Regional oil and gas yield prediction model training method, device and equipment

Abstract

The invention relates to the field of regional oil and gas yield prediction, and discloses a training method, device and equipment of a regional oil and gas yield prediction model, which can respectively determine the first N time points of a target prediction time point as sample time points, and obtain the parameter value of each key parameter at each sample time point. And obtaining the regional oil and gas yield at each sample time point. And for any sample time point, preprocessing the parameter value of each key parameter at the sample time point to obtain a preprocessed parameter value, and according to a set sequence, arranging the regional oil and gas yield at the sample time point and the preprocessed parameter value of each key parameter at the sample time point to obtain a corresponding parameter value array. And training the model by taking the parameter value array corresponding to each sample time point as training data. According to the trained regional oil and gas yield prediction model, reliable prediction of the regional oil and gas yield can be realized, and reliable prediction modes of the regional oil and gas yield are diversified.

Description

Technical Field

[0001] The present invention relates to the technical field of regional oil and gas production prediction, and in particular to a training method, device and equipment for a regional oil and gas production prediction model. Background Art

[0002] Accurate and efficient regional oil and gas production forecasting is the fundamental basis for establishing oilfield regional development strategies, project adjustment decisions, and economic investment portfolios.

[0003] Given that regional production covers many oil and gas fields, the development stages and levels of each oil field are inconsistent, and it is affected by multiple factors such as exploration resources and reserve quality, prediction is difficult, and there are few reliable regional oil and gas production prediction methods in related technologies. Therefore, a reliable regional oil and gas production prediction method is urgently needed. Summary of the invention

[0004] The present invention provides a training method, device and equipment for a regional oil and gas production prediction model, which are used to solve the defect that there are few reliable regional oil and gas production prediction methods in the related technology and diversify the regional oil and gas production prediction methods.

[0005] In a first aspect, the present invention provides a method for training a regional oil and gas production prediction model, comprising: According to the oil and gas production related data of the target area, multiple key parameters having an important impact on the regional oil and gas production prediction are determined from among the multiple regional oil and gas production related parameters set; and N time points before the target prediction time point are respectively determined as a sample time point; wherein N is an integer greater than 1; Obtaining the parameter value of each of the key parameters at each of the sample time points, and obtaining the regional oil and gas production at each of the sample time points; For any of the sample time points, preprocess the parameter value of each of the key parameters at the sample time point to obtain the preprocessed parameter value of each of the key parameters at the sample time point, and arrange the regional oil and gas production at the sample time point and the preprocessed parameter value of each of the key parameters at the sample time point according to the set parameter sequence to obtain a parameter value array corresponding to the sample time point; The parameter value array corresponding to each of the sample time points is used as training data, and is input into the regional oil and gas production prediction model to be trained to predict the parameter value array corresponding to the target prediction time point, and the corresponding prediction result is obtained. The regional oil and gas production prediction model to be trained is updated according to the difference between the prediction result and the corresponding true result to obtain a trained regional oil and gas production prediction model.

[0006] Optionally, the method of determining, based on the oil and gas production related data of the target area, a plurality of key parameters having an important impact on the regional oil and gas production prediction from among the set plurality of regional oil and gas production related parameters, includes: According to the oil and gas production related data of the target area, among the set multiple regional oil and gas production related parameters, a plurality of first parameters whose correlation with the regional oil and gas production exceeds a preset correlation threshold are determined; Determine an influence score of each of the first parameters on target model performance, where the target model performance is the model performance of the regional oil and gas production prediction model to be trained; According to the impact score of each of the first parameters on the target model performance, determine, among the plurality of first parameters, a plurality of second parameters whose impact scores are not less than a first preset score and not greater than a second preset score; Each of the second parameters is respectively used as a key parameter.

[0007] Optionally, preprocessing the parameter value of each of the key parameters at the sample time point to obtain the preprocessed parameter value of each of the key parameters at the sample time point includes: Normalizing the impact score of each second parameter on the target model performance to a preset interval, obtaining a normalized score corresponding to each second parameter, and using the scores as the impact weights of the corresponding key parameters on regional oil and gas production; Performing differential processing on the parameter value of each of the key parameters at the sample time point to obtain the differential value of each of the key parameters at the sample time point; Normalizing the differential value of each of the key parameters at the sample time point to obtain a normalized differential value of each of the key parameters at the sample time point; The normalized difference value of each of the key parameters at the sample time point is multiplied by the corresponding influence weight to obtain the preprocessed parameter value of each of the key parameters at the sample time point.

[0008] Optionally, performing differential processing on the parameter value of each of the key parameters at the sample time point to obtain the differential value of each of the key parameters at the sample time point includes: For any of the key parameters, the parameter value of the key parameter at the sample time point is subtracted from the parameter value of the key parameter at the sample time point before the sample time point to obtain the differential value of the key parameter at the sample time point.

[0009] Optionally, normalizing the differential value of each of the key parameters at the sample time point to obtain the normalized differential value of each of the key parameters at the sample time point includes: For any of the key parameters, determine the maximum differential value and the minimum differential value corresponding to the key parameter, normalize the differential value of the key parameter at the sample time point according to the maximum differential value and the minimum differential value, and obtain the normalized differential value of the key parameter at the sample time point.

[0010] Optionally, when the time point is year, the multiple regional oil and gas production-related parameters include at least two of the annual newly added geological reserves, remaining geological reserves, proportion of low-grade reserves, number of oil wells opened, number of water injection wells opened, water content, single well water drive recoverable reserves and oil price.

[0011] Optionally, when N is 2, after obtaining the trained regional oil and gas production prediction model, the method further includes: Obtain parameter value arrays corresponding to the T-2th time point and the T-1th time point, and input the parameter value arrays corresponding to the T-2th time point and the T-1th time point into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the parameter value array corresponding to the Tth time point; wherein T is the time point at which the regional oil and gas production needs to be predicted; The parameter value array corresponding to the T-1th time point and the parameter value array corresponding to the Tth time point are input into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the parameter value array corresponding to the T+1th time point, until the parameter value array corresponding to the T+mth time point is obtained; wherein m is an integer not less than 1.

