Thickened oil temperature working condition diagnosis method, device and system and storage medium

Abstract

The invention relates to the technical field of thickened oil development, in particular to a thickened oil temperature working condition diagnosis method, device and system and a storage medium. The thickened oil temperature working condition diagnosis method comprises the steps that the real value of a key temperature parameter at the current moment in actual thickened oil production is obtained, and the real value of the key temperature parameter at the current moment is input into a temperature prediction model; and outputting a key temperature parameter prediction value at the next moment, inputting the key temperature parameter prediction value into the working condition identification model, and outputting a working condition identification result. According to the method, the temperature prediction model and the working condition identification model are constructed, the key temperature parameter prediction value at the next moment is predicted through the temperature prediction model by using the key temperature parameter true value at the current moment in actual thickened oil production, and the abnormal working condition is identified in combination with the key temperature parameter prediction value, so that technicians are assisted in process tracing; advanced treatment of abnormal working conditions is achieved, effective control and reduction of the temperature of each node are achieved, invalid circulation of steam on the ground is reduced, and the purposes of increasing the thickened oil production heat energy utilization rate and reducing energy consumption are achieved.

Description

Technical Field

[0001] The invention relates to the technical field of heavy oil development, and relates to a heavy oil temperature condition diagnosis method, device, system and storage medium. Background Art

[0002] Steam injection is the lifeline of heavy oil development. Heavy oil itself has high viscosity and solidification, which greatly increases the difficulty of extraction. Thermal oil recovery occupies a dominant position in the heavy oil extraction process. No matter which thermal recovery technology is used, the purpose is to inject the most heat into the reservoir. Heavy oil accounts for 32% of Xinjiang Oilfield, with an annual steam injection volume of nearly 39 million tons and steam costs of nearly 3 billion yuan. However, due to the temperature difference between the high-temperature fluid and the surrounding formation, the hot fluid will inevitably produce heat loss during the injection of the reservoir through the wellbore, and the temperature will also change accordingly. Therefore, it is very important to monitor the temperature conditions of the entire process of heavy oil.

[0003] Currently, most methods rely on manual experience to identify abnormal temperature conditions in the entire heavy oil process and manually adjust core parameters. This method of manual experience to identify abnormal temperature conditions in the entire heavy oil process and adjust core parameters has problems such as long data preparation cycle and uncontrollable identification accuracy. It also has certain safety risks and is difficult to meet the requirements of efficient and stable identification. Summary of the invention

[0004] The present invention provides a method, which overcomes the above-mentioned deficiencies of the prior art and can effectively solve the problems of low efficiency and uncontrollable accuracy in the existing method of monitoring the temperature condition of the whole process of heavy oil by relying on manual experience.

[0005] One of the technical solutions of the present invention is achieved by the following measures: A method for diagnosing heavy oil temperature conditions, comprising: Obtain the real values ​​of key temperature parameters at the current moment in actual heavy oil production, where the real values ​​of key temperature parameters include metering line temperature, oil gathering line temperature, steam injection line temperature, steam injection line pressure, boiler temperature, and closed station temperature; Input the real value of the key temperature parameter at the current moment into the temperature prediction model, and output the predicted value of the key temperature parameter at the next moment, wherein the temperature prediction model is obtained by training and learning using multiple sets of sample data, and each set of sample data in the multiple sets of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production; The predicted value of the key temperature parameter is input into the operating condition identification model, and the operating condition identification result is output, wherein the operating condition identification model is obtained by training and learning using multiple groups of sample data, and each group of sample data in the multiple groups of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production and identification information of corresponding operating condition levels.

[0006] The following are further optimizations and / or improvements to the above technical solutions: The construction process of the above temperature prediction model includes: Construct a historical data set, pre-process the data in the historical data set, and fit the discrete data in the historical data set using a gradient boosting algorithm, wherein the historical data set includes historical values ​​of several key temperature parameters in the entire process of actual heavy oil production; Divide the historical data set into a training data subset and a test data subset; The long short-term memory recurrent neural network model is trained using the training data subset to obtain a temperature prediction model; The temperature prediction model is tested using a subset of test data, and the long short-term memory recurrent neural network model is adjusted and trained accordingly based on the test results until the optimal temperature prediction model is obtained.

