Method and device for determining production stage of heavy oil well and electronic equipment

Abstract

The invention discloses a heavy oil well production stage determination method and device and electronic equipment. The method comprises the steps that a target time sequence diagram, first production data and second production data of a target heavy oil well in the current sampling time period are obtained, the first production data are production data of a text type, and the second production data are production data of a numerical value type; event feature extraction processing is conducted on the first production data, and target event features corresponding to the target heavy oil well are obtained; performing time sequence feature extraction processing on the second production data to obtain a target time sequence feature corresponding to the target heavy oil well; based on the target time sequence diagram, the target event features and the target time sequence features, the production stage of the target heavy oil well in the current sampling time period is determined, and the production stage comprises a low-yield stage and a high-yield stage. According to the method, the technical problem of low accuracy caused by incomplete consideration factors of a heavy oil well production stage determination method in related technologies is solved.

Description

Technical Field

[0001] The present invention relates to the field of oil production, and in particular to a method, a device and an electronic device for determining the production stage of a heavy oil well. Background Art

[0002] A heavy oil well refers to a situation where the crude oil produced from the oil well has a high viscosity and viscosity. Accurately judging that the heavy oil well is in the low-production stage helps to determine whether the heavy oil well is a well to be operated by a turbine, and to determine the heavy oil well that needs turbine steam injection, so as to help reduce unnecessary steam injection and effectively utilize the formation capacity and steam injection heat for production. The method for determining the production stage of heavy oil wells in related technologies only considers changes in production, pressure, etc., and the factors considered are not comprehensive, resulting in low accuracy in the determination results of the heavy oil well generation stage.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present invention provide a method, device and electronic device for determining the production stage of a heavy oil well, so as to at least solve the technical problem that the method for determining the production stage of a heavy oil well in the related art does not consider all factors, resulting in low accuracy.

[0005] According to one aspect of an embodiment of the present invention, a method for determining a production stage of a heavy oil well is provided, comprising: obtaining a target time series diagram, first production data, and second production data of a target heavy oil well in a current sampling period, wherein the first production data is text-type production data, and the second production data is numerical-type production data; performing event feature extraction processing on the first production data to obtain a target event feature corresponding to the target heavy oil well; performing time series feature extraction processing on the second production data to obtain a target time series feature corresponding to the target heavy oil well; and determining the production stage of the target heavy oil well in the current sampling period based on the target time series diagram, the target event feature, and the target time series feature, wherein the production stage includes a low-production stage and a high-production stage.

[0006] According to another aspect of an embodiment of the present invention, a device for determining the production stage of a heavy oil well is also provided, comprising: a first acquisition module, used to acquire a target time series diagram, first production data, and second production data of a target heavy oil well in a current sampling period, wherein the first production data is text-type production data, and the second production data is numerical-type production data; a first processing module, used to perform event feature extraction processing on the first production data to obtain target event features corresponding to the target heavy oil well; a second processing module, used to perform time series feature extraction processing on the second production data to obtain target time series features corresponding to the target heavy oil well; a first determination module, used to determine the production stage of the target heavy oil well in the current sampling period based on the target time series diagram, the target event features, and the target time series features, wherein the production stage includes a low-production stage and a high-production stage.

[0007] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the methods for determining the production stage of a heavy oil well.

[0008] In an embodiment of the present invention, a target time series diagram, first production data, and second production data of a target heavy oil well in a current sampling period are obtained, wherein the first production data is text type production data, and the second production data is numerical type production data; event feature extraction processing is performed on the first production data to obtain a target event feature corresponding to the target heavy oil well; time series feature extraction processing is performed on the second production data to obtain a target time series feature corresponding to the target heavy oil well; based on the target time series diagram, the target event feature, and the target time series feature, the production stage of the target heavy oil well in the current sampling period is determined, wherein the production stage includes a low production stage and a high production stage, thereby achieving the purpose of accurately determining the production stage of the target heavy oil based on the event feature and time series feature corresponding to the target heavy oil well, thereby achieving the technical effect of improving the accuracy of the heavy oil well production stage determination result, thereby solving the technical problem of low accuracy caused by the incomplete consideration of factors in the heavy oil well production stage determination method in the related art. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0010] Figure 1is a schematic diagram of a method for determining a production stage of a heavy oil well according to an embodiment of the present invention;

[0011] Figure 2 is a schematic diagram of an optional liquid production curve according to an embodiment of the present invention;

[0012] Figure 3 is a schematic diagram of another optional liquid production curve according to an embodiment of the present invention;

[0013] Figure 4 is a schematic diagram of an optional event feature classification according to an embodiment of the present invention;

[0014] Figure 5 is a schematic diagram of an optional event feature extraction according to an embodiment of the present invention;

[0015] Figure 6 is a schematic diagram of an optional timing curve problem according to an embodiment of the present invention;

[0016] Figure 7 is a schematic diagram of an optional supervised learning transformation according to an embodiment of the present invention;

[0017] Figure 8 is a schematic diagram of an optional decision tree model according to an embodiment of the present invention;

[0018] Fig. 9 is a schematic diagram of a device for determining a production stage of a heavy oil well according to an embodiment of the present invention;

[0019] Fig.10 is a schematic diagram of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0020] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0021] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0022] According to an embodiment of the present invention, a method embodiment for determining the production stage of a heavy oil well is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0023] Figure 1 is a flow chart of a method for determining a production stage of a heavy oil well according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0024] Step S102, obtaining a target time series diagram, first production data, and second production data of a target heavy oil well in a current sampling period, wherein the first production data is text-type production data, and the second production data is numerical-type production data.

[0025] Optionally, the target timing diagram may be a timing curve diagram of a production indicator, for example, a liquid production curve diagram, such as Figure 2 and Figure 3 As shown, Figure 2 is a schematic diagram of an optional liquid production curve according to an embodiment of the present invention, Figure 3 It is a schematic diagram of another optional liquid production curve according to an embodiment of the present invention. The first production data is the remark information data in the daily production report or the daily steam injection report, which is text type data. The second production data is the production index data, which is numerical type data.

[0026] Step S104: performing event feature extraction processing on the first production data to obtain target event features corresponding to the target heavy oil well.

