Well cementation pump fault diagnosis method and device

Abstract

The invention discloses a well cementation pump fault diagnosis method and device. The method comprises the steps that historical data of a well cementation pump rotating speed sensor, historical data of a vibration sensor and historical operation working conditions in well cementation operation are obtained; extracting multi-dimensional historical features of the vibration sensor based on the time domain and the frequency domain; the multi-dimensional historical characteristics of the vibration sensor and historical data of the rotating speed sensor are combined to obtain historical input characteristics of the well cementation pump; calculating the relative inter-class dispersion distance of the historical input features of the well cementation pump, constructing feature importance factors, and performing weighting processing; and inputting the weighted historical input features and historical operation conditions of the well cementation pump into a preset weighted width learning model, training the preset weighted width learning model, and obtaining the trained preset weighted width learning model as a fault classifier. According to real-time data of the well cementation pump rotating speed sensor and the vibration sensor in the well cementation operation process, a fault diagnosis result is obtained through prediction.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of safety monitoring of offshore oil cementing operation equipment, and particularly to a method and device for diagnosing faults of a cementing pump. Background Art

[0002] In the process of offshore oil production, the cementing operation is a key link connecting drilling and oil production. It plays a crucial role in ensuring the safe, efficient and environmental protection production of oil and gas wells, and has a direct impact on the stable operation and economic benefits of the entire oil and gas industrial chain.

[0003] The core of the cementing pump injection system is the cementing pump, whose main function is to inject cement slurry into the wellbore to fix the wellbore wall and prevent oil and gas leakage. The continuous and stable operation of the cementing pump helps to shorten the operation time, improve efficiency and reduce costs. The working environment on the offshore platform is harsh, and there are many wearing parts during the operation of the cementing pump (such as valves, plungers, packing, etc.). Once equipment failures occur (such as valve leakage, severe packing wear, etc.), it may lead to uneven injection of cement slurry and unstable pressure, affecting the fixing effect of the entire wellbore, and even may cause important marine ecological disasters. Therefore, it is very important to monitor the fault conditions of the cementing pump in real time and accurately to avoid serious losses caused by sudden shutdown during actual use.

[0004] In recent years, with the construction of the informatization technology of cementing operation equipment, a large amount of operation data of the cementing pump can be collected in real time at the operation site and stored in the database for the research on the fault diagnosis of the cementing pump based on data driving. However, there are deficiencies in the existing fault diagnosis of offshore cementing pumps based on neural networks: on the one hand, the traditional neural network adopts an iterative optimization method, and the training efficiency is low; on the other hand, the existing methods ignore the differences in data features and process all features in the same way, making it difficult to give full play to the prominent role of some effective features. Summary of the Invention

[0005] In view of the above problems, the embodiments of the present invention are proposed to provide a method and device for diagnosing faults of a cementing pump that overcome the above problems or at least partially solve the above problems.

[0006] According to one aspect of the embodiments of the present invention, a method for diagnosing faults of a cementing pump is provided. The method includes:

[0007] Obtaining historical data of a rotational speed sensor of the cementing pump, historical data of a vibration sensor and historical operating conditions in the cementing operation; the historical operating conditions include normal conditions and fault conditions;

[0008] For the historical data of any vibration sensor, multi-dimensional historical features of the vibration sensor are extracted based on the time domain and frequency domain; the multi-dimensional historical features of the vibration sensor are combined with the historical data of the rotational speed sensor to obtain the historical input features of the cementing pump;

[0009] Calculate the relative between-class scatter distance of each historical input feature in the historical input features of the cementing pump, and construct a feature importance factor according to the relative between-class scatter distance, so as to perform weighted processing on the historical input features of the cementing pump according to the feature importance factor;

[0010] Input the weighted historical input features of the cementing pump and the historical operating conditions into a preset weighted width learning model, train the preset weighted width learning model, and obtain the trained preset weighted width learning model as a fault classifier, so as to predict the fault diagnosis result according to the real-time data of the rotational speed sensor and vibration sensor of the cementing pump during the cementing operation.

[0011] According to another aspect of the embodiments of the present invention, there is provided a cementing pump fault diagnosis device, which includes:

[0012] A historical data acquisition module, adapted to acquire the historical data of the rotational speed sensor of the cementing pump, the historical data of the vibration sensor, and the historical operating conditions during the cementing operation; the historical operating conditions include normal conditions and fault conditions;

[0013] A multi-dimensional extraction module, adapted to extract multi-dimensional historical features of the vibration sensor based on the time domain and frequency domain for the historical data of any vibration sensor; combine the multi-dimensional historical features of the vibration sensor with the historical data of the rotational speed sensor to obtain the historical input features of the cementing pump;

[0014] A weighted processing module, adapted to calculate the relative between-class scatter distance of each historical input feature in the historical input features of the cementing pump, and construct a feature importance factor according to the relative between-class scatter distance, so as to perform weighted processing on the historical input features of the cementing pump according to the feature importance factor;

[0015] A model training and prediction module, adapted to input the weighted historical input features of the cementing pump and the historical operating conditions into a preset weighted width learning model, train the preset weighted width learning model, and obtain the trained preset weighted width learning model as a fault classifier, so as to predict the fault diagnosis result according to the real-time data of the rotational speed sensor and vibration sensor of the cementing pump during the cementing operation.

