Shale oil well yield prediction method and system based on error weight

Abstract

The invention provides a shale oil well yield prediction method and system based on an error weight. The prediction method comprises the steps of obtaining related shale oil well production data; regularization processing is conducted on shale oil well production data, and a Blasingame typical plate is used for recognizing a flowing stage; selecting corresponding m empirical production decline models according to the flowing stage and fitting the shale oil well production data to obtain an empirical production decline model fitting value; according to the empirical production decline model fitting value, calculating an error function between each empirical production decline model and shale oil well production data; and obtaining a new combined model according to the error function of the yield decline model of each experience, and realizing accurate prediction of the yield of the shale oil well. The method not only can be used for predicting shale oil well production data, but also can be popularized and applied to other types of unconventional oil and gas reservoirs, is low in calculation cost, and has wide application value.

Description

Technical Field

[0001] The present invention relates to the technical field of unconventional oil and gas exploitation, and in particular to a shale oil well production prediction method and system based on error weights. Background Art

[0002] The empirical production decline model is a method for studying the dynamic production law of oil wells. Its parameters are easy to solve and can quickly and accurately calculate the recoverable reserves (EUR) of oil wells. It is widely used in major oil and gas reservoirs at home and abroad. Arps (1944) first summarized three classic empirical production decline models based on the decline law of oil and gas wells. Subsequently, many scholars proposed modified Arps models based on the application of Arps models in different blocks, such as: Logistic Growth model (1956), SEPD model (2009), etc. With the depletion of conventional oil reservoirs, unconventional oil reservoirs have become the focus of exploration and development. Scholars have established empirical production decline models based on the flow characteristics of unconventional fluids, such as: PLE model (2008), Duong model (2010), RB-LGM model (2020), etc. In view of the rapid decline stage and slow decline stage of unconventional oil wells, the combined model has become the research focus of the empirical production decline model, such as: Duong+Arps (2018), extended Duong+improved power law exponential model (2021), etc. The combination model generally takes equal production or equal decline rate as the combination point, but there are multiple solutions for the combination point. The essence of the combination model is segmented fitting. The EUR forecast benefits from the accurate fitting of the decline data in the latter segment, and the empirical production decline model fitted in the former segment has little effect. Summary of the invention

[0003] In view of the above problems, the present invention is proposed to provide a shale oil development method and system based on ternary storage and permeability that overcomes the above problems or at least partially solves the above problems.

[0004] According to one aspect of the present invention, a shale oil well production prediction method based on error weight is provided, and the prediction method comprises:

[0005] Step S1: Obtain relevant shale oil well production data;

[0006] Step S2: normalizing the production data of the shale oil well and identifying the flow stage using the Blasingame typical plate;

[0007] Step S3: selecting corresponding m empirical production decline models according to the flow stage and fitting the production data of the shale oil well to obtain the fitting value of the empirical production decline model;

[0008] Step S4: calculating the error function between each empirical production decline model and the shale oil well production data according to the empirical production decline model fitting value;

[0009] Step S5: Based on the error functions of each empirical production decline model, a new combined model is obtained to achieve accurate prediction of shale oil well production.

[0010] Optionally, the shale oil well production data includes production date, wellhead pressure and actual daily oil production.

[0011] Optionally, the step S5: obtaining a new combined model based on the error functions of each empirical production decline model to achieve accurate prediction of shale oil well production specifically includes: performing normalization processing on the error functions of each empirical production decline model to obtain the weights of each empirical production decline model, obtaining a new combined model to achieve accurate prediction of shale oil well production.

[0012] Optionally, in step S2, the production data processing method for identifying the flow stage using the Blasingame typical plate is as follows:

[0013]

[0014] Among them, q o is the actual daily oil production; Q o is the actual cumulative oil production; t c is the material equilibrium time; is the normalized output; p i is the original formation pressure; p wf is the bottom hole flowing pressure;

[0015] is the regularized production integral; is the normalized yield integral derivative.

[0016] Optionally, the step S2 uses the Blasingame typical plate to identify the flow stage by plotting the regularized production, the regularized production integral and the regularized production integral derivative on the Blasingame typical plate, and judging the flow stage of the shale oil well according to the position of the actual data points on the plate.