[0012] Optionally, when N is 2, after obtaining the trained regional oil and gas production prediction model, the method further includes: Obtain parameter value arrays corresponding to the T-2th time point and the T-1th time point, and input the parameter value arrays corresponding to the T-2th time point and the T-1th time point into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the first parameter value array corresponding to the Tth time point; wherein T is the time point at which the regional oil and gas production needs to be predicted; Determining the predicted regional oil and gas production at the Tth time point in the first parameter value array corresponding to the Tth time point; Acquire the true parameter value of each of the key parameters at the Tth time point, and preprocess the true parameter value of each of the key parameters at the Tth time point to obtain the preprocessed parameter value of each of the key parameters at the Tth time point; Arranging the predicted regional oil and gas production at the Tth time point and the preprocessed parameter value of each key parameter at the Tth time point according to the parameter sequence to obtain a second parameter value array corresponding to the Tth time point; The parameter value array corresponding to the T-1th time point and the second parameter value array corresponding to the Tth time point are input into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the first parameter value array corresponding to the T+1th time point, until the first parameter value array corresponding to the T+mth time point is obtained, and the predicted regional oil and gas production at the T+mth time point is determined; wherein m is an integer not less than 1.

[0013] In a second aspect, the present invention provides a training device for a regional oil and gas production prediction model, comprising: A first determination unit is used to determine, based on the oil and gas production related data of the target area, a plurality of key parameters having an important influence on the regional oil and gas production prediction from among the set plurality of regional oil and gas production related parameters; A second determining unit, configured to determine each of the N time points before the target prediction time point as a sample time point; wherein N is an integer greater than 1; An acquisition unit, used to acquire the parameter value of each of the key parameters at each of the sample time points, and to acquire the regional oil and gas production at each of the sample time points; An array construction unit is used to preprocess the parameter value of each key parameter at any of the sample time points to obtain the preprocessed parameter value of each key parameter at the sample time point, and arrange the regional oil and gas production at the sample time point and the preprocessed parameter value of each key parameter at the sample time point according to a set parameter sequence to obtain a parameter value array corresponding to the sample time point; An input unit, used to input the entire parameter value array corresponding to each of the sample time points as training data into the regional oil and gas production prediction model to be trained to predict the parameter value array corresponding to the target prediction time point, and obtain the corresponding prediction result; The updating unit is used to update the regional oil and gas production prediction model to be trained according to the difference between the prediction result and the corresponding true result to obtain a trained regional oil and gas production prediction model.

[0014] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to execute the training method of the regional oil and gas production prediction model of the first aspect or any corresponding embodiment thereof.

[0015] The training method, device and equipment of the regional oil and gas production prediction model provided by the present invention can determine multiple key parameters that have an important impact on the regional oil and gas production prediction, and respectively determine the first N time points of the target prediction time point as a sample time point, obtain the parameter value of each key parameter at each sample time point, and obtain the regional oil and gas production at each sample time point. For any sample time point, the parameter value of each key parameter at the sample time point is preprocessed to obtain the preprocessed parameter value of each key parameter at the sample time point, and the regional oil and gas production at the sample time point and the preprocessed parameter value of each key parameter at the sample time point are arranged according to the set parameter order to obtain the parameter value array corresponding to the sample time point; the parameter value array corresponding to each sample time point is used as training data as a whole, and is input into the parameter value array corresponding to the target prediction time point in the regional oil and gas production prediction model to be trained to obtain the corresponding prediction result, and the regional oil and gas production prediction model to be trained is updated according to the difference between the prediction result and the corresponding true result to obtain a trained regional oil and gas production prediction model. The regional oil and gas production prediction model trained by the present invention can predict the regional oil and gas production of the target area at the target time point, realize the reliable prediction of the regional oil and gas production, and diversify the reliable prediction methods of the regional oil and gas production. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 A flowchart of a method for training a regional oil and gas production prediction model provided by an embodiment of the present invention; Figure 2 A flowchart of another method for training a regional oil and gas production prediction model provided by an embodiment of the present invention; Figure 3 A schematic diagram of the structure of a training device for a regional oil and gas production prediction model provided by an embodiment of the present invention; Figure 4 A schematic diagram of the structure of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] Combine the following Figure 1-Figure 2 The present invention describes the training method of the regional oil and gas production prediction model.

[0020] like Figure 1 As shown, this embodiment proposes a first method for training a regional oil and gas production prediction model, which may include the following steps: S101. Based on the oil and gas production related data of the target area, multiple key parameters having an important impact on the regional oil and gas production prediction are determined from among multiple set regional oil and gas production related parameters.

[0021] The target area may be a region that includes multiple oil and gas fields being produced and requires regional oil and gas production forecasting.

[0022] The oil and gas production-related data may be data related to regional oil and gas production in the target area, such as the number of oil wells opened, the number of water injection wells opened, oil prices and geological reserves in the target area.

[0023] Specifically, the regional oil and gas production related parameters may be parameters related to the regional oil and gas production during the oil and gas production process. Optionally, the multiple regional oil and gas production related parameters include at least two of the annual newly added geological reserves, the remaining geological reserves, the proportion of low-grade reserves, the number of oil wells opened, the number of water injection wells opened, the water content, the recoverable reserves of a single well by water drive, and the oil price.

[0024] It is understandable that regional oil and gas production forecasts are different from single oil field or single well production forecasts. They are composed of a comprehensive number of oil fields and are not only affected by the different development levels of each oil field, but also involve macro factors such as exploration resources, regional development strategies, oil quality and oil market economy. It requires comprehensive consideration of multiple disciplines. Starting from the three major directions of exploration, development and economy, this time, we selected regional oil and gas production-related parameters, including 3 exploration indicators (i.e., annual new geological reserves, remaining geological reserves, and proportion of low-grade reserves), 4 development indicators (i.e., the number of oil wells opened, the number of water injection wells opened, water content, and recoverable reserves of single well water drive) and 1 economic indicator (i.e., oil price).

