitting on all parameters of the well test interpretation model after the analysis and calculation by a genetic algorithm to obtain a fitting result of the primary fitting; performing a secondary fitting on the fitting result of the primary fitting by a Levenberg-Marquarelt algorithm to obtain the fitting parameter result of the well test interpretation model, and obtaining accurate interpretation parameters of the well test interpretation model. The present invention solves the problem of low efficiency and poor quality of well test interpretation caused by human factors and fitting difficulties in the current well test interpretation parameter fitting, can obtain accurate fitting parameters within a limited time, is applicable to a variety of well test interpretation models, has good practicality, and improves the efficiency and quality of well test interpretation.

Description

Technical Field

[0001] The invention relates to the technical field of oil and gas field development, and in particular to an automatic parameter fitting method for a well test interpretation model. Background Art

[0002] Well test interpretation model can evaluate the characteristic parameters of wells and oil and gas reservoirs. As an important part of analyzing the characteristics of oil and gas reservoir models, the well test interpretation model mainly includes two parts: model identification and parameter fitting. Well test interpretation parameter fitting is essentially to seek the best fitting of theoretical curve values ​​and actual curve values ​​under the best parameters and achieve the minimum residual square sum. For well testing, forward models and inverse models can be constructed. The forward model is the pressure transient model function obtained by solving the equation, while the inverse model is the value of the unknown model parameter inferred based on the actual data observed.

[0003] The efficient acquisition of accurate fitting parameters depends on the automatic fitting system of well test interpretation parameters. At present, although there are some relatively mature automatic fitting algorithms, there are still problems such as slow fitting speed, inaccurate fitting results, unstable fitting effect, etc. In general, the current algorithms for automatic fitting of well test interpretation parameters have certain limitations. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a method for automatically fitting parameters of a well test interpretation model, which solves the problems of low efficiency and poor quality of well test interpretation caused by human factors and fitting difficulties in the current well test interpretation parameter fitting. Accurate fitting parameters can be obtained within a limited time, and the method is applicable to a variety of well test interpretation models, has good practicality, and improves the efficiency and quality of the well test interpretation model.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A method for automatically fitting parameters of a well test interpretation model comprises the following steps:

[0007] Obtain the main influencing parameters of different well test interpretation model fitting;

[0008] Analyze and calculate the main influencing parameters of the fitting of the different well test interpretation models;

[0009] All parameters of the well test interpretation model after analysis and trial calculation are initially fitted by genetic algorithm to obtain the fitting results of the initial fitting;

[0010] The fitting result of the initial fitting is subjected to a secondary fitting by using the Levenberg-Marquarelt algorithm to obtain the fitting parameter result of the well test interpretation model, thereby obtaining accurate interpretation parameters of the well test interpretation model.

[0011] Preferably, the main influencing parameters of the fitting of different well test interpretation models are obtained, including:

[0012] According to the theories and characteristics of different well test interpretation models, the parameters of different well test interpretation model fitting are determined; the parameters include permeability, skin coefficient, wellbore storage coefficient, inner zone radius, mobility ratio, pressure conductivity ratio, fracture conductivity, fracture half length, fracture storage ratio, cross-flow coefficient, and boundary distance;

[0013] According to the influence of different well test interpretation model fitting parameters on the flow stage, the main influencing parameters of the different well test interpretation model fittings are obtained.

[0014] Preferably, the main influencing parameters of the fitting of the different well test interpretation models are analyzed and calculated, including:

[0015] Perform wavelet transformation and polishing on the original well test data to obtain a smooth interpretation curve;

[0016] The smooth interpretation curve is randomly segmented in a manner of random segmentation of intervals to obtain characteristic points and flow stage data of main influencing parameter control;

[0017] Based on the flow stage data controlled by the main influencing parameters, the main influencing parameters are analyzed and calculated through DTS technology. If the flow stage data controlled by the main influencing parameters do not exist, the main influencing parameters are analyzed and calculated through error analysis.

[0018] Preferably, all parameters of the well test interpretation model after analysis and trial calculation are initially fitted by a genetic algorithm to obtain the fitting results of the initial fitting, including:

[0019] According to the number of fitting parameters of different well test interpretation models, an adaptive genetic algorithm or a parallel adaptive genetic algorithm is selected;

[0020] Configuring basic parameters in the selected genetic algorithm and dynamically adjusting the selected genetic algorithm;

[0021] Roulette wheel method is used as the selection method, iteration and termination conditions are set in the selected genetic algorithm, and the preliminary fitting results of the genetic algorithm are obtained.