[0017] Optionally, the error function in step S4 is:

[0018]

[0019] Where: q i is the oil production on the i-th day, i = 1, 2, 3, …, n; is the fitted value of the empirical production decline model on the i-th day; MAE is the mean absolute error; MAPE is the mean absolute percentage error; RMSE is the root mean square error.

[0020] Optionally, the normalization process is performed such that the larger the error function of the empirical yield decline model, the smaller the weight assigned, and the sum of the weights is 1, and the formula is as follows:

[0021]

[0022] Where: er j is the error function value of the jth empirical yield decline model; ar j is the intermediate variable; j is the weight of the jth empirical yield decline model, and

[0023] Optionally, the combination model formula in step S5 is:

[0024]

[0025] Where: q i,comb is the output forecast value of the combined model on the i-th day; ω j is the weight of the jth empirical yield decline model; is the output prediction value of the jth empirical output decline model on the i-th day; m is the number of empirical output decline models selected in step S3.

[0026] Optionally, the empirical production decline model includes: FDC model, Duong model, Matthews-Leflcovits model and Li decline model.

[0027] Optionally, the FDC model is:

[0028]

[0029] Where: m is the maximum daily oil production of shale oil wells in the decline stage, t / d; λ is an undetermined coefficient, describing the trend of the slow decline stage; α is the diffusion coefficient; t is the production time, d.

[0030] Optionally, the Duong model is:

[0031]

[0032] Where, qi is the decreasing initial output, t / d; Q o is the cumulative oil production, t; t is the production time, d; a and m are unknown coefficients, describing the linear relationship between the ratio of production and cumulative production in shale oil wells and time, a is usually less than 0, and m is usually greater than 1.

[0033] Optionally, the Matthews-Leflcovits model is:

[0034]

[0035] Where: qi is the initial production of decline, t / d; t is the production time, d; a is the coefficient to be determined. Optionally, the Li decline model is:

[0036] q=q i ·exp[-λ(lnt) 2 ]

[0037] In the formula, q i is the decreasing initial output, t / d; λ is the decreasing exponent; t is the production time, d.

[0038] The present invention also provides a shale oil well production prediction system based on error weights, which uses the above-mentioned shale oil well production prediction method based on error weights. The prediction system includes:

[0039] A shale oil well production data acquisition module is used to acquire relevant shale oil well production data;

[0040] A regularization processing module, used for regularizing the shale oil well production data and identifying the flow stage using a Blasingame typical plate;

[0041] A data fitting module, used for selecting corresponding m empirical production decline models according to the flow stage and fitting the production data of the shale oil well to obtain the fitting value of the empirical production decline model;

[0042] An error function calculation module, used to calculate the error function of each empirical production decline model and the shale oil well production data according to the empirical production decline model fitting value;

[0043] The combined model calculation module is used to obtain a new combined model based on the error functions of each empirical production decline model to achieve accurate prediction of shale oil well production.

[0044] The present invention provides a shale oil well production prediction method and system based on error weights, the prediction method comprising: step S1: obtaining relevant shale oil well production data; step S2: regularizing the shale oil well production data, and using the Blasingame typical plate to identify the flow stage; step S3: selecting m corresponding empirical production decline models according to the flow stage and fitting the shale oil well production data to obtain the empirical production decline model fitting value; step S4: calculating the error function of each empirical production decline model and the shale oil well production data according to the empirical production decline model fitting value; step S5: obtaining a new combination model according to the error function of each empirical production decline model to achieve accurate prediction of shale oil well production. In addition to being used to predict shale oil well production data, it can also be extended to other types of unconventional oil and gas reservoirs, and the calculation cost is relatively low, and it has a wide range of application value.

[0045] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0047] Figure 1 A flowchart of a shale oil well production prediction method based on error weight provided by an embodiment of the present invention;

[0048] Figure 2 This is the production diagram of Well 1 in FYP Well Area of ​​Shengli Oilfield;

[0049] Figure 3 Schematic diagram for identifying flow phases for the Blasingame method;

[0050] Figure 4 It is a schematic diagram of the fitting effect of four empirical yield decline models;

[0051] Figure 5 It is a comparison diagram of the fitting effect of the actual oil production of four single models and the model of the present invention;

[0052] Figure 6 It is a comparison diagram of the fitting effect of the actual accumulated oil volume of four single models and the model of the present invention;

[0053] Figure 7This is a comparison chart of the mean absolute percentage errors of four single models and the model of the present invention under the validation set. DETAILED DESCRIPTION

[0054] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0055] The terms "comprises" and "having" and any variations thereof in the description embodiments, claims and drawings of the present invention are intended to cover non-exclusive inclusions, for example, including a series of steps or units.

