CN112785091A - Method for performing fault prediction and health management on oil field electric submersible pump - Google Patents
Method for performing fault prediction and health management on oil field electric submersible pump Download PDFInfo
- Publication number
- CN112785091A CN112785091A CN202110239188.5A CN202110239188A CN112785091A CN 112785091 A CN112785091 A CN 112785091A CN 202110239188 A CN202110239188 A CN 202110239188A CN 112785091 A CN112785091 A CN 112785091A
- Authority
- CN
- China
- Prior art keywords
- submersible pump
- electric submersible
- fault
- prediction
- oil field
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/25—Design optimisation, verification or simulation using particle-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Mining
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/06—Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]
Abstract
The invention provides a method for carrying out fault prediction and health management on an oil field electric submersible pump, which comprises the following steps: the system comprises a fault mode analysis module, a state monitoring module, a health state evaluation module, a fault prediction module and a maintenance strategy module; the state monitoring module is used for acquiring characteristic parameters of the electric submersible pump; the health state evaluation module completes the real-time evaluation of the state of the electric submersible pump through the processing and analysis of the monitoring data and the historical data; the fault prediction module adopts a combined prediction method and is combined with the fault mode analysis module to estimate the residual service life of the electric submersible pump; and the maintenance strategy module analyzes and works out a corresponding maintenance strategy to realize a guarantee decision on the electric submersible pump. According to the method, firstly, a complete fault prediction and health management system framework of the oil field electric submersible pump is constructed, and secondly, a fault prediction method based on the combination of a time sequence and a BP neural network is adopted in a fault prediction module, so that the defects of a single prediction method can be effectively overcome, the prediction precision is improved, and the prediction time is effectively shortened.
Description
Technical Field
The invention relates to the technical field of prediction and health management, in particular to a method for performing fault prediction and health management on an oil field electric submersible pump.
Background
With the rapid development of scientific technology, the integration level, complexity and intelligent degree of mechanical equipment are increased rapidly, and the traditional fault diagnosis and maintenance support technology is difficult to adapt to new requirements gradually.
In order to meet the requirements of rapid, reliable and stable operation of equipment, the technology is transformed to informatization, digitization and intelligence, and autonomous guarantee and autonomous diagnosis are realized, the PHM (prediction and health management) technology has come into force at the end of the last century, and the PHM technology quickly gains high attention of the Western strong country represented by the United states. Compared with the traditional maintenance after fault or regular maintenance, the fault detection and diagnosis based on the current health state is the prediction of the future health state, the passive maintenance activity is changed into the leading maintenance and guarantee activity, and the safety and the efficiency of the equipment are greatly improved. The technology can provide effective technical support for realizing autonomous logistics, improving system 'sexuality' (reliability, maintainability, testability, safety, supportability and environmental adaptability) and reducing the whole life cycle cost of the modern oil field electric submersible pump.
At present, various fault prediction methods based on monitoring data exist for an electric submersible pump, for example, the vibration state of the electric submersible pump is predicted, alarmed and stored by adopting a grey theory at the peak of Beijing university of industry; the Yao reclamation of China Petroleum university adopts a PSO-BP neural network method to predict the well lying time of the submersible electric pump well, and can judge the fault of the submersible electric pump well by combining other pump parameters to conjecture the occurring working condition; the working condition of the pump is predicted by an artificial neural network self-learning method and the oil well yield of the electric submersible pump is improved through optimization of production parameters in Sandalong university of China; the torpedo steel of Harbin engineering university provides a state monitoring method based on RBF network time sequence prediction for an underground sensor of an electric submersible pump unit.
In the above researches, a set of complete fault prediction and health management system is not formed for the oil field electric submersible pump, and the fault prediction methods all adopt single prediction models. When using different single models to predict device performance, the accuracy of the prediction is much different than the model runtime: if the time series model is suitable for processing the linear relation in the data, the long-term data can interfere the prediction, and the accuracy can be reduced sharply; the neural network model is suitable for extracting nonlinear relations and interaction among variables in data, and the training process is slow, so that the problems of inaccurate single-step prediction and invalid multi-step prediction of complex equipment are generally caused.
In order to overcome the defects of the prior art, the invention firstly designs a fault prediction and health management system architecture of the oilfield electric submersible pump, and secondly provides an oilfield electric submersible pump fault prediction method based on the combination of a time sequence and a BP neural network. The neural network is used as a data-driven soft computing method, can learn the rules of a complex system according to specific input and output samples, and is very suitable for being combined with other prediction methods to construct a combined model with strong robustness. The method can effectively make up for the defects of a single prediction method, can process the linear relation, the nonlinear relation and the interaction among variables in data, captures the characteristics and the variation mode of equipment, weakens the defect that the defects of the single model are amplified when in use so as to increase the result error, improves the prediction precision and the performance of the model, and shortens the operation time of model prediction.