[0022] Preferably, configuring basic parameters in the selected genetic algorithm and dynamically adjusting the selected genetic algorithm includes:

[0023] According to the error expression of the objective function in the selected genetic algorithm, the error value of the objective function is calculated;

[0024] The fitness function is defined by the error value of the objective function;

[0025] The crossover and variation coefficient expressions in the selected genetic algorithm are dynamically adjusted by using the fitness function and setting the population size as basic parameters.

[0026] Preferably, the objective function error expression is:

[0027]

[0028] Where, E is the fitting error; Δp(measured) is the actual observed pressure value; Δp(calculated) is the theoretical calculated pressure value; Δp'(measured) is the actual observed pressure derivative value; Δp'(calculated) is the theoretical calculated pressure derivative value; a is the calculation coefficient, which is 0.2;

[0029] The fitness function is:

[0030] f=E -1 ;

[0031] Among them, f is the fitness function and E is the fitting error.

[0032] Preferably, the crossover and coefficient of variation expressions in the selected genetic algorithm are dynamically adjusted by using the fitness function and setting the population size as the basic parameter, including:

[0033] The population size is set to N×10, where N is the number of basic parameters. The crossover and variation coefficient expressions are:

[0034]

[0035] Among them, P c is the cross coefficient; P m is the coefficient of variation; f is the fitness function; f max is the best fitness value in each generation of the population; f ave is the average fitness value of the population in each generation; f' is the larger fitness value among the individuals undergoing crossover; f" is the fitness value of the individuals undergoing mutation; k1, k2, k3, k4 are all correlation coefficients, k1=k2=0.8, k3=k4=0.1.

[0036] Preferably, the roulette wheel method is used as the selection method, and the iteration and termination conditions are set in the selected genetic algorithm to obtain the preliminary fitting results of the genetic algorithm, including:

[0037] The selection method of genetic algorithm is determined to be roulette method, the maximum number of iterations is set to 15, the minimum number of iterations is set to 5, and the fitting is performed through the well test interpretation model fitting error expression:

[0038]

[0039] Among them, ε M is the model fitting error; N is the number of fitting parameters; a n,c is the fitting parameter value of the optimal individual in the current generation; a n,p are the fitting parameter values ​​of the best individuals of the previous generation;

[0040] The above formula is used to determine the fitting error of the interpretation model of each two adjacent generations of well tests, and the fitting error is set to be less than 0.01 as the termination condition to obtain the preliminary fitting results of the genetic algorithm.

[0041] Preferably, performing a secondary fitting on the fitting result of the initial fitting by using the Levenberg-Marquarelt algorithm to obtain the fitting parameter result of the well test interpretation model, and determining the accurate interpretation parameters of the well test interpretation model, including:

[0042] The fitting result of the initial fitting is used as the initial value of the Levenberg-Marquarelt algorithm to perform secondary parameter fitting. When the relative fitting error is less than 0.01%, the interpretation parameters of the well test interpretation model are finally obtained.

[0043] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0044] The invention provides a parameter automatic fitting method for a well test interpretation model. Based on a selected well test interpretation model, a feature point extraction technology is used and combined with a DTS technology or an error analysis to calculate the main influencing parameters of the fitting. A genetic algorithm or a parallel genetic algorithm is combined according to the number of fitting parameters to obtain a better fitting solution. Then, a Levenberg-Marguarelt algorithm is combined for secondary fitting to obtain accurate interpretation parameters of the model, so that the actual observation value and the theoretical model calculation value are within the allowable error range. At the same time, the method solves the problem of low efficiency and poor quality of well test interpretation caused by human factors and fitting difficulties in the current well test interpretation parameter fitting. Accurate fitting parameters can be obtained within a limited time, and the method is applicable to a variety of well test interpretation models. The method has good practicability and improves the efficiency and quality of the well test interpretation model. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0046] Figure 1 A flow chart of a method for automatically fitting parameters of a well test interpretation model according to the present invention;

[0047] Figure 2 A schematic diagram of a process of extracting data in a flow phase corresponding to the main influencing parameters provided in the first embodiment of the present invention;

[0048] Figure 3 A schematic diagram of the process of obtaining the corresponding stages and feature points of the main influencing parameters provided in the first embodiment of the present invention;

[0049] Figure 4 A comparison chart of permeability and inner zone radius error analysis results provided in Example 1 of the present invention;

[0050] Figure 5 The genetic algorithm fitting result and double logarithmic graph of the first generation optimal solution provided in the first embodiment of the present invention;

[0051] Figure 6 It is the fitting result and double logarithmic graph of the third generation optimal solution of the genetic algorithm provided in the first embodiment of the present invention.