[0056] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.

[0057] Example 1

[0058] like Figure 1 As shown, a shale oil well production prediction method based on error weight, the specific calculation method includes the following steps:

[0059] Step S1: preparing relevant shale oil well production data, wherein the shale oil well production data includes production date, wellhead pressure and actual daily oil production;

[0060] Step S2: Regularize the shale oil well production data and identify the flow stage using the Blasingame typical plate;

[0061] The normalization process in step S2 is the production data processing method proposed by Blasingame under variable bottom hole pressure production, and the formula is as follows:

[0062]

[0063] Where: q o is the actual daily oil production; Q o is the actual cumulative oil production; t c is the material equilibrium time; is the normalized output; p i is the original formation pressure; p wf is the bottom hole flowing pressure; is the regularized production integral; is the regularized yield integral derivative.

[0064] The identification of the flow stage in step S2 is to plot the calculated normalized production, normalized production integral and normalized production integral derivative on the Blasingame typical plate, and determine the flow stage of the shale oil well according to the position of the actual data points on the plate.

[0065] Step S3: selecting corresponding m empirical production decline models according to the flow stage and fitting them to the shale oil well production data;

[0066] Step S4: based on the empirical production decline model fitting value obtained in S3, calculating the error function between each empirical production decline model and the shale oil well production data;

[0067] The error functions in step S4 are mean absolute error, mean absolute percentage error and root mean square error, and the formula is as follows:

[0068]

[0069] Where: q i is the oil production on the i-th day, i = 1, 2, 3, …, n; is the fitted value of the empirical production decline model on the i-th day; MAE is the mean absolute error; MAPE is the mean absolute percentage error; RMSE is the root mean square error.

[0070] Step S5: Based on the error functions of each empirical production decline model obtained in step S4, the function is normalized to obtain the weights of each empirical production decline model, thereby obtaining a new combined model to achieve accurate prediction of shale oil well production.

[0071] The normalization process in step S5 is that the larger the error function of the empirical yield decline model, the smaller the weight assigned, and the sum of the weights is 1, and the formula is as follows:

[0072]

[0073] Where: er j is the error function value of the jth empirical yield decline model; ar j is the intermediate variable; j is the weight of the jth empirical yield decline model, and

[0074] The combined model formula in step S5 is as follows:

[0075]

[0076] Where: q i,comb is the output forecast value of the combined model on the i-th day; ω j is the weight of the jth empirical yield decline model; is the output prediction value of the jth empirical output decline model on the i-th day; m is the number of empirical output decline models selected in step S3.

[0077] Embodiment 2:

[0078] The shale oil well production data used in this experimental example comes from Shengli Oilfield. The data comes from Well 1 in FYP Well Area of ​​Shengli Oilfield. The data sample includes production date, wellhead pressure and actual daily oil production, such as Figure 2 shown.

[0079] First, the production data of shale oil wells are normalized and the processed data are plotted on the Blasingame typical plate to identify the flow stage of well 1, such as Figure 2 :

[0080]

[0081] Where: q o is the actual daily oil production; Q o is the actual cumulative oil production; t c is the material equilibrium time; is the normalized output; p i is the original formation pressure; p wf is the bottom hole flowing pressure; is the regularized production integral; is the normalized yield integral derivative.

[0082] according to Figure 3 , indicating that the flow stage of Well 1 is in the fracture linear flow and transition flow stage. The empirical production decline models applicable to this flow stage are FDC, Duong, Matthews-Leflcovits and Li decline models. The four different empirical production decline models are as follows:

[0083] (1) FDC model:

[0084]

[0085] Where: m is the maximum daily oil production of shale oil wells in the decline stage, t / d; λ is an undetermined coefficient, describing the trend of the slow decline stage; α is the diffusion coefficient; t is the production time, d.

[0086] (2) Duong model

[0087]

[0088] Where: qi is the decreasing initial output, t / d; Q ois the cumulative oil production, t; t is the production time, d; a and m are unknown coefficients, describing the linear relationship between the ratio of production and cumulative production in shale oil wells and time, a is usually less than 0, and m is usually greater than 1.

[0089] (3) Matthews-Leflcovits model

[0090]

[0091] In the formula: qi is the decreasing initial output, t / d; t is the production time, d; a is the unknown coefficient.

[0092] (4) Li decline model

[0093] q=q i ·exp[-λ(lnt) 2 ]

[0094] In the formula: qi is the decreasing initial output, t / d; λ is the decreasing index; t is the production time, d.

[0095] To verify the effectiveness of the present invention, the production data of Well 1 is divided into a fitting set and a validation set, with a ratio of 8:2. The fitting set is used to determine the fitting parameters of the empirical production decline model, and the validation set is used to compare the prediction reliability of different empirical production decline models. In this embodiment, the production data of Well 1 is first fitted using the above four empirical production decline models, and then the error function of each empirical production decline model is calculated, the weight of each model is calculated using the formula and the combined model is obtained, and finally the combined model is compared and analyzed with the four empirical production decline models.

[0096] Schematic diagram of the fitting effect of four empirical yield decline models Figure 4 As shown, the error function formula in step S4 is used to calculate the error functions of the four empirical yield decline models in the fitting set, and the calculation results are shown in Table 1.

[0097] Table 1 Error function calculation results of four empirical production decline models in the fitting set

[0098] Model Name MAE MAPE RMSE FDC 4.6293 0.1072 6.8770 Duong 3.8002 0.0838 6.2967 ML 4.2623 0.0973 6.6403 Li 3.8109 0.0824 7.2296

[0099] The weight calculation formula in step S5 is used to calculate the weights of the four empirical yield decline models under the three error functions. The calculation results are shown in Table 2.

[0100] Table 2 Weight calculation results of four empirical yield decline models

[0101] Model Name MAE MAPE RMSE FDC 0.2213 0.2136 0.2452 Duong 0.2696 0.2734 0.2677 ML 0.2403 0.2352 0.2539 Li 0.2688 0.2778 0.2332

[0102] According to the calculation results in Table 2, three combination models can be obtained. The schematic diagram of the model fitting effect is shown in Figure 5 and Figure 6 As shown, where: combination model 1 represents the combination model obtained by error function MAE, combination model 2 represents the combination model obtained by error function MAPE, and combination model 3 represents the combination model obtained by error function RMSE. Figure 5 and Figure 6 It shows that the prediction results of the combined model of the present invention are more accurate than those of the four single empirical production decline models. Figure 7 The mean absolute percentage errors of four single empirical yield decline models and the combined model of the present invention under the validation set were compared, indicating that the combined model of the present invention is more reliable in prediction.

[0103] Beneficial effects: Based on the Blasingame theory, the flow stage of shale oil wells is identified, the corresponding empirical production decline model is selected and fitted using the flow stage, the error function of each empirical model is calculated, and the error function is normalized to obtain a new combination model, thereby realizing accurate prediction of shale oil well production. In addition to being used to predict shale oil well production data, the present invention can also be extended to other types of unconventional oil and gas reservoirs, and the calculation cost is relatively low, and it has a wide range of application value.

[0104] The above specific implementation methods further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above are only specific implementation methods of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A shale oil well production prediction method based on error weight, characterized in that: The prediction method includes. Step S1: Obtain relevant shale oil well production data; Step S2: normalizing the production data of the shale oil well and identifying the flow stage using the Blasingame typical plate; Step S3: selecting corresponding m empirical production decline models according to the flow stage and fitting the production data of the shale oil well to obtain the fitting value of the empirical production decline model; Step S4: calculating the error function between each empirical production decline model and the shale oil well production data according to the empirical production decline model fitting value; Step S5: Based on the error functions of each empirical production decline model, a new combined model is obtained to achieve accurate prediction of shale oil well production.

2. The shale oil well production prediction method based on error weight according to claim 1 is characterized in that: The shale oil well production data includes production date, wellhead pressure and actual daily oil production.

3. The shale oil well production prediction method based on error weight according to claim 1 is characterized in that: The step S5: obtaining a new combined model based on the error functions of each empirical production decline model to achieve accurate prediction of shale oil well production specifically includes: performing normalization processing on the error functions of each empirical production decline model to obtain the weights of each empirical production decline model, obtaining a new combined model to achieve accurate prediction of shale oil well production.

4. The shale oil well production prediction method based on error weight according to claim 1 is characterized in that: The production data processing method of using the Blasingame typical plate to identify the flow stage in step S2 is as follows: Among them, q o is the actual daily oil production; Q o is the actual cumulative oil production; t c is the material equilibrium time; is the normalized output; p i is the original formation pressure; p wf is the bottom hole flowing pressure; is the regularized production integral; is the normalized yield integral derivative.

5. The shale oil well production prediction method based on error weight according to claim 1 is characterized in that: The step S2 uses the Blasingame typical plate to identify the flow stage by plotting the normalized production, the normalized production integral and the normalized production integral derivative on the Blasingame typical plate, and judging the flow stage of the shale oil well according to the position of the actual data point on the plate.

6. The shale oil well production prediction method based on error weight according to claim 1 is characterized in that: The error function in step S4 is: Where: q i is the oil production on the i-th day, i = 1, 2, 3, …, n; is the fitted value of the empirical production decline model on the i-th day; MAE is the mean absolute error; MAPE is the mean absolute percentage error; RMSE is the root mean square error.

7. The shale oil well production prediction method based on error weight according to claim 3 is characterized in that: The normalization process is that the larger the error function of the empirical yield decline model, the smaller the weight assigned, and the sum of the weights is 1, and the formula is as follows: Where: er j is the error function value of the jth empirical yield decline model; ar j is the intermediate variable; j is the weight of the jth empirical yield decline model, and 8. The shale oil well production prediction method based on error weight according to claim 1 is characterized in that: The combination model formula in step S5 is: Where: q i,comb is the output forecast value of the combined model on the i-th day; ω j is the weight of the jth empirical yield decline model; is the output prediction value of the jth empirical output decline model on the i-th day; m is the number of empirical output decline models selected in step S3.

9. The shale oil well production prediction method based on error weight according to claim 1, characterized in that: The empirical production decline models include: FDC model, Duong model, Matthews-Leflcovits model and Li decline model.

10. The shale oil well production prediction method based on error weight according to claim 9, characterized in that: The FDC model is: Where: m is the maximum daily oil production of shale oil wells in the decline stage, t / d; λ is an undetermined coefficient, describing the trend of the slow decline stage; α is the diffusion coefficient; t is the production time, d.

11. The method for predicting shale oil well production based on error weight according to claim 9, characterized in that: The Duong model is: Where, qi is the decreasing initial output, t / d; Q o is the cumulative oil production, t; t is the production time, d; a and m are unknown coefficients, describing the linear relationship between the ratio of production and cumulative production in shale oil wells and time, a is usually less than 0, and m is usually greater than 1.

12. The method for predicting shale oil well production based on error weight according to claim 9, characterized in that: The Matthews-Leflcovits model is: In the formula: qi is the decreasing initial output, t / d; t is the production time, d; a is the unknown coefficient.

13. The method for predicting shale oil well production based on error weight according to claim 9, characterized in that: The Li decline model is: q=q i ·exp[-λ(lnt) 2 ] In the formula, q i is the decreasing initial output, t / d; λ is the decreasing exponent; t is the production time, d.

14. A shale oil well production prediction system based on error weight, using a shale oil well production prediction method based on error weight as described in any one of claims 1 to 13, characterized in that: The prediction system comprises: A shale oil well production data acquisition module is used to acquire relevant shale oil well production data; A regularization processing module, used for regularizing the shale oil well production data and identifying the flow stage using a Blasingame typical plate; A data fitting module, used for selecting corresponding m empirical production decline models according to the flow stage and fitting the production data of the shale oil well to obtain the fitting value of the empirical production decline model; An error function calculation module, used to calculate the error function of each empirical production decline model and the shale oil well production data according to the empirical production decline model fitting value; The combined model calculation module is used to obtain a new combined model based on the error functions of each empirical production decline model to achieve accurate prediction of shale oil well production.

Images