Oil-water well field fault diagnosis system and method based on abnormal feature recognition

Abstract

The invention belongs to the field of well field fault diagnosis, and relates to the technical field of data analysis, in particular to an oil-water well field fault diagnosis system and method based on abnormal feature recognition, and the system comprises a model distribution module, a diagnosis evaluation module and a distribution optimization module, and the model distribution module, the diagnosis evaluation module and the distribution optimization module are in communication connection in sequence. The model distribution module is used for performing distribution analysis on the fault diagnosis model of the oil-water well field: generating a distribution period, randomly distributing the fault diagnosis model for an analysis object, marking the fault diagnosis model as a matching model of the analysis object, and performing fault diagnosis analysis by adopting the matching model as the analysis object in the distribution period; according to the method, the fault diagnosis processing effect of the oil-water well field is evaluated and analyzed, a matching model is adopted for an analysis object through an evaluation coefficient, and the fault diagnosis analysis effect is evaluated and fed back in a distribution period.

Description

Technical Field

[0001] The present invention belongs to the field of well site fault diagnosis and relates to data analysis technology, and specifically to an oil and water well site fault diagnosis system and method based on abnormal feature recognition. Background Art

[0002] Oil and water well sites are important facilities in the process of oil extraction, used to extract underground oil resources and conduct preliminary processing and management. Oil and water well site fault diagnosis is a vital part of oilfield production management, which aims to ensure the smooth operation and efficient development of oilfield production by analyzing and diagnosing abnormal changes in oil and water wells.

[0003] The invention patent with announcement number CN117196586A discloses a well site inspection method, device, system and electronic equipment, which sends the target inspection results to the scheduling system, which is used for the scheduling system to determine the maintenance task corresponding to the target well site based on the target inspection results, and control the robot to perform fault repair on the target fault point in the target well site based on the maintenance task; however, the well site inspection method cannot count the effect parameters of the fault diagnosis process, and it is difficult to perform allocation optimization analysis of the fault diagnosis model based on the diagnosis effect, resulting in the failure to improve the well site fault diagnosis efficiency and recognition rate.

[0004] In view of the above technical problems, this application proposes a solution. Summary of the invention

[0005] The purpose of the present invention is to provide an oil and water well field fault diagnosis system and method based on abnormal feature recognition, which is used to solve the problem in the prior art that it is impossible to perform distribution optimization analysis of the fault diagnosis model according to the diagnosis effect.

[0006] The technical problem to be solved by the present invention is: how to provide an oil and water well field fault diagnosis system and method based on abnormal feature recognition, which can perform distribution optimization analysis of fault diagnosis models according to diagnosis results.

[0007] The object of the present invention can be achieved by the following technical solutions: an oil and water well site fault diagnosis system based on abnormal feature recognition, comprising a model allocation module, a diagnosis evaluation module and an allocation optimization module; The model allocation module is used to allocate and analyze the fault diagnosis model of the oil and water well field: generate an allocation cycle, mark the oil and water well field as an analysis object, obtain environmental parameters of the analysis object at the beginning of the allocation cycle, randomly allocate a fault diagnosis model to the analysis object and mark it as a matching model of the analysis object, and use the matching model to perform fault diagnosis analysis for the analysis object within the allocation cycle; The diagnostic evaluation module is used to evaluate and analyze the fault diagnosis and processing effect of the oil and water well field: obtain the efficiency data XS and identification data SS of the analysis object in the allocation cycle and perform numerical calculations to obtain the evaluation coefficient PG of the fault diagnosis analysis of the analysis object using the matching model in the allocation cycle; The allocation optimization module is used to optimize and analyze the fault diagnosis model allocation process of the oil and water well field: generate an optimization cycle, and at the beginning of the optimization cycle, the management personnel selects an allocation mode, which includes an efficiency priority mode, an identification priority mode, and a comprehensive priority mode.

[0008] Furthermore, the environmental parameters of the analysis object include temperature and pressure values; the fault diagnosis model includes a current comparison model, a machine learning model, a time series analysis model and a SVM model.

[0009] Furthermore, the process of acquiring the efficiency data XS includes: when an abnormality is detected in the analysis object, fault diagnosis is performed through the matching model, and the start time of the fault diagnosis is marked as the diagnosis start time; the time when the matching model completes the fault type classification or determines that the fault type cannot be identified is marked as the diagnosis end time; the diagnosis start time and the diagnosis end time constitute a fault diagnosis process, and the average duration of all fault diagnosis processes of the analysis object within the allocation period is marked as the efficiency data XS.

[0010] Furthermore, the process of acquiring the identification data SS includes: marking the fault diagnosis process that completes the fault type classification as the identification process, and marking the ratio of the number of identification processes of the analysis object within the allocation cycle to the number of fault diagnosis processes as the identification data SS.

[0011] Furthermore, the specific process of using the efficiency priority mode to allocate the fault diagnosis model includes: marking the matching model of the analysis object with the smallest efficiency data XS in the allocation cycle as the efficiency priority model, and using the efficiency priority model to perform fault diagnosis analysis in the optimization cycle.

[0012] Furthermore, the specific process of using the recognition priority mode to allocate the fault diagnosis model includes: marking the matching model of the analysis object with the largest recognition data SS in the allocation cycle as the recognition priority model, and using the recognition priority model to perform fault diagnosis analysis in the optimization cycle.

[0013] Furthermore, the specific process of using the comprehensive priority mode to assign fault diagnosis models includes: generating several environmental data groups, marking the analysis objects whose temperature values ​​and pressure values ​​are respectively within the temperature range and pressure range in the environmental data group as compliant objects of the environmental data group, and marking the matching model of the compliant object with the smallest evaluation coefficient PG in the environmental data group as a compliant model of the environmental data group; obtaining the temperature value and pressure value of the analysis object at the beginning of the optimization cycle and marking them as temperature analysis values ​​and pressure analysis values ​​respectively, searching for the corresponding environmental data group through the temperature analysis value and pressure analysis value and marking it as a compliant data group, marking the compliant model of the compliant data group as a comprehensive priority model, and using the comprehensive priority model for fault diagnosis analysis within the optimization cycle.

[0014] Furthermore, the generation process of the environmental data group includes: forming a temperature range from the maximum and minimum temperature values ​​of all analysis objects obtained at the start of the allocation cycle, dividing the temperature range into a number of temperature intervals; forming a pressure range from the maximum and minimum pressure values ​​of all analysis objects obtained at the start of the allocation cycle, dividing the pressure range into a number of pressure intervals; and arbitrarily combining temperature intervals and pressure intervals to obtain a number of environmental data groups.

[0015] The present invention also proposes an oil and water well field fault diagnosis method based on abnormal feature recognition, comprising the following steps: Step 1: Perform distribution analysis on the fault diagnosis model of the oil and water well field: generate a distribution cycle, mark the oil and water well field as the analysis object, obtain the environmental parameters of the analysis object at the beginning of the distribution cycle, randomly assign the fault diagnosis model to the analysis object and mark it as the matching model of the analysis object; Step 2: Evaluate and analyze the fault diagnosis and treatment effect of the oil and water well field: obtain the efficiency data XS and identification data SS of the analysis object in the allocation period and perform numerical calculation to obtain the evaluation coefficient PG of the analysis object in the allocation period; Step 3: Generate an optimization cycle. At the beginning of the optimization cycle, the manager selects an allocation mode and allocates the fault diagnosis model to the analysis object according to the selected allocation mode.

[0016] The present invention has the following beneficial effects: The model allocation module can be used to allocate and analyze the fault diagnosis models of oil and water wells, randomly allocate the fault diagnosis models in the allocation cycle, and collect statistics on the environmental parameters of the oil and water wells to provide data support for the evaluation and optimization analysis process; The diagnostic evaluation module can be used to evaluate and analyze the fault diagnosis and treatment effects of oil and water well sites, and to comprehensively analyze and calculate multiple effect parameters of all fault diagnosis processes of the analysis object within the allocation cycle to obtain the evaluation coefficient. The evaluation coefficient can be used to evaluate and provide feedback on the effect of the fault diagnosis analysis of the analysis object using the matching model within the allocation cycle. The allocation optimization module can be used to optimize and analyze the fault diagnosis model allocation process of the oil and water well site. Users can select the allocation mode according to their own needs. The fault diagnosis models obtained by different allocation modes have different diagnostic tendencies, and the different needs of different users can be solved by multi-mode allocation, thereby improving user satisfaction. At the same time, an optimal allocation method that matches the environmental parameters of the oil and water well site is provided through a comprehensive priority mode. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 is a system block diagram of Embodiment 1 of the present invention; Figure 2 This is a flow chart of the method of Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0019] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] Embodiment 1: Figure 1 As shown, the oil and water well field fault diagnosis system based on abnormal feature recognition includes a model allocation module, a diagnosis evaluation module and an allocation optimization module, and the model allocation module, the diagnosis evaluation module and the allocation optimization module are communicatively connected in sequence.

[0021] The model allocation module is used to allocate and analyze the fault diagnosis model of the oil and water well field: generate an allocation cycle, mark the oil and water well field as the analysis object, obtain the environmental parameters of the analysis object at the beginning of the allocation cycle, the environmental parameters of the analysis object include temperature and pressure values, randomly allocate a fault diagnosis model to the analysis object and mark it as the matching model of the analysis object. The fault diagnosis model includes a current comparison model, a machine learning model, a time series analysis model and a SVM model.

[0022] The current comparison model diagnoses oil well faults by comparing the current changes of the motor under different working conditions. This method is suitable for detecting the working conditions of the motor under different slip rates, such as the changes in the states of the motor, generator and electromagnetic brake. ‌Machine learning model‌ is a technology that allows computers to automatically extract rules and patterns from data to complete specific tasks. It allows computers to automatically interact with the environment and learn how to maximize rewards. Reinforcement learning is suitable for scenarios that require decision-making and strategy optimization. Machine learning models can process large amounts of data and learn from them the operating modes and fault characteristics of equipment, providing new solutions for fault prediction and diagnosis‌‌; Time series analysis model Time series analysis is an important fault diagnosis technology. It processes and analyzes continuous signals collected at equal intervals to establish statistical laws. The model includes time domain analysis and frequency domain analysis. Time domain analysis evaluates the periodicity and randomness of signals by observing their time history, while frequency domain analysis decomposes complex signals into superpositions of simple signals through Fourier transform. Time series models can capture the time dependence in oil and water well production data and use historical data to predict future states. In fault diagnosis, these models can identify abnormal patterns or trends by analyzing time series data such as electrical parameter signals and production data of oil or water wells, and warn of potential faults in advance. The SVM model extracts the HOG features of the normal dynamometer diagram of the oil well and uses the support vector machine (SVM) model for intelligent diagnosis. This method can identify minor sanding, upper collision and lower hanging faults with high accuracy and is suitable for diagnosis under complex working conditions. During the allocation period, the matching model is used as the analysis object for fault diagnosis analysis. The fault diagnosis model of the oil and water well field is allocated and analyzed, and the fault diagnosis model is randomly allocated during the allocation period. At the same time, the environmental parameters of the oil and water well field are statistically analyzed to provide data support for the evaluation and optimization analysis process.

[0023] The diagnosis evaluation module is used to evaluate and analyze the fault diagnosis and processing effect of the oil and water well field: obtain the efficiency data XS and identification data SS of the analysis object within the allocation cycle, and the acquisition process of the efficiency data XS includes: when the analysis object is detected to be abnormal, the fault diagnosis is performed through the matching model, and the start time of the fault diagnosis is marked as the diagnosis start time, and the time when the matching model completes the fault type classification or determines that the fault type cannot be identified is marked as the diagnosis end time. The diagnosis start time and the diagnosis end time constitute a fault diagnosis process, and the average time of all fault diagnosis processes of the analysis object within the allocation cycle is marked as the efficiency data XS; The process of obtaining the identification data SS includes: marking the fault diagnosis process that completes the fault type classification as the identification process, marking the ratio of the number of identification processes of the analysis object in the allocation cycle to the number of fault diagnosis processes as the identification data SS; obtaining the evaluation coefficient PG of the analysis object using the matching model for fault diagnosis analysis in the allocation cycle through the formula PG=k1×XS-k2×SS, the evaluation coefficient PG is a value reflecting the overall effect of the analysis object using the matching model for fault diagnosis in the allocation cycle, the smaller the value of the evaluation coefficient PG, the better the overall effect of the analysis object using the matching model for fault diagnosis in the allocation cycle; Among them, k1 and k2 are both proportional coefficients, and k1>k2>1. The fault diagnosis and treatment effects of the oil and water well fields are evaluated and analyzed, and the multiple effect parameters of all fault diagnosis processes of the analysis object within the allocation period are comprehensively analyzed and calculated to obtain the evaluation coefficient. The evaluation coefficient is used to evaluate and provide feedback on the effect of fault diagnosis analysis of the analysis object using the matching model within the allocation period.

[0024] The allocation optimization module is used to optimize and analyze the fault diagnosis model allocation process of the oil and water well field: generate an optimization cycle, and at the beginning of the optimization cycle, the management personnel select the allocation mode, which includes efficiency priority mode, identification priority mode and comprehensive priority mode.

[0025] The specific process of using the efficiency priority mode to allocate the fault diagnosis model includes: marking the matching model of the analysis object with the smallest efficiency data XS in the allocation cycle as the efficiency priority model, and using the efficiency priority model to perform fault diagnosis analysis in the optimization cycle; The specific process of using the recognition priority mode to allocate the fault diagnosis model includes: marking the matching model of the analysis object with the largest recognition data SS in the allocation cycle as the recognition priority model, and using the recognition priority model to perform fault diagnosis analysis in the optimization cycle; The specific process of using the comprehensive priority mode to allocate the fault diagnosis model includes: the maximum and minimum temperature values ​​of all analysis objects obtained at the beginning of the allocation cycle constitute a temperature range, and the temperature range is divided into a number of temperature intervals; the maximum and minimum pressure values ​​of all analysis objects obtained at the beginning of the allocation cycle constitute a pressure range, and the pressure range is divided into a number of pressure intervals; and the temperature intervals and pressure intervals are arbitrarily combined to obtain a number of environmental data groups; Mark the analysis objects whose temperature values ​​and pressure values ​​are respectively within the temperature interval and pressure interval in the environmental data group as the conforming objects of the environmental data group, and mark the matching model of the conforming object with the smallest evaluation coefficient PG in the environmental data group as the conforming model of the environmental data group; At the beginning of the optimization cycle, the temperature value and pressure value of the analysis object are obtained and marked as temperature analysis value and pressure analysis value respectively; the corresponding environmental data group is searched through the temperature analysis value and pressure analysis value and marked as a conforming data group; the conforming model of the conforming data group is marked as a comprehensive priority model; and the comprehensive priority model is used to perform fault diagnosis analysis within the optimization cycle; The fault diagnosis model allocation process of oil and water well sites is optimized and analyzed. Users can select the allocation mode according to their own needs. The fault diagnosis models obtained by different allocation modes have different diagnostic tendencies. The different needs of different users are solved in a multi-mode allocation manner to improve user satisfaction. At the same time, an optimal allocation method that matches the environmental parameters of the oil and water well site is provided through a comprehensive priority mode.

[0026] Embodiment 2: Figure 2 As shown, the present invention also proposes an oil and water well field fault diagnosis method based on abnormal feature recognition, comprising the following steps: Step 1: Perform distribution analysis on the fault diagnosis model of the oil and water well field: generate a distribution cycle, mark the oil and water well field as the analysis object, obtain the environmental parameters of the analysis object at the beginning of the distribution cycle, randomly assign the fault diagnosis model to the analysis object and mark it as the matching model of the analysis object; Step 2: Evaluate and analyze the fault diagnosis and treatment effect of the oil and water well field: obtain the efficiency data XS and identification data SS of the analysis object in the allocation period and perform numerical calculation to obtain the evaluation coefficient PG of the analysis object in the allocation period; Step 3: Generate an optimization cycle. At the beginning of the optimization cycle, the manager selects an allocation mode and allocates the fault diagnosis model to the analysis object according to the selected allocation mode.

[0027] The oil and water well field fault diagnosis system and method based on abnormal feature recognition, when working, generates an allocation cycle, marks the oil and water well field as an analysis object, obtains the environmental parameters of the analysis object at the beginning of the allocation cycle, randomly allocates a fault diagnosis model to the analysis object and marks it as a matching model for the analysis object; obtains the efficiency data XS and identification data SS of the analysis object in the allocation cycle and performs numerical calculations to obtain the evaluation coefficient PG of the analysis object in the allocation cycle; generates an optimization cycle, and at the beginning of the optimization cycle, the management personnel selects the allocation mode, and allocates the fault diagnosis model to the analysis object according to the selected allocation mode.

[0028] The above formulas are obtained by collecting a large amount of data for software simulation and selecting a formula close to the real value. The coefficients in the formula are set by technicians in this field according to the actual situation; for example: formula PG=k1×XS-k2×SS; technicians in this field collect multiple groups of sample data and set corresponding evaluation coefficients for each group of sample data; substitute the set evaluation coefficients and the collected sample data into the formula, any two formulas constitute a set of two-variable linear equations, screen the calculated coefficients and take the average, and obtain the values ​​of k1 and k2 as 3.52 and 2.19 respectively; The size of the coefficient is to quantify each parameter to obtain a specific value for subsequent comparison. The size of the coefficient depends on the amount of sample data and the preliminary setting of the corresponding evaluation coefficient for each set of sample data by technical personnel in this field; as long as it does not affect the proportional relationship between the parameter and the quantified value, such as the evaluation coefficient is proportional to the value of the efficiency data.

[0029] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

Claims

1. Oil and water well field fault diagnosis system based on abnormal feature recognition, characterized in that: It includes a model allocation module, a diagnosis and evaluation module, and an allocation optimization module; The model allocation module is used to allocate and analyze the fault diagnosis model of the oil and water well field: generate an allocation cycle, mark the oil and water well field as an analysis object, obtain environmental parameters of the analysis object at the beginning of the allocation cycle, randomly allocate a fault diagnosis model to the analysis object and mark it as a matching model of the analysis object, and use the matching model to perform fault diagnosis analysis for the analysis object within the allocation cycle; The diagnostic evaluation module is used to evaluate and analyze the fault diagnosis and processing effect of the oil and water well field: obtain the efficiency data XS and identification data SS of the analysis object in the allocation cycle and perform numerical calculations to obtain the evaluation coefficient PG of the fault diagnosis analysis of the analysis object using the matching model in the allocation cycle; The allocation optimization module is used to optimize and analyze the fault diagnosis model allocation process of the oil and water well field: generate an optimization cycle, and at the beginning of the optimization cycle, the management personnel selects an allocation mode, which includes an efficiency priority mode, an identification priority mode, and a comprehensive priority mode.

2. The oil and water well site fault diagnosis system based on abnormal feature recognition according to claim 1 is characterized in that: The environmental parameters of the analysis object include temperature and pressure values; the fault diagnosis model includes a current comparison model, a machine learning model, a time series analysis model and a SVM model.

3. The oil and water well site fault diagnosis system based on abnormal feature recognition according to claim 2 is characterized in that: The process of obtaining the efficiency data XS includes: when an abnormality of the analysis object is monitored, fault diagnosis is performed through the matching model, and the start time of the fault diagnosis is marked as the diagnosis start time. The time when the matching model completes the fault type classification or determines that the fault type cannot be identified is marked as the diagnosis end time. The diagnosis start time and the diagnosis end time constitute a fault diagnosis process, and the average duration of all fault diagnosis processes of the analysis object within the allocation period is marked as the efficiency data XS.

4. The oil and water well site fault diagnosis system based on abnormal feature recognition according to claim 3 is characterized in that: The process of obtaining the identification data SS includes: marking the fault diagnosis process that completes the fault type classification as the identification process, and marking the ratio of the number of identification processes of the analysis object within the allocation cycle to the number of fault diagnosis processes as the identification data SS.

5. The oil and water well site fault diagnosis system based on abnormal feature recognition according to claim 4 is characterized in that: The specific process of using the efficiency priority mode to allocate fault diagnosis models includes: marking the matching model of the analysis object with the smallest efficiency data XS in the allocation cycle as the efficiency priority model, and using the efficiency priority model to perform fault diagnosis analysis in the optimization cycle.

6. The oil and water well site fault diagnosis system based on abnormal feature recognition according to claim 5 is characterized in that: The specific process of using the recognition priority mode to allocate the fault diagnosis model includes: marking the matching model of the analysis object with the largest recognition data SS in the allocation cycle as the recognition priority model, and using the recognition priority model to perform fault diagnosis analysis in the optimization cycle.

7. The oil and water well site fault diagnosis system based on abnormal feature recognition according to claim 6 is characterized in that: The specific process of using the comprehensive priority mode to assign fault diagnosis models includes: generating several environmental data groups, marking the analysis objects whose temperature values ​​and pressure values ​​are respectively within the temperature range and pressure range in the environmental data group as compliant objects of the environmental data group, and marking the matching model of the compliant object with the smallest evaluation coefficient PG in the environmental data group as a compliant model of the environmental data group; obtaining the temperature value and pressure value of the analysis object at the beginning of the optimization cycle and marking them as temperature analysis values ​​and pressure analysis values ​​respectively, searching for the corresponding environmental data group through the temperature analysis value and pressure analysis value and marking it as a compliant data group, marking the compliant model of the compliant data group as a comprehensive priority model, and using the comprehensive priority model for fault diagnosis analysis within the optimization cycle.

8. The oil and water well site fault diagnosis system based on abnormal feature recognition according to claim 7 is characterized in that: The process of generating an environmental data group includes: forming a temperature range from the maximum and minimum temperature values ​​of all analysis objects obtained at the start of the allocation cycle, dividing the temperature range into several temperature intervals; forming a pressure range from the maximum and minimum pressure values ​​of all analysis objects obtained at the start of the allocation cycle, dividing the pressure range into several pressure intervals; and arbitrarily combining temperature intervals and pressure intervals to obtain several environmental data groups.

9. The oil and water well field fault diagnosis method based on abnormal feature recognition is characterized by: The following steps are involved: Step 1: Perform distribution analysis on the fault diagnosis model of the oil and water well field: generate a distribution cycle, mark the oil and water well field as the analysis object, obtain the environmental parameters of the analysis object at the beginning of the distribution cycle, randomly assign the fault diagnosis model to the analysis object and mark it as the matching model of the analysis object; Step 2: Evaluate and analyze the fault diagnosis and treatment effect of the oil and water well field: obtain the efficiency data XS and identification data SS of the analysis object in the allocation period and perform numerical calculation to obtain the evaluation coefficient PG of the analysis object in the allocation period; Step 3: Generate an optimization cycle. At the beginning of the optimization cycle, the manager selects an allocation mode and allocates the fault diagnosis model to the analysis object according to the selected allocation mode.

Images