[0052] Figure 7 It is the fitting result and double logarithmic graph of the fifth generation optimal solution of the genetic algorithm provided in the first embodiment of the present invention.

[0053] Figure 8 It is the fitting result and double logarithmic graph of the 7th generation optimal solution of the genetic algorithm provided in the first embodiment of the present invention.

[0054] Fig. 9 It is the fitting result and double logarithmic graph of the 9th generation optimal solution of the genetic algorithm provided in the first embodiment of the present invention.

[0055] Fig.10 It is the fitting result and double logarithmic graph of the 11th generation optimal solution of the genetic algorithm provided in the first embodiment of the present invention.

[0056] Fig.11 This is the final fitting result and double logarithmic graph provided by the first embodiment of the present invention. DETAILED DESCRIPTION

[0057] The technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all of the embodiments. Based on the embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0058] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0059] like Figure 1As shown, the present invention provides a method for automatically fitting parameters of a well test interpretation model, comprising the following steps:

[0060] Step 100, obtaining main influencing parameters of fitting of different well test interpretation models;

[0061] Step 200, analyzing and calculating the main influencing parameters of the fitting of the different well test interpretation models;

[0062] Step 300, performing a preliminary fitting on all parameters of the well test interpretation model after analysis and trial calculation by means of a genetic algorithm, and obtaining a fitting result of the preliminary fitting;

[0063] Step 400: Perform secondary fitting on the fitting result of the initial fitting by using the Levenberg-Marquarelt algorithm to obtain fitting parameter results of the well test interpretation model, thereby obtaining accurate interpretation parameters of the well test interpretation model.

[0064] Specifically, in step 100, the main influencing parameters of the fitting of different well test interpretation models are obtained, including:

[0065] S101. Determine the fitting parameters of different well test interpretation models according to the theories and characteristics of different well test interpretation models; the parameters include permeability, skin coefficient, wellbore storage coefficient, inner zone radius, mobility ratio, pressure conductivity ratio, fracture conductivity, fracture half-length, fracture storage ratio, crossflow coefficient, boundary distance, etc.

[0066] S102. Obtaining main influencing parameters of the fitting of different well test interpretation models according to the influence of the fitting parameters of different well test interpretation models on the flow stage.

[0067] Specifically, in step 200, the main influencing parameters of the fitting of the different well test interpretation models are analyzed and calculated, including:

[0068] S201, performing wavelet transformation and polishing on the original well test data to obtain a smoother interpretation curve;

[0069] Firstly, the original well test data is subjected to wavelet transform, Daubechies 4 (db4) is used as the orthogonal wavelet transform function for wavelet decomposition, Sqtwolog is used as the threshold calculation rule for quantization, and finally wavelet reconstruction is performed to generate denoised pressure data.

[0070] Secondly, the pressure derivative data is obtained to obtain p i In t i The derivative at time satisfies |log(t)-log(t i)|≤d, calculate their difference and then calculate, d is the polishing coefficient, the parameter range is 0-1, and the parameter range is continuously adjusted until the pressure derivative curve is smoother and the correlation coefficient of the polynomial fitting result is higher. The derivative formula is expressed as:

[0071]

[0072] in, is the pressure derivative value at the calculation point; t i is the time corresponding to the calculation point; t1 is the time corresponding to calculation point 1; t2 is the time corresponding to calculation point 2; is the pressure corresponding to the calculation point; is the pressure corresponding to calculation point 1; is the pressure corresponding to calculation point 2; d is the polishing coefficient.

[0073] S202, randomly segmenting the relatively smooth interpretation curve in a manner of random segmentation of intervals to obtain characteristic points and flow stage data of main influencing parameter control;

[0074] S203. Based on the flow stage data controlled by the main influencing parameters, the main influencing parameters are analyzed and calculated by DTS technology. If the flow stage data controlled by the main influencing parameters do not exist, the main influencing parameters are analyzed and calculated by error analysis.