[0025] Among them, the key parameters may be parameters that have an important impact on the regional oil and gas production prediction.

[0026] Optionally, step S101 may include: According to the oil and gas production related data of the target area, among the set multiple regional oil and gas production related parameters, a plurality of first parameters whose correlation with the regional oil and gas production exceeds a preset correlation threshold are determined; Determine the impact score of each first parameter on the target model performance, where the target model performance is the model performance of the regional oil and gas production prediction model to be trained; According to the impact score of each first parameter on the target model performance, determine, among the multiple first parameters, multiple second parameters whose impact scores are not less than the first preset score and not greater than the second preset score; Treat each second parameter as a key parameter separately.

[0027] Specifically, this embodiment can determine the correlation between each regional oil and gas production related parameter and the regional oil and gas production based on the Pearson correlation coefficient method, and then determine the regional oil and gas production related parameter whose correlation with the regional oil and gas production exceeds a preset correlation threshold as the first parameter.

[0028] Among them, the Pearson correlation coefficient method formula is:

[0029] in, and They are the regional oil and gas production related parameters and the regional oil and gas production parameter values, and They are the regional oil and gas production related parameters and the average parameter values ​​of regional oil and gas production, respectively. is the Pearson correlation coefficient.

[0030] The preset correlation threshold may be set by a technician according to actual conditions, for example, it may be set to 0.5, and this embodiment does not limit this.

[0031] Specifically, this embodiment can evaluate the importance of each first parameter to the model performance based on the random forest method, obtain the impact score of each first parameter on the model performance, determine which first parameters are most important to the model performance, and then determine the first parameter whose impact score is not less than the first preset score and not greater than the second preset score, and use it as the second parameter.

[0032] The first preset score and the second preset score can be set by the technician according to actual conditions. For example, the first preset score can be set to 0.05, and the second preset score can be set to 0.8.

[0033] It should be noted that the condition of not less than the first preset score in this embodiment can be used to screen out the second parameter that is of high importance to the model performance, that is, has a greater impact. The condition of not greater than the second preset score can be used to eliminate the second parameter that is too important to the model performance, that is, has too much impact. Because if a second parameter has a very large impact on the model performance, then the model may rely too much on this second parameter. Although it performs well on the training data, it may perform poorly on unseen data, resulting in overfitting. At the same time, it will also lead to increased model instability and will be too sensitive to this second parameter. Therefore, the condition of not greater than the second preset score in this embodiment can effectively guarantee the training effect and model performance of the regional oil and gas production prediction model. Therefore, this embodiment selects the second parameter used to constrain the regional oil and gas production prediction by setting the impact score to be not less than the first preset score and not greater than the second preset score, which can effectively reduce noise, reduce dimensionality and avoid overfitting risks.

[0034] S102, respectively determining the first N time points of the target prediction time point as a sample time point; wherein N is an integer greater than 1.

[0035] Among them, the target prediction time point is the time point to be predicted. It should be noted that the time point can be a time point of a time scale such as a day, a month, and a year. For example, when the time point is an annual time point, the target prediction time point is the year to be predicted, and the first N time points of the target prediction time point include the first year before the target prediction time point, the second year before the target prediction time point, and the Nth year before the target prediction time point. Further, when N is 2, the first two time points of the target prediction time point include the first year before the target prediction time point and the second year before the target prediction time point.

[0036] Specifically, the specific size of N can be set by technical personnel according to actual conditions, and this embodiment does not limit it.

[0037] It should be noted that step S102 can be performed after step S101. Figure 1 Step S102 may also be performed before step S101, and of course, step S102 and step S101 may also be performed simultaneously.

[0038] S103, obtaining the parameter value of each key parameter at each sample time point, and obtaining the regional oil and gas production at each sample time point.

[0039] Among them, the target prediction time point is the time point to be predicted. It should be noted that the time point can be a time point of a time scale such as a day, a month, and a year. For example, when the time point is an annual time point, the target prediction time point is the year to be predicted, and the first N time points of the target prediction time point include the first year before the target prediction time point, the second year before the target prediction time point, and the Nth year before the target prediction time point. Further, when N is 2, the first two time points of the target prediction time point include the first year before the target prediction time point and the second year before the target prediction time point.

[0040] It should be noted that the specific size of N can be set by technicians according to actual conditions and is not limited in this embodiment.

[0041] It is understandable that the number of sample time points is N. The parameter value of the key parameter at each sample time point includes the parameter value of the key parameter at each of the N time points before the target prediction time point. For example, when N is 2, the parameter value of the key parameter at each sample time point includes the parameter value of the key parameter at the first time point before the target prediction time point and the parameter value of the key parameter at the second time point before the target prediction time point.

[0042] S104. For any sample time point, preprocess the parameter value of each key parameter at the sample time point to obtain the preprocessed parameter value of each key parameter at the sample time point, and arrange the regional oil and gas production at the sample time point and the preprocessed parameter value of each key parameter at the sample time point according to the set parameter sequence to obtain the parameter value array corresponding to the sample time point.

[0043] Among them, the above-mentioned parameter order can be the arrangement order of each key parameter set. In this embodiment, the pre-processed parameter values ​​of each key parameter at the sample time point can be arranged according to the parameter order to obtain the corresponding parameter value array. For example, it is determined that the key parameters include three key parameters A, B and C, and the pre-processed parameter values ​​of the three key parameters A, B and C at the sample time point are A1, B1 and C1 respectively. When the set parameter order is C, B and A, the parameter value array can be {C1, B1, A1}, when the set parameter order is C, A and B, the parameter value array can be {C1, A1, B1}, and when the set parameter order is B, C and A, the parameter value array can be {B1, C1, A1}.

[0044] It should be noted that the parameter order can be set by technicians according to actual conditions and is not limited in this embodiment.