[0016] According to yet another aspect of the embodiments of the present invention, there is provided a computing device, including: a processor, a memory, a communication interface, and a communication bus, and the processor, the memory, and the communication interface complete communication with each other through the communication bus;

[0017] The memory is used to store at least one executable instruction, and the executable instruction causes the processor to perform operations corresponding to the above-mentioned cementing pump fault diagnosis method.

[0018] According to another aspect of the embodiments of the present invention, there is provided a computer storage medium, in which at least one executable instruction is stored, and the executable instruction causes the processor to perform operations corresponding to the above-mentioned cementing pump fault diagnosis method.

[0019] According to still another aspect of the embodiments of the present invention, there is provided a computer program product, including at least one executable instruction, and the executable instruction causes the processor to perform operations corresponding to the above-mentioned cementing pump fault diagnosis method.

[0020] According to the cementing pump fault diagnosis method and device provided by the embodiments of the present invention, starting from time-frequency and frequency-domain, multi-dimensional data features of the vibration sensor of the cementing pump are fully extracted. When determining the feature importance factor according to the multi-dimensional features for weighted processing, important features closely related to the fault condition can be assigned larger weights, and secondary features with low correlation are given smaller weights, making full use of the information differences contained in different features to improve the modeling effect of the preset weighted width learning model, thereby improving the accuracy of predicting the fault classification of the cementing pump.

[0021] The above description is only an overview of the technical solutions of the embodiments of the present invention. In order to be able to understand the technical means of the embodiments of the present invention more clearly, it can be implemented according to the content of the description. And in order to make the above and other purposes, features and advantages of the embodiments of the present invention more obvious and understandable, the following specifically describes the specific embodiments of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] By reading the detailed description of the preferred embodiments below, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the embodiments of the present invention. And throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0023] Figure 1 A flowchart of a cementing pump fault diagnosis method according to an embodiment of the present invention is shown;

[0024] Figure 2 A schematic structural diagram of a preset weighted width learning model is shown;

[0025] Figure 3 A confusion matrix of the classification results of a traditional learning model is shown;

[0026] Figure 4 A confusion matrix of the classification of a feature weighted width learning model is shown;

[0027] Figure 5 Shows a schematic structural diagram of a cementing pump fault diagnosis device according to an embodiment of the present invention;

[0028] Figure 6 Shows a schematic structural diagram of a computing device according to an embodiment of the present invention. Detailed implementation manners

[0029] Hereinafter, exemplary embodiments of the present invention will be described in more detail with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present invention can be more thoroughly understood and the scope of the present invention can be completely conveyed to those skilled in the art.

[0030] Figure 1 Shows a flowchart of a cementing pump fault diagnosis method according to an embodiment of the present invention, as Figure 1 shown, the method includes the following steps:

[0031] Step S101, obtain historical data of a rotational speed sensor of a cementing pump, historical data of a vibration sensor, and historical operating conditions during a cementing operation.

[0032] During a historical cementing operation, the operating data of the cementing pump during the cementing operation can be recorded and stored, that is, the historical operating data of the cementing pump during the historical cementing operation is stored in a database. The historical data of the rotational speed sensor of the cementing pump, the historical data of the vibration sensor, and the historical operating conditions can be obtained from the database. Among them, if there are multiple vibration sensors of the cementing pump, the historical data of each vibration sensor needs to be obtained. The historical operating conditions include normal conditions and fault conditions, and different numerical classifications can be set for various fault conditions.

[0033] Specifically, obtain the historical operating data of the cementing pump {s(t), v1(t), v2(t),... v i (t),... v m (t), y(t)} including n samples from the database, 1 ≤ t ≤ n, where t represents the sample serial number, s(t) represents the historical data of the rotational speed sensor, v1(t), v2(t),..., v m(t) represents the historical data of m vibration sensors installed at different positions, and y(t) represents the type of working condition to which the historical operation data of the cementing pump belongs. Its value can be an integer such as {0, 1, …, C}, corresponding to the normal working condition and various fault working conditions respectively. The data values of the historical working conditions are for illustrative purposes only and are specifically set according to the implementation situation, which is not limited here. Among them, the historical working conditions correspond to the historical data of the rotational speed sensor and the historical data of the vibration sensor in the sample. The historical working conditions can be marked with the historical data of the rotational speed sensor and the historical data of the vibration sensor for use in the subsequent training of the preset weighted width learning model.