Disclosure of Invention
The invention provides a method for performing fault prediction and health management on an oil field electric submersible pump, which mainly aims to monitor, manage and analyze data information acquired by the oil field electric submersible pump through state monitoring to realize the health state evaluation of the oil field electric submersible pump, perform fault prediction and health management on the oil field electric submersible pump before a system fails, assume that the past trend of the electric submersible pump extends to the future, predict the data according to the irregularity, set aside the causal relationship, and achieve the purpose of integrating fault detection, isolation, health prediction and evaluation and maintenance decision into a whole.
To achieve the above object, according to one aspect of the present invention, there is provided a method for performing fault prediction and health management on an oilfield electric submersible pump, comprising:
the fault mode analysis module is used for analyzing the functions of the oil field electric submersible pump, analyzing all fault modes influencing the performance of the oil field electric submersible pump and obtaining key components influencing the functions of the electric submersible pump;
the state monitoring module is used for selecting a sensor to continuously monitor the running state of the oil field electric submersible pump in real time so as to obtain the exact information of images, parameters and the like of the state characteristic change of the electric submersible pump and know the running state of the electric submersible pump;
the health state evaluation module is used for evaluating the health state of the electric submersible pump of the oil field in real time by comparing the data of the health state of the system with the historical data of maintenance;
the fault prediction module is used for predicting the fault trend of the electric submersible pump, and estimating the residual service life of the electric submersible pump in the oil field by using a fault prediction method based on the combination of a time sequence and a BP neural network;
and the maintenance strategy module is used for analyzing and making a corresponding maintenance strategy according to the fault mode and the fault prediction result of the electric submersible pump so as to realize a guarantee decision.
Further, a failure mode analysis module, comprising:
the system comprises a fault mode analysis (FMECA), an influence analysis (FMECA) and a hazard analysis (FMECA), wherein the fault mode analysis (FMECA) is used for analyzing the functions of the oil field electric submersible pump to obtain a fault mode influencing the performance of the oil field electric submersible pump and analyzing various reasons and influences which may cause the fault mode;
and the Fault Tree Analysis (FTA) is used for deducing from top to bottom, analyzing key components which have the greatest influence on the performance of the oil field electric submersible pump and finding out the main reasons and the fault mechanism of the fault.
Further, a health status assessment module, comprising:
(1) weights are calculated and a weight matrix is constructed. And calculating the weight of each component and the electric submersible pump subsystem according to the system composition structure. Comparing every two by adopting a 9-scale analytic hierarchy process to establish a judgment matrix, calculating weight vectors and carrying out consistency check to finally obtain a weight matrix;
(2) and establishing a component-level degradation degree fuzzy judgment matrix. Substituting the degradation degrees of all the components into a ridge-shaped distribution membership function to obtain a component-level degradation degree fuzzy judgment matrix;
(3) and calculating a system-level health state evaluation matrix. And performing matrix multiplication operation by using the component-level weight matrix and the component-level degradation degree fuzzy judgment matrix to obtain a subsystem-level health state evaluation matrix. And by analogy, obtaining a health state evaluation matrix of the previous level;
(4) and judging the health state of the electric submersible pump according to the maximum membership principle.
Further, the step (1) comprises:
(1-1) carrying out failure mode, influence and hazard analysis (FMECA) on the electric submersible pump of the oil field to obtain the severity of failure of each system;
and (1-2) corresponding the 9 scales to the severity, and determining weights according to the electric submersible pump grading evaluation model to establish a weight matrix.
The corresponding relation between the 9 scales and the severity in the health state evaluation module comprises the following steps:
further, a fault prediction method based on the combination of the time series and the BP neural network is adopted, and the method comprises the following steps:
(1) establishing a characteristic parameter time sequence model for monitoring the oil field electric submersible pump;
(2) determining the number of input characteristic parameters, carrying out normalization processing on the data sequence, calculating a correlation coefficient between variables in the sequence, and determining the characteristic parameters with high correlation with the next moment;
(3) taking the historical characteristic parameters and the residual errors determined by the time series model in the step (1) and the historical characteristic parameters determined in the step (2) as input variables of the neural network together, thereby determining the structure of the neural network;
(4) establishing a characteristic parameter by adopting a particle swarm optimization neural network to obtain a prediction model;
(5) evaluating the quality of prediction by using indexes of Mean Absolute Error (MAE), mean relative error (MAPE) and Root Mean Square Error (RMSE);
(6) and carrying out estimation prediction by using a failure prediction model which passes the test and is based on the combination of the time series and the BP neural network.
Further, a maintenance strategy module comprising: task preparation, task execution, maintenance planning and maintenance execution.
The embodiment of the invention constructs a complete fault prediction and health management system framework of the oil field electric submersible pump, and realizes the fault prediction corresponding to the equipment fault mode through processing and analyzing the monitoring data and the historical data, thereby achieving the purpose of making a maintenance decision in time. The failure prediction method based on the combination of the time sequence and the BP neural network is adopted, so that the defect of a single prediction method can be effectively overcome, the prediction precision and performance of the model are improved, and the operation time of model prediction is effectively shortened.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings used in the technical description will be briefly introduced below.
FIG. 1 is a flow chart of a method for fault prediction and health management of an electric submersible pump in an oil field according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a hierarchical evaluation model for an electrical submersible pump according to an embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating a health status assessment according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a method for predicting a fault of an oilfield electrical submersible pump based on a combination of a time series and a BP neural network according to an embodiment of the present invention;
fig. 5 is a schematic view of a maintenance strategy according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some embodiments, but not all embodiments, of the present invention. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
A method for fault prediction and health management of an oil field electric submersible pump comprises the following steps:
FIG. 1 is a flow chart of a method for fault prediction and health management of an electric submersible pump in an oil field according to an embodiment of the present invention;
the fault mode analysis module is used for analyzing the functions of the oil field electric submersible pump, analyzing all fault modes influencing the performance of the oil field electric submersible pump and obtaining key components influencing the functions of the electric submersible pump;
the state monitoring module is used for selecting a sensor to continuously monitor the running state of the oil field electric submersible pump in real time so as to obtain the exact information of images, parameters and the like of the state characteristic change of the electric submersible pump and know the running state of the electric submersible pump;
the health state evaluation module is used for evaluating the health state of the electric submersible pump of the oil field in real time by comparing the data of the health state of the system with the historical data of maintenance;
the fault prediction module is used for predicting the fault trend of the electric submersible pump, and estimating the residual service life of the electric submersible pump in the oil field by using a fault prediction method based on the combination of a time sequence and a BP neural network;
and the maintenance strategy module is used for analyzing and making a corresponding maintenance strategy according to the fault mode and the fault prediction result of the electric submersible pump so as to realize a guarantee decision.
Preferably, the specific implementation of the failure mode analysis module is as follows:
(1) the failure mode of the electric submersible pump and the FMEA of the influence analysis table are obtained by comprehensively analyzing the system of the electric submersible pump in the oil field, and the method mainly comprises the following steps: location, function, failure mode, failure cause, failure impact, severity category and possible measures;
(2) further, the oil field electric submersible pump is analyzed to obtain a hazard analysis table CA which mainly comprises the following steps: location, failure rate λpFault mode frequency ratio alpha (j), fault influence probability beta (j), working time t and fault mode harmfulness Cmi(j) Degree of harmfulness Cr(j)；
(3) The method comprises the following steps of (1) and (2), realizing failure mode, influence and hazard analysis (FMECA) of the oil field electric submersible pump, analyzing the functions of the oil field electric submersible pump to obtain a failure mode influencing the performance of the oil field electric submersible pump, and analyzing various reasons and influences which may cause the failure mode;
(4) further, Fault Tree Analysis (FTA) is carried out for deduction from top to bottom, key components influencing the functions of the electric submersible pump are obtained through analysis, and main reasons and fault mechanisms of the electric submersible pump are found out.
Preferably, the specific implementation of the state monitoring module is as follows:
the running state of the electric submersible pump in the oil field is continuously monitored in real time by determining the installation sensor of the test point, so that the exact information of the image, the parameter and the like of the state characteristic change of the electric submersible pump is obtained, the running state of the electric submersible pump is known, and the data is stored in real time.
Preferably, the health status evaluation module is implemented as follows:
FIG. 2 is a schematic diagram of a hierarchical evaluation model for an electric submersible pump according to an embodiment of the present invention;
the hierarchical structure is divided into four layers, wherein the first layer is a target layer and is the health state of the electric submersible pump; the second layer is the ability layer, which is a factor of the health assessment index; the third layer is a index layer and is a health evaluation index system; the fourth layer is a parameter layer and is an evaluation substrate composed of health status parameters or parameter characteristics.
Fig. 3 is a schematic diagram illustrating a health status evaluation according to an embodiment of the invention;
table 1 shows the correspondence between 9-level tables and severity
After fault mode, influence and hazard analysis (FMECA) is carried out on the electric submersible pump of the oil field, the severity of faults of all systems is obtained; table 1 associates the 9 scale with the severity described above to facilitate determining weights and establishing a weight matrix.
(1) Weights are calculated and a weight matrix is constructed. And calculating the weight of each component and the electric submersible pump subsystem according to the system composition structure. And performing pairwise comparison by adopting a 9-scale analytic hierarchy process to establish a judgment matrix, calculating weight vectors and performing consistency check to finally obtain a weight matrix.
TABLE 2 average random consistency index RI
| Order of the scale | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
| RI | 0 | 0 | 0.52 | 0.89 | 1.12 | 1.26 | 1.36 | 1.41 |
| Order of the scale | 9 | 10 | 11 | 12 | 13 | 14 | 15 | |
| RI | 1.46 | 1.49 | 1.52 | 1.54 | 1.56 | 1.58 | 1.59 |
Index of consistency
When CI is 0, the complete consistency is realized, CI is close to 0, the satisfactory consistency is realized, and the larger CI is, the more serious the inconsistency is. Wherein: n is the order of the decision matrix.
Consistency ratio
And (3) utilizing the consistency index, the consistency ratio less than 0.1 and a numerical table (table 2) of the random consistency index RI to carry out a test process on the weight matrix A. Since λ (characteristic root of A) is continuously dependent on aijIf λ is larger than n, the inconsistency of a is more serious and the judgment error is larger. The magnitude of the λ -n value can be used to measure the degree of inconsistency of a. In general, when the consistency ratio CR is CI/RI < 0.1, A is considered to beThe degree of inconsistency of (2) is within an allowable range, and satisfactory consistency is achieved through consistency check. Otherwise, it needs to be reconfigured into the comparison matrix A, pair aijTo be adjusted.
(2) And establishing a component-level degradation degree fuzzy judgment matrix. And substituting the degradation degrees of all the components into a ridge-shaped distribution membership function to obtain a component-level degradation degree fuzzy judgment matrix.
The system state characteristic data obtained in the state monitoring module is used as input in the degradation degree calculation process.
The deterioration degree has a value range of [0, 1], 0 represents no deterioration, and 1 represents a failure or a failure. For the ith state characteristic parameter, the degradation degree calculation formula is as follows:
wherein A isiThe normal value of the ith state characteristic parameter is obtained; b isiThe limiting value of the ith state characteristic parameter; ciThe actual measured value of the ith state characteristic parameter; k is an index which reflects the influence degree of the change of the ith state parameter on the functions of the electric submersible pump of the oil field, and 2 is taken under the condition of general faults; a. thei、BiThe value of (a) is taken from prepared detection criteria, which is determined based on oilfield electrical submersible pump design use and maintenance instructions or based on actual experience.
(3) And calculating a system-level health state evaluation matrix. And performing matrix multiplication operation by using the component-level weight matrix and the component-level degradation degree fuzzy judgment matrix to obtain a subsystem-level health state evaluation matrix. And by analogy, obtaining a health state evaluation matrix of the previous level.
(4) And judging the health state of the electric submersible pump according to the maximum membership principle.
The principle of maximum membership, set AiE f (u) (i 1,2, …, n), for u0E.g. U if there is i0So that: a. thei0(u0)＝max{A1(u0),A2(u0),…,An(u0) Is then u is considered0Relatively belonging to Ai0。
TABLE 3 health evaluation chart for electric submersible pump in four-dimensional oil field
According to the degradation degree value range of [0, 1], related evaluation standards and expert experience are referred to, and a change process of a ridge-shaped distribution membership function along with the degradation degree is combined, wherein a normal state adopts a small distribution (data in a processing interval is small), a sub-health state and an abnormal state select an intermediate distribution (data in the middle of the processing interval), and a fault state selects a large distribution (data in the processing interval). And constructing the membership vector grade of the health state of the electric submersible pump of the four-dimensional oil field as shown in Table 3. The health state of the electric submersible pump is judged according to the maximum membership degree principle, so that the state of the component is basically known.