[0045] Optionally, the preprocessing of the parameter value of each key parameter at the sample time point to obtain the preprocessed parameter value of each key parameter at the sample time point includes: Normalize the impact score of each second parameter on the target model performance to a preset interval to obtain the normalized score corresponding to each second parameter, and use them as the impact weight of the corresponding key parameter on the regional oil and gas production; Performing differential processing on the parameter value of each key parameter at the sample time point to obtain the differential value of each key parameter at the sample time point; Normalizing the differential value of each key parameter at the sample time point to obtain the normalized differential value of each key parameter at the sample time point; The normalized difference value of each key parameter at the sample time point is multiplied by the corresponding influence weight to obtain the preprocessed parameter value of each key parameter at the sample time point.

[0046] Among them, the preset interval can be set by the technician according to the actual situation, for example, it can be set to an interval from 0 to 1, and this embodiment does not limit it.

[0047] Optionally, performing differential processing on the parameter value of each key parameter at the sample time point to obtain the differential value of each key parameter at the sample time point includes: For any key parameter, the parameter value of the key parameter at the sample time point is subtracted from the parameter value of the key parameter at the previous sample time point at the sample time point to obtain the differential value of the key parameter at the sample time point.

[0048] It should be noted that this embodiment can better capture the dynamic changes in the data through differential processing, while reducing the noise and fluctuation of the data and improving the prediction accuracy of the model.

[0049] Optionally, the above normalizing the differential value of each key parameter at the sample time point to obtain the normalized differential value of each key parameter at the sample time point includes: For any key parameter, determine the maximum difference value and the minimum difference value corresponding to the key parameter, normalize the difference value of the key parameter at the sample time point according to the maximum difference value and the minimum difference value, and obtain the normalized difference value of the key parameter at the sample time point.

[0050] The maximum differential value corresponding to the key parameter may be the maximum value among the differential values ​​of the key parameter at each sample time point, and the minimum differential value corresponding to the key parameter may be the minimum value among the differential values ​​of the key parameter at each sample time point.

[0051] Specifically, this embodiment can scale the differential value of the key parameter at a time point to the interval [0,1] based on the maximum-minimum normalization formula, the maximum differential value and the minimum differential value corresponding to the key parameter. The normalization formula can be: .

[0052] in, is the normalized difference value, is the differential value of the key parameter at the sample time point, and are the maximum and minimum difference values ​​corresponding to the key parameters respectively.

[0053] Specifically, this embodiment multiplies the normalized difference value of the key parameter at the sample time point by the corresponding influence weight, which can make the model focus more on the information that is more critical to the current task, reduce the attention to other information, and even filter out irrelevant information, which helps to extract and process important information in long sequences.

[0054] S105, taking the entire parameter value array corresponding to each sample time point as training data, and inputting it into the parameter value array corresponding to the prediction target prediction time point in the regional oil and gas production prediction model to be trained, to obtain the corresponding prediction result.

[0055] Among them, the regional oil and gas production prediction model can be a long short-term memory network (Long Short-Term Memory, LSTM), or it can be other types of time series data prediction networks.

[0056] Specifically, the prediction result is an array of parameter values ​​corresponding to the target prediction time point predicted by the regional oil and gas production prediction model to be trained.

[0057] S106. Update the regional oil and gas production prediction model to be trained according to the difference between the prediction result and the corresponding actual result to obtain a trained regional oil and gas production prediction model.

[0058] The actual result may be an array of actual parameter values ​​corresponding to the target prediction time point. In this embodiment, the actual result may be obtained based on the actual oil and gas production related data in the target area.

[0059] Specifically, when it is determined that the difference between the predicted result and the actual result is large, the present embodiment can update the model parameters in the regional oil and gas production prediction model to be trained according to the difference to obtain an updated regional oil and gas production prediction model. Afterwards, the present embodiment can obtain new training data and input it into the updated regional oil and gas production prediction model to continue predicting the parameter value array corresponding to the new target prediction time point until the difference between the predicted result and the actual result converges, and determine that the regional oil and gas production prediction model at this time meets the model performance requirements, and use it as the trained regional oil and gas production prediction model.

[0060] Optionally, in this embodiment, when the difference between the predicted results and the actual results converges, the test data set and the verification data set can be used to continue to test and verify the current regional oil and gas production prediction model. If it fails the test or verification, the training data can continue to be used to train the regional oil and gas production prediction model until the difference between the predicted results and the actual results converges, and the regional oil and gas production prediction model passes the test and verification and is used as the trained regional oil and gas production prediction model.

[0061] It should be noted that the regional oil and gas production prediction methods in related technologies, such as the life cycle method and the decline curve method (DCA), have low prediction accuracy due to the single consideration of factors. The most widely used production composition method not only has high labor costs, but also has poor prediction effects on future production, and cannot achieve rapid regional production prediction. In addition, the production composition method relies on the numerical simulation results of the reservoirs of each oil field. The process is cumbersome, consumes a lot of manpower and financial resources, and the results are easily affected by the subjective judgment of engineers. In order to capture the time series characteristics of regional production, statistical models such as regression moving average and autoregressive integrated moving average are introduced, but the statistical models are based on single historical data, and the prediction effect needs to be improved. Although machine learning methods such as support vector machines, neural networks, and random forests can process complex data, they ignore the temporal characteristics of production and have limitations in prediction under specific production scenarios.

[0062] The inventors of the present invention have found that deep learning models such as long short-term memory networks and gated recurrent units can not only effectively capture temporal and nonlinear relationships, but also show excellent performance in the field of regional oil and gas production prediction. However, these models are extremely sensitive to input feature selection and model hyperparameter configuration, especially in the complex task of regional production prediction, which integrates exploration, development, economy and other multi-dimensional factors, and puts forward strict requirements on model construction. The model can comprehensively consider exploration, development, and economic factors, and is committed to achieving the comprehensiveness and objectivity of feature screening, optimizing hyperparameter configuration, and enhancing model stability and prediction accuracy. This embodiment can quickly and accurately predict regional oil and gas production, provide strong decision-making support for oil and gas field development planning, and has far-reaching strategic significance.