[0007] The construction process of the above working condition identification model includes: Construct a historical data set, preprocess the data in the historical data set, and fit the discrete data in the historical data set in combination with a gradient boosting algorithm, wherein the historical data set includes historical values ​​of several key temperature parameters in the entire process of actual heavy oil production and identification information of corresponding operating conditions; Divide the historical data set into a training data subset and a test data subset; The deep decision tree model is trained using the training data subset to obtain the working condition recognition model; The operating condition recognition model is tested using a subset of test data, and the deep decision tree model is adjusted and trained accordingly based on the test results until the optimal temperature prediction model is obtained.

[0008] The above-mentioned preprocessing of the data in the historical data set includes: removing unique attributes, processing missing values, identifying and deleting abnormal data, data standardization, data regularization, and extracting eigenvalues ​​through principal component analysis.

[0009] The above also includes inputting the predicted value of the key temperature parameter into the working condition identification model, outputting the working condition identification result, and outputting corresponding alarm information for the abnormal working condition identification result.

[0010] The second technical solution of the present invention is achieved by the following measures: a device for diagnosing heavy oil temperature conditions, comprising: A data acquisition unit, which acquires the real values ​​of key temperature parameters at the current moment in actual heavy oil production, wherein the real values ​​of key temperature parameters include metering line temperature, oil gathering line temperature, steam injection line temperature, steam injection line pressure, boiler temperature, and closed station temperature; The temperature prediction unit inputs the real value of the key temperature parameter at the current moment into the temperature prediction model, and outputs the predicted value of the key temperature parameter at the next moment, wherein the temperature prediction model is obtained by training and learning using multiple sets of sample data, and each set of sample data in the multiple sets of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production; The operating condition identification unit inputs the predicted value of the key temperature parameter into the operating condition identification model and outputs the operating condition identification result, wherein the operating condition identification model is obtained by training and learning using multiple groups of sample data, and each group of sample data in the multiple groups of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production and identification information of corresponding operating condition levels.

[0011] The following are further optimizations and / or improvements to the above technical solutions: The above also includes a first model building unit and a second model building unit; The first model building unit builds a temperature prediction model, including: Construct a historical data set, pre-process the data in the historical data set, and fit the discrete data in the historical data set using a gradient boosting algorithm, wherein the historical data set includes historical values ​​of several key temperature parameters in the entire process of actual heavy oil production; Divide the historical data set into a training data subset and a test data subset; The long short-term memory recurrent neural network model is trained using the training data subset to obtain a temperature prediction model; The temperature prediction model is tested using a subset of test data, and the long short-term memory recurrent neural network model is adjusted and trained accordingly based on the test results until the optimal temperature prediction model is obtained.

[0012] The second model building unit builds a working condition identification model, including: Construct a historical data set, preprocess the data in the historical data set, and fit the discrete data in the historical data set in combination with a gradient boosting algorithm, wherein the historical data set includes historical values ​​of several key temperature parameters in the entire process of actual heavy oil production and identification information of corresponding operating conditions; Divide the historical data set into a training data subset and a test data subset; The deep decision tree model is trained using the training data subset to obtain the working condition recognition model; The operating condition recognition model is tested using a subset of test data, and the deep decision tree model is adjusted and trained accordingly based on the test results until the optimal temperature prediction model is obtained.

[0013] The third technical solution of the present invention is achieved by the following measures: a heavy oil temperature condition diagnosis system, comprising: A heavy oil temperature condition diagnosis device, wherein the heavy oil temperature condition diagnosis device is the heavy oil temperature condition diagnosis device as described in the above technical solution; The interactive device is used for the operator to communicate with the heavy oil temperature condition diagnosis device, so that the operator provides the heavy oil temperature condition diagnosis device with an interactive interface of the data to be predicted through the interactive unit.

[0014] The present invention introduces machine learning, utilizes historical data accumulated in heavy oil production, constructs a temperature prediction model and an operating condition identification model, and uses the temperature prediction model to utilize the real value of a key temperature parameter at the current moment in actual heavy oil production to predict the predicted value of the key temperature parameter at the next moment, and identifies abnormal operating conditions in combination with the predicted value of the key temperature parameter, thereby assisting technical personnel in tracing the process, locating abnormal nodes, and implementing advance disposal of abnormal operating conditions, effectively alleviating the pressure of back-end process processing, thereby scientifically and effectively guiding the staff to better perform heavy oil production equipment maintenance, realize effective control and reduction of the temperature of each node, reduce the ineffective circulation of steam on the ground, achieve the purpose of improving the thermal energy utilization rate of heavy oil production and reducing energy consumption, and realize more efficient and scientific mining. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Attached Figure 1 It is a schematic diagram of the method flow of the present invention.

