A method and system for predicting the productivity of fractured solution reservoir

Abstract

The present invention discloses a method for predicting the productivity of fractured-vuggy reservoir oil layers, which includes: according to the well test interpretation data of carbonate fractured-vuggy reservoirs, using the entropy weight method to screen out multiple characteristic parameters related to productivity, and the characteristic parameters include effective permeability, reservoir fluctuation coefficient, and formation coefficient; according to the original data samples of various characteristic parameters, using the improved DBSCAN algorithm to obtain the optimal sample data corresponding to each type of characteristic parameter; using a preset multiple regression equation model, from all the optimal sample data, obtaining a prediction model for predicting the productivity of fractured-vuggy reservoir oil layers; based on the real-time sample data set of a well to be predicted regarding multiple characteristic parameters, using the prediction model to conduct well test interpretation productivity prediction. The present invention realizes real-time, flexible, and accurate prediction of the productivity of fractured-vuggy oil and gas wells, and solves the problem of great difficulty in productivity prediction caused by the inapplicability of conventional seepage theory and the difficulty in obtaining key basic parameters.

Description

Technical Field

[0001] The present invention relates to the technical field of well test interpretation, and more particularly to a method and system for predicting the productivity of fractured-vuggy reservoir oil layers. Background Art

[0002] At present, carbonate fractured-vuggy reservoirs are widely developed in the Tarim Basin. Since the reservoir media of fractured-vuggy reservoirs consist of vugs, fractures, and matrix, among which vugs are the main reservoir spaces; fractures are not only reservoir spaces but also flow channels, and the pressure change laws of the two are different. Due to the good permeability and high production of fractured-vuggy reservoirs, the previous research methods mainly focused on the inter-channeling ability between porous and permeable media are obviously no longer applicable to such oil and gas reservoirs. Currently, developers tend to study more practical reservoir parameters such as the fracture-vug fabric pattern and the development scale of vugs in the reservoir to reasonably formulate development plans.

[0003] Due to the extremely strong heterogeneity of fractured-vuggy reservoirs, the currently applied static methods such as the study of seismic, logging, and mud logging are all affected by accuracy or detection range, and little is known about the reservoir properties outside the wellbore. Therefore, it is of great significance to apply the well test data collected during development in the prior art to study the change characteristics in the wellbore, vugs, and fractures, and explore a method for dynamic productivity prediction that can rely on a large number of well test interpretation parameters to improve the productivity prediction efficiency and accuracy. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides a method for predicting the productivity of fractured-vuggy reservoir oil layers, the method comprising: according to the well test interpretation data of carbonate fractured-vuggy reservoirs, using the entropy weight method to screen out multiple characteristic parameters related to productivity, the characteristic parameters including effective permeability, reservoir fluctuation coefficient, and formation factor; according to the original data samples of each type of the characteristic parameters, using an improved DBSCAN algorithm to obtain the optimal sample data corresponding to each type of characteristic parameter; using a preset multiple regression equation model, obtaining a prediction model for predicting the productivity of fractured-vuggy reservoir oil layers from the optimal sample data of all types of characteristic parameters; and based on the real-time sample data set of a well to be predicted with respect to the multiple characteristic parameters, using the prediction model to perform well test interpretation productivity prediction.

[0005] Preferably, in the step of screening out multiple characteristic parameters related to productivity by using the entropy weight method according to the well test interpretation data of carbonate fracture-vuggy reservoirs and the geological structure of the oilfield area where they are located, the following steps are included: determining multiple original factors related to the productivity of fault-karst reservoir oil layers, and collecting original data samples corresponding to each type of original factor from the well test interpretation data; calculating the information entropy of each type of original factor by using the entropy weight method according to multiple types of original data samples, and then obtaining the entropy weight value of the corresponding type of factor; according to the entropy weight values of all types, taking the original factors corresponding to the entropy weight values that meet the preset preferred range of weight values as the multiple characteristic parameters.

[0006] Preferably, before the step of calculating the information entropy of each type of original factor by using the entropy weight method according to multiple types of original data samples and then obtaining the entropy weight value of the corresponding type of factor, the following steps are also included: sequentially performing standardization and normalization processing on each type of original data sample, so that all sample point data are converted into data with the same dimension level and distributed within the same interval range.

[0007] Preferably, the original factors include: fluctuation coefficient, cavern storage coefficient, effective permeability, formation pressure, formation factor, reservoir permeability, fracture volume, and original formation pressure; the preset preferred range of weight values is 0.2 to 0.3.

[0008] Preferably, in the step of obtaining the optimal sample data corresponding to each type of characteristic parameter by using the improved DBSCAN algorithm according to the original data samples of each type of the characteristic parameters, the following steps are included: setting the Minpts value and Eps value in the improved DBSCAN algorithm, and determining the corresponding density search range; taking the original data samples as a sample set, accessing the sample set according to the density search range and the preset access conditions, and after completing the access to all sample points, taking the currently obtained core points as the optimal sample data.

[0009] In addition, the present invention also provides a system for predicting the productivity of fault-karst reservoir oil layers, and the system includes: a characteristic parameter generation module, which screens out multiple characteristic parameters related to productivity by using the entropy weight method according to the well test interpretation data of carbonate fracture-vuggy reservoirs, and the characteristic parameters include effective permeability, reservoir fluctuation coefficient, and formation factor; an optimal sample generation module, which obtains the optimal sample data corresponding to each type of characteristic parameter by using the improved DBSCAN algorithm according to the original data samples of each type of the characteristic parameters; a model construction module, which uses a preset multiple regression equation model to obtain a prediction model for predicting the productivity of fault-karst reservoir oil layers from the optimal sample data of all types of characteristic parameters; a real-time prediction module, which performs well test interpretation productivity prediction by using the prediction model based on the real-time sample data set of the multiple characteristic parameters of the well to be predicted.

[0010] Preferably, the feature parameter generation module includes: a data acquisition unit that determines a plurality of original factors related to the productivity of the fracture-vug reservoir and collects the original data samples corresponding to each type of original factor from the well test interpretation data; a factor weight calculation unit that calculates the information entropy of each type of original factor by using the entropy weight method based on multiple types of original data samples, and then obtains the entropy weight value of the corresponding type of factor; a parameter screening unit that, based on the entropy weight values of all types, uses the original factors corresponding to the entropy weight values that meet the preset preferred weight range as the plurality of feature parameters.

[0011] Preferably, the feature parameter generation module further includes: a preprocessing unit that sequentially performs standardization and normalization processing on each type of original data sample, so that all sample point data is converted into data with the same dimension level and distributed within the same interval range.

[0012] Preferably, the original factors include: fluctuation coefficient, cavern storage coefficient, effective permeability, formation pressure, formation factor, reservoir permeability, fracture volume, and original formation pressure; the preset preferred weight range is 0.2 to 0.3.

[0013] Preferably, the optimal sample generation module includes: a parameter setting unit that sets the Minpts value and Eps value in the improved DBSCAN algorithm to determine the corresponding density search range; an optimal value search unit that uses the original data samples as a sample set, accesses the sample set according to the density search range and the preset access conditions, and after completing the access to all sample points, uses the currently obtained core points as the optimal sample data.

[0014] Compared with the prior art, one or more of the above embodiments may have the following advantages or beneficial effects:

[0015] The present invention provides a method and system for predicting the productivity of fractured solution reservoir oil layers. The method and system obtain relevant parameters of the fractured solution reservoir from relevant oil and gas wells in the fractured solution oilfield area, select parameters with high correlation with the productivity of the fractured solution reservoir by using the entropy weight method, determine the selected optimal parameters, screen out the optimal data set from the original sample data corresponding to the optimal parameters by improving the DBSCAN clustering algorithm. Based on this, using the multiple regression fitting method, an equation model for predicting the productivity of the fractured solution reservoir is constructed to use this model to predict the dynamic production of single-well well test interpretation in real time. In this way, the present invention forms a productivity prediction method applicable to the fractured solution reservoir, provides reliable productivity prediction for the oilfield, and can greatly reduce the dependence on reservoir model establishment, description of complex fluid flow characteristics, etc. compared with the conventional reservoir productivity prediction method. By means of a large number of in-depth analyses of highly correlated parameters, real-time, flexible and accurate prediction of the productivity of fractured-vuggy oil and gas wells can be achieved, solving the problem of difficult productivity prediction of fractured-vuggy oil and gas reservoirs caused by the inapplicability of conventional seepage theory and the difficulty in obtaining key basic parameters.

[0016] Other advantages, objectives, and features of the present invention will be described to some extent in the subsequent specification, and to some extent, will be obvious to those skilled in the art based on the study of the following text, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the structures specifically pointed out in the following specification, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, and do not constitute a limitation to the present invention. In the drawings:

[0018] Figure 1 It is a step diagram of the method for predicting the productivity of fractured solution reservoir oil layers according to the embodiment of the present application.

[0019] Figure 2 It is a flowchart of the characteristic parameter generation step in the method for predicting the productivity of fractured solution reservoir oil layers according to the embodiment of the present application.

[0020] Figure 3 It is a schematic diagram of the effect during the model construction process in the method for predicting the productivity of fractured solution reservoir oil layers according to the embodiment of the present application.

[0021] Figure 4 It is a block diagram of the system for predicting the productivity of fractured solution reservoir oil layers according to the embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] The embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings and embodiments, so as to fully understand how the present invention uses technical means to solve technical problems and achieve the implementation process of technical effects and implement accordingly. It should be noted that as long as there is no conflict, the various embodiments in the present invention and the various features in each embodiment can be combined with each other, and the formed technical solutions are all within the protection scope of the present invention.

[0023] At present, carbonate fracture-vuggy reservoirs are widely developed in the Tarim Basin. Since the reservoir media of fracture-type oil and gas reservoirs are composed of vugs, fractures and matrix, among which, vugs are the main reservoir spaces; fractures are not only reservoir spaces but also flow channels, and the pressure change laws of the two are different. Due to the good permeability and high production of fracture-vuggy reservoirs, the previous research methods mainly focused on the inter-channeling ability between pore-permeable media are obviously no longer applicable to such oil and gas reservoirs. At present, developers tend to study more practical reservoir parameters such as the fracture-vug structure model of the reservoir and the development scale of vugs in order to reasonably formulate development plans.

[0024] Due to the extremely strong heterogeneity of fracture-vuggy reservoirs, the current static methods such as the research on seismic, logging and mud logging are affected by accuracy or detection range and have little knowledge of the reservoir properties outside the wellbore. Therefore, it is necessary to apply the well test data collected during development in the prior art to study the change characteristics in the wellbore, vugs and fractures, and explore a method for dynamic production prediction that can rely on a large number of well test interpretation parameters, so as to improve the efficiency and accuracy of production prediction, which is of great significance.

[0025] To solve the above technical problems, the present invention proposes a method and system for predicting the production capacity of fault-karst oil reservoirs. The method and system first use the entropy weight method to determine various characteristic parameters that have a greater impact on the production capacity prediction results of fault-karst oil reservoirs, and calculate the entropy weight value corresponding to each type of characteristic parameter; according to the historical sample data of these characteristic parameters, an improved DBSCAN algorithm is used to obtain the optimal sample data corresponding to each type of characteristic parameter; then, using a preset multiple regression equation model, according to the optimal sample data corresponding to each type of characteristic parameter and the corresponding entropy weight value, a prediction model for predicting the production capacity of fault-karst oil reservoirs is obtained; finally, based on the real-time sample data group of multiple characteristic parameters of the well to be predicted, the production capacity is predicted using the prediction model. In this way, the present invention forms a production capacity prediction method applicable to fault-karst reservoirs. Compared with the conventional reservoir production capacity prediction method, it can reduce the dependence on reservoir model establishment, description of complex fluid flow characteristics, etc. By means of a large number of in-depth analyses of highly correlated parameters, it can achieve real-time and flexible accurate prediction of the production capacity of fracture-vuggy oil and gas wells.

[0026] Figure 1 It is a step diagram of the method for predicting the production capacity of fault-karst oil reservoirs according to the embodiments of the present application. Below, refer toFigure 1 , a detailed description of the method for predicting the productivity of fracture-vuggy reservoir oil layers (hereinafter referred to as "productivity prediction method") of the present invention is given. First, in step S110, according to the well test interpretation data of all historical wells in the fracture-vuggy reservoir oil field area where the well to be predicted is located and the geological structure data of the current oil field area, the entropy weight method is used to screen out multiple types of characteristic parameters related to the productivity prediction result of the fracture-vuggy reservoir oil layer, and the entropy weight value of each type of characteristic parameter is calculated.

[0027] In the embodiment of the present invention, the well to be predicted can be a historical well that has been drilled and produced in the current fracture-vuggy reservoir oil field area. For such a well, it is necessary to predict the productivity at a future time. For an oil field area with a fracture-vuggy reservoir, such a reservoir has reservoir bodies with different combinations of matrix, fractures, and fracture-vugs. When predicting productivity, there are many original factors (original elements) that can affect the productivity prediction result of the fracture-vuggy reservoir oil layer. However, among these numerous original factors, only a part of the characteristic factors have a greater impact on the productivity prediction result. Therefore, in step S110, it is necessary to select multiple characteristic factors (characteristic parameters) that are most relevant to the productivity prediction result of the fracture-vuggy reservoir oil layer (factors that have a greater impact on the productivity prediction result).

[0028] Figure 2 This is a flowchart of the characteristic parameter generation step in the method for predicting the productivity of fracture-vuggy reservoir oil layers according to the embodiment of the present application. The following combines Figure 1 and Figure 2 to give a detailed description of the characteristic parameter screening process. First, in step S201, multiple original factors related to the productivity of the fracture-vuggy reservoir oil layer are determined, and the original data samples corresponding to each type of original factor are collected from the currently obtained well test interpretation data. After obtaining the well test interpretation data of the fracture-vuggy reservoir and the geological structure data of the current reservoir area, based on these data, the well test interpretation parameters, formation conditions of the reservoir area, reservoir properties, and well testing conditions that affect the single-well productivity of the current fracture-vuggy reservoir are analyzed to obtain multiple types of original factors. In the embodiment of the present invention, the original factors include the following 8 types: fluctuation coefficient, cavern storage coefficient, effective permeability of formation rock, formation pressure, formation factor, reservoir permeability, fracture volume, and original formation pressure. The cavern storage coefficient represents the ability of the cavern to store oil and gas after shutting in the well.

[0029] Thus, after determining the original factors, it is necessary to collect the original data samples corresponding to each type of original factor from the currently obtained well test interpretation data, so as to enter step S202. Preferably, in the embodiment of the present invention, the original data sample of each type of characteristic parameter is the historical data of the same type of characteristic parameter of multiple historical wells within 10 years in the oil field area where the current fracture-vuggy reservoir is located, where the original data sample of the same type of characteristic parameter includes multiple sample point data.

[0030] Due to the complex underground geological conditions, when optimizing parameters, the conventional method needs to comprehensively consider information such as the structural factors of the current reservoir, reservoir properties, and oil testing conditions. Based on the comprehensive analysis of geological factors, reasonable selection of optimization parameters (parameters most relevant to the production capacity prediction results) is made. However, in the embodiments of the present invention, it is not necessary to conduct a detailed analysis of the geological factors of the current fuse body reservoir. Only the information entropy is calculated according to the historical sample data of each type of original factor, so as to use the information entropy of each type of original factor to quantify the influence degree of this type of factor on the production capacity prediction result, and the entropy weight value is used to represent the quantified influence degree of the current factor on the production capacity prediction result. However, to obtain an accurate calculation result of the quantified influence degree, it is necessary to use step S202 to preprocess the parameters corresponding to different types of original factors.

[0031] In step S202, each type of original data sample is sequentially subjected to standardization and normalization processing, so that all sample point data within the same type of original data sample are converted into the same dimension level and all sample point data are distributed within the same interval range. Specifically, in step S202, first, the original data sample corresponding to each type of original factor is rewritten in matrix form; then, each sample point data within the sample matrix corresponding to each type of original factor is standardized, so that all sample point data within this type of original data sample have the same dimension level (for example: for the original factor of fracture volume type, the unit of all sample point data is unified to m 3 ); then, each sample point data within each sample matrix that has been standardized is further normalized (for example: each sample point data is divided by the maximum sample point data within the corresponding type of sample matrix), so that all sample point data within this type of original data sample are distributed within the same interval range. Thus, each sample point data is converted into a numerical value with the same dimension unit at the same level, and the numerical value is between the intervals [-1, 1].

[0032] After the preprocessing of each type of original data sample is completed, it enters step S203. According to the multiple types of preprocessed original data samples obtained in step S202, the entropy weight method is adopted. First, the information entropy of each type of original factor is calculated, and then, according to the information entropy of each type of original factor, using the entropy weight value calculation formula, the entropy weight value of the corresponding category factor is obtained. In the actual application process, since the technology adopted in the process of calculating the information entropy is relatively mature and the method is not unique, the present invention does not make specific limitations on the method adopted in this process. Among them, the above entropy weight value calculation formula is expressed by the following expression:

[0033]

[0034] In the formula, W j$E_j$ represents the entropy weight value of each type of original factor, $Ej$ represents the calculation result of the information entropy of each type of original factor, and $n$ represents the total number of original factors. In this way, a corresponding quantification value of the influence degree, that is, the entropy weight value, is obtained for each type of original factor.

[0035] For example: Based on the well test interpretation data of oil and gas wells in areas such as Tahe, Yuejin, and Shunbei in the Tarim Basin, and collecting the original data samples of each original factor from them, the entropy weight value corresponding to each original factor is calculated by using the above formula (1). Table 1 shows the entropy weight value table of 8 types of original factors.

[0036] Table 1 Entropy weight value table of each original factor of the fault-karst reservoir

[0037]

[0038] In this way, after obtaining the entropy weight value corresponding to each type of original factor, it enters step S204. Step S204 selects the original factors corresponding to the entropy weight values that meet the preset weight optimization range according to the entropy weight values of all categories, and uses them as multiple characteristic parameters that have a greater impact on the prediction result of the fault-karst reservoir productivity. In the embodiment of the present invention, preferably, the above preset weight optimization range is 0.2 to 0.3. Therefore, the present invention needs to select the original factors that meet the conditions of the above preset weight optimization range according to the weights of the sample data of different original factors to evaluate the productivity of the fault-karst reservoir. Referring to Table 1, according to the entropy weight value data of each original factor, select the fault-karst reservoir parameters with the weight in the range of [0.2, 0.3] for productivity prediction. Further, the selected optimal fault-karst reservoir parameters (characteristic parameters) include: effective permeability, formation factor, and fluctuation coefficient.

[0039] After optimizing multiple characteristic parameters related to the prediction result of the fault-karst oil reservoir productivity, step S110 ends and enters step S120. Step S120 uses the improved DBSCAN algorithm according to the original data samples of each type of characteristic parameter to obtain the optimal sample data corresponding to each type of characteristic parameter. It should be noted that in the embodiment of the present invention, since the optimal sample data screening process of the original data samples of each type of characteristic parameter needs to be processed in the same way. Therefore, in the embodiment of the present invention, the optimal sample data screening process of the original data samples of one type of characteristic parameter is taken as an example to illustrate step S120.

[0040] First, (step S1201) set the Minpts value (the number of samples within the search radius) and the Eps value (the radius) in the improved DBSCAN clustering algorithm to determine the density search range required for the current search process. Then, (step S1202) randomly select an initial access point from the original data samples of the current type of characteristic parameters, use the current historical sample data as the sample set, and starting from the initial access point, access the current sample set according to the set density search range and the preset access conditions. After accessing all sample points, take the currently obtained core point or the set of core points as the optimal sample data (the optimal sample data here may be a single sample point data or multiple sample point data). Thus, the density search clustering process is completed, and the core point or the set of core points is used as the characteristic point of the current category of characteristic parameters (i.e., the optimal sample data of the current category of characteristic parameters), and then enter the above step S130.

[0041] Among them, the above access condition is that the set corresponding to the density search range is set as A, and b and c respectively represent the boundary point and the core point of the set to which the current density search range belongs. When and only when A satisfies the following constraint conditions:

[0042] (1) If b ∈ A and b is density-reachable from b, then c ∈ A;

[0043] (2) If b ∈ A and c ∈ A, then b and c are density-connected.

[0044] In step S130, using the preset multiple regression equation model, from the optimal sample data of all category characteristic parameters and the corresponding entropy weight values, a prediction model for predicting the productivity of the fault dissolution reservoir is obtained. In step S130, according to the optimal sample data of each type of characteristic parameter obtained in step S120, using the multiple regression fitting method, the characteristic points of each parameter are used as the independent variable x m , and the productivity value of the fault dissolution reservoir is used as the dependent variable F, so as to perform a linear fit on the multiple regression equation, and then obtain the above prediction model. Since there is a positive correlation between various types of characteristic parameters and the productivity prediction results, the above multiple regression equation model is expressed by the following expression:

[0045]

[0046] Among them, F represents the prediction result of the productivity of the fault dissolution reservoir, α0, α1, ……, α m respectively represent the fitting coefficients, x1 …… x m respectively represent the optimal sample point data corresponding to different types of characteristic parameters, and σ 2Denote the variance corresponding to the disturbance random error ε (which has a zero mean) following a normal distribution, and N represents the normal distribution symbol. Further, it is assumed here that the random error ε is independent of x1, x2, and x3 and follows a normal distribution with a mean of zero. Then, only the parameter coefficients α need to be calculated. m to determine the relationship of the productivity prediction solution reservoir parameters. Figure 3 This is a schematic diagram of the effect of the model construction process in the method for predicting the productivity of a fault dissolution reservoir in the embodiment of the present application. Numerically fitting the above multiple regression equation model using matlab, as Figure 3 shown, we get α0 = 14.8064, α1 = 0.1210, α2 = -0.0002, α3 = 44.2725. Therefore, in the embodiment of the present invention, preferably, the multiple regression equation model for predicting the productivity of the fault dissolution reservoir is:

[0047] F = 14.8064 + 0.1210x1 + (-0.0002)x2 + 44.2725x3.

[0048] After completing the fitting construction of the prediction model, it enters step S140 to predict the well test interpretation production of the well to be predicted in the oilfield area where the current fuse body reservoir is located. Since in the well test interpretation stage, to predict the productivity of the well to be predicted, it is necessary to obtain the real-time sample data group of the well to be predicted regarding multiple characteristic parameters. Further, step S140, based on the real-time sample data group of the well to be predicted regarding multiple characteristic parameters, inputs the data groups of these 3 types of real-time characteristic parameters (after the above standardization and normalization processing) into the prediction model constructed in step S130 to perform real-time well test interpretation productivity prediction using this regression fitting model and output the productivity prediction result of the current single well regarding the well to be predicted.

[0049] On the other hand, based on the above method for predicting the productivity of a fault dissolution reservoir, the present invention also proposes a system for predicting the productivity of a fault dissolution reservoir (hereinafter referred to as "productivity prediction system"). Figure 4 This is a block diagram of the module for predicting the productivity of a fault dissolution reservoir in the embodiment of the present application. As Figure 4As shown in the figure, the production capacity prediction system of the present invention includes: a characteristic parameter generation module 41, an optimal sample generation module 42, a model construction module 43, and a real-time prediction module 44. Among them, the characteristic parameter generation module 41 is implemented according to the method described in the above step S110, and is configured to screen out multiple characteristic parameters related to the production capacity according to the well test interpretation data of carbonate fracture-vuggy reservoirs by using the entropy weight method. Among them, the characteristic parameters include effective permeability, reservoir fluctuation coefficient, and formation coefficient. The optimal sample generation module 42 is implemented according to the method described in the above step S120, and is configured to obtain the optimal sample data corresponding to each type of characteristic parameter by using an improved DBSCAN algorithm based on the original data samples of various characteristic parameters. The model construction module 43 is implemented according to the method described in the above step S130, and is configured to obtain a prediction model for predicting the production capacity of fault-karst oil layers by using a preset multiple regression equation model from the optimal sample data of all types of characteristic parameters. The real-time prediction module 44 is implemented according to the method described in the above step S140, and is configured to perform well test interpretation production capacity prediction by using the prediction model based on the real-time sample data group of the well to be predicted under multiple characteristic parameters.

[0050] Further, the above-mentioned characteristic parameter generation module 41 includes: a data acquisition unit 411, a preprocessing unit 412, a factor weight calculation unit 413, and a parameter screening unit 414. Among them, the data acquisition unit 411 is configured to determine multiple original factors related to the production capacity of the fault-karst oil layer, and collect the original data samples corresponding to each type of original factor from the well test interpretation data. The preprocessing unit 412 is configured to perform standardization and normalization processing on each type of original data sample in sequence, so that all sample point data is converted into data with the same dimension level and distributed within the same interval range. The factor weight calculation unit 413 is configured to calculate the information entropy of each type of original factor by using the entropy weight method based on multiple types of original data samples, and then obtain the entropy weight value of the corresponding type of factor. The parameter screening unit 414 is configured to use the entropy weight values of all types as the above-mentioned multiple characteristic parameters for the original factors corresponding to the entropy weight values that meet the preset weight optimization range.

[0051] Further, the above-mentioned original factors include: fluctuation coefficient, cavern storage coefficient, effective permeability, formation pressure, formation coefficient, reservoir permeability, fracture volume, and original formation pressure. The above-mentioned preset weight optimization range is 0.2 to 0.3.

[0052] Further, the optimal sample generation module 42 includes: a parameter setting unit 421 and an optimal value search unit 422. Among them, the parameter setting unit 421 is configured to set the Minpts value and the Eps value in the improved DBSCAN algorithm, and determine the corresponding density search range. The optimal value search unit 422 is configured to use the original data samples as a sample set, access the sample set according to the density search range and the preset access conditions, and after all sample points are accessed, use the currently obtained core points as the optimal sample data.

[0053] The present invention discloses a method and a system for predicting the productivity of a fractured solution reservoir. The method and the system obtain the relevant parameters of the fractured solution reservoir from the relevant oil and gas wells in the fractured solution oilfield area, select the parameters with high correlation with the productivity of the fractured solution reservoir by using the entropy weight method, determine the selected optimal parameters, screen out the optimal data set from the original sample data corresponding to the optimal parameters by using the improved DBSCAN clustering algorithm. Based on this, a regression equation model for predicting the productivity of the fractured solution reservoir is constructed by using the multiple regression fitting method, so as to use the model to predict the dynamic production of a single well test interpretation in real time. In this way, the present invention forms a productivity prediction method applicable to the fractured solution reservoir, provides reliable productivity prediction for the oilfield, and compared with the conventional reservoir productivity prediction method, can greatly reduce the dependence on reservoir model establishment, description of complex fluid flow characteristics, etc. By means of a large number of in-depth analyses of highly correlated parameters, the real-time and flexible prediction of the productivity of fractured-vuggy oil and gas wells can be realized, and the problem of difficult productivity prediction of fractured-vuggy oil and gas reservoirs caused by the inapplicability of conventional seepage theory and the difficulty in obtaining key basic parameters is solved.

[0054] In addition, the present invention fits the regression model between each characteristic parameter and the productivity based on a large number of highly correlated data sets between the fractured solution reservoir parameters and the productivity, so as to achieve a better prediction effect of the productivity of the fractured solution reservoir. In addition, the present invention uses the productivity prediction method of the fractured solution reservoir based on the entropy weight method and the DBSCAN clustering algorithm to predict the productivity of the fractured solution reservoirs of 12 wells in blocks such as Tahe, Yuejin, and Shunbei, and good results can be obtained therefrom.

[0055] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for predicting the productivity of a fractured solution reservoir, characterized in that, The method includes: According to the well test interpretation data of carbonate fracture-vuggy reservoirs, using the entropy weight method, multiple characteristic parameters related to productivity are screened out. The characteristic parameters include effective permeability, reservoir fluctuation coefficient, and formation coefficient; According to the original data samples of various types of the characteristic parameters, using the improved DBSCAN algorithm, the optimal sample data corresponding to each type of characteristic parameter is obtained. This includes: setting the Minpts value and Eps value in the improved DBSCAN algorithm, determining the corresponding density search range, and then using the original data samples as the sample set. According to the density search range and the preset access conditions, the sample set is accessed. After all sample points are accessed, the currently obtained core points or the set of core points are used as the optimal sample data; Using a preset multiple regression equation model, taking the optimal sample data of all category characteristic parameters as independent variables and the productivity value of the fault-karst reservoir as the dependent variable, a prediction model for predicting the productivity of the fault-karst oil reservoir is obtained; In the well test interpretation stage of the well to be predicted, based on the real-time sample data group of the well to be predicted regarding the multiple characteristic parameters, the prediction model is used to predict the well test interpretation productivity.