[0027] Optionally, heavy oil is currently usually produced in the form of injection-stemming-production turbines. During a turbine cycle of 70 to 100 days, production wells may experience equipment anomalies, production anomalies, environmental interference and other events, which are recorded in the form of text in the notes of the daily report, i.e., the first production data. This information is very important for production analysis, production measures optimization, and turbine planning. Therefore, it is necessary to extract event features from the notes of the daily report data. For example, abnormal event features, production management event features, new measures event features, etc. Among them, the event features of heavy oil wells can be extracted into more than 2,000 entities, classified into a hierarchical structure of 6 major categories, 20 medium categories, 62 minor categories, and 406 items, such as Figure 4 As shown, Figure 4 It is a schematic diagram of an optional event feature classification according to an embodiment of the present invention. The event feature classification result is used to assist in determining the production stage of the heavy oil well, thereby improving the accuracy of determining the production stage of the heavy oil well.

[0028] In an optional embodiment, event feature extraction processing is performed on the first production data to obtain target event features corresponding to the target heavy oil well, including: based on the first production data, an event feature extraction model is used to obtain the target event features, wherein the event feature extraction model is based on the fifth production data of the second heavy oil well in the second cycle, and is trained by machine training, and the fifth production data is text-type production data.

[0029] Optional, Figure 5 is a schematic diagram of an optional event feature extraction according to an embodiment of the present invention, such as Figure 5 As shown, in the application stage of the event extraction model, firstly, based on the domain dictionary, the first production data is subjected to keyword extraction, and string matching and other methods can be used to determine whether there are keywords and extract relevant information; based on the keyword extraction results, phrase extraction is performed according to the phrase pattern in the phrase pattern library to form a regular expression, wherein pattern matching and other methods can be used to determine whether there are phrases that meet the pattern and extract relevant information; according to the classification tree, event feature extraction is performed on the data after phrase extraction, such as the time feature, well name feature, event feature, etc. corresponding to the typical event, to obtain the target event feature and merge and summarize the target event feature according to the rotation cycle. Among them, the domain dictionary, phrase pattern library and classification tree are based on the fifth production data of the second heavy oil well in the second rotation cycle, and are trained by machine training.

[0030] In an optional embodiment, before obtaining the target event feature by using an event feature extraction model based on the first production data, the method also includes: acquiring fifth production data; performing word segmentation processing on the fifth production data to obtain sixth production data; performing word frequency statistics processing on the target event included in the sixth production data to obtain word frequency statistics results, wherein the word frequency statistics results are used to indicate the frequency of occurrence of the target event included in the fifth production data; determining a phrase pattern group based on the sixth production data and the word frequency statistics results, wherein the phrase pattern group includes at least the target event in the fifth production data and the event type corresponding to the target event; performing clustering processing on the phrase pattern group to obtain clustering results; and training the initial model based on the clustering results to obtain an event feature extraction model.

[0031] Optional, still as Figure 5 As shown, in the training stage of the event extraction model, firstly, the daily production data of the second heavy oil well, i.e., the fifth production data, is obtained; the fifth production data can be segmented by a word segmentation engine (e.g., jieba word segmentation) to obtain the fifth production data after segmentation, i.e., the sixth production data, and the words included in the sixth production data are organized into a dictionary to form a domain dictionary for the convenience of subsequent use and expansion; the target events included in the sixth production data are processed by a word frequency statistics tool (e.g., word frequency-inverse document frequency (TF-IDF)) to obtain word frequency statistics results; based on the word frequency statistics results, a phrase extraction algorithm (e.g., association rule mining, statistical-based text analysis method (n-gram) model) can be used to select words or phrases with higher frequency as Common phrase pattern groups, and organize their regular expressions into phrase pattern libraries; based on common phrase pattern groups, clustering algorithms (such as K-means algorithm) can be used for cluster analysis, and similar phrase pattern groups can be classified into one category to form the results of cluster analysis; based on the results of cluster analysis, decision tree algorithms (such as decision tree algorithm based on information gain (ID3 algorithm), decision tree algorithm based on Gini index (CART algorithm)) can be used to build a classification tree, summarize the typical events of heavy oil production and the corresponding event types, including abnormal events (as shown in Table 1), production management events (such as parameter adjustment), new measures events (such as microbial measures), daily well operation and maintenance events (such as new well maintenance), heavy oil production sub-layer switching and cycle conversion events, external interference events (such as power outages). Classification tree is a decision tree model for classification problems, which classifies input data by dividing the data set into different categories. The root node of the classification tree represents the most important feature, the branches represent different values ​​of the feature, and the leaf nodes represent the final classification results. For example, for the abnormal event data in Table 1, the root node can be the pump to be inspected, and the leaf nodes corresponding to the root node can be the pump to be inspected (pump stuck), pump to be inspected (rod stuck), pump to be inspected (scale stuck), pump to be inspected (rod detached / pump detached), pump to be inspected (pump leakage), and pump to be inspected (steam injection failure).

[0032] Table 1

[0033]

[0034]

[0035]

[0036] Step S106, performing time series feature extraction processing on the second production data to obtain target time series features corresponding to the target heavy oil well.

[0037] In an optional embodiment, the second production data is subjected to time series feature extraction processing to obtain target time series features, including: based on the second production data, using a time series feature extraction model to obtain target time series features, wherein the time series feature extraction model is based on the seventh production data of the second heavy oil well in the second cycle, and is trained by machine training, and the seventh production data is numerical type production data.

[0038] Optional, because the actual production curve has various quality problems, e.g. Figure 6 is a schematic diagram of an optional timing curve problem according to an embodiment of the present invention, such as Figure 6 As shown, the data may be discontinuous, the time series may be incomplete, there may be strong fluctuations in individual cycles, and some cycles are overall low. Therefore, it is difficult to adapt to these actual quality problems by relying solely on single-point values ​​or simple trends of indicators such as liquid production, wellhead temperature, and water content. Therefore, it is necessary to extract time series features with higher robustness (the higher the robustness, the better the performance can be maintained under extreme conditions). The time series feature extraction model is based on the seventh production data of the second heavy oil well in the second cycle, and is trained by machine training. The time series feature extraction model can be an autoregressive model, a recurrent neural network (RNN), a convolutional neural network (CNN), etc. For example, the RNN network can be trained using the seventh production data through the following steps:

[0039] Step S1061, preprocessing the seventh production data, which may include cleaning the data, removing noise, standardizing the data, etc.