Preferably, the fault prediction module is implemented in a specific manner as follows:
fig. 4 is a schematic diagram illustrating a method for predicting a fault of an oilfield electrical submersible pump based on a combination of a time series and a BP neural network according to an embodiment of the present invention;
(1) establishing a characteristic parameter time sequence model for monitoring the oil field electric submersible pump;
analyzing the time sequence data, checking whether the time sequence is stable or not, wherein the stability of the time sequence and the BP neural network combination is an important premise for modeling, and if the time sequence is non-stable sequence data, performing differential processing to enable the data to be stable sequence;
LB statistic analysis is adopted for white noise test with less observation period, when the delay period m of the differential sequence is respectively 5, 10 and 15, the p values are all less than alpha which is 0.05, the original hypothesis is rejected with 95% confidence level, namely the sequence is a non-white noise sequence, and the white noise test is passed;
after the sequence is determined to be a stable sequence of non-white noise, a corresponding time sequence model is determined through the autocorrelation and partial autocorrelation function graph of the sequence.
(2) Determining the number of input characteristic parameters, carrying out normalization processing on the data sequence, calculating a correlation coefficient between variables in the sequence, and determining the characteristic parameters with high correlation with the next moment;
(3) taking the historical characteristic parameters and the residual errors determined by the time series model in the step (1) and the historical characteristic parameters determined in the step (2) as input variables of the neural network together, thereby determining the structure of the neural network;
(4) establishing a characteristic parameter by adopting a particle swarm optimization neural network to obtain a prediction model;
taking the weight and the threshold of the network as position vectors of the particles, wherein the number of the weight and the threshold is the dimension of the particles, and then setting the population scale of the particles, the maximum iteration times, the learning factor, the inertial weight and the initial speed and position of each particle;
selecting a training sample set of a neural network, starting training the network after normalizing samples, continuously updating input information by adopting a rolling optimization strategy, carrying out inverse normalization on data obtained after training, taking the mean square error between the data and an actual value as a fitness function of particles, and calculating the fitness value of individual particles and population of the particles, wherein the smaller the fitness value is, the better the particle position is, and otherwise, the worse the particle position is. And continuously updating the positions of the particles according to the fitness value until an optimal weight threshold is found, stopping the algorithm if the optimal weight threshold is not found when the maximum iteration times is reached, and determining the weight threshold of the network according to the currently found optimal position.
(5) Evaluating the quality of prediction by using indexes of Mean Absolute Error (MAE), mean relative error (MAPE) and Root Mean Square Error (RMSE);
wherein, N is the total number of the test samples, Xobs, i is the estimated value of the ith test sample, and Xmodel, i is the actual value of the ith test sample;
the method is used for carrying out fault prediction by mounting a vibration sensor on the electric submersible pump of the oil field to obtain monitoring data.
TABLE 4 Fault prediction method calculation results based on time series and BP neural network combination
As can be seen from table 4, the Mean Absolute Error (MAE), the mean relative error (MAPE), and the Root Mean Square Error (RMSE) of the fault prediction of the time series and BP neural network combination are respectively reduced by 63%, 61%, and 65% as compared with the time series method, and are respectively reduced by 48%, 39%, and 37% as compared with the BP neural network method, which indicates that the prediction accuracy of the combination model is higher, and meanwhile, the training time of the combination model is reduced by one order of magnitude as compared with the time series method, and is reduced by 57% as compared with the BP neural network method.
(6) And carrying out estimation prediction by using a failure prediction model which passes the test and is based on the combination of the time series and the BP neural network. And taking 30% of data as modeling historical data, predicting the later 70% of data, observing the trend effect between the data and actual data, and detecting the model prediction effect. Meanwhile, the method realizes the short-term prediction in the later period, and the prediction result is output and stored in a text form, and comprises the following steps: data type, failure mode, failure cause, failure impact, processing mode and remaining service life.
Preferably, the specific implementation manner of the maintenance strategy module is as follows:
fig. 5 is a schematic view of a maintenance strategy according to an embodiment of the present invention;
the method comprises the steps of carrying out diagnosis and analysis on received faults and abnormal conditions of the oil field electric submersible pump and system function state information of the oil field electric submersible pump, giving reasoning decisions and emergency treatment schemes according to conditions, judging whether the problems affect the task of the oil field electric submersible pump, giving task suggestions of continuing to execute the task or stopping the oil field electric submersible pump, further analyzing data, carrying out fault deep diagnosis fault prediction and performance trend analysis of the oil field electric submersible pump, and carrying out maintenance planning, dynamic resource scheduling, execution and maintenance activities according to the task suggestions. And finally, triggering corresponding maintenance support activities before the oil field electric submersible pump completes tasks according to the analysis result.
Claims (6)
1. A method for fault prediction and health management of an oil field electric submersible pump is characterized by comprising the following steps:
the fault mode analysis module is used for analyzing the functions of the oil field electric submersible pump, analyzing all fault modes influencing the performance of the oil field electric submersible pump and obtaining key components influencing the functions of the electric submersible pump;
the state monitoring module is used for selecting a sensor to continuously monitor the running state of the oil field electric submersible pump in real time so as to obtain the exact information of images, parameters and the like of the state characteristic change of the electric submersible pump and know the running state of the electric submersible pump;
the health state evaluation module is used for evaluating the health state of the electric submersible pump of the oil field in real time by comparing the data of the health state of the system with the historical data of maintenance;
the fault prediction module is used for predicting the fault trend of the electric submersible pump, and estimating the residual service life of the electric submersible pump in the oil field by using a fault prediction method based on the combination of a time sequence and a BP neural network;
and the maintenance strategy module is used for analyzing and making a corresponding maintenance strategy according to the fault mode and the fault prediction result of the electric submersible pump so as to realize a guarantee decision.
2. The method of claim 1, wherein the failure mode analysis module comprises:
the system comprises a fault mode analysis (FMECA), an influence analysis (FMECA) and a hazard analysis (FMECA), wherein the fault mode analysis (FMECA) is used for analyzing the functions of the oil field electric submersible pump to obtain a fault mode influencing the performance of the oil field electric submersible pump and analyzing various reasons and influences which may cause the fault mode;
and the Fault Tree Analysis (FTA) is used for deducing from top to bottom, analyzing key components which have the greatest influence on the performance of the oil field electric submersible pump and finding out the main reasons and the fault mechanism of the fault.
3. The method of fault prediction and health management of an oilfield electrical submersible pump of claim 1, wherein the health status assessment module comprises:
(1) weights are calculated and a weight matrix is constructed. And calculating the weight of each component and the electric submersible pump subsystem according to the system composition structure. Comparing every two by adopting a 9-scale analytic hierarchy process to establish a judgment matrix, calculating weight vectors and carrying out consistency check to finally obtain a weight matrix;
(2) and establishing a component-level degradation degree fuzzy judgment matrix. Substituting the degradation degrees of all the components into a ridge-shaped distribution membership function to obtain a component-level degradation degree fuzzy judgment matrix;
(3) and calculating a system-level health state evaluation matrix. And performing matrix multiplication operation by using the component-level weight matrix and the component-level degradation degree fuzzy judgment matrix to obtain a subsystem-level health state evaluation matrix. And by analogy, obtaining a health state evaluation matrix of the previous level;
(4) and judging the health state of the electric submersible pump according to the maximum membership principle.
4. The health status assessment module of claim 3, wherein said step (1) comprises:
(1-1) carrying out failure mode, influence and hazard analysis (FMECA) on the electric submersible pump of the oil field to obtain the severity of failure of each system;
and (1-2) corresponding the 9 scales to the severity, and determining weights according to the electric submersible pump grading evaluation model to establish a weight matrix.
The correspondence between the 9 scales and the severity in the health status evaluation module comprises the following steps:
5. the method for fault prediction and health management of the oilfield electric submersible pump according to claim 1, wherein the fault prediction module adopts a fault prediction method based on a combination of a time series and a BP neural network, and comprises the following steps:
(1) establishing a characteristic parameter time sequence model for monitoring the oil field electric submersible pump;
(2) determining the number of input characteristic parameters, carrying out normalization processing on the data sequence, calculating a correlation coefficient between variables in the sequence, and determining the characteristic parameters with high correlation with the next moment;
(3) taking the historical characteristic parameters and the residual errors determined by the time series model in the step (1) and the historical characteristic parameters determined in the step (2) as input variables of the neural network together, thereby determining the structure of the neural network;
(4) establishing a characteristic parameter by adopting a particle swarm optimization neural network to obtain a prediction model;
(5) evaluating the quality of prediction by using indexes of Mean Absolute Error (MAE), mean relative error (MAPE) and Root Mean Square Error (RMSE);
(6) and carrying out estimation prediction by using a failure prediction model which passes the test and is based on the combination of the time series and the BP neural network.