[0075] Finally, randomly select points at intervals from the pressure and pressure derivative curves, control the flow stage according to the main influencing parameters of the interpretation model, obtain the flow stage data corresponding to the parameters, and use the DTS technology to obtain the characteristic points and characteristic lines, calculate the main influencing parameter values, and use error analysis to obtain the parameter range if there is no corresponding data for this stage. Among them, DTS (Direct Type-curve Synthesis) technology refers to the extraction of key characteristic points and characteristic lines from the double logarithmic curve of the pressure derivative by analyzing and processing the well test data, and verifying the effectiveness of these characteristic points and characteristic lines, so as to more accurately build and optimize the well test interpretation model.

[0076] Specifically, in step 300, all parameters of the well test interpretation model after analysis and trial calculation are initially fitted by genetic algorithm to obtain the fitting results of the initial fitting, including:

[0077] S301, selecting an adaptive genetic algorithm or a parallel adaptive genetic algorithm according to the number of fitting parameters of different well test interpretation models;

[0078] S302, configuring basic parameters in the selected genetic algorithm, and dynamically adjusting the selected genetic algorithm; wherein, according to the error expression of the objective function in the selected genetic algorithm, calculating the error value of the objective function; defining the fitness function by the error value of the objective function, and dynamically adjusting the crossover and variation coefficient expressions in the selected genetic algorithm by taking the fitness function and setting the population size as basic parameters.

[0079] First, the objective function error expression is:

[0080]

[0081] Where, E is the fitting error; Δp(measured) is the actual observed pressure value; Δp(calculated) is the theoretical calculated pressure value; Δp'(measured) is the actual observed pressure derivative value; Δp'(calculated) is the theoretical calculated pressure derivative value; a is the calculation coefficient, which is 0.2;

[0082] The fitness function is:

[0083] f=E -1 ;

[0084] Among them, f is the fitness function and E is the fitting error.

[0085] Secondly, by using the fitness function and setting the population size as the basic parameter, the crossover and variation coefficient expressions in the selected genetic algorithm are dynamically adjusted, including:

[0086] The population size is set to N×10, where N is the number of basic parameters. The crossover and variation coefficient expressions are:

[0087]

[0088] Among them, P c is the cross coefficient; P m is the coefficient of variation; f is the fitness function; f max is the best fitness value in each generation of the population; f ave is the average fitness value of the population in each generation; f' is the larger fitness value among the individuals undergoing crossover; f" is the fitness value of the individuals undergoing mutation; k1, k2, k3, k4 are all correlation coefficients, k1=k2=0.8, k3=k4=0.1.

[0089] S303. Finally, the roulette wheel method is used as a selection method to set iteration and termination conditions in the selected genetic algorithm to obtain the preliminary fitting results of the genetic algorithm. Including:

[0090] The selection method of genetic algorithm is determined to be roulette method, the maximum number of iterations is set to 15, the minimum number of iterations is set to 5, and the fitting is performed through the well test interpretation model fitting error expression:

[0091]

[0092] Among them, ε M is the model fitting error; N is the number of fitting parameters; a n,c is the fitting parameter value of the optimal individual in the current generation; a n,p are the fitting parameter values ​​of the best individuals of the previous generation;

[0093] The above formula is used to determine the fitting error of the interpretation model of each two adjacent generations of well tests, and the fitting error is set to be less than 0.01 as the termination condition to obtain the preliminary fitting results of the genetic algorithm.

[0094] Specifically, in step 400, a second fitting is performed on the fitting result of the first fitting by using the Levenberg-Marquarelt algorithm to obtain the fitting parameter result of the well test interpretation model, and determine the accurate interpretation parameters of the well test interpretation model, including:

[0095] S401, taking the fitting result of the initial fitting as the initial value of the Levenberg-Marquarelt algorithm, performing secondary parameter fitting, and finally obtaining the interpretation parameters of the well test interpretation model when the relative fitting error is less than 0.01%.