[0040] Step S1062, divide the seventh production data into a training set, a validation set and a test set. Usually, 80% of the data set is used as a training set for training, 10% as a validation set for verifying the performance of the model, and 10% as a test set for evaluating the accuracy of the model.

[0041] Step S1063, select an appropriate recurrent neural network (RNN) structure, establish the input and output shapes of the model, and add the required layers (such as input layer, hidden layer, output layer in sequence, with recurrent units (such as LSTM or GRU) as the basic units of the hidden layer, and the output layer is a fully connected layer) and activation functions to obtain the model to be trained.

[0042] Step S1064, configure the optimizer, loss function and evaluation index of the model to be trained. Select an appropriate optimization algorithm, such as stochastic gradient descent, etc.; select an appropriate loss function, such as mean square error, cross entropy loss, etc.; select an appropriate evaluation index according to task requirements, such as accuracy, precision, recall, etc.

[0043] Step S1065, train the model using the training set. Input the training set into the model to be trained, and update the parameters of the model through back propagation and optimization algorithm. Depending on the size and complexity of the training set, a suitable training strategy can be selected, such as batch gradient descent or stochastic gradient descent.

[0044] Step S1066, using the validation set to evaluate the performance of the model. Calculate the loss and evaluation index of the model on the validation set, and observe whether overfitting or underfitting occurs during the training process of the model.

[0045] Step S1067, adjust and optimize the model according to the results of the validation set. You can try different hyperparameter configurations, such as learning rate, batch size, hidden layer size, etc., to improve the performance of the model.

[0046] Step S1068, using the test set to evaluate the final performance of the model. The loss and evaluation index of the model on the test set are calculated to obtain a final evaluation of the accuracy of the model in practical applications, and finally a trained time series feature extraction model is obtained.

[0047] A time series feature extraction model is used to extract time series features from the second production data to obtain more robust time series features. For example, the initial liquid production of each rotor can be extracted, as can the average liquid production in the first 2 to 8 days, the cumulative number of days of continuous mining, the average liquid production in the past 3 or 5 days, the relative liquid production in the past 3 or 5 days, the slope of the liquid production in the past 5 days, the type of operating events in the past 3 or 5 days, etc. For example, the specific steps for using the trained RNN to extract time series features from the second production data are as follows: loading the trained RNN model; preprocessing the second production data to ensure that it is processed in the same way as the training data, which may include data standardization, normalization, and padding; inputting the preprocessed second production data into the loaded RNN model, and running the model to obtain the time series feature extraction results.

[0048] Step S108, based on the target time series diagram, the target event characteristics, and the target time series characteristics, determine the production stage of the target heavy oil well in the current sampling period, wherein the production stage includes a low production stage and a high production stage.

[0049] Optionally, the production stage of the target heavy oil well in the current sampling period can be accurately determined based on the target time series diagram, the target event characteristics, and the target time series characteristics, so that a suitable production strategy can be determined according to the production stage of the heavy oil well; the production potential and recoverable reserves of the heavy oil well can be evaluated to provide a basis for subsequent development decisions; the reservoir management strategy can be optimized, including water injection, fracturing and other measures, to improve the reservoir recovery rate and economic benefits.

[0050] In an optional embodiment, based on the target time series diagram, the target event characteristics, and the target time series characteristics, determining the production stage of the target heavy oil well in the current sampling period includes: determining a target prediction model corresponding to the target time series diagram, wherein the target prediction model is based on the third production data of the first heavy oil well in the first rotation cycle, and the fourth production data, trained by machine training, the third production data is text type production data, and the fourth production data is numerical type production data; based on the target event characteristics and the target time series characteristics, using the target prediction model to determine the production stage of the target heavy oil well in the current sampling period.

[0051] Optionally, based on the third production data of the first heavy oil well in the first rotation cycle and the fourth production data, a target prediction model is obtained by machine training, and then the target event features and the target time series features are input into the target prediction model, so that the production stage of the target heavy oil well in the current sampling period can be determined. Through the multi-feature prediction method based on event features and time series features, the prediction target can be described more comprehensively, and various aspects of the target can be better captured, thereby improving the accuracy of the prediction.

[0052] In an optional embodiment, based on the target event characteristics and the target time series characteristics, a target prediction model is used to determine the production stage of the target heavy oil well in the current sampling period, including: based on the target event characteristics and the target time series characteristics, a target prediction model is used to obtain a probability value that the target heavy oil well is in a low production stage in the current sampling period; if the probability value is greater than or equal to a first preset threshold, it is determined that the target heavy oil well is in a low production stage, wherein the low production stage is used to indicate a period of time when the liquid production of the heavy oil well is lower than a second preset threshold, and the oil production is lower than a third preset threshold; if the probability value is less than the first preset threshold, it is determined that the target heavy oil well is in a high production stage, wherein the high production stage is used to indicate a period of time when the liquid production of the heavy oil well is higher than the second preset threshold, or the oil production is higher than the third preset threshold.

[0053] Optionally, the target event characteristics and the target time series characteristics are input into the target prediction model to obtain the probability value of the target heavy oil well being in the low-production stage during the current sampling period. If the probability value is greater than or equal to the first preset threshold, the target heavy oil well is determined to be in the low-production stage. If the probability value is less than the first preset threshold, the target heavy oil well is determined to be in the high-production stage. For example, if the preset threshold is 0.5, when the probability value is greater than or equal to 0.5, the target heavy oil well is determined to be in the low-production stage; when the probability value is less than 0.5, the target heavy oil well is determined to be in the high-production stage. The above method can accurately determine the stage of the heavy oil well, which helps to determine the probability that the heavy oil well is a well to be treated as a turbine, so as to help optimize the steam injection scheme for the well to be treated as a turbine, adjust the injection-production balance, and improve the efficiency and effect of steam injection.

[0054] In an optional embodiment, determining a target prediction model corresponding to a target time series diagram includes: obtaining time series diagrams corresponding to multiple heavy oil wells in multiple rotation cycles, respectively, to obtain multiple time series diagrams; performing grouping processing on the multiple time series diagrams according to preset conditions, to obtain N groups of curve data; performing clustering processing on the time series diagrams respectively included in the N groups of curve data, to obtain time series clusters respectively included in the N groups of curve data; based on the preset conditions, determining target group curve data matching the target time series diagram from the N groups of curve data; determining the similarity respectively corresponding to the target time series diagram and the time series clusters included in the target group curve data; and determining the target prediction model based on the similarity respectively corresponding to the time series clusters included in the target group curve data.

[0055] Optionally, based on domain knowledge, the time series performance of indicators of different well types (vertical wells, horizontal wells), sub-layers, and different runner cycles (or runner stages, such as the 1st to 5th runner) may vary greatly. However, in actual production, even in the same well type, the same sub-layer, and the same runner cycle, the time series performance of indicators such as liquid production may also vary greatly. For example, Figure 2 and Figure 3 As shown, Figure 2This is the liquid production curve of well F3006. Figure 3 This is the liquid production curve of well F340424. Figure 2 and Figure 3 Each subgraph in is a daily liquid production time series curve of the rotor cycle. It can be seen that in the same rotor cycle, the performance of heavy oil wells of the same well type and sublayer is very different. Therefore, it is necessary to group multiple time series graphs according to well type (vertical well, horizontal well), sublayer, and rotor cycle to obtain N groups of curve data; cluster the time series graphs included in the N groups of curve data, for example, dynamic time warping (DTW) and cosine similarity can be used for clustering to obtain time series clusters included in the N groups of curve data; based on the well type, sublayer, and rotor cycle, determine the target group of curve data matching the target time series graph from the N groups of curve data, that is, the curve data of the same well type, the same sublayer, and the same rotor cycle; determine the similarity corresponding to the target time series graph and the time series clusters included in the target group of curve data, for example, the Euclidean distance can be used to calculate the straight-line distance between the time series graph and the center of the time series cluster; and then determine the target prediction model based on the similarity corresponding to the time series clusters included in the target group of curve data.

[0056] In an optional embodiment, a target prediction model is determined based on the similarities corresponding to the time series clusters included in the target group curve data, including: based on the similarities corresponding to the time series clusters included in the target group curve data, determining a first time series cluster corresponding to the maximum similarity value from the time series clusters included in the target group curve data; determining a first heavy oil well corresponding to the first time series cluster, and a first rotation cycle corresponding to the first time series cluster; and establishing a target prediction model based on the first heavy oil well and the first rotation cycle.

[0057] Optionally, the first time series cluster corresponding to the maximum similarity value is determined from the time series clusters included in the target group curve data, indicating that the target time series diagram is closer to the characteristics and attributes of the first time series cluster, and has the highest similarity with the first time series cluster. Therefore, it is necessary to establish a target prediction model based on the first heavy oil well corresponding to the first time series cluster and the first rotation cycle. The target prediction model established in the above manner can more accurately predict the probability of the target heavy oil well being in the production stage.

[0058] In an optional embodiment, based on the first heavy oil well and the first rotation cycle, a target prediction model is established, including: acquiring third production data and fourth production data; performing event feature extraction processing on the third production data to obtain a first event feature corresponding to the first heavy oil well; performing time series feature extraction processing on the fourth production data to obtain a first time series feature corresponding to the first heavy oil well; based on the first event feature and the first time series feature, an initial prediction model is trained to determine a target prediction model.

[0059] Optionally, obtain the third production data and the fourth production data; input the third production data into the event feature extraction model to obtain the first event feature corresponding to the first heavy oil well; input the fourth production data into the time series feature extraction model to obtain the first time series feature corresponding to the first heavy oil well; based on the first event feature and the first time series feature, train the initial prediction model to determine the target prediction model. Since the third production data only contains the start and end dates of each wheel and there is no low-production mark, such data cannot support the machine learning model. Therefore, the data can be converted into a supervised learning problem through two strategies: 1) Convert it into a classification problem, and consider the last N days (for example, 10 days) of each wheel in the data as a low-production stage, and the other days are non-low-production stages; 2) Convert it into a regression problem, and set the low-production probability of the last day of the rotation cycle to 1. According to the number of days from the last day, the probability gradually becomes 0. Here, smoothing functions such as Sigmod can be used for transition.

[0060] For example, when training the initial prediction model, according to the second strategy, the last 10 days of the first rotation period are marked as the low-yield stage, and the remaining days are marked as the high-yield stage, such as Figure 7 As shown, Figure 7 is a schematic diagram of an optional supervised learning transformation according to an embodiment of the present invention; the first event feature and 70% of the first time series feature are used as a training set to train an initial prediction model based on a decision tree, and 30% are used as a test set to test the trained prediction model. If the test result meets the preset condition, the trained prediction model is used as the target prediction model. Figure 8 is a schematic diagram of an optional decision tree model according to an embodiment of the present invention, such as Figure 8 The target prediction model based on the decision tree is shown. When the rotation cycle is less than 45 days, the probability of being in the low-yield stage is 0.14, accounting for 26% of the total; when the rotation cycle is greater than or equal to 45 days, it is judged whether it is greater than or equal to 202 days. If it is greater than or equal to 202 days, the probability of being in the low-yield stage is 0.34, accounting for 9% of the total. Similarly, it is finally concluded that when the rotation cycle is less than 131 days and the daily liquid production is greater than or equal to 9.3, the probability of being in the low-yield stage is 0.56, accounting for 14% of the total. When the rotation cycle is greater than or equal to 60 days, the daily liquid production is less than 9.3, and the relative liquid production in the past 5 days is greater than or equal to 0.056, the probability of being in the low-yield stage is the highest, 0.67, accounting for 32% of the total.

[0061] Through the above steps S102 to S108, the purpose of accurately determining the production stage of the target heavy oil well based on the event characteristics and time series characteristics corresponding to the target heavy oil well can be achieved, thereby achieving the technical effect of improving the accuracy of the heavy oil well production stage determination result, and further solving the technical problem of low accuracy caused by the incomplete consideration of factors in the heavy oil well production stage determination method in the related art.

[0062] Based on the above embodiments and optional embodiments, the present invention proposes an optional implementation mode, the method includes a model training stage and a model application stage, wherein the model training stage includes three parts:

[0063] The event feature extraction model is established in the following way:

[0064] Step S11, obtaining fifth production data of the second heavy oil well in the second cycle, where the fifth production data is text-type production data.