[0063] It should be noted that this embodiment achieves more comprehensive parameter selection and more objective feature screening. From multiple professional perspectives such as exploration, development, and economy, variable parameters closely related to regional production are comprehensively screened and obtained; through the comprehensive application of correlation analysis and importance analysis, it effectively solves the problem of poor prediction results that may be caused by manual feature extraction.

[0064] The training method of the regional oil and gas production prediction model proposed in this embodiment can determine multiple key parameters that have an important impact on the regional oil and gas production prediction, determine the first N time points of the target prediction time point as a sample time point, obtain the parameter value of each key parameter at each sample time point, and obtain the regional oil and gas production at each sample time point. For any sample time point, the parameter value of each key parameter at the sample time point is preprocessed to obtain the preprocessed parameter value of each key parameter at the sample time point, and the regional oil and gas production at the sample time point and the preprocessed parameter value of each key parameter at the sample time point are arranged according to the set parameter order to obtain the parameter value array corresponding to the sample time point; the parameter value array corresponding to each sample time point is used as training data as a whole, and is input into the parameter value array corresponding to the target prediction time point in the regional oil and gas production prediction model to be trained to obtain the corresponding prediction result, and the regional oil and gas production prediction model to be trained is updated according to the difference between the prediction result and the corresponding true result to obtain a trained regional oil and gas production prediction model. The regional oil and gas production prediction model trained by the present invention can predict the regional oil and gas production of the target area at the target time point, realize the reliable prediction of the regional oil and gas production, and diversify the reliable prediction methods of the regional oil and gas production.

[0065] based on Figure 1 This embodiment proposes a second training method for a regional oil and gas production prediction model. In this method, this embodiment can construct a supervised data set including multiple training data, and use the parameter value arrays corresponding to the t-1 time point and the t time point to predict the parameter value array corresponding to the t+1 time point.

[0066] Specifically, this embodiment can divide the supervised data set into a training data set and a test data set. For example, this embodiment can use 85% of the training data in the supervised data set for training the model and 15% of the training data for testing the model.

[0067] Specifically, in this embodiment, the root mean square error formula can be used to define the error function. The root mean square error formula is: .

[0068] in, is the root mean square error, is the total number of observations, is the sequence number of the observed sample, For the The predicted value of the observed sample, For the The actual value of the observed sample.

[0069] Specifically, this embodiment can initialize hyperparameters based on the tent chaotic mapping strategy, including learning rate, number of LSTM neural units, batch size, number of iterations, number of particle swarms, etc.

[0070] Specifically, in the regional oil and gas production prediction model, the LSTM module receives the processed multivariate feature data, introduces the Nadam optimizer, and constructs an irregular time series model. After being processed by two LSTM layers and two random inactivation layers, the model multiplies the hidden layer output vectors of different time nodes with the corresponding weights and sums them to finally obtain the predicted value.

[0071] In this embodiment, the particle swarm optimization algorithm (PSO) can be used to optimize the hyperparameters of the LSTM neural network model. Based on the objective function of minimizing the prediction error of LSTM, the hyperparameters of the neural network are updated and optimized based on the PSO optimization algorithm. When the stopping condition is reached or the number of iterations reaches the maximum, the iteration is stopped, thereby obtaining the best neural network and prediction results. The specific steps include: (1) Based on the tent chaotic mapping strategy, set the optimization variables such as learning rate, number of LSTM neural units, batch size, number of iterations, number of particle swarms, etc.; at the same time, set fixed variables such as the number of hidden layers and loss function type.

[0072] (2) The parameters are introduced into the LSTM network model, i.e., the regional oil and gas production prediction model to be trained, and data training and prediction are introduced.

[0073] (3) Based on the PSO optimization algorithm, the minimum RMSE is the goal, and the parameter combination is updated to obtain the global optimal fitness and the individual optimal fitness. ; .

[0074] For the Particles in The update speed of the moment; For the Particles in The location at the moment; For the Particles in Individual optimum at the moment; for The global optimum at the moment; is the inertia weight, 、 are cognitive coefficient and social coefficient respectively. is the coefficient to be determined.

[0075] (4) After the iteration, the optimal parameter combination and the LSTM network model constructed with the optimal parameter combination are obtained.

[0076] The model establishment, training and optimization process is as follows Figure 2 shown.

[0077] exist Figure 2 In this embodiment, data screening can be started first, and multiple first parameters whose correlation with regional oil and gas production exceeds a preset correlation threshold are selected from multiple regional oil and gas production related parameters. Then, among the multiple first parameters, first parameters whose impact score on model performance is less than a first preset score are eliminated, and first parameters whose impact score on model performance is greater than a second preset score are eliminated, and the remaining first parameters are determined as key parameters of regional oil and gas production.

[0078] Afterwards, the present embodiment can perform data preprocessing, that is, preprocess the parameter values ​​of each key parameter at different time points to construct parameter value arrays corresponding to different time points. The data preprocessing includes processes such as differential processing, normalization processing, construction of supervision data, and data set division.

[0079] Afterwards, this embodiment can perform hyperparameter initialization, and then perform model training based on the training data, the LSTM layer in the LSTM network model, i.e., the regional oil and gas production prediction model to be trained, the random deactivation layer, the LSTM layer, the random deactivation layer, the Nadam optimizer, and the error function, and update the model parameters in the LSTM network model. Among them, during the training process, this embodiment can optimize the hyperparameters based on the PSO algorithm when it is determined that the training process does not meet the iteration termination conditions, specifically including updating the PSO particle parameter combination, solving the fitness, and updating the individual optimum and the global optimum. When it is determined that the training process meets the iteration termination conditions, the training can be terminated.

[0080] It should be noted that this embodiment establishes a more efficient, fast and accurate regional production prediction model. Based on the LSTM algorithm, the model hyperparameters are optimized based on PSO. The Nadam optimizer is introduced to solve the problem of gradient flattening or slowing down the search speed and reducing the search step size in the gradient descent, thereby improving the optimization efficiency and model accuracy.