6. The method of claim 1, wherein the maintenance strategy module comprises: task preparation, task execution, maintenance planning and maintenance execution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110239188.5A CN112785091A (en) | 2021-03-04 | 2021-03-04 | Method for performing fault prediction and health management on oil field electric submersible pump |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110239188.5A CN112785091A (en) | 2021-03-04 | 2021-03-04 | Method for performing fault prediction and health management on oil field electric submersible pump |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112785091A true CN112785091A (en) | 2021-05-11 |
Family
ID=75762247
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110239188.5A Pending CN112785091A (en) | 2021-03-04 | 2021-03-04 | Method for performing fault prediction and health management on oil field electric submersible pump |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112785091A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109772439A (en) * | 2017-11-14 | 2019-05-21 | 中国石油化工股份有限公司 | A kind of tert-butyl alcohol catalyst for refining and its preparation method and application |
| CN113570231A (en) * | 2021-07-22 | 2021-10-29 | 浙江幸福航天科技有限公司 | Power station health management system |
| CN113642251A (en) * | 2021-08-31 | 2021-11-12 | 佛山众陶联供应链服务有限公司 | Data analysis and prediction method and system for powder making quality of architectural ceramic spray |
| CN113836820A (en) * | 2021-10-20 | 2021-12-24 | 联想新视界(江西)智能科技有限公司 | Equipment health assessment and fault diagnosis algorithm based on autocorrelation model and multivariate monitoring method |
| CN114565161A (en) * | 2022-03-01 | 2022-05-31 | 北京九天翱翔科技有限公司 | Storage stage fault prediction and service life prolonging method based on multi-time-space dimension data fusion |
| CN114707266A (en) * | 2022-03-31 | 2022-07-05 | 江苏苏华泵业有限公司 | Industrial centrifugal pump operation stability prediction system based on artificial intelligence |
| CN115860718A (en) * | 2022-12-02 | 2023-03-28 | 深圳市黑金工业制造有限公司 | Touch all-in-one machine comprehensive maintenance management system and method based on big data |
| CN117539155A (en) * | 2024-01-09 | 2024-02-09 | 深圳市威诺达工业技术有限公司 | Optimized control method of electric submersible pump |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016089237A1 (en) * | 2014-12-02 | 2016-06-09 | Siemens Aktiengesellschaft | Monitoring of a pump |
| WO2016153895A1 (en) * | 2015-03-25 | 2016-09-29 | Schlumberger Technology Corporation | System and method for monitoring an electric submersible pump |
| CN106529145A (en) * | 2016-10-27 | 2017-03-22 | 浙江工业大学 | ARIMA-BP neutral network-based bridge monitoring data prediction method |
| CN109934422A (en) * | 2019-04-25 | 2019-06-25 | 中国人民解放军国防科技大学 | Neural network wind speed prediction method based on time series data analysis |
| CN210402440U (en) * | 2019-10-22 | 2020-04-24 | 常州机电职业技术学院 | Predictive maintenance system for marine centrifugal pump |
| CN111353236A (en) * | 2020-03-16 | 2020-06-30 | 福建省特种设备检验研究院 | Health state evaluation system of petrochemical normal-pressure oil storage tank based on multiple factors |
| CN111461551A (en) * | 2020-04-01 | 2020-07-28 | 未必然数据科技(北京)有限公司 | Electric submersible pump fault early warning method based on deep learning and SPC (statistical computer) criterion |
-
2021
- 2021-03-04 CN CN202110239188.5A patent/CN112785091A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016089237A1 (en) * | 2014-12-02 | 2016-06-09 | Siemens Aktiengesellschaft | Monitoring of a pump |
| WO2016153895A1 (en) * | 2015-03-25 | 2016-09-29 | Schlumberger Technology Corporation | System and method for monitoring an electric submersible pump |
| CN106529145A (en) * | 2016-10-27 | 2017-03-22 | 浙江工业大学 | ARIMA-BP neutral network-based bridge monitoring data prediction method |
| CN109934422A (en) * | 2019-04-25 | 2019-06-25 | 中国人民解放军国防科技大学 | Neural network wind speed prediction method based on time series data analysis |
| CN210402440U (en) * | 2019-10-22 | 2020-04-24 | 常州机电职业技术学院 | Predictive maintenance system for marine centrifugal pump |
| CN111353236A (en) * | 2020-03-16 | 2020-06-30 | 福建省特种设备检验研究院 | Health state evaluation system of petrochemical normal-pressure oil storage tank based on multiple factors |
| CN111461551A (en) * | 2020-04-01 | 2020-07-28 | 未必然数据科技(北京)有限公司 | Electric submersible pump fault early warning method based on deep learning and SPC (statistical computer) criterion |
Non-Patent Citations (7)
| Title |
|---|