[0065] Step S12, performing word segmentation processing on the fifth production data to obtain sixth production data.

[0066] Step S13, performing word frequency statistics processing on the target event included in the sixth production data to obtain a word frequency statistics result, wherein the word frequency statistics result is used to indicate the frequency of occurrence of the target event included in the fifth production data.

[0067] Step S14, determining a phrase pattern group based on the sixth production data and word frequency statistics, wherein the phrase pattern group at least includes the target event in the fifth production data and the event type corresponding to the target event.

[0068] Step S15, clustering the phrase pattern groups to obtain clustering results.

[0069] Step S16: Based on the clustering results, the initial model is trained to obtain an event feature extraction model.

[0070] The time series feature extraction model is established in the following way:

[0071] Step S21, obtaining seventh production data of the second heavy oil well in the second cycle, wherein the seventh production data is numerical type production data.

[0072] Step S22: training the model based on the seventh production data to obtain a time series feature extraction model.

[0073] The target time series prediction model is established in the following way:

[0074] Step S31, obtaining third production data and fourth production data of the first heavy oil well in the first rotation cycle, wherein the third production data is text type production data, and the fourth production data is numerical type production data.

[0075] Step S32: performing event feature extraction processing on the third production data to obtain a first event feature corresponding to the first heavy oil well.

[0076] Step S33, performing time series feature extraction processing on the fourth production data to obtain a first time series feature corresponding to the first heavy oil well.

[0077] Step S34: Based on the first event feature and the first time series feature, the initial prediction model is trained to determine the target prediction model.

[0078] Apply the target prediction model as follows:

[0079] Step S41, obtaining a target time series diagram, first production data, and second production data of a target heavy oil well in a current sampling period, wherein the first production data is text-type production data, and the second production data is numerical-type production data.

[0080] Step S42: Based on the first production data, an event feature extraction model is used to obtain target event features.

[0081] Step S43: Based on the second production data, a time series feature extraction model is used to obtain target time series features.

[0082] Step S44, obtaining the timing diagrams corresponding to the plurality of heavy oil wells in the plurality of rotation cycles, and obtaining a plurality of timing diagrams.

[0083] Step S45, grouping the multiple time sequence diagrams according to preset conditions to obtain N groups of curve data.

[0084] Step S46, clustering processing is performed on the time series graphs respectively included in the N groups of curve data to obtain the time series clusters respectively included in the N groups of curve data.

[0085] Step S47, based on preset conditions, determining a target group of curve data that matches the target timing diagram from the N groups of curve data.

[0086] Step S48, determining the similarities between the target time series diagram and the time series clusters included in the target group curve data.

[0087] Step S49: Based on the similarities respectively corresponding to the time series clusters included in the target group curve data, a first time series cluster corresponding to the maximum similarity value is determined from the time series clusters included in the target group curve data.

[0088] Step S410, determining a first heavy oil well corresponding to a first time series cluster and a first rotation period corresponding to the first time series cluster.

[0089] Step S411, determining a target prediction model based on the first heavy oil well and the first rotation period.

[0090] Step S412, based on the target event characteristics and the target time series characteristics, a target prediction model is used to obtain a probability value that the target heavy oil well is in a low production stage during the current sampling period.

[0091] Step S413, if the probability value is greater than or equal to the first preset threshold, it is determined that the target heavy oil well is in a low production stage, wherein the low production stage is used to indicate a period of time when the liquid production of the heavy oil well is lower than the second preset threshold and the oil production is lower than the third preset threshold.

[0092] Step S414: if the probability value is less than the first preset threshold, it is determined that the target heavy oil well is in a high production stage, wherein the high production stage is used to indicate a period of time when the liquid production of the heavy oil well is higher than the second preset threshold, or the oil production is higher than the third preset threshold.

[0093] The embodiments of the present invention can at least achieve the following technical effects: by matching the production curve of the heavy oil well with the trajectory, the model is selected in a targeted manner, and the operation events of the heavy oil well are combined to avoid the interference of abnormal events, so as to accurately determine the production stage of the heavy oil well.

[0094] In this embodiment, a device for determining the production stage of a heavy oil well is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the terms "module" and "device" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0095] According to an embodiment of the present invention, there is also provided an embodiment of a device for implementing the above-mentioned method for determining the production stage of a heavy oil well. Fig. 9 is a schematic diagram of the structure of a device for determining the production stage of a heavy oil well according to an embodiment of the present invention. Fig. 9 As shown, the above-mentioned device for determining the production stage of a heavy oil well includes: a first acquisition module 902, a first processing module 904, a second processing module 906, and a first determination module 908, wherein:

[0096] A first acquisition module 902 is used to acquire a target time series diagram, first production data, and second production data of a target heavy oil well in a current sampling period, wherein the first production data is text-type production data, and the second production data is numerical-type production data;

[0097] The first processing module 904 is connected to the first acquisition module 902 and is used to perform event feature extraction processing on the first production data to obtain target event features corresponding to the target heavy oil well;

[0098] The second processing module 906 is connected to the first processing module 904 and is used to perform time series feature extraction processing on the second production data to obtain target time series features corresponding to the target heavy oil well;

[0099] The first determination module 908 is connected to the second processing module 906, and is used to determine the production stage of the target heavy oil well in the current sampling period based on the target time series diagram, the target event characteristics, and the target time series characteristics, wherein the production stage includes a low production stage and a high production stage.

[0100] By setting a first acquisition module, it is used to obtain the target time series diagram, the first production data, and the second production data of the target heavy oil well in the current sampling period, wherein the first production data is text type production data, and the second production data is numerical type production data; the first processing module is used to extract the event feature of the first production data to obtain the target event feature corresponding to the target heavy oil well; the second processing module is used to extract the time series feature of the second production data to obtain the target time series feature corresponding to the target heavy oil well; the first determination module is used to determine the production stage of the target heavy oil well in the current sampling period based on the target time series diagram, the target event feature, and the target time series feature, wherein the production stage includes a low production stage and a high production stage. The purpose of accurately determining the production stage of the target heavy oil well based on the event feature and time series feature corresponding to the target heavy oil well is achieved, thereby achieving the technical effect of improving the accuracy of the determination result of the production stage of the heavy oil well, and then solving the technical problem of low accuracy caused by the incomplete consideration of factors in the production stage determination method of the heavy oil well in the related technology.