[0081] The training method of the regional oil and gas production prediction model proposed in this embodiment can establish, train and optimize the regional oil and gas production prediction model, ensure the training effect and model performance of the regional oil and gas production prediction model, realize reliable prediction of regional oil and gas production, and ensure prediction accuracy.

[0082] based on Figure 1This embodiment proposes a third method for training a regional oil and gas production prediction model. When N is 2, after obtaining the trained regional oil and gas production prediction model, the method further includes: Obtain the parameter value arrays corresponding to the T-2th time point and the T-1th time point, and input the parameter value arrays corresponding to the T-2th time point and the T-1th time point into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the parameter value array corresponding to the Tth time point; wherein T is the time point at which the regional oil and gas production needs to be predicted; The parameter value array corresponding to the T-1th time point and the parameter value array corresponding to the Tth time point are input into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the parameter value array corresponding to the T+1th time point, until the parameter value array corresponding to the T+mth time point is obtained; wherein m is an integer not less than 1.

[0083] Specifically, this embodiment can implement a prediction strategy based on rolling data. This embodiment can use continuous time series data, and use the parameter value array corresponding to each time point to predict the parameter value array corresponding to the next time point. Specifically, assuming that the regional oil and gas production at the t+1 time point is predicted, the model will use the parameter value array corresponding to the t-1 and t time points to predict the parameter value array at the t+1 time point (including the regional oil and gas production), and then use the parameter value array corresponding to the t and t+1 time points to predict the parameter value array at the t+2 time point (including the regional oil and gas production), and so on, and so on, so as to achieve recursive prediction of regional oil and gas production.

[0084] Optionally, in the training method of other regional oil and gas production prediction models proposed in this embodiment, when N is 2, after obtaining the trained regional oil and gas production prediction model, the method further includes: Obtain parameter value arrays corresponding to the T-2th time point and the T-1th time point, and input the parameter value arrays corresponding to the T-2th time point and the T-1th time point into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the first parameter value array corresponding to the Tth time point; wherein T is the time point at which the regional oil and gas production needs to be predicted; In the first parameter value array corresponding to the Tth time point, the predicted regional oil and gas production at the Tth time point is determined; Obtaining the true parameter value of each key parameter at the Tth time point, and preprocessing the true parameter value of each key parameter at the Tth time point to obtain the preprocessed parameter value of each key parameter at the Tth time point; Arrange the predicted regional oil and gas production at the Tth time point and the preprocessed parameter values ​​of each key parameter at the Tth time point according to the parameter sequence to obtain a second parameter value array corresponding to the Tth time point; The parameter value array corresponding to the T-1th time point and the second parameter value array corresponding to the Tth time point are input into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the first parameter value array corresponding to the T+1th time point, until the first parameter value array corresponding to the T+mth time point is obtained, and the predicted regional oil and gas production at the T+mth time point is determined; wherein m is an integer not less than 1.

[0085] The inventor of the present invention has found that in practical applications, the planning strategy will be constantly adjusted under the influence of macro factors such as exploration resources, oil field development level, and oil market. This embodiment proposes a prediction strategy based on planning data. Specifically, this embodiment can only use the regional oil and gas production predicted at the previous time point when predicting the production at a certain time point in the future, and does not use the pre-processed parameter values ​​corresponding to the recursive key parameters. Instead, the manually set planning strategy data is re-read, and the data is re-preprocessed to form a new input data set, so as to predict the regional oil and gas production at the next time point, and meet the actual needs of adaptive regional production prediction.

[0086] It should be noted that the regional oil and gas production prediction model trained in this embodiment comprehensively considers multiple professional factors such as exploration, development, and economy, is committed to achieving comprehensiveness and objectivity in feature screening, and optimizes hyperparameter configuration. It can enhance model stability and prediction accuracy, quickly and accurately predict regional oil and gas production, and provide strong decision-making support for oil and gas field development planning, which has far-reaching strategic significance.

[0087] The training method of the regional oil and gas production prediction model proposed in this embodiment can realize the prediction of regional oil and gas production based on the rolling data prediction strategy, and can improve the prediction efficiency. It can also establish a future production prediction method based on planning guidance, alleviate the problem of decreased prediction accuracy as the prediction time increases, improve the guidance and constraints of planning guidance on future production, and improve the scientific nature of the prediction.

[0088] like Figure 3 As shown, this embodiment proposes a training device for a regional oil and gas production prediction model, including: The first determination unit 301 is used to determine a plurality of key parameters that have an important impact on regional oil and gas production prediction from among a plurality of set regional oil and gas production related parameters according to the oil and gas production related data of the target region; The second determining unit 302 is used to determine each of the N time points before the target prediction time point as a sample time point; wherein N is an integer greater than 1; An acquisition unit 303 is used to acquire the parameter value of each key parameter at each sample time point, and to acquire the regional oil and gas production at each sample time point; The array construction unit 304 is used to preprocess the parameter value of each key parameter at any sample time point to obtain the preprocessed parameter value of each key parameter at the sample time point, and arrange the regional oil and gas production at the sample time point and the preprocessed parameter value of each key parameter at the sample time point according to the set parameter sequence to obtain the parameter value array corresponding to the sample time point; An input unit 305 is used to input the entire parameter value array corresponding to each sample time point as training data into the parameter value array corresponding to the prediction target prediction time point in the regional oil and gas production prediction model to be trained to obtain the corresponding prediction result; The updating unit 306 is used to update the regional oil and gas production prediction model to be trained according to the difference between the prediction result and the corresponding actual result, so as to obtain a trained regional oil and gas production prediction model.

[0089] It should be noted that the processing of the first determining unit 301, the second determining unit 302, the obtaining unit 303, the array building unit 304, the input unit 305 and the updating unit 306 and the beneficial effects thereof can be respectively referred to in Figure 1 Steps S101 to S106 in the above are not described in detail.