| "Applying Big Data Analytics to Detect, Diagnode, and Prevent Impending Failures in Electric Submersible Pumps", SOCIETY OF PETROLEUM ENGINEERS * |
| LU LI: "Condition monitoring and fault diagnosis of electric submersible pump based on wellhead electrical parameters and production parameters", SYSTEMS SCIENCE & CONTROL ENGINEERING * |
| 彭宇: "故障预测与健康管理技术综述", 电子测量与仪器学报, vol. 24, no. 1 * |
| 李硕: "雷达装备健康状态评价方法研究", 《火控雷达技术》, vol. 46, no. 1, pages 159 - 163 * |
| 林鹏飞: "分子泵故障诊断与寿命预测技术研究", 中国优秀硕士学位论文电子期刊, pages 14 - 23 * |
| 隋先富: "基于主成分分析的电潜泵故障诊断", 西南石油大学学报, vol. 42, no. 6, pages 107 - 114 * |
| 韩丽娟: "油田潜油电泵故障诊断系统的设计与研究", 中国优秀硕士学位论文电子期刊 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109772439A (en) * | 2017-11-14 | 2019-05-21 | 中国石油化工股份有限公司 | A kind of tert-butyl alcohol catalyst for refining and its preparation method and application |
| CN113570231A (en) * | 2021-07-22 | 2021-10-29 | 浙江幸福航天科技有限公司 | Power station health management system |
| CN113642251A (en) * | 2021-08-31 | 2021-11-12 | 佛山众陶联供应链服务有限公司 | Data analysis and prediction method and system for powder making quality of architectural ceramic spray |
| CN113836820A (en) * | 2021-10-20 | 2021-12-24 | 联想新视界(江西)智能科技有限公司 | Equipment health assessment and fault diagnosis algorithm based on autocorrelation model and multivariate monitoring method |
| CN114565161A (en) * | 2022-03-01 | 2022-05-31 | 北京九天翱翔科技有限公司 | Storage stage fault prediction and service life prolonging method based on multi-time-space dimension data fusion |
| CN114707266A (en) * | 2022-03-31 | 2022-07-05 | 江苏苏华泵业有限公司 | Industrial centrifugal pump operation stability prediction system based on artificial intelligence |
| CN115860718A (en) * | 2022-12-02 | 2023-03-28 | 深圳市黑金工业制造有限公司 | Touch all-in-one machine comprehensive maintenance management system and method based on big data |
| CN117539155A (en) * | 2024-01-09 | 2024-02-09 | 深圳市威诺达工业技术有限公司 | Optimized control method of electric submersible pump |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112785091A (en) | Method for performing fault prediction and health management on oil field electric submersible pump | |
| CN110555273B (en) | Bearing life prediction method based on hidden Markov model and transfer learning | |
| CN108984893B (en) | Gradient lifting method-based trend prediction method | |
| CN111539515B (en) | Complex equipment maintenance decision method based on fault prediction | |
| CN111274737A (en) | Method and system for predicting remaining service life of mechanical equipment | |
| CN106528975B (en) | A kind of prognostic and health management method applied to Circuits and Systems | |
| CN112083244B (en) | Integrated intelligent diagnosis system for faults of avionic equipment | |
| CN109710983B (en) | Diesel engine cylinder layered fault diagnosis method based on key performance indexes | |
| CN108304348B (en) | Bearing residual life prediction method based on binary wiener process | |
| CN111813084A (en) | Mechanical equipment fault diagnosis method based on deep learning | |
| CN110705812A (en) | Industrial fault analysis system based on fuzzy neural network | |
| US20220341996A1 (en) | Method for predicting faults in power pack of complex equipment based on a hybrid prediction model | |
| CN108959498A (en) | A kind of big data processing platform and its design method for health monitoring | |
| Daher et al. | New prognosis approach for preventive and predictive maintenance—Application to a distillation column | |
| CN116595319A (en) | Prediction method and system applied to rail transit motor health state evaluation | |
| CN111079348B (en) | Method and device for detecting slowly-varying signal | |
| CN117609836A (en) | Electromagnetic sensitivity prediction and health management method for integrated module | |
| Gandur et al. | Uncertainty Quantification in the Prediction of Remaining Useful Life Considering Multiple Failure Modes | |
| CN112380782A (en) | Rotating equipment fault prediction method based on mixed indexes and neural network | |
| CN112729878A (en) | Method for evaluating health state of CRH380 type running gear system | |
| CN114861979A (en) | Rolling bearing residual life prediction method based on LSTM and TDNN | |
| CN114861349A (en) | Rolling bearing RUL prediction method based on model migration and wiener process | |
| CN111259494B (en) | Health monitoring and analyzing method for heavy machine equipment | |
| CN110060374B (en) | Method and device for detecting abnormality of aircraft fuel system | |
| Zheng et al. | Research on Predicting Remaining Useful Life of Equipment Based on Health Index |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |