An intelligent early warning method and system for oil pump unit fault model

Abstract

The present invention provides an intelligent early warning method and system for an oil pump unit fault model, and relates to the technical field of pipeline oil transportation equipment diagnosis. The method includes: obtaining a fault signal of the oil pump unit; recording the time parameters corresponding to the fault signal in the SCADA module, and performing early warning levels and fault markings on the data in the SCADA module; analyzing the parameter status of the SCADA data at the corresponding time according to the parameter over-limit information; constructing an association rule model according to the warning level, fault marking and parameter status; and obtaining predicted fault early warning information of the oil pump unit according to the association rule model. The present invention solves the technical problem in the prior art that the fault early warning effect is poor due to the insensitivity to the change of working conditions and the change process of the unit state, and the alarm is too late. By analyzing the SCADA system monitoring data reflecting the operating state of the oil pump unit, an early abnormal state early warning of the oil pump unit fault is achieved, thereby achieving the effect of improving the pertinence and accuracy of the fault early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of pipeline oil transportation equipment diagnosis, and in particular to an intelligent early warning method and system for an oil pump unit fault model. Background Art

[0002] As a key equipment for pipeline oil transportation, the oil pump is an important factor in ensuring the normal operation of pipeline transportation. At present, the diagnosis technology of oil pumps mainly faces the following three problems: First, the current status monitoring of 80% of oil pump units relies on SCADA system warnings, and only two fixed alarm thresholds are set: high alarm and high-high alarm. This warning mechanism does not pay attention to the changes in the unit's operating conditions and the process of unit operating status changes and personalized performance, resulting in insensitivity to changes in monitoring parameters. Second, the fixed threshold alarm value design based on the manufacturer's experience and the equipment's tolerance limit has no unified standard reference. Due to the high alarm value setting, the unit alarm is too late and is often on the verge of shutdown. Third, SCADA warnings only have simple over-value alarms, and the status monitoring system based on the installation of high-frequency vibration sensors has a low degree of intelligence due to the small accumulation of fault data, and fault diagnosis usually requires manual experience.

[0003] The existing technology has technical problems such as being insensitive to changes in operating conditions and unit status, and the alarm is too late, resulting in poor fault warning effect. Summary of the invention

[0004] The embodiment of the present application provides an intelligent early warning method and system for an oil pump unit fault model, which is used to solve the technical problems in the prior art of being insensitive to changes in operating conditions and unit state changes, and alarming too late, resulting in poor fault warning effects.

[0005] In view of the above problems, an embodiment of the present application provides an intelligent early warning method and system for an oil pump unit fault model.

[0006] In the first aspect, an embodiment of the present application provides an intelligent early warning method for an oil pump unit fault model, the method comprising: obtaining a fault signal of the oil pump unit; recording the time parameters corresponding to the fault signal in the SCADA module, and performing early warning level and fault marking on the data in the SCADA module; analyzing the parameter status of the SCADA data at the corresponding time according to the parameter over-limit information in the SCADA module; constructing an association rule model according to the early warning level, the fault mark and the parameter status; and obtaining predicted fault early warning information of the oil pump unit according to the association rule model.

[0007] In the second aspect, an embodiment of the present application provides an intelligent early warning system for an oil pump unit fault model, the system comprising: a fault signal acquisition module, the fault signal acquisition module is used to obtain the fault signal of the oil pump unit; a data marking module, the data marking module is used to record the time parameters corresponding to the fault signal in the SCADA module, and to perform early warning level and fault marking on the data in the SCADA module; a parameter status analysis module, the parameter status analysis module is used to analyze the parameter status of the SCADA data at the corresponding time according to the parameter over-limit information in the SCADA module; a model construction module, the model construction module is used to construct an association rule model according to the early warning level, the fault mark and the parameter status; and an early warning information acquisition module, the early warning information acquisition module is used to obtain the predicted fault early warning information of the oil pump unit according to the association rule model.