[0090] Optionally, the first determining unit 301 is further configured to: According to the oil and gas production related data of the target area, among the set multiple regional oil and gas production related parameters, a plurality of first parameters whose correlation with the regional oil and gas production exceeds a preset correlation threshold are determined; Determine the impact score of each first parameter on the target model performance, where the target model performance is the model performance of the regional oil and gas production prediction model to be trained; According to the impact score of each first parameter on the target model performance, determine, among the multiple first parameters, multiple second parameters whose impact scores are not less than the first preset score and not greater than the second preset score; Treat each second parameter as a key parameter separately.

[0091] Optionally, the array building unit 304 is further used for: Normalize the impact score of each second parameter on the target model performance to a preset interval to obtain the normalized score corresponding to each second parameter, and use them as the impact weight of the corresponding key parameter on the regional oil and gas production; Performing differential processing on the parameter value of each key parameter at the sample time point to obtain the differential value of each key parameter at the sample time point; Normalizing the differential value of each key parameter at the sample time point to obtain the normalized differential value of each key parameter at the sample time point; The normalized difference value of each key parameter at the sample time point is multiplied by the corresponding influence weight to obtain the preprocessed parameter value of each key parameter at the sample time point.

[0092] Optionally, the array building unit 304 is further configured to: For any key parameter, the parameter value of the key parameter at the sample time point is subtracted from the parameter value of the key parameter at the previous sample time point at the sample time point to obtain the differential value of the key parameter at the sample time point.

[0093] Optionally, the array building unit 304 is further configured to: For any key parameter, determine the maximum difference value and the minimum difference value corresponding to the key parameter, normalize the difference value of the key parameter at the sample time point according to the maximum difference value and the minimum difference value, and obtain the normalized difference value of the key parameter at the sample time point.

[0094] Optionally, when the time point is year, the multiple regional oil and gas production-related parameters include at least two of the annual newly added geological reserves, remaining geological reserves, proportion of low-grade reserves, number of oil wells opened, number of water injection wells opened, water content, single well water drive recoverable reserves and oil price.

[0095] Optionally, when N is 2, the above device further includes: The first prediction unit is used to obtain the parameter value arrays corresponding to the T-2th time point and the T-1th time point after obtaining the trained regional oil and gas production prediction model, and input the parameter value arrays corresponding to the T-2th time point and the T-1th time point into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the parameter value array corresponding to the Tth time point; wherein T is a time point for predicting regional oil and gas production; the parameter value array corresponding to the T-1th time point and the parameter value array corresponding to the Tth time point are input into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the parameter value array corresponding to the T+1th time point, until the parameter value array corresponding to the T+mth time point is obtained; wherein m is an integer not less than 1.

[0096] Optionally, when N is 2, the above device further includes: The second prediction unit is used to obtain the parameter value array corresponding to the T-2th time point and the T-1th time point after obtaining the trained regional oil and gas production prediction model, and input the parameter value array corresponding to the T-2th time point and the T-1th time point into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the first parameter value array corresponding to the Tth time point; wherein T is a time point for predicting regional oil and gas production; in the first parameter value array corresponding to the Tth time point, determine the predicted regional oil and gas production at the Tth time point; obtain the true parameter value of each key parameter at the Tth time point, and preprocess the true parameter value of each key parameter at the Tth time point , obtain the preprocessed parameter value of each key parameter at the Tth time point; arrange the predicted regional oil and gas production at the Tth time point and the preprocessed parameter value of each key parameter at the Tth time point in parameter order to obtain the second parameter value array corresponding to the Tth time point; input the parameter value array corresponding to the T-1th time point and the second parameter value array corresponding to the Tth time point into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the first parameter value array corresponding to the T+1th time point, until the first parameter value array corresponding to the T+mth time point is obtained, and the predicted regional oil and gas production at the T+mth time point is determined; wherein m is an integer not less than 1.

[0097] The training device of the regional oil and gas production prediction model proposed in this embodiment can obtain a trained regional oil and gas production prediction model. The trained regional oil and gas production prediction model in this embodiment can predict the regional oil and gas production of the target area at the target time point, realize the reliable prediction of regional oil and gas production, and diversify the reliable prediction methods of regional oil and gas production.

[0098] The training device for the regional oil and gas production prediction model in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0099] The embodiment of the present invention also provides a computer device having the above Figure 3 The training device of the regional oil and gas production prediction model is shown.

[0100] See also Figure 4, a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are connected to each other using different buses for communication, and can be installed on a common motherboard or installed in other ways as needed. The processor can process instructions executed in the computer device, including instructions stored in or on the memory to display graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 4 A processor 10 is taken as an example.

[0101] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0102] The memory 20 stores instructions executable by at least one processor 10, so that at least one processor 10 executes the method shown in the above embodiment.

[0103] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function. The data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage devices. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0104] The memory 20 may include a volatile memory, such as a random access memory. The memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive. The memory 20 may also include a combination of the above-mentioned types of memory.

[0105] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0106] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0107] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A training method for a regional oil and gas production prediction model, characterized in that: include: According to the oil and gas production related data of the target area, multiple key parameters having an important impact on the regional oil and gas production prediction are determined from among the multiple regional oil and gas production related parameters set; and N time points before the target prediction time point are respectively determined as a sample time point; wherein N is an integer greater than 1; Obtaining the parameter value of each of the key parameters at each of the sample time points, and obtaining the regional oil and gas production at each of the sample time points; For any of the sample time points, preprocess the parameter value of each of the key parameters at the sample time point to obtain the preprocessed parameter value of each of the key parameters at the sample time point, and arrange the regional oil and gas production at the sample time point and the preprocessed parameter value of each of the key parameters at the sample time point according to the set parameter sequence to obtain a parameter value array corresponding to the sample time point; The parameter value array corresponding to each of the sample time points is used as training data, and is input into the regional oil and gas production prediction model to be trained to predict the parameter value array corresponding to the target prediction time point, and the corresponding prediction result is obtained. The regional oil and gas production prediction model to be trained is updated according to the difference between the prediction result and the corresponding true result to obtain a trained regional oil and gas production prediction model.