[0096] Embodiment 1

[0097] The above fitting method is applied to the interpretation of the well test data of a water injection well provided in this embodiment, specifically:

[0098] The test type of this well is the injection section pressure drop test. The basic data of this well are the well radius of 0.07m, the formation thickness of 15.5m, the porosity of 0.158, and the formation volume coefficient of 1m 3 / stm 3 , water viscosity is 1cp, comprehensive compression coefficient is 0.002MPa -1 , with a daily injection volume of 39m 3 / D, the interpretation model of this well is vertical well-fracturing (limited conductivity)-radial composite-infinite. The parameters that need to be fitted include wellbore storage coefficient, skin coefficient, fracture half-length, fracture conductivity, inner zone radius, conductivity-pressure coefficient ratio, mobility ratio, and permeability.

[0099] The data processing process is as follows Figure 2As shown, first, the pressure data is subjected to wavelet transformation to generate denoised pressure data; then the denoised pressure derivative data is polished, and finally, points are randomly selected at intervals on the pressure and pressure derivative curves, and the flow stage data corresponding to the parameters are obtained according to the flow stage controlled by the main influencing parameters of the interpretation model. Figure 3 , combined with DTS technology to obtain characteristic points, calculate the main influencing parameter values, if there is no corresponding data for this stage, use error analysis to obtain the parameter range. Figure 3 The entire processing process provided can obtain the average value of the pressure derivative of the radial flow stage in the inner zone affected by the permeability, which is 0.3296MPa, and the pressure derivative value corresponding to the linear flow stage affected by the fracture half-length in 1hr is 0.4998MPa. Combined with the DTS technology, the permeability is calculated to be 6.51mD and the fracture half-length is 10.0571m. For the inner zone radius, the error analysis is used to obtain the range, refer to Figure 4 , the range of the inner zone radius is (10~100)m.

[0100] The pressure and pressure derivative data of the actual well test interpretation data were used as the fitting target, and the trial calculation results and ranges of the main influencing parameters were used as the initial values ​​and initial ranges to automatically fit the genetic algorithm parameters. During the algorithm fitting process, the termination conditions were met when the 11th generation was reached. Figure 5-10 The fitting results and double logarithmic plots of the 1st, 3rd, 5th, 7th, 9th and 11th generations are shown respectively. The Levenberg-Marquarelt algorithm is used for quadratic fitting again. The number of iterations of the fitting operation is 5 generations. The final fitting results are shown in the figure below. Fig.11 As shown in Figure 2, the entire fitting process only took 20.24 seconds, and the fitting results were reliable. The specific fitting parameter, the wellbore storage coefficient, was 0.2684 m 3 / MPa, the skin coefficient is 0.06, the permeability is 6.02mD, the inner zone radius is 14.9m, the mobility ratio is 2.55, the pressure conductivity ratio is 10.13, the fracture half-length is 8.4m, and the fracture conductivity is 8417.53mD.m.

[0101] Therefore, the above-mentioned automatic parameter fitting method for a well test interpretation model solves the current problems of low efficiency and poor quality of well test interpretation caused by human factors and fitting difficulties in well test interpretation parameter fitting. Accurate fitting parameters can be obtained within a limited time, and it is applicable to a variety of well test interpretation models, has good practicality, and improves the efficiency and quality of the well test interpretation model.

[0102] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core ideas of the present invention. At the same time, for those skilled in the art, according to the ideas of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for automatically fitting parameters of a well test interpretation model, characterized in that: The following steps are involved: Obtaining the main influencing parameters of the fitting of different well test interpretation models; including: determining the parameters of the fitting of different well test interpretation models according to the theories and characteristics of different well test interpretation models; the parameters include permeability, skin coefficient, wellbore storage coefficient, inner zone radius, mobility ratio, pressure conductivity ratio, fracture conductivity, fracture half-length, fracture storage ratio, crossflow coefficient, and boundary distance; obtaining the main influencing parameters of the fitting of different well test interpretation models according to the influence of the fitting parameters of different well test interpretation models on the flow stage; Analyze and calculate the main influencing parameters of the different well test interpretation models; including: performing wavelet transformation and polishing on the original well test data to obtain a smooth interpretation curve; randomly segmenting the smooth interpretation curve in an interval random segmentation manner to obtain characteristic points and flow stage data controlled by the main influencing parameters; based on the flow stage data controlled by the main influencing parameters, analyze and calculate the main influencing parameters through the DTS technology, if there is no flow stage data controlled by the main influencing parameters, analyze and calculate the main influencing parameters through error analysis; wherein, DTS (Direct Type-curve Synthesis) technology refers to extracting key characteristic points and characteristic lines from the double logarithmic curve of the pressure derivative by analyzing and processing the well test data, and verifying the effectiveness of these characteristic points and characteristic lines, so as to more accurately construct and optimize the well test interpretation model; All parameters of the well test interpretation model after analysis and trial calculation are initially fitted by genetic algorithm to obtain the fitting results of the initial fitting; The fitting result of the initial fitting is subjected to a secondary fitting by using the Levenberg-Marquarelt algorithm to obtain the fitting parameter result of the well test interpretation model, thereby obtaining accurate interpretation parameters of the well test interpretation model.

2. The automatic parameter fitting method of a well test interpretation model according to claim 1, characterized in that: The genetic algorithm is used to perform the initial fitting of all parameters of the well test interpretation model after analysis and trial calculation, and the fitting results of the initial fitting are obtained, including: According to the number of fitting parameters of different well test interpretation models, an adaptive genetic algorithm or a parallel adaptive genetic algorithm is selected; Configuring basic parameters in the selected genetic algorithm and dynamically adjusting the selected genetic algorithm; Roulette wheel method is used as the selection method, iteration and termination conditions are set in the selected genetic algorithm, and the preliminary fitting results of the genetic algorithm are obtained.

3. The automatic parameter fitting method of a well test interpretation model according to claim 2, characterized in that: Configure basic parameters in the selected genetic algorithm and dynamically adjust the selected genetic algorithm, including: According to the error expression of the objective function in the selected genetic algorithm, the error value of the objective function is calculated; The fitness function is defined by the error value of the objective function; The crossover and variation coefficient expressions in the selected genetic algorithm are dynamically adjusted by using the fitness function and setting the population size as basic parameters.

4. The automatic parameter fitting method of a well test interpretation model according to claim 3, characterized in that: The objective function error expression is: Where, E is the fitting error; Δp(measured) is the actual observed pressure value; Δp(calculated) is the theoretical calculated pressure value; Δp'(measured) is the actual observed pressure derivative value; Δp'(calculated) is the theoretical calculated pressure derivative value; a is the calculation coefficient, which is 0.2; The fitness function is: f=E -1 ; Among them, f is the fitness function and E is the fitting error.

5. The automatic parameter fitting method of a well test interpretation model according to claim 4, characterized in that: By using the fitness function and setting the population size as the basic parameter, the crossover and variation coefficient expressions in the selected genetic algorithm are dynamically adjusted, including: The population size is set to N×10, where N is the number of basic parameters. The crossover and variation coefficient expressions are: Among them, P c is the cross coefficient; P m is the coefficient of variation; f is the fitness function; f max is the best fitness value in each generation of the population; f ave is the average fitness value of the population in each generation; f ’ is the larger fitness value among the individuals undergoing crossover; f ” is the individual fitness value for mutation; k1, k2, k3, k4 are all correlation coefficients, k1=k2=0.8, k3=k4=0.

1.

6. The automatic parameter fitting method of a well test interpretation model according to claim 2, characterized in that: The roulette wheel method is used as the selection method, and the iteration and termination conditions are set in the selected genetic algorithm to obtain the preliminary fitting results of the genetic algorithm, including: The selection method of genetic algorithm is determined to be roulette method, the maximum number of iterations is set to 15, the minimum number of iterations is set to 5, and the fitting is performed through the well test interpretation model fitting error expression: Among them, ε M is the model fitting error; N is the number of fitting parameters; a n,c is the fitting parameter value of the optimal individual in the current generation; a n,p are the fitting parameter values ​​of the best individuals of the previous generation; The above formula is used to determine the fitting error of the interpretation model of each two adjacent generations of well tests, and the fitting error is set to be less than 0.01 as the termination condition to obtain the preliminary fitting results of the genetic algorithm.

7. The automatic parameter fitting method of a well test interpretation model according to claim 1, characterized in that: Performing a secondary fitting on the fitting result of the initial fitting by using the Levenberg-Marquarelt algorithm to obtain the fitting parameter result of the well test interpretation model, and obtaining accurate interpretation parameters of the well test interpretation model, including: The fitting result of the initial fitting is used as the initial value of the Levenberg-Marquarelt algorithm to perform secondary parameter fitting. When the relative fitting error is less than 0.01%, the interpretation parameters of the well test interpretation model are finally obtained.

Images