[0101] In an optional embodiment, the above-mentioned first determination module includes: a first determination submodule, used to determine the target prediction model corresponding to the target time series diagram, wherein the target prediction model is based on the third production data of the first heavy oil well in the first rotation cycle, and the fourth production data, trained by machine training, the third production data is text type production data, and the fourth production data is numerical type production data; a second determination submodule, used to determine the production stage of the target heavy oil well in the current sampling period based on the target event characteristics and the target time series characteristics, using the target prediction model.

[0102] In an optional embodiment, the above-mentioned second determination submodule includes: a third determination submodule, which is used to obtain the probability value of the target heavy oil well being in the low production stage in the current sampling period by using the target prediction model based on the target event characteristics and the target time series characteristics; a fourth determination submodule, which is used to determine that the target heavy oil well is in the low production stage if the probability value is greater than or equal to the first preset threshold, wherein the low production stage is used to indicate a period of time when the liquid production of the heavy oil well is lower than the second preset threshold and the oil production is lower than the third preset threshold; a fifth determination submodule, which is used to determine that the target heavy oil well is in the high production stage if the probability value is less than the first preset threshold, wherein the high production stage is used to indicate a period of time when the liquid production of the heavy oil well is higher than the second preset threshold or the oil production is higher than the third preset threshold.

[0103] In an optional embodiment, the above-mentioned first determination submodule includes: a first acquisition submodule, which is used to obtain the timing diagrams corresponding to multiple heavy oil wells in multiple rotation cycles respectively, and obtain multiple timing diagrams; a first grouping submodule, which is used to group the multiple timing diagrams according to preset conditions to obtain N groups of curve data; a first processing submodule, which is used to cluster the timing diagrams respectively included in the N groups of curve data, and obtain the timing clusters respectively included in the N groups of curve data; a first matching submodule, which is used to determine the target group curve data matching the target timing diagram from the N groups of curve data based on preset conditions; a sixth determination submodule, which is used to determine the similarity corresponding to the target timing diagram and the timing clusters included in the target group curve data; a seventh determination submodule, which is used to determine the target prediction model based on the similarity corresponding to the timing clusters included in the target group curve data.

[0104] In an optional embodiment, the seventh determination submodule includes: an eighth determination submodule, which is used to determine the first time series cluster corresponding to the maximum similarity value from the time series clusters included in the target group curve data based on the similarities corresponding to the time series clusters included in the target group curve data; a ninth determination submodule, which is used to determine the first heavy oil well corresponding to the first time series cluster, and the first rotation cycle corresponding to the first time series cluster; and a first establishment submodule, which is used to establish a target prediction model based on the first heavy oil well and the first rotation cycle.

[0105] In an optional embodiment, the above-mentioned first establishment submodule includes: a second acquisition submodule, used to acquire third production data and fourth production data; a second processing submodule, used to perform event feature extraction processing on the third production data to obtain a first event feature corresponding to the first heavy oil well; a third processing submodule, used to perform time series feature extraction processing on the fourth production data to obtain a first time series feature corresponding to the first heavy oil well; a first training submodule, used to train the initial prediction model based on the first event feature and the first time series feature to determine the target prediction model.

[0106] In an optional embodiment, the above-mentioned first processing module includes: a tenth determination submodule, which is used to obtain target event features based on the first production data by using an event feature extraction model, wherein the event feature extraction model is based on the fifth production data of the second heavy oil well in the second cycle, and is trained by machine training, and the fifth production data is text-type production data.

[0107] In an optional embodiment, the above-mentioned device also includes: a third acquisition submodule, used to acquire the fifth production data; a fourth processing submodule, used to perform word segmentation processing on the fifth production data to obtain the sixth production data; a fifth processing submodule, used to perform word frequency statistics processing on the target event included in the sixth production data to obtain a word frequency statistics result, wherein the word frequency statistics result is used to indicate the frequency of occurrence of the target event included in the fifth production data; an eleventh determination submodule, used to determine the phrase pattern group based on the sixth production data and the word frequency statistics result, wherein the phrase pattern group at least includes the target event in the fifth production data and the event type corresponding to the target event; the sixth processing submodule, used to perform clustering processing on the phrase pattern group to obtain a clustering result; and a second training submodule, used to train the initial model based on the clustering result to obtain an event feature extraction model.

[0108] In an optional embodiment, the above-mentioned second processing module includes: a twelfth determination submodule, which is used to obtain target time series features based on the second production data by using a time series feature extraction model, wherein the time series feature extraction model is based on the seventh production data of the second heavy oil well in the second cycle, and is trained by machine training, and the seventh production data is a numerical type of production data.

[0109] It should be noted that the above modules can be implemented by software or hardware. For example, for the latter, it can be implemented in the following ways: the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0110] It should be noted that the first acquisition module 902, the first processing module 904, the second processing module 906, and the first determination module 908 correspond to steps S102 to S108 in the embodiment, and the examples and application scenarios implemented by the modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above embodiments. It should be noted that the modules as part of the device can be run in a computer terminal.

[0111] It should be noted that the optional or preferred implementation of this embodiment can refer to the relevant description in the embodiment, which will not be repeated here.

[0112] The above-mentioned heavy oil well production stage determination device may also include a processor and a memory. The above-mentioned first acquisition module 902, the first processing module 904, the second processing module 906, the first determination module 908, etc. are all stored in the memory as program modules, and the processor executes the above-mentioned program modules stored in the memory to realize the corresponding functions.

[0113] The processor includes a kernel, which retrieves the corresponding program module from the memory. The kernel may be one or more. The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. The memory includes at least one memory chip.

[0114] According to an embodiment of the present application, an embodiment of a non-volatile storage medium is also provided. Optionally, in this embodiment, the non-volatile storage medium includes a stored program, wherein when the program is executed, the device where the non-volatile storage medium is located is controlled to execute any of the above-mentioned methods for determining the production stage of a heavy oil well.

[0115] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group, and the non-volatile storage medium includes a stored program.

[0116] Optionally, when the program is running, the device where the non-volatile storage medium is located is controlled to perform the following functions: obtain a target timing diagram, first production data, and second production data of the target heavy oil well in the current sampling period, wherein the first production data is text-type production data and the second production data is numerical-type production data; perform event feature extraction processing on the first production data to obtain target event features corresponding to the target heavy oil well; perform time series feature extraction processing on the second production data to obtain target time series features corresponding to the target heavy oil well; based on the target timing diagram, target event features, and target time series features, determine the production stage of the target heavy oil well in the current sampling period, wherein the production stage includes a low-production stage and a high-production stage.

[0117] According to an embodiment of the present application, an embodiment of a processor is also provided. Optionally, in this embodiment, the processor is used to run a program, wherein when the program is run, any one of the above-mentioned methods for determining the production stage of a heavy oil well is executed.

[0118] According to an embodiment of the present application, an embodiment of a computer program product is also provided, which, when executed on a data processing device, is suitable for executing a program that initializes any of the above-mentioned steps of the method for determining the production stage of a heavy oil well.

[0119] Optionally, the above-mentioned computer program product, when executed on a data processing device, is suitable for executing a program that is initialized with the following method steps: obtaining a target timing diagram, first production data, and second production data of a target heavy oil well in a current sampling period, wherein the first production data is text-type production data and the second production data is numerical-type production data; performing event feature extraction processing on the first production data to obtain target event features corresponding to the target heavy oil well; performing time series feature extraction processing on the second production data to obtain target time series features corresponding to the target heavy oil well; based on the target time series diagram, the target event features, and the target time series features, determining the production stage of the target heavy oil well in the current sampling period, wherein the production stage includes a low-production stage and a high-production stage.

[0120] Optionally, the above-mentioned computer program product is also suitable for executing a program initialized with the following method steps: determining a target prediction model corresponding to a target time series diagram, wherein the target prediction model is based on the third production data of the first heavy oil well in the first cycle, and the fourth production data, trained by machine training, the third production data is text type production data, and the fourth production data is numerical type production data; based on target event characteristics and target time series characteristics, the target prediction model is used to determine the production stage of the target heavy oil well in the current sampling period.

[0121] Optionally, the above-mentioned computer program product is also suitable for executing a program initialized with the following method steps: based on the target event characteristics and the target time series characteristics, using the target prediction model, obtain the probability value of the target heavy oil well being in the low production stage during the current sampling period; if the probability value is greater than or equal to the first preset threshold, it is determined that the target heavy oil well is in the low production stage, wherein the low production stage is used to indicate the period when the liquid production of the heavy oil well is lower than the second preset threshold, and the oil production is lower than the third preset threshold; if the probability value is less than the first preset threshold, it is determined that the target heavy oil well is in the high production stage, wherein the high production stage is used to indicate the period when the liquid production of the heavy oil well is higher than the second preset threshold, or the oil production is higher than the third preset threshold.

[0122] Optionally, the above-mentioned computer program product is also suitable for executing a program that is initialized with the following method steps: obtaining time series diagrams corresponding to multiple heavy oil wells in multiple rotation cycles respectively, and obtaining multiple time series diagrams; grouping the multiple time series diagrams according to preset conditions to obtain N groups of curve data; clustering the time series diagrams respectively included in the N groups of curve data to obtain time series clusters respectively included in the N groups of curve data; based on the preset conditions, determining the target group of curve data that matches the target time series diagram from the N groups of curve data; determining the similarity corresponding to the target time series diagram and the time series clusters included in the target group of curve data; and determining the target prediction model based on the similarity corresponding to the time series clusters included in the target group of curve data.

[0123] Optionally, the above-mentioned computer program product is also suitable for executing an initialized program having the following method steps: based on the similarities corresponding to the time series clusters included in the target group curve data, determining the first time series cluster corresponding to the maximum similarity from the time series clusters included in the target group curve data; determining the first heavy oil well corresponding to the first time series cluster, and the first rotation cycle corresponding to the first time series cluster; and establishing a target prediction model based on the first heavy oil well and the first rotation cycle.

[0124] Optionally, the above-mentioned computer program product is also suitable for executing a program initialized with the following method steps: obtaining third production data and fourth production data; performing event feature extraction processing on the third production data to obtain a first event feature corresponding to the first heavy oil well; performing time series feature extraction processing on the fourth production data to obtain a first time series feature corresponding to the first heavy oil well; based on the first event feature and the first time series feature, training the initial prediction model to determine the target prediction model.

[0125] Optionally, the above-mentioned computer program product is also suitable for executing an initialized program having the following method steps: based on the first production data, an event feature extraction model is used to obtain target event features, wherein the event feature extraction model is based on the fifth production data of the second heavy oil well in the second cycle, and is trained by machine training, and the fifth production data is text-type production data.

[0126] Optionally, the above-mentioned computer program product is also suitable for executing an initialized program having the following method steps: obtaining fifth production data; performing word segmentation processing on the fifth production data to obtain sixth production data; performing word frequency statistics processing on the target event included in the sixth production data to obtain word frequency statistics results, wherein the word frequency statistics results are used to indicate the frequency of occurrence of the target event included in the fifth production data; based on the sixth production data and the word frequency statistics results, determining a phrase pattern group, wherein the phrase pattern group at least includes the target event in the fifth production data and the event type corresponding to the target event; performing clustering processing on the phrase pattern group to obtain clustering results; based on the clustering results, training the initial model to obtain an event feature extraction model.

[0127] Optionally, the above-mentioned computer program product is also suitable for executing an initialized program having the following method steps: based on the second production data, a time series feature extraction model is used to obtain a target time series feature, wherein the time series feature extraction model is based on the seventh production data of the second heavy oil well in the second cycle, and is trained by machine training, and the seventh production data is a numerical type of production data.

[0128] like Fig.10As shown, an embodiment of the present invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are implemented: obtaining a target timing diagram, first production data, and second production data of a target heavy oil well in a current sampling period, wherein the first production data is text-type production data, and the second production data is numerical-type production data; performing event feature extraction processing on the first production data to obtain target event features corresponding to the target heavy oil well; performing time series feature extraction processing on the second production data to obtain target time series features corresponding to the target heavy oil well; and determining the production stage of the target heavy oil well in the current sampling period based on the target timing diagram, the target event features, and the target time series features, wherein the production stage includes a low-production stage and a high-production stage.

[0129] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0130] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the above modules can be a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, modules or indirect coupling or communication connection of modules, which can be electrical or other forms.

[0131] The modules described above as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment.

[0132] In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of software functional modules.

[0133] If the above-mentioned integrated module is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable non-volatile storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a non-volatile storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the various embodiments of the present invention. The aforementioned non-volatile storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0134] The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for determining the production stage of a heavy oil well, characterized in that: include: Obtaining a target time series diagram, first production data, and second production data of a target heavy oil well in a current sampling period, wherein the first production data is text-type production data, and the second production data is numerical-type production data; Performing event feature extraction processing on the first production data to obtain target event features corresponding to the target heavy oil well; Performing time series feature extraction processing on the second production data to obtain target time series features corresponding to the target heavy oil well; Based on the target time sequence diagram, the target event characteristics, and the target time sequence characteristics, the production stage of the target heavy oil well in the current sampling period is determined, wherein the production stage includes a low-production stage and a high-production stage.

2. The method according to claim 1, characterized in that The determining, based on the target time sequence diagram, the target event feature, and the target time sequence feature, the production stage of the target heavy oil well in the current sampling period includes: Determine a target prediction model corresponding to the target time series diagram, wherein the target prediction model is obtained by training in a machine training manner based on third production data and fourth production data of the first heavy oil well in the first rotation cycle, the third production data is text-type production data, and the fourth production data is numerical-type production data; Based on the target event characteristics and the target time series characteristics, the target prediction model is adopted to determine the production stage of the target heavy oil well in the current sampling period.

3. The method according to claim 2, characterized in that The step of determining the production stage of the target heavy oil well in the current sampling period based on the target event characteristics and the target time series characteristics by using the target prediction model includes: Based on the target event characteristics and the target time series characteristics, the target prediction model is used to obtain a probability value that the target heavy oil well is in a low-production stage during the current sampling period; If the probability value is greater than or equal to a first preset threshold, it is determined that the target heavy oil well is in the low production stage, wherein the low production stage is used to indicate a period of time when the liquid production of the heavy oil well is lower than a second preset threshold and the oil production is lower than a third preset threshold; If the probability value is less than the first preset threshold, it is determined that the target heavy oil well is in the high production stage, wherein the high production stage is used to indicate a period of time when the liquid production of the heavy oil well is higher than the second preset threshold, or the oil production is higher than the third preset threshold.

4. The method according to claim 2, characterized in that: The determining of the target prediction model corresponding to the target timing diagram includes: Obtaining time sequence diagrams corresponding to a plurality of heavy oil wells in a plurality of rotation cycles, respectively, to obtain a plurality of time sequence diagrams; The plurality of time sequence diagrams are grouped according to preset conditions to obtain N groups of curve data; Performing clustering processing on the time series graphs respectively included in the N groups of curve data to obtain the time series clusters respectively included in the N groups of curve data; Based on the preset condition, determining a target group of curve data matching the target timing diagram from the N groups of curve data; Determine the similarity between the target timing diagram and the timing clusters included in the target group curve data; The target prediction model is determined based on the similarities respectively corresponding to the time series clusters included in the target group curve data.

5. The method according to claim 4, characterized in that The determining the target prediction model based on the similarities respectively corresponding to the time series clusters included in the target group curve data includes: Based on the similarities respectively corresponding to the time series clusters included in the target group curve data, determining a first time series cluster corresponding to the maximum similarity value from the time series clusters included in the target group curve data; Determining the first heavy oil well corresponding to the first time series cluster and the first rotation period corresponding to the first time series cluster; The target prediction model is established based on the first heavy oil well and the first cycle.

6. The method according to claim 5, characterized in that The establishing of the target prediction model based on the first heavy oil well and the first cycle includes: acquiring the third production data and the fourth production data; performing event feature extraction processing on the third production data to obtain a first event feature corresponding to the first heavy oil well; Performing time series feature extraction processing on the fourth production data to obtain a first time series feature corresponding to the first heavy oil well; Based on the first event feature and the first time series feature, an initial prediction model is trained to determine the target prediction model.

7. The method according to any one of claims 1 to 6, characterized in that: The performing event feature extraction processing on the first production data to obtain target event features corresponding to the target heavy oil well includes: Based on the first production data, an event feature extraction model is used to obtain the target event feature, wherein the event feature extraction model is based on the fifth production data of the second heavy oil well in the second cycle and is trained by machine training, and the fifth production data is text-type production data.

8. The method according to claim 7, characterized in that Before obtaining the target event feature by using an event feature extraction model based on the first production data, the method further includes: acquiring the fifth production data; Performing word segmentation processing on the fifth production data to obtain sixth production data; Performing word frequency statistics processing on the target event included in the sixth production data to obtain a word frequency statistics result, wherein the word frequency statistics result is used to indicate the frequency of occurrence of the target event included in the fifth production data; Based on the sixth production data and the word frequency statistics result, determining a phrase pattern group, wherein the phrase pattern group at least includes a target event in the fifth production data and an event type corresponding to the target event; Performing clustering processing on the phrase pattern group to obtain a clustering result; Based on the clustering results, the initial model is trained to obtain the event feature extraction model.

9. The method according to claim 8, characterized in that The performing time series feature extraction processing on the second production data to obtain the target time series feature includes: Based on the second production data, a time series feature extraction model is used to obtain the target time series feature, wherein the time series feature extraction model is trained by machine training based on the seventh production data of the second heavy oil well in the second cycle, and the seventh production data is numerical type production data.

10. A device for determining the production stage of a heavy oil well, characterized in that: include: A first acquisition module is used to acquire a target time series diagram, first production data, and second production data of a target heavy oil well in a current sampling period, wherein the first production data is text-type production data, and the second production data is numerical-type production data; A first processing module is used to perform event feature extraction processing on the first production data to obtain target event features corresponding to the target heavy oil well; A second processing module is used to perform time series feature extraction processing on the second production data to obtain a target time series feature corresponding to the target heavy oil well; The first determination module is used to determine the production stage of the target heavy oil well in the current sampling period based on the target time sequence diagram, the target event characteristics, and the target time sequence characteristics, wherein the production stage includes a low production stage and a high production stage.

11. An electronic device, characterized in that: It comprises one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method for determining the production stage of a heavy oil well as described in any one of claims 1 to 9.

Images