2. The method according to claim 1, characterized in that According to the oil and gas production related data of the target area, multiple key parameters having an important impact on the regional oil and gas production prediction are determined from among the multiple regional oil and gas production related parameters set, including: According to the oil and gas production related data of the target area, among the set multiple regional oil and gas production related parameters, a plurality of first parameters whose correlation with the regional oil and gas production exceeds a preset correlation threshold are determined; Determine an influence score of each of the first parameters on target model performance, where the target model performance is the model performance of the regional oil and gas production prediction model to be trained; According to the impact score of each of the first parameters on the target model performance, determine, among the plurality of first parameters, a plurality of second parameters whose impact scores are not less than a first preset score and not greater than a second preset score; Each of the second parameters is respectively used as a key parameter.

3. The method according to claim 2, characterized in that The preprocessing of the parameter value of each of the key parameters at the sample time point to obtain the preprocessed parameter value of each of the key parameters at the sample time point includes: Normalizing the impact score of each second parameter on the target model performance to a preset interval, obtaining a normalized score corresponding to each second parameter, and using the scores as the impact weights of the corresponding key parameters on regional oil and gas production; Performing differential processing on the parameter value of each of the key parameters at the sample time point to obtain the differential value of each of the key parameters at the sample time point; Normalizing the differential value of each of the key parameters at the sample time point to obtain a normalized differential value of each of the key parameters at the sample time point; The normalized difference value of each of the key parameters at the sample time point is multiplied by the corresponding influence weight to obtain the preprocessed parameter value of each of the key parameters at the sample time point.

4. The method according to claim 3, characterized in that The step of performing differential processing on the parameter value of each of the key parameters at the sample time point to obtain the differential value of each of the key parameters at the sample time point includes: For any of the key parameters, the parameter value of the key parameter at the sample time point is subtracted from the parameter value of the key parameter at the sample time point before the sample time point to obtain the differential value of the key parameter at the sample time point.

5. The method according to claim 3, characterized in that: The step of normalizing the differential value of each of the key parameters at the sample time point to obtain the normalized differential value of each of the key parameters at the sample time point includes: For any of the key parameters, determine the maximum differential value and the minimum differential value corresponding to the key parameter, normalize the differential value of the key parameter at the sample time point according to the maximum differential value and the minimum differential value, and obtain the normalized differential value of the key parameter at the sample time point.

6. The method according to claim 1, characterized in that When the time point is year, the multiple regional oil and gas production-related parameters include at least two of the annual newly added geological reserves, remaining geological reserves, proportion of low-grade reserves, number of oil wells opened, number of water injection wells opened, water content, single well water-drive recoverable reserves and oil price.

7. The method according to any one of claims 1 to 6, characterized in that When N is 2, after obtaining the trained regional oil and gas production prediction model, the method further includes: Obtain parameter value arrays corresponding to the T-2th time point and the T-1th time point, and input the parameter value arrays corresponding to the T-2th time point and the T-1th time point into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the parameter value array corresponding to the Tth time point; wherein T is the time point at which the regional oil and gas production needs to be predicted; The parameter value array corresponding to the T-1th time point and the parameter value array corresponding to the Tth time point are input into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the parameter value array corresponding to the T+1th time point, until the parameter value array corresponding to the T+mth time point is obtained; wherein m is an integer not less than 1.

8. The method according to any one of claims 1 to 6, characterized in that When N is 2, after obtaining the trained regional oil and gas production prediction model, the method further includes: Obtain parameter value arrays corresponding to the T-2th time point and the T-1th time point, and input the parameter value arrays corresponding to the T-2th time point and the T-1th time point into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the first parameter value array corresponding to the Tth time point; wherein T is the time point at which the regional oil and gas production needs to be predicted; Determining the predicted regional oil and gas production at the Tth time point in the first parameter value array corresponding to the Tth time point; Acquire the true parameter value of each of the key parameters at the Tth time point, and preprocess the true parameter value of each of the key parameters at the Tth time point to obtain the preprocessed parameter value of each of the key parameters at the Tth time point; Arranging the predicted regional oil and gas production at the Tth time point and the preprocessed parameter value of each key parameter at the Tth time point according to the parameter sequence to obtain a second parameter value array corresponding to the Tth time point; The parameter value array corresponding to the T-1th time point and the second parameter value array corresponding to the Tth time point are input into the trained regional oil and gas production prediction model, so that the trained regional oil and gas production prediction model predicts and outputs the first parameter value array corresponding to the T+1th time point, until the first parameter value array corresponding to the T+mth time point is obtained, and the predicted regional oil and gas production at the T+mth time point is determined; wherein m is an integer not less than 1.

9. A training device for a regional oil and gas production prediction model, characterized in that: include: A first determination unit is used to determine, based on the oil and gas production related data of the target area, a plurality of key parameters having an important influence on the regional oil and gas production prediction from among the set plurality of regional oil and gas production related parameters; A second determining unit, configured to determine each of the N time points before the target prediction time point as a sample time point; wherein N is an integer greater than 1; An acquisition unit, used to acquire the parameter value of each of the key parameters at each of the sample time points, and to acquire the regional oil and gas production at each of the sample time points; An array construction unit is used to preprocess the parameter value of each key parameter at any of the sample time points to obtain the preprocessed parameter value of each key parameter at the sample time point, and arrange the regional oil and gas production at the sample time point and the preprocessed parameter value of each key parameter at the sample time point according to a set parameter sequence to obtain a parameter value array corresponding to the sample time point; An input unit, used to input the entire parameter value array corresponding to each of the sample time points as training data into the regional oil and gas production prediction model to be trained to predict the parameter value array corresponding to the target prediction time point, and obtain the corresponding prediction result; The updating unit is used to update the regional oil and gas production prediction model to be trained according to the difference between the prediction result and the corresponding true result to obtain a trained regional oil and gas production prediction model.

10. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the training method of the regional oil and gas production prediction model described in any one of claims 1 to 8 by executing the computer instructions.

Images