[0034] Step S102, for the historical data of any vibration sensor, multi-dimensional historical features of the vibration sensor are extracted based on the time domain and the frequency domain, and the multi-dimensional historical features of the vibration sensor are combined with the historical data of the rotational speed sensor to obtain the historical input features of the cementing pump.

[0035] When processing the historical data of vibration sensors in the prior art, all features are treated equally and processed without distinction. Obviously, this method cannot effectively utilize the differences of each feature of the historical data of vibration sensors and cannot reflect the degree of influence of different data features on the working conditions. Considering the above problems, in this embodiment, for the historical data of any vibration sensor of the cementing pump, starting from the time domain and the frequency domain, the historical data of the vibration sensor is analyzed to extract multi-dimensional historical features of the vibration sensor. Among them, the multi-dimensional historical features include the maximum value, minimum value, mean value, standard deviation, peak-to-peak value, effective value, skewness, kurtosis, center frequency, root mean square frequency, mean frequency, and frequency standard deviation of the vibration signal. The above multi-dimensional historical features can be calculated from the historical data v i (t). Among them, 1 ≤ i ≤ m, m is the number of vibration sensors, and t is the t-th sample.

[0036] For the maximum value It is calculated using formula (1):

[0037]

[0038] Among them, are the R data in the t-th sample of the i-th vibration sensor.

[0039] For the minimum value It is calculated using formula (2):

[0040]

[0041] For the mean value It is calculated using formula (3):

[0042]

[0043] For the standard deviation It is calculated using formula (4):

[0044]

[0045] For the peak-to-peak value It is calculated using formula (5):

[0046]

[0047] For the effective value It is calculated using formula (6):

[0048]

[0049] For the skewness It is calculated using formula (7):

[0050]

[0051] For the kurtosis It is calculated using formula (8):

[0052]

[0053] For the center frequency It is calculated using formula (9):

[0054]

[0055] where d i,j (1 < j < J) are J frequency values in the spectrum of v i (t), and P(d i,j ) is the power spectrum of the vibration signal.

[0056] For the root mean square frequency It is calculated using formula (10):

[0057]

[0058] For the mean frequency It is calculated using formula (11):

[0059] [[ID=7,6]]

[0060] For the frequency standard deviation It is calculated using formula (12):

[0061]

[0062] The historical data of each vibration sensor are calculated through the above formulas (1)-(12) to obtain the multi-dimensional historical features of each cementing pump vibration sensor. The multi-dimensional historical features of the vibration sensor and the historical data of the rotational speed sensor are combined to obtain the historical input features of the cementing pump. Among them, the historical input features are obtained by combining the multi-dimensional historical features of the vibration sensor and the historical data of the rotational speed sensor. Its total quantity is related to the dimension number a of the multi-dimensional historical features and the number m of vibration sensors. The total quantity M = am + 1, and f k (t) is the k-th historical input feature, where 1 ≤ k ≤ M.

[0063] Step S103: Calculate the relative between-class scatter distance of each historical input feature in the historical input features of the cementing pump, and construct a feature importance factor based on the relative between-class scatter distance, so as to perform weighted processing on the historical input features of the cementing pump according to the feature importance factor.

[0064] The historical input features of the cementing pump include the multi-dimensional historical features of the vibration sensor and the historical features of the rotational speed sensor. However, the importance of each historical feature is different in the working conditions. If the historical input features of the cementing pump are directly used for model training, the differences in different data features are ignored, resulting in some data features being unable to play an effective role during model training. Therefore, it is necessary to further process the differences in important features in the historical output features to improve the accuracy of the model.

[0065] Specifically, the relative scatter analysis method is used to calculate the first between-class scatter and the second between-class scatter of the cementing pump under each historical operating condition according to the mean value of the historical input features of the cementing pump under each historical operating condition and the covariance matrix under all historical operating conditions. Among them, the first between-class scatter is the overall between-class scatter of the cementing pump, and the second between-class scatter is the between-class scatter of the cementing pump after removing any one feature.

[0066] Specifically, the first between-class scatter disp all is the overall between-class scatter of all historical features including various working conditions, and is calculated according to formula (13):

[0067]

[0068] Among them, mf x and mf y are the mean values of the historical feature data f(t) corresponding to the x-th and y-th historical working conditions respectively, and are both row vectors. covf is the covariance matrix of the historical feature data f(t) of all historical working condition data. The historical working condition data takes integer values {0, 1, …, C}, with a total of C + 1 kinds of working condition data.

[0069] The between-class scatter of the second type, disp k is the between-class scatter after removing the k-th feature, where k is any feature and is calculated according to formula (14):

[0070]

[0071] where respectively represent the means of the historical feature data corresponding to the historical working conditions of the x-th and y-th classes after removing the k-th feature, both of which are row vectors. covf k is the covariance matrix of all working condition data after removing the k-th feature.

[0072] The relative between-class scatter distance dist of each feature in the historical input features of the cementing pump can be calculated based on the between-class scatter of the first type and the between-class scatter of the second type k and is calculated according to formula (15):

[0073]

[0074] The relative between-class scatter distance dist k The larger it is, the more important the corresponding feature is. The feature importance factor imp k is obtained according to formula (16):

[0075]

[0076] where, according to the comparison result of the between-class scatter of the first type and the between-class scatter of the second type, the feature importance factor is determined. When the between-class scatter of the first type is greater than the between-class scatter of the second type, the feature importance factor is determined by the relative between-class scatter distance; otherwise, the feature importance factor is determined by a preset factor. The preset factor γ can be set according to the actual situation and is not limited here.

[0077] After obtaining the feature importance factor, a diagonal matrix Imp is constructed according to the feature importance factor, as shown in formula (17):

[0078] Imp = diag{[imp1 … imp k … imp M} (17)

[0079] where, diag represents the operator for constructing a diagonal matrix, M is the total amount of the historical input features of the cementing pump, and M is specifically determined with reference to the description in step S102.

[0080] The historical input features f(t) of the cementing pump are weighted using the diagonal matrix Imp to obtain the weighted historical input features f w (t), as shown in the following formula (18):

[0081] f w (t) = f(t) × Imp (18)

[0082] Step S104: Input the weighted historical input features of the cementing pump and the historical operating conditions into the preset weighted width learning model, and train the preset weighted width learning model to obtain the trained preset weighted width learning model as a fault classifier, so as to predict the fault diagnosis result based on the real-time data of the rotational speed sensor and vibration sensor of the cementing pump during the cementing operation.

[0083] Construct an input matrix according to the weighted historical input features of the cementing pump. Among them, the input matrix F is obtained by using the weighted historical input features of its n samples w = [f w (1); f w (2); …; f w (n)]; n is the number of samples. Perform one-hot encoding on the historical operating condition y(t) to obtain the output matrix Y.

[0084] Perform primary mapping on the input matrix F w to obtain primary mapping nodes. Specifically, according to the non-linear function φ of the preset primary mapping and the preset first weight parameter group which can be randomly generated. Perform primary mapping on the input matrix F w to obtain the primary mapping nodes Z p , where there are multiple primary mapping nodes, such as P, 1 ≤ p ≤ P, as shown in formula (19):

[0085]

[0086] Merge multiple primary mapping nodes, Z all = [Z1, Z2, … Z p …, Z P . According to the non-linear function ξ of the preset enhanced mapping and the preset second weight parameter group which can be randomly generated. Perform enhanced mapping on Z all to obtain the enhanced mapping nodes H q . Among them, there are multiple enhanced mapping nodes, such as Q, 1 ≤ q ≤ Q, as shown in formula (20):

[0087]

[0088] Merge the primary mapping nodes and the enhanced mapping nodes to obtain the first matrix [Z all H1 H2 … H Q, calculate the generalized inverse processing result of the first matrix, and calculate the output weight parameter B based on the generalized inverse processing result and the output matrix Y. As shown in formula (21):

[0089] B = [Z all H1 H2 … H Q * Y (21)

[0090] where, * represents finding the generalized inverse of the matrix.

[0091] The preset weighted width WBL (Web - Based Learning, learning model) is constructed through the above formulas (19) - (21), and its model expression is:

[0092]

[0093] where, g() represents the overall function of the model.

[0094] The structure diagram of the preset weighted width learning model is as Figure 2 shown. According to the historical data of the rotational speed sensor and vibration sensor of n samples, such as rotational speed data and vibration data, the feature vector of the historical input features is obtained through steps S101 - S102. The feature importance factor is constructed according to the relative inter - class scatter distance, and the feature vector is weighted to obtain the weighted feature vector. Based on the non - linear function of the preset primary mapping and the random weight parameter that is, the preset first weight parameter group, the primary mapping nodes are obtained. For the combined primary mapping nodes, according to the non - linear function of the enhanced mapping and the random weight parameter that is, the preset second weight parameter group for enhanced mapping, the enhanced mapping nodes are obtained. According to the primary mapping nodes, enhanced mapping nodes, and the working condition classification result, that is, the output matrix Y, the optimized output weight parameter B is obtained, and the training of the preset weighted width learning model is completed, and the trained preset weighted width learning model is used as a fault classifier. The fault classifier can be obtained by offline training using historical data. Further, the fault classifier can continue to collect historical data for optimization, etc., which is not limited here.

[0095] After obtaining the fault classifier, it is convenient to use the fault classifier to perform real - time working condition prediction on the cementing operation, diagnose the faults of the working conditions, and determine the fault results. For real - time prediction, the real - time data of the cementing pump rotational speed sensor and vibration sensor during the cementing operation can be obtained {s(t new ), v1(t new ), v2(t new ), …, v m (t new )}, t new ​Represents the serial number of real-time data. For the real-time data of any vibration sensor, referring to step S102, multi-dimensional real-time features of the vibration sensor are extracted based on the time domain and frequency domain, and the multi-dimensional real-time features of the vibration sensor are combined with the real-time features of the real-time data of the rotational speed sensor to obtain the real-time input feature f(t new ) of the cementing pump. A diagonal matrix Imp is constructed according to the feature importance factor, the real-time input feature of the cementing pump is weighted, and then the weighted real-time input feature of the cementing pump is input into the fault classifier to predict the fault diagnosis result.

[0096] In this embodiment, fault prediction and diagnosis are realized through two stages: offline and online. In the offline training stage, historical data of the cementing pump under various working conditions are collected. For the historical data of the vibration sensor, time domain and frequency domain features are extracted, the importance of each feature is analyzed by using the relative dispersion analysis method, each feature is weighted, and an offline training is carried out to obtain a feature weighted width learning model to establish a fault classifier. In the online testing stage, real-time data of the cementing pump are collected, corresponding time domain and frequency domain features are extracted, the features are weighted by using the feature importance factor and then input into the fault classifier, and the classification result of the working condition is obtained according to the output.

[0097] Taking the cementing pump device of a certain company as the test object, four different working conditions are tested during the operation of the device: normal working condition, wear of the upper valve of cylinder 1, wear of the upper valve of cylinder 3, and wear of the lower valve of cylinder 3, which are respectively labeled as working condition 1, working condition 2, working condition 3, and working condition 4. There are a total of 3582 samples of data for all working conditions, which are divided into two groups. One group contains 60% of the samples as the model training data set, and the other group contains 40% of the samples as the model test data set. A fault classifier is established by using a traditional learning model (such as the width learning model) and the feature weighted width learning model of this embodiment respectively, and Figure 3 the confusion matrix of the classification result of the traditional learning model as shown is obtained, Figure 4 and the classification confusion matrix of the feature weighted width learning model as shown is obtained. Figure 3 、 Figure 4 In, the horizontal axis represents the real working conditions of the classes, and the vertical axis represents the output classes using the model as the classifier. Figure 3 In, the traditional learning model correctly identifies 192 samples for the situation of working condition 2. Figure 4 In, the feature weighted width learning model can correctly identify 216 samples for working condition 1, with an obvious improvement. Figure 4 The recognition result for working condition 4 in is also significantly improved compared with Figure 3 , producing a better improvement effect. The comparison of the recognition accuracy rates of the traditional model and the model of this embodiment under different working conditions is shown in Table 1 below:

[0098] Table 1

[0099]

[0100] In Table 1, the overall recognition accuracy of the traditional model is 89.92%, while the recognition accuracy of the feature-weighted width learning model of this embodiment can reach 92.02%, an increase of 2.1%, indicating that this embodiment can effectively improve the fault classification accuracy of the cementing pump.

[0101] According to the cementing pump fault diagnosis method provided by the embodiment of the present invention, starting from time-frequency and domain-frequency, multi-dimensional data features of the vibration sensor of the cementing pump are fully extracted. When determining the feature importance factor according to the multi-dimensional features for weighting processing, the important features closely related to the fault conditions can be assigned larger weights, and the secondary features with low correlation are given smaller weights, making full use of the information differences contained in different features to improve the modeling effect of the preset weighted width learning model, thereby improving the accuracy of predicting the fault classification of the cementing pump.

[0102] Figure 5 The structural schematic diagram of the cementing pump fault diagnosis device provided by the embodiment of the present invention is shown. As Figure 5 shown, the device includes:

[0103] A historical data acquisition module 510, adapted to acquire the historical data of the rotational speed sensor of the cementing pump, the historical data of the vibration sensor, and the historical operating conditions during the cementing operation; the historical operating conditions include normal conditions and fault conditions;

[0104] A multi-dimensional extraction module 520, adapted to extract multi-dimensional historical features of the vibration sensor based on the time domain and the frequency domain for the historical data of any vibration sensor; combine the multi-dimensional historical features of the vibration sensor with the historical data of the rotational speed sensor to obtain the historical input features of the cementing pump;

[0105] A weighting processing module 530, adapted to calculate the relative between-class scatter distance of each historical input feature in the historical input features of the cementing pump, and construct a feature importance factor according to the relative between-class scatter distance, so as to perform weighting processing on the historical input features of the cementing pump according to the feature importance factor;

[0106] A model training and prediction module 540, adapted to input the weighted historical input features of the cementing pump and the historical operating conditions into a preset weighted width learning model, train the preset weighted width learning model, and obtain the trained preset weighted width learning model as a fault classifier, for predicting a fault diagnosis result according to the real-time data of the rotational speed sensor and the vibration sensor of the cementing pump during the cementing operation.

[0107] Optionally, the multi-dimensional historical features of the vibration sensor include:

[0108] Maximum value, minimum value, mean value, standard deviation, peak-to-peak value, effective value, skewness, kurtosis, center frequency, root mean square frequency, mean frequency, frequency standard deviation.

[0109] Optionally, the weighting processing module 530 is further adapted to:

[0110] Calculate the first type of between-class scatter and the second type of between-class scatter of the cementing pump under each historical operating condition based on the historical input characteristics of the cementing pump, the mean value of each historical operating condition, and the covariance matrix of all historical operating conditions by using the relative scatter analysis method; wherein, the first type of between-class scatter is the overall between-class scatter of the cementing pump; the second type of between-class scatter is the between-class scatter of the cementing pump after removing any one characteristic;

[0111] Calculate the relative between-class scatter distance of each characteristic in the historical input characteristics of the cementing pump according to the first type of between-class scatter and the second type of between-class scatter;

[0112] Determine the feature importance factor according to the comparison result of the first type of between-class scatter and the second type of between-class scatter; wherein, the feature importance factor is determined by the relative between-class scatter distance or a preset factor according to the comparison result;

[0113] Construct a diagonal matrix according to the feature importance factor, and use the diagonal matrix to perform weighting processing on the historical input characteristics of the cementing pump.

[0114] Optionally, the model training and prediction module 540 is further adapted to:

[0115] Construct an input matrix according to the weighted historical input characteristics of the cementing pump, perform one-hot encoding processing on the historical operating conditions, and obtain an output matrix;

[0116] Perform primary mapping on the input matrix to obtain primary mapping nodes, and perform enhanced mapping on the primary mapping nodes to obtain enhanced mapping nodes;

[0117] Merge the primary mapping nodes and the enhanced mapping nodes, and calculate the output weight parameters according to the output matrix to complete the training of the preset weighted width learning model, and obtain the trained preset weighted width learning model as a fault classifier.

[0118] Optionally, the model training and prediction module 540 is further adapted to:

[0119] Perform primary mapping on the input matrix according to the nonlinear function of the preset primary mapping and the preset first weight parameter group to obtain primary mapping nodes; there are multiple primary mapping nodes;

[0120] Merge the multiple primary mapping nodes and perform enhanced mapping according to the nonlinear function of the preset enhanced mapping and the preset second weight parameter group to obtain enhanced mapping nodes; there are multiple enhanced mapping nodes;

[0121] Combining the primary mapping node and the enhanced mapping node, and calculating the output weight parameter according to the output matrix further includes:

[0122] Combining the primary mapping node and the enhanced mapping node to obtain a first matrix, calculating the generalized inverse processing result of the first matrix, and calculating the output weight parameter according to the generalized inverse processing result and the output matrix.

[0123] Optionally, the model training and prediction module 540 is further adapted to:

[0124] Obtain the real-time data of the cementing pump speed sensor and the real-time data of the vibration sensor during the cementing operation;

[0125] For the real-time data of any vibration sensor, extract the multi-dimensional real-time features of the vibration sensor based on the time domain and frequency domain; combine the multi-dimensional real-time features of the vibration sensor with the real-time features of the real-time data of the speed sensor to obtain the real-time input features of the cementing pump;

[0126] Perform weighted processing on the real-time input features of the cementing pump according to the feature importance factor, and input the weighted real-time input features of the cementing pump into the fault classifier to predict the fault diagnosis result.

[0127] The descriptions of the above modules refer to the corresponding descriptions in the method embodiments and will not be repeated here.

[0128] The embodiment of the present invention also provides a non-volatile computer storage medium, and the computer storage medium stores at least one executable instruction, and the executable instruction can execute the operations corresponding to the cementing pump fault diagnosis method in any of the above method embodiments.

[0129] The embodiment of the present application provides a computer program product, and the computer program product includes at least one executable instruction or computer program, and the executable instruction or computer program can enable the processor to execute the operations corresponding to the cementing pump fault diagnosis method in any of the above method embodiments.

[0130] Figure 6 A schematic structural diagram of a computing device according to an embodiment of the present invention is shown, and the specific embodiments of the embodiment of the present invention do not limit the specific implementation of the computing device.

[0131] As Figure 6 shown, the computing device may include: a processor 602, a communication interface 604, a memory 606, and a communication bus 608.

[0132] Wherein:

[0133] The processor 602, the communication interface 604, and the memory 606 communicate with each other via the communication bus 608.

[0134] The communication interface 604 is used to communicate with network elements of other devices such as clients or other servers.

[0135] The processor 602 is used to execute the program 610, and specifically can execute the relevant steps in the above-mentioned embodiments of the cementing pump fault diagnosis method.

[0136] Specifically, the program 610 may include program code, and the program code includes computer operation instructions.

[0137] The processor 602 may be a central processing unit CPU, or a specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present invention. One or more processors included in the computing device may be of the same type of processor, such as one or more CPUs; or may be of different types of processors, such as one or more CPUs and one or more ASICs.

[0138] The memory 606 is used to store the program 610. The memory 606 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk memory.

[0139] The program 610 is specifically used to cause the processor 602 to execute the cementing pump fault diagnosis method in any of the above method embodiments. For the specific implementation of each step in the program 610, reference may be made to the corresponding steps and descriptions in the corresponding units in the above-mentioned embodiments of the cementing pump fault diagnosis, which will not be elaborated here. Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the above-described devices and modules can refer to the corresponding process descriptions in the foregoing method embodiments, which will not be repeated here.

[0140] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. Based on the above description, the structure required to construct such a system is obvious. In addition, the embodiments of the present invention are not directed to any specific programming language. It should be understood that the content of the embodiments of the present invention described herein can be implemented using various programming languages, and the description of the specific language above is to disclose the preferred embodiments of the embodiments of the present invention.

[0141] In the specification provided herein, a number of specific details are set forth. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail so as not to obscure an understanding of the present specification.

[0142] Similarly, it should be understood that in order to streamline the embodiments of the present invention and assist in understanding one or more of the various inventive aspects, in the above description of the exemplary embodiments of the present invention, the various features of the embodiments of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, the disclosed method should not be construed as reflecting an intention that the claimed embodiments of the present invention require more features than are expressly recited in each claim. Rather, as reflected in the following claims, the inventive aspects lie in less than all the features of the single embodiments disclosed previously. Thus, the claims following the detailed description are hereby expressly incorporated into the detailed description, where each claim stands on its own as a separate embodiment of the present invention.

[0143] Those skilled in the art will appreciate that the modules in the devices in the embodiments can be adaptively changed and disposed in one or more devices different from the embodiments. The modules or units or components in the embodiments can be combined into one module or unit or component, and in addition, they can be divided into multiple sub-modules or sub-units or sub-components. Except that at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all the features disclosed in this specification (including the accompanying claims, abstract and drawings) and all the processes or units of any method or device so disclosed. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstract and drawings) can be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0144] In addition, those skilled in the art will be able to understand that although some of the embodiments herein include certain features included in other embodiments but not other features, the combination of the features of different embodiments means that it is within the scope of the present invention and forms different embodiments. For example, in the following claims, any one of the claimed embodiments can be used in any combination.

[0145] Each component embodiment of the present invention can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art should understand that a microprocessor or a digital signal processor (DSP) can be used in practice to implement some or all of the functions of some or all of the components according to the embodiments of the present invention. Embodiments of the present invention can also be implemented as a device or apparatus program (e.g., a computer program and a computer program product) for performing part or all of the methods described herein. Such a program implementing the embodiments of the present invention can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, or provided on a carrier signal, or in any other form.

[0146] It should be noted that the above embodiments illustrate the embodiments of the present invention rather than limit the present invention, and those skilled in the art can design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. Embodiments of the present invention can be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In the unit claims listing several devices, several of these devices can be embodied by the same item of hardware. The use of the words first, second, and third, etc. does not denote any order. These words can be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be construed as limiting the order of execution.

Claims

1. A method for diagnosing the failure of a cementing pump, characterized in that, The method includes: Obtaining the historical data of the cementing pump speed sensor, the historical data of the vibration sensor, and the historical operating conditions during the cementing operation; the historical operating conditions include normal conditions and fault conditions; For the historical data of any vibration sensor, multi-dimensional historical features of the vibration sensor are extracted based on the time domain and frequency domain; the multi-dimensional historical features of the vibration sensor are combined with the historical data of the speed sensor to obtain the historical input features of the cementing pump; Calculating the relative between-class scatter distance of each historical input feature in the historical input features of the cementing pump, and constructing a feature importance factor according to the relative between-class scatter distance, so as to perform weighted processing on the historical input features of the cementing pump according to the feature importance factor; Inputting the weighted historical input features of the cementing pump and the historical operating conditions into a preset weighted width learning model, training the preset weighted width learning model, and obtaining the trained preset weighted width learning model as a fault classifier, so as to predict a fault diagnosis result based on the real-time data of the cementing pump speed sensor and the vibration sensor during the cementing operation.

2. The method according to claim 1, wherein The multi-dimensional historical features of the vibration sensor include: Maximum value, minimum value, mean value, standard deviation, peak-to-peak value, effective value, skewness, kurtosis, center frequency, root mean square frequency, mean frequency, frequency standard deviation.

3. The method according to claim 1, characterized in that, The calculating the relative between-class scatter distance of each historical input feature in the historical input features of the cementing pump, and constructing a feature importance factor according to the relative between-class scatter distance, so as to perform weighted processing on the historical input features of the cementing pump according to the feature importance factor further includes: Using the relative scatter analysis method to calculate the first between-class scatter and the second between-class scatter of the cementing pump under each historical operating condition according to the mean value of the historical input features of the cementing pump under each historical operating condition and the covariance matrix under all historical operating conditions; wherein, the first between-class scatter is the overall between-class scatter of the cementing pump; the second between-class scatter is the between-class scatter of the cementing pump after removing any feature; Calculating the relative between-class scatter distance of each feature in the historical input features of the cementing pump according to the first between-class scatter and the second between-class scatter; Determining a feature importance factor according to the comparison result of the first between-class scatter and the second between-class scatter; wherein, the feature importance factor is determined by the relative between-class scatter distance or a preset factor according to the comparison result; Constructing a diagonal matrix according to the feature importance factor, and using the diagonal matrix to perform weighted processing on the historical input features of the cementing pump.

4. The method according to claim 1, wherein The inputting the weighted historical input features of the cementing pump and the historical operating conditions into a preset weighted width learning model, and training the preset weighted width learning model further includes: Constructing an input matrix according to the weighted historical input features of the cementing pump, and performing one-hot encoding processing on the historical operating conditions to obtain an output matrix; Performing primary mapping on the input matrix to obtain primary mapping nodes, and performing enhanced mapping on the primary mapping nodes to obtain enhanced mapping nodes; Merge the primary mapping nodes and the enhanced mapping nodes, and calculate output weight parameters based on the output matrix to complete the training of the preset weighted width learning model, and obtain the trained preset weighted width learning model as a fault classifier.

5. The method according to claim 4, wherein The step of performing primary mapping on the input matrix to obtain primary mapping nodes and performing enhanced mapping on the primary mapping nodes to obtain enhanced mapping nodes further includes: Perform primary mapping on the input matrix according to a preset non-linear function of primary mapping and a preset first weight parameter group to obtain primary mapping nodes; there are multiple primary mapping nodes; After merging the multiple primary mapping nodes, perform enhanced mapping according to a preset non-linear function of enhanced mapping and a preset second weight parameter group to obtain enhanced mapping nodes; there are multiple enhanced mapping nodes; The step of merging the primary mapping nodes and the enhanced mapping nodes and calculating output weight parameters based on the output matrix further includes: Merge the primary mapping nodes and the enhanced mapping nodes to obtain a first matrix, calculate the generalized inverse processing result of the first matrix, and calculate output weight parameters based on the generalized inverse processing result and the output matrix.

6. The method according to claim 1, wherein The step of predicting a fault diagnosis result based on the real-time data of the cementing pump speed sensor and the vibration sensor during the cementing operation further includes: Obtain the real-time data of the cementing pump speed sensor and the real-time data of the vibration sensor during the cementing operation; For the real-time data of any vibration sensor, extract multi-dimensional real-time features of the vibration sensor based on the time domain and the frequency domain; combine the multi-dimensional real-time features of the vibration sensor with the real-time data of the speed sensor to obtain the real-time input features of the cementing pump; Perform weighted processing on the real-time input features of the cementing pump according to the feature importance factor, and input the weighted real-time input features of the cementing pump into the fault classifier to predict a fault diagnosis result.

7. A cementing pump fault diagnosis device, characterized in that, The device includes: A historical data acquisition module, adapted to acquire the historical data of the cementing pump speed sensor, the historical data of the vibration sensor, and the historical operating conditions during the cementing operation; the historical operating conditions include normal conditions and fault conditions; A multi-dimensional extraction module, adapted to extract multi-dimensional historical features of the vibration sensor based on the time domain and the frequency domain for the historical data of any vibration sensor; combine the multi-dimensional historical features of the vibration sensor with the historical data of the speed sensor to obtain the historical input features of the cementing pump; A weighted processing module, adapted to calculate the relative between-class scatter distance of each historical input feature in the historical input features of the cementing pump, and construct a feature importance factor based on the relative between-class scatter distance, so as to perform weighted processing on the historical input features of the cementing pump according to the feature importance factor; The model training and prediction module is adapted to input the historical input features of the cementing pump after weighting processing and the historical operating conditions into a preset weighted width learning model, train the preset weighted width learning model, and obtain the trained preset weighted width learning model as a fault classifier, so as to predict a fault diagnosis result according to the real-time data of the cementing pump speed sensor and vibration sensor during the cementing operation.

8. A computing device, characterized in that, It includes: A processor, a memory, a communication interface and a communication bus. The processor, the memory and the communication interface complete communication with each other through the communication bus; The memory is used to store at least one executable instruction, and the executable instruction causes the processor to perform the operations corresponding to the cementing pump fault diagnosis method according to any one of claims 1-6.

9. A computer storage medium, characterized in that, At least one executable instruction is stored in the storage medium, and the executable instruction causes the processor to perform the operations corresponding to the cementing pump fault diagnosis method according to any one of claims 1-6.

10. A computer program product, characterized in that, It includes at least one executable instruction, and the executable instruction causes the processor to perform the operations corresponding to the cementing pump fault diagnosis method according to any one of claims 1-6.

Images