[0016] Attached Figure 2 This is a schematic diagram of the results of the LSTM model in the present invention.

[0017] Attached Figure 3 The figure is a schematic diagram of the structure of a device of the present invention.

[0018] Attached Figure 4 This is a schematic diagram of another device structure of the present invention.

[0019] Attached Figure 5 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0020] The present invention is not limited by the following embodiments, and specific implementation methods can be determined based on the technical solution of the present invention and actual conditions.

[0021] The present invention will be further described below in conjunction with embodiments and drawings: Embodiment 1: As attached Figure 1 As shown, the embodiment of the present invention discloses a method for diagnosing heavy oil temperature conditions, comprising: Step S110, obtaining the real values ​​of key temperature parameters at the current moment in actual heavy oil production, wherein the real values ​​of key temperature parameters include metering line temperature, oil gathering line temperature, steam injection line temperature, steam injection line pressure, boiler temperature, and closed station temperature; Step S120, inputting the real value of the key temperature parameter at the current moment into the temperature prediction model, and outputting the predicted value of the key temperature parameter at the next moment, wherein the temperature prediction model is obtained by training and learning using multiple sets of sample data, and each set of sample data in the multiple sets of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production; Step S130, inputting the predicted value of the key temperature parameter into the operating condition identification model, and outputting the operating condition identification result, wherein the operating condition identification model is obtained by training and learning using multiple groups of sample data, and each group of sample data in the multiple groups of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production and identification information of corresponding operating condition levels.

[0022] In the above step S130, the predicted value of the key temperature parameter is input into the operating condition identification model, and after the operating condition identification result is output, the corresponding alarm information is output for the abnormal operating condition identification result.

[0023] The embodiment of the present invention discloses a method for diagnosing temperature working conditions of heavy oil. The method introduces machine learning, utilizes historical data accumulated in heavy oil production, constructs a temperature prediction model and a working condition identification model, utilizes the real value of a key temperature parameter at the current moment in actual heavy oil production through the temperature prediction model, predicts the predicted value of the key temperature parameter at the next moment, and identifies abnormal working conditions in combination with the predicted value of the key temperature parameter, assists technical personnel in tracing the process, locates abnormal nodes, realizes advanced disposal of abnormal working conditions, effectively reduces the pressure of back-end process processing, thereby scientifically and effectively guiding the staff to better perform maintenance of heavy oil mining equipment, realizes effective control and reduction of the temperature of each node, reduces the ineffective circulation of steam on the ground, achieves the purpose of improving the thermal energy utilization rate of heavy oil production and reducing energy consumption, and realizes more efficient and scientific mining.

[0024] Embodiment 2: The embodiment of the present invention discloses a method for diagnosing heavy oil temperature conditions, comprising: Step S210, obtaining the real values ​​of key temperature parameters at the current moment in actual heavy oil production, wherein the real values ​​of key temperature parameters include metering line temperature, oil gathering line temperature, steam injection line temperature, steam injection line pressure, boiler temperature, and closed station temperature.

[0025] In the above step S210, the key links in the actual heavy oil production process, namely, boilers, single wells, manifolds, closed gathering and transmission transfer stations, are combined to obtain the real values ​​of key temperature parameters at the boiler outlet, single well produced liquid, manifold oil collection line, closed outlet, and processing station inlet at the current moment in the actual heavy oil production, specifically including metering line temperature, oil collection line temperature, steam injection line temperature, steam injection line pressure, boiler temperature, and closed station temperature.

[0026] After obtaining the true values ​​of the key temperature parameters at the current moment in actual heavy oil production, the true values ​​of the key temperature parameters can also be preprocessed, and the preprocessing includes: removing unique attributes, processing missing values, identifying and deleting abnormal data, data standardization, data regularization, and extracting eigenvalues ​​through principal component analysis.

[0027] Step S220, inputting the real value of the key temperature parameter at the current moment into the temperature prediction model, and outputting the predicted value of the key temperature parameter at the next moment, wherein the temperature prediction model is obtained by training and learning using multiple groups of sample data, and each group of sample data in the multiple groups of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production.