2. The method according to claim 1, characterized in that In the step of screening out multiple characteristic parameters related to productivity according to the well test interpretation data of carbonate fracture-vuggy reservoirs and the geological structure of the oil field area where it is located, using the entropy weight method, it includes: Determining multiple original factors related to the productivity of the fault-karst oil reservoir, and collecting the original data samples corresponding to each type of original factor from the well test interpretation data; According to multiple types of original data samples, using the entropy weight method, calculating the information entropy of each type of original factor, and then obtaining the entropy weight value of the corresponding category factor; According to the entropy weight values of all categories, the original factors corresponding to the entropy weight values that meet the preset weight range are used as the multiple characteristic parameters.

3. The method according to claim 2, wherein Before the step of calculating the information entropy of each type of original factor using the entropy weight method according to multiple types of original data samples and then obtaining the entropy weight value of the corresponding category factor, it also includes: Successively performing standardization and normalization processing on each type of original data sample, so that all sample point data is converted into data with the same dimension level and distributed within the same interval range.

4. According to the method described in claim 2 or 3, characterized in that The original factors include: fluctuation coefficient, cave storage coefficient, effective permeability, formation pressure, formation coefficient, reservoir permeability, fracture volume, and original formation pressure; The preset weight range is 0.2 to 0.

3.

5. A system for predicting the productivity of a fracture-cavity reservoir oil layer, characterized in that, The system includes: A characteristic parameter generation module, which according to the well test interpretation data of carbonate fracture-vuggy reservoirs, uses the entropy weight method to screen out multiple characteristic parameters related to productivity. The characteristic parameters include effective permeability, reservoir fluctuation coefficient, and formation coefficient; An optimal sample generation module, which according to the original data samples of various types of the characteristic parameters, uses the improved DBSCAN algorithm to obtain the optimal sample data corresponding to each type of characteristic parameter; A model construction module, which uses a preset multiple regression equation model, takes the optimal sample data of all category feature parameters as independent variables and the productivity value of the fracture-vug reservoir as the dependent variable, and obtains a prediction model for predicting the productivity of the fracture-vug oil reservoir; A real-time prediction module, which, during the well test interpretation stage of the well to be predicted, uses the prediction model to predict the well test interpretation productivity based on the real-time sample data set of the well to be predicted regarding the multiple feature parameters. Among them, The optimal sample generation module includes: A parameter setting unit, which sets the Minpts value and the Eps value in the improved DBSCAN algorithm to determine the corresponding density search range; An optimal value search unit, which takes the original data sample as the sample set, accesses the sample set according to the density search range and the preset access conditions, and after completing the access to all sample points, takes the currently obtained core point or the set of core points as the optimal sample data.

6. The system according to claim 5, wherein The feature parameter generation module includes: A data acquisition unit, which determines multiple original factors related to the productivity of the fracture-vug oil reservoir and acquires the original data samples corresponding to each type of original factor from the well test interpretation data; A factor weight calculation unit, which calculates the information entropy of each type of original factor using the entropy weight method based on multiple types of original data samples, and then obtains the entropy weight value of the corresponding category factor; A parameter screening unit, which takes the original factors corresponding to the entropy weight values that meet the preset weight range as the multiple feature parameters according to the entropy weight values of all categories.

7. The system according to claim 6, wherein, The feature parameter generation module further includes: A preprocessing unit, which sequentially performs standardization and normalization processing on each type of original data sample, so that all sample point data is converted into data with the same dimension level and distributed within the same interval range.

8. The system according to claim 6 or 7, wherein The original factors include: fluctuation coefficient, cave storage coefficient, effective permeability, formation pressure, formation factor, reservoir permeability, fracture volume, and original formation pressure; The preset weight range is 0.2 to 0.3.

Images