[0008] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0009] An intelligent early warning method for an oil pump unit fault model provided in an embodiment of the present application relates to the technical field of pipeline oil transportation equipment diagnosis, and obtains a fault signal of the oil pump unit; records the time parameters corresponding to the fault signal in the SCADA module, and performs early warning levels and fault markings on the data in the SCADA module; analyzes the parameter status of the SCADA data at the corresponding time according to the parameter over-limit information in the SCADA module; constructs an association rule model according to the early warning level, the fault mark and the parameter status; obtains the predicted fault early warning information of the oil pump unit according to the association rule model. The technical problem of poor fault early warning effect caused by insensitivity to changes in operating conditions and unit status changes and too late alarms in the prior art is solved. By analyzing the SCADA system monitoring data reflecting the operating status of the oil pump unit, early abnormal state early warning of the oil pump unit fault is achieved, thereby achieving the effect of improving the pertinence and accuracy of fault early warning.

[0010] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 A schematic flow chart of an intelligent early warning method for an oil pump unit fault model is provided for an embodiment of the present application;

[0012] Figure 2 A schematic diagram of a process flow for obtaining a fault signal in an intelligent early warning method for a fault model of an oil pump unit is provided in an embodiment of the present application;

[0013] Figure 3 A schematic diagram of a process flow for obtaining parameter states in an intelligent early warning method for an oil pump unit fault model is provided for an embodiment of the present application;

[0014] Figure 4 A schematic diagram of the structure of an intelligent early warning system for an oil pump unit fault model is provided for an embodiment of the present application.

[0015] Explanation of reference numerals: fault signal acquisition module 10 , data marking module 20 , parameter state analysis module 30 , model building module 40 , warning information acquisition module 50 . DETAILED DESCRIPTION

[0016] The embodiment of the present application provides an intelligent early warning method for an oil pump unit fault model, which is used to solve the technical problems in the prior art that the method is insensitive to changes in operating conditions and unit state changes, and the alarm is too late, resulting in poor fault warning effect.

[0017] Embodiment 1

[0018] like Figure 1 As shown, an embodiment of the present application provides an intelligent early warning method for an oil pump unit fault model, the method is applied to an intelligent early warning system for an oil pump unit fault model, the system includes a SCADA module, and the method includes:

[0019] Step S100: obtaining a fault signal of the oil pump unit;

[0020] Specifically, an intelligent early warning method for an oil pump unit fault model provided in an embodiment of the present application is applied to an intelligent early warning system for an oil pump unit fault model. The intelligent early warning system for the oil pump unit fault model includes a SCADA module, and the SCADA module is used for data collection and monitoring and control of the oil pump unit.

[0021] Furthermore, if Figure 2 As shown, step S100 of the present application also includes:

[0022] Step S110: collecting the operating status and operating parameters of the oil pump unit to obtain unit status data;

[0023] Furthermore, in step S110 of the present application, the unit status data includes an oil pump vibration signal and / or an oil pump temperature signal and / or an oil pump flow signal and / or an oil pump motor voltage signal and / or an oil pump motor current signal / or upstream and downstream regulating valve position information / or motor bearing temperature / or motor winding temperature / or mechanical seal leakage / or pump inlet pressure / or pump outlet pressure.

[0024] Specifically, the detection device is used to collect the oil pump vibration signal and / or the oil pump temperature signal and / or the oil pump flow signal and / or the oil pump motor voltage signal and / or the oil pump motor current signal / or the upstream and downstream regulating valve position information / or the motor bearing temperature / or the motor winding temperature / or the mechanical seal leakage / or the pump inlet pressure / or the pump outlet pressure of the oil pump unit. To collect data, the measuring points must be determined first. For the selection of measuring points, the principle of obtaining the most information with the least measuring points should be followed under the premise of ensuring accuracy. When selecting measuring points, in addition to the placement of the instrument close to the position to be measured, it is also necessary to consider whether the installation and disassembly of the sensor is convenient. Generally, the bearings, pump casings and monitoring points of the oil pump are preferably selected. For example, a vibration meter is used to select multiple vibration monitoring points to monitor the vibration signal. For example, for a horizontal machine pump, its vibration monitoring points can be selected on four bearing seats, and the vibration information values ​​from the vertical and horizontal directions are aligned with the axis. The collected oil pump vibration signal and / or oil pump temperature signal and / or oil pump flow signal and / or oil pump motor voltage signal and / or oil pump motor current signal / or upstream and downstream regulating valve position information / or motor bearing temperature / or motor winding temperature / or mechanical seal leakage / or pump inlet pressure / or pump outlet pressure of the oil pump unit are added to the unit status data.

[0025] Step S120: pre-processing the unit status data to obtain unit status standard data;

[0026] Furthermore, step S120 of the present application also includes:

[0027] Step S121: performing data cleaning on the unit status data to obtain unit status characteristic data;

[0028] Step S122: normalizing the unit state characteristic data to obtain the unit state standard data.

[0029] Specifically, data cleaning is the process of streamlining the database to remove duplicate records and converting the remaining part into a standard acceptable format. Data cleaning is processed through the following steps: first, select the first data column according to the unit status data. The first data column is the data that needs to be analyzed. To avoid interference, other data columns that do not participate in the analysis can be hidden. For example, the oil pump temperature signal is selected as the first data column; second, obtain the duplicate data values ​​in the first data column, retain the first data of the duplicate data, and delete other duplicate data; third, data values ​​may be missing in the original data, and the missing values ​​are deleted. When the data is small, the missing data can be manually supplemented. For example, the average value of the first data column is calculated, and the average value is used as the missing value of the first data column to supplement the first data column, or the data with the highest frequency in the first data column is counted and used as the missing value to supplement the first data column; fourth, when there is data in the first data column that is inconsistent with the standard or the naming rules are inconsistent, the data values ​​in the inconsistent data column are split using the column splitting function; fifth, the data in the first data column is filtered, sorted, ascending, descending, and other functions are applied to complete data cleaning. Through data cleaning, the data is reviewed and verified, and then duplicate data is deleted, erroneous data is corrected, and data consistency is provided to generate unit status characteristic data.

[0030] Normalization processing generally maps data to a specified range to remove the dimensions and dimensional units of data of different dimensions. For example, the present application prefers the Min-Max normalization method to map the result value to between [0, 1]. The conversion function is as follows: Among them, x new is the normalized data, x is the sample data, x max is the maximum value of the sample data, x min The minimum value of the sample data is obtained by normalization processing to obtain the standard data of the unit state. Through normalization processing, the data is mapped to a specified range, so that indicators of different units or magnitudes can be weighted and compared, achieving the effect of simplifying calculation.

[0031] Step S130: compressing the unit state standard data to obtain unit state standard information;

[0032] Step S140: extracting abnormal features from the standard information of the unit status to obtain the fault signal.

[0033] Specifically, the unit state standard data is compressed by the principal component analysis algorithm. First, a feature data set matrix is ​​constructed according to the unit state standard data, and a first feature data set is obtained. Then, each feature data in the first feature data set is centralized. First, the average value of each feature in the first feature data set is solved, and then for all samples, each feature is subtracted from its own mean value, and then a new eigenvalue is obtained. The second feature data set is formed by the new feature data set, and the second feature data set is a data matrix. The second feature data set is operated by the covariance formula to obtain the first covariance matrix of the second feature data set, and then the eigenvalue and eigenvector of the first covariance matrix are obtained by matrix operation, and each eigenvalue corresponds to an eigenvector. Among the first eigenvectors obtained, the largest first J eigenvalues ​​and their corresponding eigenvectors are selected, and the original features in the first feature data set are projected onto the selected eigenvectors to obtain the first feature data set after dimensionality reduction, which is used as the unit state standard information. The fault data in the unit state standard data is obtained, and it is marked, and the information with fault marks in the unit state standard information is extracted to obtain the fault signal.

[0034] Step S200: recording the time parameters corresponding to the fault signal in the SCADA module, and performing warning level and fault marking on the data in the SCADA module;

[0035] Specifically, the SCADA module is a computer-based production process control and scheduling automation module, which can monitor and control the operating equipment on site, and record the time parameters corresponding to the fault signal through the SCADA module.

[0036] According to the standard information of the unit status, J characteristic values ​​of the first characteristic data set are obtained, wherein each characteristic value corresponds to a temperature data interval, and each characteristic value is used as a warning level. The closer the characteristic value is to the front, the higher the warning level is. The maximum value of the temperature data interval corresponding to each characteristic value is used as the maximum threshold of the corresponding warning level, and the minimum value of the temperature data interval corresponding to each characteristic value is used as the minimum threshold of the corresponding warning level. And the fault data corresponding to all temperature data in the temperature data interval corresponding to the characteristic value is used as the fault data within the warning level. Similarly, the other characteristic data sets in the standard information of the unit status are divided into warning levels and faults, and the data in the SCADA module are marked with warning levels and faults according to the division results.

[0037] Step S300: Analyze the parameter status of the SCADA data at the corresponding time according to the parameter over-limit information in the SCADA module;

[0038] Furthermore, if Figure 3As shown, step S300 of the present application also includes:

[0039] Step S310: extracting abnormal warning features and real-time data of the oil pump status based on the SCADA module;

[0040] Step S320: constructing an abnormal state model based on the abnormal warning characteristics and the real-time data of the oil pump state;

[0041] Specifically, based on the SCADA module, the abnormal warning features of different warning levels and the real-time data of the oil pump status corresponding to the abnormal warning features are extracted. Based on the abnormal warning features and the real-time data of the oil pump status, the K-means algorithm is used to build an abnormal state model. Specifically, the real-time data of the oil pump status is used as a sample, K points are randomly selected from the sample as the initial centroid, the distance from each sample to each centroid is calculated, and the samples are divided into clusters with the nearest centroid. The mean of all samples in each cluster is calculated, and the centroid of the cluster is updated using the mean. Iteration is performed until the position change of the centroid is less than the specified threshold or the maximum number of iterations is reached. At this time, a clustering result is generated, and each clustering result is matched with the abnormal warning feature to build an abnormal state model.

[0042] Step S330: Perform multi-parameter correlation characteristic analysis of the operating condition and the state of the oil pump unit by using the operating condition-state characteristic analysis method to construct an early warning model for the oil pump unit;

[0043] Furthermore, step S330 of the present application also includes:

[0044] Step S331: construct an oil pump unit early warning model using one or more algorithms of cluster analysis, long short-term memory, neural network, regression modeling, and outlier detection.

[0045] Specifically, cluster analysis is an analysis process that groups a collection of sample data into multiple classes consisting of similar objects. The goal of cluster analysis is to collect data on a similar basis for classification and to build models based on the classification results. Long short-term memory is a variant of recurrent neural network, which is used to process sequence data. It can accept input sequences of any length and process each element in the sequence through internal recurrent connections. It can also solve the gradient vanishing and gradient exploding problems of traditional RNN in processing long sequence data. Neural network is a mathematical model or computational model that imitates the structure and function of biological neural network and is used to estimate or approximate functions. Regression modeling is a predictive modeling technique that studies the relationship between dependent and independent variables. It can perform predictive analysis, build time series models, and discover causal relationships between variables. Outlier detection is the process of finding objects whose behavior is different from the expected object. Such objects are called outliers or anomalies. Model building is performed using cluster analysis, long short-term memory, neural network, regression modeling, outlier detection, or one or more of these analysis methods.

[0046] Preferably, the present application adopts cluster analysis to analyze the multi-parameter correlation characteristics of the operating condition and the state of the oil pump unit. First, the normal parameters and abnormal parameters in the state parameters of the oil pump unit are obtained, and the fault level is used as the horizontal coordinate and the number of faults is used as the vertical coordinate to construct a state parameter coordinate system, and the state parameter coordinate system is regionalized and labeled. Different regions correspond to different label classification results, that is, normal parameters and abnormal parameters. Input the operating condition parameters into the state parameter coordinate system, obtain the state parameter vector corresponding to the operating condition parameters, calculate the distance of the parameters according to the Euclidean distance, obtain the Euclidean distance data set between the state parameter vector and other parameters, and obtain the parameter classification data set according to the Euclidean distance data set. The parameter classification data set is the shortest k distances in the Euclidean distance data set, and the k value is a part of the Euclidean distance data set and can be set by itself. Mapping and matching is performed based on the parameter classification data set and the label classification results, and the classification label corresponding to the state parameter vector is obtained, so as to perform an abnormal parameter warning for the oil pump unit.

[0047] Step S340: Record the time when the fault occurs in the SCADA module and construct a fault tag;

[0048] Step S350: Based on the abnormal state model, the oil pump unit early warning model and the fault label, the oil pump unit operating state parameter data in the SCADA data at the corresponding time is analyzed to obtain the parameter state.

[0049] Specifically, the node where the fault occurs, the fault information, and the operating status parameter data of the node are marked according to the fault label, that is, the oil pump unit is in a normal state before this node, and the oil pump unit is in an abnormal state at and after this node, and the corresponding operating status parameter data is also abnormal data.

[0050] The operating status parameter data of the oil pump unit in the SCADA data at the corresponding moment is input into the abnormal state model, the distance between the operating status parameter data and each centroid is calculated, the cluster where the nearest centroid is located is obtained, the operating status parameter data is divided into the cluster, and the abnormal warning feature corresponding to the cluster is obtained. The abnormal warning feature is the operating status parameter data corresponding to the operating status parameter data, and the warning level of the abnormal warning feature is used as the warning level of the operating status parameter data.

[0051] The operating status parameter data of the oil pump unit in the SCADA data at the corresponding moment is input into the oil pump unit early warning model to obtain the parameter vector corresponding to the operating status parameter data. The distance calculation is performed to obtain the shortest k distances between the parameter vector and other parameters as the state parameter classification data set. The state parameter classification data set is mapped and matched with the label classification result to obtain the classification label corresponding to the parameter vector, so as to perform abnormal parameter early warning of the oil pump unit.

[0052] According to the fault label, the node where the fault occurred, the fault information, and the operating status parameter data of the oil pump unit are obtained. Combined with the warning level and warning items, the parameter status of the oil pump unit at the node is constructed.

[0053] Step S400: constructing an association rule model according to the warning level, the fault mark and the parameter state;

[0054] Furthermore, step S400 of the present application also includes:

[0055] Step S410: dividing the warning level, the fault mark and the parameter state by time to obtain a parameter limit-exceeding information table;

[0056] Step S420: Based on a preset period, the parameter limit-exceeding information table is analyzed by an association rule mining algorithm to obtain an association rule model.

[0057] Specifically, a parameter-time coordinate system is constructed with time as the horizontal axis and the operating status parameter as the vertical axis. A curve y=y(t) showing the operating status parameter changing with time is drawn in the coordinate system. The vertical axis of the coordinate system is divided according to the warning level. For example, for the warning parameter of the first warning level, [a, b] is drawn. Two straight lines y=a and y=b are drawn in the parameter-time coordinate system. When the operating status parameter increases and the curve y=y(t) reaches the first warning level, the curve y=y(t) exceeds the straight line y=a, and the intersection node t of the curve y=y(t) and the straight line y=a is obtained. 1 ; Then, when the operating state parameter decreases and the curve y=y(t) is lower than the first warning level, the curve y=y(t) is lower than the straight line y=a, and the intersection node t of the curve y=y(t) and the straight line y=a is obtained. 2 . Set node t 1 With node t 2 The operating status parameters between are called the over-limit parameters for the first warning level. Over-limit means that the target value reaches the upper and lower limits of the over-limit value. The operating status parameters are monitored based on the over-limit value. When the over-limit value is reached, it means that the warning level has been reached, so an early warning is issued. The warning level, fault mark and parameter status of the over-limit parameter are obtained, and a parameter over-limit information table is constructed.

[0058] Association rules reflect the interdependence between an object and other objects. If there is a certain association between multiple objects, then one object can be predicted through other objects. Set a preset period, exemplarily set to 24 hours, extract data within 24 hours, obtain a data set, and obtain one or more parameter limit-crossing information tables in the extracted data set. Set I = (i 1 ,i 2 ,i 3 ...i k ,) is a set of all items in the data set. The parameter over-limit information table is set as D. In the parameter over-limit information table, the situation where the warning level, fault mark and parameter status appear at the same time is marked as thing t, and each thing t is a subset of I. The support is calculated, that is, the probability of the over-limit parameter appearing in the entire parameter over-limit information table, the ratio of the number of times a certain combination appears to the total number of times, and the formula is For combinations that do not appear many times, we do not need to pay attention to them and set the minimum support Sup min , the minimum support is a threshold parameter used to identify the minimum support of a data set. Finally, the confidence is calculated, that is, the probability of the B parameter appearing after the A parameter appears, which can be regarded as the conditional probability P(B|A)=P(AB) / P(A). The association relationship of the parameters is calculated based on the support, minimum support and confidence, and the association rule model is constructed.

[0059] Step S500: obtaining predicted fault warning information of the oil pump unit according to the association rule model;

[0060] Specifically, the association relationship between the warning level, fault mark and parameter status is obtained according to the association rule model, the collected real-time parameter status is input into the association rule model, and the corresponding warning level, fault mark and parameter status are matched according to the association relationship, so as to quickly detect the parameters of the oil pump unit. When abnormal parameters appear, the corresponding fault type and warning level are quickly matched, so as to realize the early abnormal status warning of the oil pump unit failure, thereby achieving the effect of improving the pertinence and accuracy of the fault warning.

[0061] In summary, the intelligent early warning method and system for the oil pump unit fault model provided by the embodiment of the present application have the following technical effects:

[0062] Obtain the fault signal of the oil pump unit, record the time parameters corresponding to the fault signal in the SCADA module, and perform warning level and fault marking on the data in the SCADA module. Analyze the parameter status of the SCADA data at the corresponding time according to the parameter over-limit information in the SCADA module, and build an association rule model according to the warning level, the fault mark and the parameter status; obtain the predicted fault warning information of the oil pump unit according to the association rule model. The technical problem of poor fault warning effect caused by insensitivity to changes in working conditions and unit status changes and too late alarm in the prior art is solved. By analyzing the SCADA system monitoring data reflecting the operating status of the oil pump unit, early abnormal state warning of the oil pump unit fault is realized, so as to improve the pertinence and accuracy of fault warning.

[0063] Embodiment 2

[0064] Based on the same inventive concept as the intelligent early warning method for an oil pump unit fault model in the aforementioned embodiment, Figure 4 As shown, the present application provides an intelligent early warning system for an oil pump unit fault model, the system comprising:

[0065] A fault signal acquisition module 10, wherein the fault signal acquisition module 10 is used to obtain a fault signal of the oil pump unit;

[0066] A data marking module 20, which records the time parameters corresponding to the fault signal in the SCADA module, and the data marking module 20 is used to mark the data in the SCADA module with a warning level and a fault;

[0067] A parameter status analysis module 30, wherein the parameter status analysis module 30 is used to analyze the parameter status of the SCADA data at the corresponding time according to the parameter over-limit information in the SCADA module;

[0068] A model building module 40, the model building module 40 is used to build an association rule model according to the warning level, the fault mark and the parameter state;

[0069] The warning information acquisition module 50 is used to obtain the predicted fault warning information of the oil pump unit according to the association rule model.

[0070] Furthermore, the system further comprises:

[0071] A unit status data acquisition module, which is used to collect the operating status and operating parameters of the oil pump unit to obtain unit status data;

[0072] A preprocessing module, the preprocessing module is used to preprocess the unit status data to obtain unit status standard data;

[0073] A compression module, the compression module is used to compress the unit state standard data to obtain unit state standard information;

[0074] An abnormal feature extraction module is used to extract abnormal features from the unit status standard information to obtain the fault signal.

[0075] Furthermore, the system further comprises:

[0076] A data cleaning module, the data cleaning module is used to clean the unit status data to obtain unit status characteristic data;

[0077] A normalization processing module is used to perform normalization processing on the unit state characteristic data to obtain the unit state standard data.

[0078] Furthermore, the system further comprises:

[0079] A real-time data extraction module, which is used to extract abnormal warning features and real-time data of the oil pump status based on the SCADA module;

[0080] An abnormal state model building module, the abnormal state model building module is used to build an abnormal state model based on the abnormal warning characteristics and the real-time data of the oil pump state;

[0081] A correlation characteristic analysis module, which is used to perform a multi-parameter correlation characteristic analysis between the operating condition and the state of the oil pump unit through a condition-state characteristic analysis method, and to construct an early warning model for the oil pump unit;

[0082] A fault label building module, wherein the fault label building module is used to record the time when the fault occurs in the SCADA module and build a fault label;

[0083] An operating status parameter data analysis module is used to analyze the operating status parameter data of the oil pump unit in the SCADA data at the corresponding time based on the abnormal state model, the oil pump unit early warning model and the fault label to obtain the parameter state.

[0084] Furthermore, the system further comprises:

[0085] The early warning model construction module is used to construct an early warning model for the oil pump unit by using one or more algorithms of cluster analysis, long short-term memory, neural network, regression modeling, and outlier detection.

[0086] Furthermore, the system further comprises:

[0087] A time division module, the time division module is used to divide the warning level, the fault mark and the parameter state into time divisions to obtain a parameter limit-crossing information table;

[0088] The information table analysis module is used to analyze the parameter limit-crossing information table based on a preset period by using an association rule mining algorithm to obtain an association rule model.

[0089] Through the above-mentioned detailed description of the intelligent early warning method of an oil pump unit fault model in this specification, those skilled in the art can clearly know the intelligent early warning method and system of the oil pump unit fault model in this embodiment. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0090] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An intelligent early warning method for an oil pump unit fault model, characterized in that: The method is applied to an intelligent early warning system for a fault model of an oil pump unit, the system comprising a SCADA module, and the method comprising: Obtain the fault signal of the oil pump unit; The SCADA module records the time parameters corresponding to the fault signal, and performs warning level and fault marking on the data in the SCADA module; Analyze the parameter status of SCADA data at the corresponding time according to the parameter limit-exceeding information in the SCADA module; Constructing an association rule model according to the warning level, the fault mark and the parameter state; According to the association rule model, obtaining predicted fault warning information of the oil pump unit; According to the warning level, the fault mark and the parameter state, an association rule model is constructed, including: Dividing the warning level, the fault mark and the parameter status by time to obtain a parameter limit-exceeding information table; Based on a preset period, the parameter limit-exceeding information table is analyzed by an association rule mining algorithm to obtain an association rule model.

2. The method according to claim 1, characterized in that The method of obtaining a fault signal of the oil pump unit further comprises: Collecting the operating status parameters of the oil pump unit to obtain unit status data; Preprocessing the unit status data to obtain unit status standard data; compressing the unit state standard data to obtain unit state standard information; Abnormal features are extracted from the standard information of the unit status to obtain the fault signal.

3. The method according to claim 2, characterized in that The unit status data includes an oil pump vibration signal and / or an oil pump temperature signal and / or an oil pump flow signal and / or an oil pump motor voltage signal and / or an oil pump motor current signal.

4. The method according to claim 2, characterized in that Preprocessing the unit status data to obtain unit status standard data, the method further includes: Performing data cleaning on the unit status data to obtain unit status characteristic data; The unit state characteristic data is normalized to obtain the unit state standard data.

5. The method according to claim 1, characterized in that The method further comprises: Based on the SCADA module, extract abnormal warning features and real-time data of oil pump status; Based on the abnormal warning characteristics and the real-time data of the oil pump status, construct an abnormal status model; The multi-parameter correlation characteristics of the operating conditions and the oil pump unit status are analyzed through the operating condition-state characteristic analysis method, and the oil pump unit early warning model is constructed; Record the time when the fault occurs in the SCADA module and build a fault tag; Based on the abnormal state model, the oil pump unit early warning model and the fault label, the oil pump unit operating state parameter data in the SCADA data at the corresponding time is analyzed to obtain the parameter state.

6. The method according to claim 5, characterized in that The working condition-state characteristic analysis method includes: One or more algorithms including cluster analysis, long short-term memory, neural network, regression modeling, and outlier detection are used to construct an early warning model for the oil pump unit.

7. An intelligent early warning system for an oil pump unit fault model, characterized in that: The system comprises a SCADA module, the system comprising: A fault signal acquisition module, wherein the fault signal acquisition module is used to obtain a fault signal of the oil pump unit; A data marking module, in which the time parameters corresponding to the fault signal are recorded in the SCADA module, and the data marking module is used to mark the data in the SCADA module with a warning level and a fault; A parameter status analysis module, wherein the parameter status analysis module is used to analyze the parameter status of the SCADA data at the corresponding time according to the parameter over-limit information in the SCADA module; A model building module, the model building module is used to build an association rule model according to the warning level, the fault mark and the parameter state; A warning information acquisition module, the warning information acquisition module is used to obtain predicted fault warning information of the oil pump unit according to the association rule model; A time division module, the time division module is used to divide the warning level, the fault mark and the parameter state into time divisions to obtain a parameter limit-crossing information table; The information table analysis module is used to analyze the parameter limit-crossing information table based on a preset period by using an association rule mining algorithm to obtain an association rule model.

Images