[0028] The process of constructing the temperature prediction model in step S220 includes: (1) Construct a historical data set, preprocess the data in the historical data set, and fit the discrete data in the historical data set using a gradient boosting algorithm. The historical data set includes the historical values ​​of several key temperature parameters in the entire process of actual heavy oil production. The above preprocessing of the data in the historical data set ensures the accuracy of subsequent analysis and prediction, including removing unique attributes, processing missing values, identifying and deleting abnormal data, data standardization, data regularization, and extracting eigenvalues ​​by principal component analysis. Among them, the processing of missing values ​​can be carried out by clustering filling model, regression filling model, modeling prediction, multiple interpolation and other methods; abnormal data identification and deletion can be combined with the application of statistical test methods, setting the detection level, specifying the significance level α of detecting abnormal values ​​to be 0.05, setting the elimination level, and specifying the significance level α of detecting highly abnormal abnormal values ​​to be 0.01; data standardization is achieved by using the min-max standardization model; data regularization is achieved by using the z-score standardization model; and the principal component analysis (PCA) algorithm is used to extract the main eigenvalues ​​of the data.

[0029] The above-mentioned combination of gradient boosting algorithm to fit the discrete data in the historical data set is to use the gradient boosting (GBDT) algorithm to establish a GBDT model, use the GBDT model to classify and identify discrete data, and then use the continuous gradient function to fit and regress the discrete data, providing a basis for data feature mining.

[0030] The gradient boosting (GBDT) algorithm used here is one of the best algorithms for fitting the true distribution in traditional machine learning algorithms. Whether GBDT is used for classification or regression, it has always used the CART regression tree. GBDT is an algorithm that achieves data classification or regression by adopting an additive model, that is, a linear combination of basis functions, and continuously reducing the residuals generated by the training process. GBDT goes through multiple rounds of iterations, and each round of iteration generates a weak classifier, and each classifier is trained based on the residuals of the previous round of classifiers. The specific GBDT classification process is an existing well-known technology and will not be repeated here.

[0031] (2) Divide the historical data set into a training data subset and a test data subset. Specifically, 80% of the historical data set can be divided into a training data subset, and 20% of the historical data set can be divided into a test data subset.

[0032] (3) The long short-term memory recurrent neural network model is trained using the training data subset to obtain the temperature prediction model.

[0033] In this step, a long short-term memory recurrent neural network model is constructed using the long short-term memory recurrent neural network algorithm, and a temperature prediction model is obtained using the long short-term memory recurrent neural network model. The input value of the temperature prediction model is the true value of the key temperature parameter at the current moment, and the output value is the predicted value of the key temperature parameter at the next moment.

[0034] The Long Short-Term Memory Recurrent Neural Network (LSTM) algorithm is a special RNN that is mainly used to solve the gradient vanishing and gradient exploding problems in the long sequence training process. The difference between the two is that the ordinary RNN has only one weight matrix in a single loop structure, while the LSTM has four weight matrices in a single loop structure (also called a cell). Compared with RNN, the LSTM loop structure maintains a persistent cell state and continuously passes it on, which can perform better in longer sequences. The structure of LSTM is shown in the attached figure. Figure 2As shown in the figure, LSTM contains a forget gate (determines what information to discard from the previous state), an input gate (determines what information to save in the current state), and an output gate (determines what information to output from the state). The forget gate mainly performs linear classification on the previous information and the current information, and uses sigmoid to control the value between [0,1], and multiplies the value with the previous state, so that it can be determined which states are retained and which are forgotten; the input gate has two branches, one branch is similar to the forget gate (with different parameters), and also controls the value between [0,1], and the other branch uses the tanh activation function to control the value between [-1,1], in order to select the required information, and extract the information state after multiplying the two; the output gate adds the structures of the first two gates to obtain the final information C, and the output gate is mainly used to generate the final result. After processing in the same way, the previously obtained C is processed by the tanh activation function (mainly to ensure the consistency of data size), and the two are multiplied to obtain the final result.

[0035] (4) The temperature prediction model is tested using a subset of the test data. The long short-term memory recurrent neural network model is adjusted and trained accordingly based on the test results until the optimal temperature prediction model is obtained.

[0036] The above-mentioned combination of test results corresponds to adjusting and training the long short-term memory recurrent neural network model until the optimal temperature prediction model is obtained. Specifically, the temperature prediction model is tested using a subset of test data, and the prediction accuracy of the temperature prediction model is evaluated in combination with the test results. In response to the prediction accuracy meeting the set value, the temperature prediction model is output. In response to the prediction accuracy not meeting the set value, the long short-term memory recurrent neural network model parameters are adjusted, and the long short-term memory recurrent neural network model is retrained using a subset of training data.

[0037] Step S230, inputting the predicted value of the key temperature parameter into the operating condition identification model, and outputting the operating condition identification result, wherein the operating condition identification model is obtained by training and learning using multiple groups of sample data, and each group of sample data in the multiple groups of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production and identification information of corresponding operating condition levels.

[0038] In the above step S230, constructing the operating condition identification model includes: (1) Construct a historical data set, preprocess the data in the historical data set, and fit the discrete data in the historical data set using a gradient boosting algorithm. The historical data set includes the historical values ​​of several key temperature parameters in the entire process of actual heavy oil production and the identification information of the corresponding operating conditions. (2) Divide the historical data set into a training data subset and a test data subset; (3) Use the training data subset to train the deep decision tree model to obtain the working condition recognition model; (4) Use the test data subset to test the operating condition identification model, and adjust and train the deep decision tree model accordingly based on the test results until the optimal temperature prediction model is obtained.

[0039] The training process of the above steps (1) to (4) is the same as the model training process in step S220. In this step, the deep decision tree model is established by the deep decision tree algorithm. The deep decision tree algorithm (DDT) is an integrated learning method that achieves the fusion of multiple models by building a deep tree to achieve better performance. In machine learning, a deep decision tree is a classifier that contains multiple decision trees, and the category of its output is determined by the majority of the categories output by individual trees. The specific process of building each tree is as follows: (a) N represents the number of training cases (samples), and M represents the number of features; (b) Input the number of features m, which is used to determine the decision result of a node on the decision tree; m should be much smaller than M; (c) Sample N times with replacement from N training cases (samples) to form a training set (i.e. bootstrap sampling), and use the unsampled cases (samples) to make predictions and evaluate their errors; (d) For each node, m features are randomly selected, and the decision of each node in the decision tree is determined based on these m features. Based on these m features, the best splitting method is calculated; (e) Each tree is grown to its full size without pruning, which is what might be done after building a normal tree classifier.)

[0040] Embodiment 3: As attached Figure 3 As shown, the embodiment of the present invention discloses a device for diagnosing heavy oil temperature conditions, comprising: A data acquisition unit, which acquires the real values ​​of key temperature parameters at the current moment in actual heavy oil production, wherein the real values ​​of key temperature parameters include metering line temperature, oil gathering line temperature, steam injection line temperature, steam injection line pressure, boiler temperature, and closed station temperature; The temperature prediction unit inputs the real value of the key temperature parameter at the current moment into the temperature prediction model, and outputs the predicted value of the key temperature parameter at the next moment, wherein the temperature prediction model is obtained by training and learning using multiple sets of sample data, and each set of sample data in the multiple sets of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production; The operating condition identification unit inputs the predicted value of the key temperature parameter into the operating condition identification model and outputs the operating condition identification result, wherein the operating condition identification model is obtained by training and learning using multiple groups of sample data, and each group of sample data in the multiple groups of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production and identification information of corresponding operating condition levels.

[0041] Embodiment 4: As attached Figure 4 As shown, the embodiment of the present invention discloses a device for diagnosing heavy oil temperature conditions, comprising: The first model building unit builds a temperature prediction model, including: Construct a historical data set, pre-process the data in the historical data set, and fit the discrete data in the historical data set using a gradient boosting algorithm, wherein the historical data set includes historical values ​​of several key temperature parameters in the entire process of actual heavy oil production; Divide the historical data set into a training data subset and a test data subset; The long short-term memory recurrent neural network model is trained using the training data subset to obtain a temperature prediction model; The temperature prediction model is tested using a subset of test data, and the long short-term memory recurrent neural network model is adjusted and trained accordingly based on the test results until the optimal temperature prediction model is obtained.

[0042] The second model building unit builds a working condition identification model, including: Construct a historical data set, preprocess the data in the historical data set, and fit the discrete data in the historical data set in combination with a gradient boosting algorithm, wherein the historical data set includes historical values ​​of several key temperature parameters in the entire process of actual heavy oil production and identification information of corresponding operating conditions; Divide the historical data set into a training data subset and a test data subset; The deep decision tree model is trained using the training data subset to obtain the working condition recognition model; Use the test data subset to test the working condition recognition model, and adjust and train the deep decision tree model accordingly based on the test results until the optimal temperature prediction model is obtained; A data acquisition unit, which acquires the real values ​​of key temperature parameters at the current moment in actual heavy oil production, wherein the real values ​​of key temperature parameters include metering line temperature, oil gathering line temperature, steam injection line temperature, steam injection line pressure, boiler temperature, and closed station temperature; The temperature prediction unit inputs the real value of the key temperature parameter at the current moment into the temperature prediction model, and outputs the predicted value of the key temperature parameter at the next moment, wherein the temperature prediction model is obtained by training and learning using multiple sets of sample data, and each set of sample data in the multiple sets of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production; The operating condition identification unit inputs the predicted value of the key temperature parameter into the operating condition identification model and outputs the operating condition identification result, wherein the operating condition identification model is obtained by training and learning using multiple groups of sample data, and each group of sample data in the multiple groups of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production and identification information of corresponding operating condition levels.

[0043] Embodiment 5: As attached Figure 5 As shown, the embodiment of the present invention discloses a heavy oil temperature condition diagnosis system, which is characterized by comprising: A heavy oil temperature condition diagnosis device, wherein the heavy oil temperature condition diagnosis device is the heavy oil temperature condition diagnosis device as described in the above embodiment; The interactive device is used for the operator to communicate with the heavy oil temperature condition diagnosis device, so that the operator provides the heavy oil temperature condition diagnosis device with an interactive interface of the data to be predicted through the interactive unit.

[0044] The above-mentioned interactive devices include computers, tablet computers, etc., which provide network pages for operators to enter information, view results, and modify parameters.

[0045] Embodiment 6: The embodiment of the present invention discloses a storage medium, on which a computer program readable by a computer is stored, and the computer program is configured to execute a method for diagnosing a heavy oil temperature condition when running.

[0046] The above storage medium may include, but is not limited to: a USB flash drive, a read-only memory, a mobile hard disk, a magnetic disk or an optical disk, and other media that can store computer programs.

[0047] Embodiment 7: The embodiment of the present invention discloses an electronic device, including a processor and a memory, wherein the memory stores a computer program, and the computer program is loaded and executed by the processor to implement a method for diagnosing the temperature condition of heavy oil.

[0048] The processor may be a central processing unit (CPU), a general purpose processor, a digital signal processor (DSP), an ASIC, an FPGA or other programmable logic device, a transistor logic device, a hardware component or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. It may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like. The memory may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory, a mobile hard disk, a magnetic disk or an optical disk.

[0049] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The schemes in the embodiments of the present application may be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0050] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0051] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0052] The above technical features constitute the best embodiment of the present invention, which has strong adaptability and best implementation effect. Non-essential technical features can be added or reduced according to actual needs to meet the requirements of different situations.

Claims

1. A method for diagnosing heavy oil temperature conditions, characterized in that: include: Obtain the real values ​​of key temperature parameters at the current moment in actual heavy oil production, where the real values ​​of key temperature parameters include metering line temperature, oil gathering line temperature, steam injection line temperature, steam injection line pressure, boiler temperature, and closed station temperature; Input the real value of the key temperature parameter at the current moment into the temperature prediction model, and output the predicted value of the key temperature parameter at the next moment, wherein the temperature prediction model is obtained by training and learning using multiple sets of sample data, and each set of sample data in the multiple sets of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production; The predicted value of the key temperature parameter is input into the operating condition identification model, and the operating condition identification result is output, wherein the operating condition identification model is obtained by training and learning using multiple groups of sample data, and each group of sample data in the multiple groups of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production and identification information of corresponding operating condition levels.

2. The method for diagnosing heavy oil temperature conditions according to claim 1, characterized in that: The process of constructing the temperature prediction model includes: Construct a historical data set, pre-process the data in the historical data set, and fit the discrete data in the historical data set using a gradient boosting algorithm, wherein the historical data set includes historical values ​​of several key temperature parameters in the entire process of actual heavy oil production; Divide the historical data set into a training data subset and a test data subset; The long short-term memory recurrent neural network model is trained using the training data subset to obtain a temperature prediction model; The temperature prediction model is tested using a subset of test data, and the long short-term memory recurrent neural network model is adjusted and trained accordingly based on the test results until the optimal temperature prediction model is obtained.

3. The method for diagnosing heavy oil temperature conditions according to claim 1 or 2, characterized in that: The construction process of the working condition identification model includes: Construct a historical data set, preprocess the data in the historical data set, and fit the discrete data in the historical data set in combination with a gradient boosting algorithm, wherein the historical data set includes historical values ​​of several key temperature parameters in the entire process of actual heavy oil production and identification information of corresponding operating conditions; Divide the historical data set into a training data subset and a test data subset; The deep decision tree model is trained using the training data subset to obtain the working condition recognition model; The operating condition recognition model is tested using a subset of test data, and the deep decision tree model is adjusted and trained accordingly based on the test results until the optimal temperature prediction model is obtained.

4. The method for diagnosing heavy oil temperature conditions according to claim 2 or 3, characterized in that: The preprocessing of the data in the historical data set includes: removing unique attributes, processing missing values, identifying and deleting abnormal data, data standardization, data regularization, and extracting eigenvalues ​​through principal component analysis.

5. The method for diagnosing heavy oil temperature conditions according to any one of claims 1 to 4, characterized in that: It also includes inputting the predicted value of the key temperature parameter into the working condition identification model, outputting the working condition identification result, and outputting corresponding alarm information for the abnormal working condition identification result.

6. A device for diagnosing heavy oil temperature conditions using the method according to any one of claims 1 to 5, characterized in that: include: A data acquisition unit, which acquires the real values ​​of key temperature parameters at the current moment in actual heavy oil production, wherein the real values ​​of key temperature parameters include metering line temperature, oil gathering line temperature, steam injection line temperature, steam injection line pressure, boiler temperature, and closed station temperature; The temperature prediction unit inputs the real value of the key temperature parameter at the current moment into the temperature prediction model, and outputs the predicted value of the key temperature parameter at the next moment, wherein the temperature prediction model is obtained by training and learning using multiple sets of sample data, and each set of sample data in the multiple sets of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production; The operating condition identification unit inputs the predicted value of the key temperature parameter into the operating condition identification model and outputs the operating condition identification result, wherein the operating condition identification model is obtained by training and learning using multiple groups of sample data, and each group of sample data in the multiple groups of sample data includes: historical values ​​of key temperature parameters in actual heavy oil production and identification information of corresponding operating condition levels.

7. The device for diagnosing heavy oil temperature conditions according to claim 6, characterized in that: Also includes a first model building unit and a second model building unit; The first model building unit builds a temperature prediction model, including: Construct a historical data set, pre-process the data in the historical data set, and fit the discrete data in the historical data set using a gradient boosting algorithm, wherein the historical data set includes historical values ​​of several key temperature parameters in the entire process of actual heavy oil production; Divide the historical data set into a training data subset and a test data subset; The long short-term memory recurrent neural network model is trained using the training data subset to obtain a temperature prediction model; The temperature prediction model is tested using a subset of test data, and the long short-term memory recurrent neural network model is adjusted and trained accordingly based on the test results until the optimal temperature prediction model is obtained; The second model building unit builds a working condition identification model, including: Construct a historical data set, preprocess the data in the historical data set, and fit the discrete data in the historical data set in combination with a gradient boosting algorithm, wherein the historical data set includes historical values ​​of several key temperature parameters in the entire process of actual heavy oil production and identification information of corresponding operating conditions; Divide the historical data set into a training data subset and a test data subset; The deep decision tree model is trained using the training data subset to obtain the working condition recognition model; The operating condition recognition model is tested using a subset of test data, and the deep decision tree model is adjusted and trained accordingly based on the test results until the optimal temperature prediction model is obtained.

8. A heavy oil temperature condition diagnosis system, characterized in that: include: A device for diagnosing heavy oil temperature conditions, wherein the device for diagnosing heavy oil temperature conditions is the device for diagnosing heavy oil temperature conditions as claimed in claims 6 to 7; The interactive device is used for the operator to communicate with the heavy oil temperature condition diagnosis device, so that the operator provides the heavy oil temperature condition diagnosis device with an interactive interface of the data to be predicted through the interactive unit.

9. A storage medium, characterized in that: The storage medium stores a computer program that can be read by a computer, and the computer program is configured to execute the method for diagnosing the temperature condition of heavy oil according to any one of claims 1 to 5 when running.

10. An electronic device, characterized in that: The method comprises a processor and a memory, wherein a computer program is stored in the memory, and the computer program is loaded and executed by the processor to implement the method for diagnosing the temperature condition of heavy oil according to any one of claims 1 to 5.

Images