CN112785091B - Method for carrying out fault prediction and health management on oil field electric submersible pump - Google Patents
Method for carrying out fault prediction and health management on oil field electric submersible pump Download PDFInfo
- Publication number
- CN112785091B CN112785091B CN202110239188.5A CN202110239188A CN112785091B CN 112785091 B CN112785091 B CN 112785091B CN 202110239188 A CN202110239188 A CN 202110239188A CN 112785091 B CN112785091 B CN 112785091B
- Authority
- CN
- China
- Prior art keywords
- submersible pump
- electric submersible
- fault
- prediction
- module
- 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.)
- Active
Links
- 230000036541 health Effects 0.000 title claims abstract description 53
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000013528 artificial neural network Methods 0.000 claims abstract description 31
- 238000012423 maintenance Methods 0.000 claims abstract description 28
- 238000011156 evaluation Methods 0.000 claims abstract description 26
- 238000004458 analytical method Methods 0.000 claims abstract description 24
- 238000012544 monitoring process Methods 0.000 claims abstract description 21
- 238000012545 processing Methods 0.000 claims abstract description 11
- 239000011159 matrix material Substances 0.000 claims description 39
- 230000015556 catabolic process Effects 0.000 claims description 18
- 238000006731 degradation reaction Methods 0.000 claims description 18
- 230000006870 function Effects 0.000 claims description 18
- 238000007726 management method Methods 0.000 claims description 17
- 239000002245 particle Substances 0.000 claims description 12
- 238000012360 testing method Methods 0.000 claims description 12
- 238000009826 distribution Methods 0.000 claims description 7
- 230000003862 health status Effects 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 230000008859 change Effects 0.000 claims description 6
- 239000013598 vector Substances 0.000 claims description 5
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 claims description 4
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 claims description 4
- 238000013210 evaluation model Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 230000007246 mechanism Effects 0.000 claims description 3
- 238000002360 preparation method Methods 0.000 claims description 2
- BTCSSZJGUNDROE-UHFFFAOYSA-N gamma-aminobutyric acid Chemical compound NCCCC(O)=O BTCSSZJGUNDROE-UHFFFAOYSA-N 0.000 claims 1
- 230000007547 defect Effects 0.000 abstract description 6
- 238000010586 diagram Methods 0.000 description 9
- 230000000694 effects Effects 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 4
- 238000003745 diagnosis Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000011058 failure modes and effects analysis Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 241000221035 Santalaceae Species 0.000 description 1
- 235000008632 Santalum album Nutrition 0.000 description 1
- 238000005311 autocorrelation function Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 230000001364 causal effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000003129 oil well Substances 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
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—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; 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]
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Quality & Reliability (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Computer Hardware Design (AREA)
- Educational Administration (AREA)
- Geometry (AREA)
- Game Theory and Decision Science (AREA)
- Primary Health Care (AREA)
- Mining & Mineral Resources (AREA)
- Marine Sciences & Fisheries (AREA)
- Animal Husbandry (AREA)
- Agronomy & Crop Science (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
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 real-time evaluation of the state of the electric submersible pump through processing and analyzing the monitoring data and the historical data; the fault prediction module adopts a combined prediction method and estimates the residual service life of the electric submersible pump by combining with the fault mode analysis module; the maintenance strategy module analyzes and prepares a corresponding maintenance strategy to realize the guarantee decision of the electric submersible pump. According to the invention, 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 defect 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 science and technology, the integration level, complexity and intelligent degree of mechanical equipment are rapidly increased, and the traditional fault diagnosis and maintenance support technology is gradually difficult to adapt to new requirements.
In order to meet the requirements of quick, reliable and stable operation of equipment, the equipment is changed to informatization, digitalization and intellectualization, the autonomous guarantee and the autonomous diagnosis are realized, and the PHM (prediction and health management) technology at the end of the last century is developed and is rapidly paid high importance. Compared with traditional maintenance or regular overhaul after faults, the fault detection and diagnosis based on the current health state is prediction of the future health state, and the passive maintenance activity is changed into the pilot maintenance and guarantee activity, so that the safety and the high efficiency of the equipment are greatly improved. The technology can provide effective technical support for realizing autonomous logistics, improving the six-degree performance (reliability, maintainability, testability, safety, guarantee and environmental adaptability) of the system and reducing the cost of the whole life cycle for the modern oil field electric submersible pump.
At present, various fault prediction methods based on monitoring data exist for the electric submersible pump, for example, the peak of Beijing industrial university adopts a gray theory to predict, alarm, store and the like the vibration state of the electric submersible pump; the method for the Yao reclamation of the Chinese petroleum university adopts a PSO-BP neural network method to predict the well-lying time of the submersible electric pump well, can combine other pump parameters to judge the fault of the submersible electric pump well, and presumes the working condition; the method comprises the steps that the artificial neural network self-learning method is utilized to predict the working condition of the pump in the eastern direction of sandalwood of China university of petroleum, and the yield of an oil well of the electric submersible pump is improved through production parameter optimization; the Harbin engineering university Dong Zhengang provides a state monitoring method based on RBF network time sequence prediction for a downhole sensor of an electric submersible pump unit.
The above researches do not form a complete fault prediction and health management system for the oil field electric submersible pump, and the fault prediction methods all adopt a single prediction model. In predicting device performance using a different single model, the accuracy of the prediction is quite different from the model run time: if the time sequence model is suitable for processing the linear relation in the data, the long-term data can interfere with the prediction, and the accuracy can be drastically reduced; the neural network model is suitable for extracting nonlinear relations and interactions among variables in data, and the training process is slow, so that the problems of inaccurate single-step prediction and invalid multi-step prediction for the prediction of complex equipment generally occur.
In order to overcome the defects of the prior art, the invention firstly designs a fault prediction and health management system architecture of the oil field electric submersible pump, and secondly provides a fault prediction method of the oil field electric submersible pump based on the combination of a time sequence and a BP neural network. The neural network is used as a soft computing method driven by data, and can learn the rules of a complex system according to specific input and output samples, so that the neural network is very suitable for constructing a combined model with strong robustness with other prediction methods. The method can effectively make up the defects of a single prediction method, can process the interaction among linear relation, nonlinear relation and variables in data, capture the characteristics and variation modes of equipment, weaken the defect that the defects of a single model are amplified when the single model is used so as to increase the result error, improve the prediction precision and performance of the model, and shorten the running time of model prediction.
Disclosure of Invention
The invention provides a method for carrying out fault prediction and health management on an oil field electric submersible pump, which mainly aims at carrying out monitoring, management and analysis on data information acquired by state monitoring of the oil field electric submersible pump to realize self health state evaluation of the oil field electric submersible pump, carrying out fault prediction and health management on the oil field electric submersible pump before a system breaks down, and assuming that the trend of the electric submersible pump going through can extend to the future, the prediction has irregularity according to data, and a causal relationship is set aside, so that the aim of integrating fault detection, isolation, health prediction, evaluation and maintenance decision is fulfilled.
To achieve the above object, according to one aspect of the present invention, there is provided a method for fault prediction and health management of an oilfield electric submersible pump, comprising:
The fault mode analysis module is used for analyzing the functions of the electric submersible pump in the oil field and analyzing all fault modes affecting the performance of the electric submersible pump in the oil field to obtain key components affecting the functions of the electric submersible pump;
The state monitoring module is used for continuously monitoring the running state of the electric submersible pump in real time by using a sensor so as to obtain the exact information of the state characteristic change of the electric submersible pump, such as images, parameters and the like, and know the running state of the electric submersible pump;
The health state evaluation module is used for evaluating the health state of the oil field electric submersible pump in real time by comparing the data of the health state of the system with the maintained historical data;
The fault prediction module is used for predicting the fault trend of the electric submersible pump, adopting a fault prediction method based on the combination of a time sequence and the BP neural network, and estimating the residual service life of the electric submersible pump of the oil field by using the fault trend prediction result;
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, the failure mode analysis module includes:
fault mode, impact and hazard analysis (FMECA) for analyzing the function of the oil field electric submersible pump to obtain fault mode affecting the performance of the oil field electric submersible pump and analyzing various reasons and impacts possibly causing the fault mode;
Fault Tree Analysis (FTA) for deducting from top to bottom, analyzing the key components with the greatest influence on the performance of the electric submersible pump in the oil field, and finding out the main reasons and fault mechanisms of the fault occurrence.
Further, the health status assessment module includes:
(1) Weights are calculated and a weight matrix is constructed. And calculating the weights of all the components and the electric submersible pump subsystem according to the system composition structure. Performing pairwise comparison by adopting a 9-scale analytic hierarchy process to establish a judgment matrix, calculating a weight vector, and performing consistency test to finally obtain the weight matrix;
(2) And establishing a component-level degradation degree fuzzy judgment matrix. Substituting the degradation degree of each component into a ridge-shaped distribution membership function to obtain a component-level degradation degree fuzzy judgment matrix;
(3) And calculating a subsystem-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 the like, obtaining a health state evaluation matrix of the previous stage;
(4) Judging the health state of the submersible pump according to the maximum membership principle.
Further, step (1) includes:
(1-1) performing fault mode, influence and hazard analysis (FMECA) on the oil field electric submersible pump to obtain the severity of faults of each system;
(1-2) associating the scale 9 with the above-mentioned severity, and further determining a weight-establishing weight matrix based on the electric submersible pump hierarchical evaluation model.
The corresponding relation between the 9 scale and the severity in the health state evaluation module comprises the following steps:
further, a fault prediction method based on a combination of time series and BP neural network is adopted, and the fault prediction 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 characteristic parameter inputs, carrying out normalization processing on a data sequence, calculating correlation coefficients among variables in the sequence, and determining characteristic parameters with larger correlation with the next moment;
(3) The historical characteristic parameters determined by the time sequence model in the step (1) and the residual errors and the historical characteristic parameters determined by the step (2) are used as input variables of the neural network together, so that the structure of the neural network is determined;
(4) Establishing characteristic parameters by adopting a neural network optimized by particle swarm to obtain a prediction model;
(5) Evaluating the prediction quality by adopting an average absolute error (MAE), an average relative error (MAPE) and a Root Mean Square Error (RMSE) index;
(6) And performing estimation prediction by using a fault prediction model which is based on the time sequence and BP neural network combination and passes the verification.
Further, the maintenance policy module includes: task preparation, task execution, maintenance planning, maintenance execution.
The embodiment of the invention constructs a complete fault prediction and health management system architecture 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 fault prediction method based on the combination of the time sequence and the BP neural network can effectively make up for the defect of a single prediction method, improves the prediction precision and performance of the model, and effectively shortens the running time of model prediction.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings used in the technical description will be briefly described below.
FIG. 1 is a flow chart of a method for fault prediction and health management of an oilfield electric submersible pump according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a method for establishing a hierarchical evaluation model for an electric submersible pump according to an embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating health status evaluation according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a fault prediction method for an oilfield electric submersible pump based on a combination of time series and BP neural network according to an embodiment of the present invention;
fig. 5 is a schematic diagram of a maintenance strategy according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
A method of fault prediction and health management of an oilfield electric submersible pump comprising:
FIG. 1 is a flow chart of a method for fault prediction and health management of an oilfield electric submersible pump according to an embodiment of the present invention;
The fault mode analysis module is used for analyzing the functions of the electric submersible pump in the oil field and analyzing all fault modes affecting the performance of the electric submersible pump in the oil field to obtain key components affecting the functions of the electric submersible pump;
The state monitoring module is used for continuously monitoring the running state of the electric submersible pump in real time by using a sensor so as to obtain the exact information of the state characteristic change of the electric submersible pump, such as images, parameters and the like, and know the running state of the electric submersible pump;
The health state evaluation module is used for evaluating the health state of the oil field electric submersible pump in real time by comparing the data of the health state of the system with the maintained historical data;
The fault prediction module is used for predicting the fault trend of the electric submersible pump, adopting a fault prediction method based on the combination of a time sequence and the BP neural network, and estimating the residual service life of the electric submersible pump of the oil field by using the fault trend prediction result;
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 mode of the fault mode analysis module is as follows:
(1) The fault mode and the influence analysis table FMEA of the electric submersible pump are obtained by comprehensively analyzing the oil field electric submersible pump, and the fault mode and the influence analysis table FMEA mainly comprise the following steps: location, function, failure mode, cause of failure, failure impact, severity category, and possible measures;
(2) Further, analysis of the oilfield electric submersible pump results in a hazard analysis table CA, which mainly includes: the part, the failure rate lambda p, the failure mode frequency ratio alpha (j), the failure influence probability beta (j), the working time t, the failure mode hazard degree C mi (j) and the hazard degree C r (j);
(3) The fault mode, the influence and the hazard analysis (FMECA) of the oil field electric submersible pump are realized through the steps (1) and (2), the fault mode which influences the performance of the oil field electric submersible pump is obtained by analyzing the functions of the oil field electric submersible pump, and various reasons and influences which possibly cause the fault mode are analyzed;
(4) Further, fault Tree Analysis (FTA) is performed for deducting from top to bottom, analyzing to obtain key components affecting the function of the electric submersible pump, and finding out the main cause and fault mechanism of the fault occurrence.
Preferably, the specific implementation mode of the state monitoring module is as follows:
And continuously monitoring the running state of the electric submersible pump in real time by determining a test point installation sensor so as to obtain the exact information of the state characteristic change image, parameters and the like of the electric submersible pump, know the running state of the electric submersible pump and store data in real time.
Preferably, the specific implementation mode of the health state evaluation module is as follows:
FIG. 2 is a schematic diagram of a method for establishing 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 a capability layer which is a factor of the health evaluation index; the third layer is an index layer and is a health evaluation index system; the fourth layer is a parameter layer and is an evaluation substrate for health status or parameter characteristic composition.
Fig. 3 is a schematic diagram illustrating health status evaluation according to an embodiment of the present invention;
table 1 shows a correspondence of 9 scales and severity
After fault mode, influence and hazard analysis (FMECA) are carried out on the oil field electric submersible pump, the severity of faults of each system is obtained; table 1 corresponds the 9 scale to the severity described above, facilitating the determination of the weight-building weight matrix.
(1) Weights are calculated and a weight matrix is constructed. And calculating the weights of all the components and the electric submersible pump subsystem according to the system composition structure. And comparing every two by adopting a 9-scale analytic hierarchy process to establish a judgment matrix, calculating a weight vector, and carrying out consistency test to finally obtain the weight matrix.
TABLE 2 average random uniformity index RI
Consistency index
Ci=0, with complete consistency, CI close to 0, with satisfactory consistency, the larger the CI, the more serious the inconsistency. Wherein: n is the order of the judgment matrix.
Consistency ratio
And (3) checking the weight matrix A by using a consistency index, a consistency ratio < 0.1 and a numerical table (table 2) of the random consistency index RI. Since λ (characteristic root of a) continuously depends on a ij, the more λ is larger than n, the more serious the inconsistency of a fails, and the larger the judgment error is caused. The degree of inconsistency of a can thus be measured by the magnitude of the lambda-n value. Generally, when the consistency ratio cr=ci/RI < 0.1, the degree of inconsistency of a is considered to be within the allowable range, and there is satisfactory consistency, passing the consistency test. Otherwise, the comparison matrix A needs to be reconfigured to adjust a ij.
(2) And establishing a component-level degradation degree fuzzy judgment matrix. Substituting the degradation degree of each component into the ridge-shaped distribution membership function to obtain the component-level degradation degree fuzzy judgment matrix.
The system state characteristic data obtained in the state monitoring module is input in the degradation degree calculation process.
The degradation degree is within the range of [0,1],0 represents no degradation, and 1 represents failure or failure. For the ith state characteristic parameter, the degradation degree calculation formula is as follows: Wherein A i is the normal value of the ith state characteristic parameter; b i is the limit value of the ith state characteristic parameter; c i is the 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 general fault condition; the value of A i、Bi is taken from a prepared test standard, which is determined based on the design use and maintenance instructions of the electric submersible pump in the oil field or based on actual experience.
(3) And calculating a subsystem-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 the like, obtaining the upper-level health state evaluation matrix.
(4) Judging the health state of the submersible pump according to the maximum membership principle.
The maximum membership principle, let a i e F (U) (i=1, 2, …, n), for U 0 e U, if i 0 is present, let ：Ai0(u0)＝max{A1(u0),A2(u0),…,An(u0)}, consider U 0 relatively affiliated to a i0.
Table 3 four-dimensional table for evaluating health status of electric submersible pump in oil field
According to the degradation degree value range of [0,1], the normal state adopts a smaller distribution (smaller data in a processing interval), the sub-health state and the abnormal state select an intermediate distribution (middle data in the processing interval), and the fault state adopts a larger distribution (larger data in the processing interval) by referring to the related evaluation standard and expert experience and combining the change process of the ridge distribution membership function along with the degradation degree. The four-dimensional oilfield electric submersible pump health status membership vector grades are constructed as shown in table 3. Judging the health state of the submersible pump according to the membership maximum principle, so that the state of the component can be seen.
Preferably, the specific implementation mode of the fault prediction module is as follows:
Fig. 4 is a schematic diagram of a fault prediction method for an oilfield electric submersible pump based on a combination of 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 time sequence data, checking whether the time sequence is stable or not, wherein the stability of the combination of the time sequence and the BP neural network is an important precondition of modeling, and if the time sequence is non-stable sequence data, carrying out differential processing to ensure that the data is stable sequence;
adopting LB statistic analysis aiming at white noise test with fewer observation period numbers, when the delay period numbers m of the differential sequence are 5, 10 and 15 respectively, p values are smaller than alpha=0.05, rejecting the original hypothesis with 95% confidence level, namely the sequence is a non-white noise sequence, and passing the white noise test;
After the sequence is determined to be a stationary sequence which is not white noise, a corresponding time sequence model is determined by an autocorrelation and partial autocorrelation function diagram of the sequence.
(2) Determining the number of characteristic parameter inputs, carrying out normalization processing on a data sequence, calculating correlation coefficients among variables in the sequence, and determining characteristic parameters with larger correlation with the next moment;
(3) The historical characteristic parameters determined by the time sequence model in the step (1) and the residual errors and the historical characteristic parameters determined by the step (2) are used as input variables of the neural network together, so that the structure of the neural network is determined;
(4) Establishing characteristic parameters by adopting a neural network optimized by particle swarm to obtain a prediction model;
Taking the weight and the threshold value of the network as position vectors of particles, wherein the number of the weight and the threshold value is the dimension of the particles, and then setting the particle population scale, the maximum iteration number, the learning factor, the inertia weight and the initial speed and position of each particle;
And after the samples are normalized, starting the training network, continuously updating the input information by adopting a rolling optimization strategy, performing inverse normalization on the data obtained after training, taking the mean square error between the data and the actual value as an fitness function of particles, and calculating fitness values of individual particles and groups, wherein the smaller the fitness value is, the better the particle position is indicated, and the worse the particle position is indicated. 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 target value is not reached when the maximum iteration number is reached, and determining the weight threshold of the network according to the optimal position found at present.
(5) Evaluating the prediction quality by adopting an average absolute error (MAE), an average relative error (MAPE) and a Root Mean Square Error (RMSE) index;
Wherein N-the total number of test samples, xobs, i-the estimated value of the ith test sample, xmodel, i-the actual value of the ith test sample;
and (3) installing a vibration sensor to the oil field electric submersible pump to obtain monitoring data, and carrying out fault prediction by adopting the method.
Table 4 calculation results of the failure prediction method based on the combination of time series and BP neural network
As can be seen from Table 4, the Mean Absolute Error (MAE), mean relative error (MAPE) and Root Mean Square Error (RMSE) of the fault predictions of the time series and BP neural network combination are respectively reduced by 63%, 61% and 65% compared with the time series method, and respectively reduced by 48%, 39% and 37% compared with the BP neural network method, which shows that the prediction accuracy of the combined model is higher, and the training time of the combined model is shortened by one order of magnitude compared with the time series method, and is shortened by 57% compared with the BP neural network method.
(6) And performing estimation prediction by using a fault prediction model which is based on the time sequence and BP neural network combination and passes the verification. And taking 30% of data as historical data of modeling, predicting the last 70% of data, observing trend effect between the data and actual data, and detecting model prediction effect. Meanwhile, the method realizes the later-period short-term prediction, 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 useful life.
Preferably, the specific implementation mode of the maintenance strategy module is as follows:
FIG. 5 is a schematic diagram of a maintenance strategy according to an embodiment of the present invention;
And diagnosing and analyzing the received fault and abnormal conditions of the electric submersible pump in the oil field and the system function state information of the electric submersible pump in the oil field, giving an inference decision according to the conditions, judging whether the problems affect the electric submersible pump task in the oil field, giving task suggestions such as whether to continue to execute the task or stop the task, further analyzing the data, predicting fault deep diagnosis faults and analyzing the performance trend of the electric submersible pump in the oil field, and accordingly carrying out maintenance planning, dynamically scheduling resources, executing and completing maintenance activities. And finally triggering corresponding maintenance and guarantee activities before the electric submersible pump in the oil field completes the task according to the analysis result.
Claims (3)
1. A method for fault prediction and health management of an oilfield electric submersible pump, comprising the steps of:
The fault mode analysis module is used for analyzing the functions of the electric submersible pump in the oil field and analyzing all fault modes affecting the performance of the electric submersible pump in the oil field to obtain key components affecting the functions of the electric submersible pump;
The state monitoring module is used for continuously monitoring the running state of the electric submersible pump in real time by using a sensor so as to obtain the exact information of the state characteristic change of the electric submersible pump, such as images, parameters and the like, and know the running state of the electric submersible pump; the health state evaluation module is used for evaluating the health state of the oil field electric submersible pump in real time by comparing the data of the health state of the system with the maintained historical data;
The fault prediction module is used for predicting the fault trend of the electric submersible pump, adopting a fault prediction method based on the combination of a time sequence and the BP neural network, and estimating the residual service life of the electric submersible pump of the oil field by using the fault trend prediction result;
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;
The fault mode analysis module comprises:
fault mode, impact and hazard analysis (FMECA) for analyzing the function of the oil field electric submersible pump to obtain fault mode affecting the performance of the oil field electric submersible pump and analyzing various reasons and impacts possibly causing the fault mode;
Fault Tree Analysis (FTA) for deducting from top to bottom, analyzing key components having the greatest influence on the performance of the electric submersible pump in the oil field, and finding out the main cause and fault mechanism of the fault occurrence;
the health status assessment module comprises:
(1) Calculating weights and constructing a weight matrix: according to the system composition structure, calculating the weight of each component and the electric submersible pump subsystem; performing pairwise comparison by adopting a 9-scale analytic hierarchy process to establish a judgment matrix, calculating a weight vector, and performing consistency test to finally obtain the weight matrix;
(2) Establishing a component-level degradation degree fuzzy judgment matrix: substituting the degradation degree of each component into a ridge-shaped distribution membership function to obtain a component-level degradation degree fuzzy judgment matrix;
(3) Calculating a subsystem-level health state evaluation matrix: 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 the like, obtaining a health state evaluation matrix of the previous stage;
(4) Judging the health state of the submersible pump according to the principle of maximum membership;
The step (1) comprises:
(1-1) performing fault mode, influence and hazard analysis (FMECA) on the oil field electric submersible pump to obtain the severity of faults of each system;
(1-2) associating the scale 9 with the above-mentioned severity, and further determining weights according to the electric submersible pump classification evaluation model to establish a weight matrix;
The corresponding relation between the 9 scale and the severity in the health state evaluation module comprises the following steps:
Scale 1 represents two element comparisons, equally important, corresponding to severity class V class level-equivalent; scale 3 represents a two element comparison, one element being slightly more important than the other, corresponding to severity class IV level-mild; scale 5 represents a two element comparison, one element being significantly more important than the other, corresponding to severity class III level-critical; scale 7 represents a two element comparison, one element being very important than the other element, corresponding to severity class II class-deadly; scale 9 represents a two element comparison, one element being absolutely more important than the other, corresponding to severity class I level-disaster; when the scales are 2,4, 6 and 8, the two adjacent judging median values are represented, and the corresponding severity intermediate states are obtained; if the weight of factor i is a ij as compared to the weight of j, then the weight of factor j is a ji=1/aij as compared to the weight of i.
2. The method for fault prediction and health management of an oilfield electric submersible pump according to claim 1, wherein the fault prediction module adopts a fault prediction method based on a combination of time series and a BP neural network, and comprises the following steps:
(2-1) establishing a characteristic parameter time series model for monitoring the oilfield electric submersible pump;
(2-2) determining the number of characteristic parameter inputs, carrying out normalization processing on a data sequence, calculating correlation coefficients among variables in the sequence, and determining characteristic parameters with larger correlation with the next moment;
(2-3) using the historical characteristic parameters determined by the time series model in the step (2-1) and the residual errors and the historical characteristic parameters determined in the step (2-2) together as input variables of the neural network, so as to determine the structure of the neural network;
(2-4) establishing characteristic parameters by adopting a neural network optimized by particle swarm to obtain a prediction model;
(2-5) evaluating the prediction quality by using an average absolute error (MAE), an average relative error (MAPE) and a Root Mean Square Error (RMSE) index;
(2-6) performing fault prediction by using a fault prediction model based on a time series combination with the BP neural network, which has passed the test.
3. The method for fault prediction and health management of an oilfield electric submersible pump of claim 1, the maintenance strategy module is characterized by comprising: task preparation, task execution, maintenance planning, maintenance execution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110239188.5A CN112785091B (en) | 2021-03-04 | 2021-03-04 | Method for carrying out fault prediction and health management on oil field electric submersible pump |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110239188.5A CN112785091B (en) | 2021-03-04 | 2021-03-04 | Method for carrying out fault prediction and health management on oil field electric submersible pump |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112785091A CN112785091A (en) | 2021-05-11 |
| CN112785091B true CN112785091B (en) | 2024-08-23 |
Family
ID=75762247
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110239188.5A Active CN112785091B (en) | 2021-03-04 | 2021-03-04 | Method for carrying out fault prediction and health management on oil field electric submersible pump |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112785091B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109772439B (en) * | 2017-11-14 | 2022-02-08 | 中国石油化工股份有限公司 | Tert-butyl alcohol refining catalyst and preparation method and application thereof |
| 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 |
| CN113642251B (en) * | 2021-08-31 | 2024-05-28 | 佛山众陶联供应链服务有限公司 | Data analysis and prediction method and system for building ceramic spray powder preparation quality |
| CN115726764B (en) * | 2021-09-01 | 2024-10-29 | 中国石油化工股份有限公司 | Electric pump well fault diagnosis system and method based on fuzzy logic |
| 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 |
| CN114707266B (en) * | 2022-03-31 | 2023-04-21 | 江苏苏华泵业有限公司 | Industrial centrifugal pump operation stability prediction system based on artificial intelligence |
| CN115860718B (en) * | 2022-12-02 | 2023-07-25 | 深圳市黑金工业制造有限公司 | Integrated 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 (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106529145A (en) * | 2016-10-27 | 2017-03-22 | 浙江工业大学 | Bridge monitoring data prediction method based on ARIMA-BP neural network |
| 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 |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10458416B2 (en) * | 2014-12-02 | 2019-10-29 | Siemens Aktiengesellschaft | Apparatus and method for monitoring a pump |
| WO2016153895A1 (en) * | 2015-03-25 | 2016-09-29 | Schlumberger Technology Corporation | System and method for monitoring an electric submersible pump |
| CN109934422B (en) * | 2019-04-25 | 2021-06-29 | 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 |
| CN111461551B (en) * | 2020-04-01 | 2023-05-02 | 未必然数据科技(北京)有限公司 | Deep learning and SPC criterion-based electric submersible pump fault early warning method |
| CN111817880A (en) * | 2020-06-17 | 2020-10-23 | 安徽创米信息技术有限公司 | Oil and gas field production equipment health management system and implementation method |
-
2021
- 2021-03-04 CN CN202110239188.5A patent/CN112785091B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106529145A (en) * | 2016-10-27 | 2017-03-22 | 浙江工业大学 | Bridge monitoring data prediction method based on ARIMA-BP neural network |
| 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 |
Non-Patent Citations (3)
| Title |
|---|
| 分子泵故障诊断与寿命预测技术研究;林鹏飞;中国优秀硕士学位论文电子期刊;第1-24页 * |
| 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.2021,第14-23页. * |
| 雷达装备健康状态评价方法研究;李硕;《火控雷达技术》;第46卷(第1期);第12-15页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112785091A (en) | 2021-05-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112785091B (en) | Method for carrying out fault prediction and health management on oil field electric submersible pump | |
| WO2021042935A1 (en) | Bearing service life prediction method based on hidden markov model and transfer learning | |
| Byington et al. | Data-driven neural network methodology to remaining life predictions for aircraft actuator components | |
| CN111274737A (en) | Method and system for predicting remaining service life of mechanical equipment | |
| CN114297918B (en) | Aero-engine residual life prediction method based on full-attention depth network and dynamic ensemble learning | |
| CN112083244B (en) | Integrated intelligent diagnosis system for faults of avionic equipment | |
| CN118154174B (en) | Intelligent operation and maintenance cloud platform for industrial equipment | |
| CN114266278B (en) | Dual-attention network-based equipment residual service life prediction method | |
| CN111813084A (en) | Mechanical equipment fault diagnosis method based on deep learning | |
| US20220341996A1 (en) | Method for predicting faults in power pack of complex equipment based on a hybrid prediction model | |
| CN110757510B (en) | Method and system for predicting remaining life of robot | |
| CN117556347A (en) | Power equipment fault prediction and health management method based on industrial big data | |
| CN112257914B (en) | Aviation safety causal prediction method based on random forest | |
| CN116109039A (en) | Data-driven anomaly detection and early warning system | |
| CN114611372A (en) | Industrial equipment health prediction method based on Internet of things edge calculation | |
| CN117032165A (en) | Industrial equipment fault diagnosis method | |
| CN113989550A (en) | Electric vehicle charging pile operation state prediction method based on CNN and LSTM hybrid network | |
| CN111079348B (en) | Method and device for detecting slowly-varying signal | |
| CN117521512A (en) | Bearing residual service life prediction method based on multi-scale Bayesian convolution transducer model | |
| CN114861349B (en) | Rolling bearing RUL prediction method based on model migration and wiener process | |
| CN118188342A (en) | Fan-oriented fault early warning and life prediction method | |
| CN114548701B (en) | Full-measurement-point-oriented coupling structure analysis and estimation process early warning method and system | |
| CN116933003A (en) | Remaining service life prediction method of unmanned aerial vehicle engine based on DaNet | |
| CN114330515A (en) | Bridge monitoring data abnormity diagnosis and repair method | |
| 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |