Method and system for predicting equipment failures in radiation monitoring stations based on small sample training models

Abstract

The present invention relates to the field of fault prediction technology, specifically to a method and system for predicting equipment faults of a radiation monitoring station based on a small sample training model, comprising the following steps: based on the operating environment of the radiation monitoring station, collecting historical data of radiation detectors to obtain historical archives of radiation monitoring equipment. In the present invention, a small sample learning model is used to effectively utilize limited data resources to build a fault prediction model, thereby improving the accuracy of the prediction. Through continuous monitoring of radiation intensity and equipment operating status, data patterns matching known fault patterns can be identified in a timely manner, and warnings can be issued in a timely manner before the problem becomes serious. Through quantitative evaluation of the probability of failure, real-time setting and adjustment of risk levels are performed, making maintenance more targeted, effectively extending the service life of the equipment and reducing operating costs. By dynamically adjusting the fault detection threshold, matching the current operating status and environment, the response speed and sensitivity are optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of fault prediction, and in particular to a method and system for predicting faults of radiation monitoring station equipment based on a small sample training model. Background Art

[0002] The field of fault prediction technology involves monitoring the health status of equipment, aiming to pre-identify signals or anomalies that lead to system failures. By analyzing historical and real-time data, applying machine learning and data analysis methods, and establishing predictive models to predict equipment failures. The core advantage of fault prediction technology is that it can significantly reduce unexpected downtime and maintenance costs, and improve system reliability and efficiency. In practical applications, fault prediction technology is widely used in heavy industry, transportation, power systems, and high-tech equipment. By monitoring equipment performance parameters in real time, the fault prediction system can issue warnings before the problem becomes serious, assisting managers to carry out timely maintenance or replacement, thereby ensuring the continuous operation of equipment and systems.

[0003] Among them, the method of predicting equipment failures in radiation monitoring stations based on small sample training models focuses on using machine learning technology to predict equipment failures in radiation monitoring stations through small sample learning models. This method is particularly suitable for scenarios with few data samples but high-precision predictions. Radiation monitoring station equipment involves high risks and high safety standards, and failures can lead to serious health and environmental consequences. By using a small sample training model, this method can effectively predict failures with the support of less historical failure data, greatly improving the early warning capability and response speed of the monitoring system. The main purpose of this technology is to ensure that the equipment is in optimal operating condition before radiation leakage or other critical performance problems occur, thereby protecting personnel safety and environmental health.

[0004] Traditional fault prediction methods rely on a large amount of historical data and complex machine learning models. They are not efficient or accurate enough when the amount of data is insufficient or the types of equipment are wide. Traditional methods lack sufficient flexibility when dealing with real-time data and rapidly changing environmental conditions. They cannot adjust the prediction model in time to adapt to new operating conditions, resulting in excessively high false alarm and missed alarm rates. The static risk assessment of traditional methods ignores the impact of equipment aging and maintenance quality, and cannot accurately reflect the actual operating risks of equipment, affecting the timeliness and effectiveness of maintenance decisions. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings existing in the prior art and propose a method and system for predicting radiation monitoring station equipment failure based on a small sample training model.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: a method for predicting equipment failures of radiation monitoring stations based on a small sample training model, comprising the following steps:

[0007] S1: Based on the operating environment of the radiation monitoring station, the historical data of the radiation detectors are collected to obtain the historical archives of the radiation monitoring equipment;

[0008] S2: Based on the historical archives of the radiation monitoring equipment, a small sample learning framework is trained, and by analyzing the radiation threshold and radiation safety standard associated data, a fault type classification prediction model matching the radiation monitoring station equipment is constructed to obtain an equipment fault prediction model;

[0009] S3: Based on the equipment failure prediction model, combined with the real-time monitored radiation intensity and the state and operation data of the radiation monitoring equipment, the matching degree between the data and the known failure mode is analyzed to obtain the predicted failure matching result;

[0010] S4: Based on the predicted fault matching result, the probability of the predicted fault under the current operating conditions is evaluated to obtain the monitoring equipment failure probability information;

[0011] S5: Based on the failure probability information of the monitoring equipment, the risk level is set and sent to the fault management device of the radiation monitoring station to generate a radiation risk management update result;

[0012] S6: Based on the radiation risk management update result, the detection threshold is adjusted to match the current operating state and radiation environment of the radiation monitoring equipment to obtain a fault threshold adjustment plan.

[0013] As a further solution of the present invention, the radiation monitoring equipment history archive includes the usage time, maintenance times and fault records of the radiation monitoring equipment, the equipment fault prediction model includes radiation intensity abnormality, equipment temperature fluctuation and power supply interruption fault categories, the predicted fault matching result is specifically a matching score between the current equipment state and multiple fault types in the prediction model, the monitoring equipment failure probability information includes the probability value of the fault type, potential impact assessment and emergency response level, the radiation risk management update result includes the risk level, matched maintenance measures and scheduled equipment inspection time point, and the fault threshold adjustment scheme includes adjusted detection threshold parameters and detection threshold adjustment frequency.

[0014] As a further solution of the present invention, based on the historical archives of the radiation monitoring equipment, a small sample learning framework training is performed, and by analyzing the radiation threshold and radiation safety standard associated data, a fault type classification prediction model matching the radiation monitoring station equipment is constructed. The steps of obtaining the equipment fault prediction model are specifically as follows:

[0015] S201: extracting key influencing factors based on the historical archive of the radiation monitoring equipment, the key influencing factors including the failure time exceeding a preset threshold and the maintenance activity record, combining with the abnormal radiation reading data, screening the key parameters causing the failure, and obtaining the influencing factor screening result;

[0016] S202: Based on the influencing factor screening result, a small sample learning framework is trained through influencing factor training, a training cycle and a parameter adjustment strategy are set, and the model is trained and adjusted to match the fault prediction under the target environment of the radiation monitoring station, so as to obtain a training optimization model;

[0017] S203: Based on the training optimization model, the parameter optimization model is verified through the historical archives of the radiation monitoring equipment to verify the accuracy of the training optimization model in predicting equipment fault classification, and the model parameters are adjusted to optimize the accuracy of the model in predicting fault classification, thereby generating an equipment fault prediction model.

[0018] As a further solution of the present invention, based on the equipment fault prediction model, combined with the real-time monitored radiation intensity and the state and operation data of the radiation monitoring equipment, the matching degree between the analysis data and the known fault mode is analyzed, and the steps of obtaining the predicted fault matching result are specifically as follows:

[0019] S301: Based on the equipment fault prediction model, traverse the radiation intensity recorded in the real-time monitoring data, compare it with the historical intensity data recorded in the fault prediction model, calculate the difference and abnormal points of the radiation intensity, and generate a radiation intensity difference analysis result;

[0020] S302: Based on the radiation intensity difference analysis result, extract the operating temperature and pressure data of the radiation monitoring equipment, evaluate the deviation between the parameters and the typical values ​​in the fault prediction model, and obtain the operating parameter deviation analysis result;

[0021] S303: Based on the operation parameter deviation analysis result, the fault mode in the fault prediction model is compared, the fit between the current data mode and the known fault is analyzed, the expected fault type is identified, and the predicted fault matching result is obtained.

[0022] As a further solution of the present invention, based on the predicted fault matching result, the probability of the predicted fault under the current operating conditions is evaluated to obtain the monitoring equipment failure probability information in the following steps:

[0023] S401: Based on the predicted fault matching result, aging data and historical maintenance quality records of the radiation monitoring equipment are collected, the potential impact of the aging and maintenance data on the current fault diagnosis is analyzed, and an aging and maintenance impact assessment result is generated;

[0024] S402: Based on the aging and maintenance impact assessment results, collect current environmental radiation, temperature and humidity data, assess the impact of the current environmental data on the performance of the radiation monitoring equipment, and obtain environmental impact analysis results;

[0025] S403: Based on the environmental impact analysis results, the predicted fault data, environmental data and radiation monitoring equipment status information are integrated, and the predicted failure probability of the radiation monitoring equipment is quantitatively analyzed through the Bayesian network algorithm to evaluate the probability of the predicted failure under the current operating conditions to obtain the monitoring equipment failure probability information.

[0026] As a further solution of the present invention, the Bayesian network algorithm is according to the formula:

[0027] ;

[0028] Calculate the failure probability value, where, For environmental data, is the device status, To maintain records, For fault prediction, For a given environment data , Device Status and maintenance records Under the condition of The posterior probability of occurrence, To predict failure Occurs, considering the device status Under the condition of environmental data The probability of For existing maintenance records Equipment status under the condition The probability of Consider maintaining records Failure prediction in the case of The probability of occurrence, To maintain the probability of record occurrence, is the marginal probability of the environmental data, is the marginal probability of the device state.

[0029] As a further solution of the present invention, based on the failure probability information of the monitoring equipment, the risk level is set and sent to the fault management device of the radiation monitoring station, and the steps of generating the radiation risk management update result are specifically as follows:

[0030] S501: Based on the monitoring equipment failure probability information, risk levels are set according to the occurrence probability and potential impact of each predicted failure category, and low risk, medium risk and high risk levels are set respectively, and a risk level assessment table is generated;

[0031] S502: Based on the risk level assessment table, a matching maintenance strategy is formulated for each risk level, the maintenance strategy including regular inspection cycle adjustment, preventive maintenance measures and emergency repair measures, to obtain a maintenance strategy solution;

[0032] S503: Based on the maintenance strategy, the solution is sent to the fault management device of the radiation monitoring station, and the fault management device automatically updates the fault management and preventive measures settings according to the maintenance strategy to obtain a radiation risk management update result.

[0033] As a further solution of the present invention, based on the radiation risk management update result, the detection threshold is adjusted to match the current operating state and radiation environment of the radiation monitoring equipment, and the steps of obtaining the fault threshold adjustment solution are specifically as follows:

[0034] S601: Based on the radiation risk management update result, organize and analyze the performance monitoring results and historical radiation data of the radiation monitoring equipment, evaluate the adaptability of the current state of the radiation monitoring equipment to the environmental conditions, and generate equipment state analysis results;

[0035] S602: Based on the device status analysis result, identify the fault detection threshold that needs to be adjusted in the current operating environment, update the fault detection threshold, set matching adjustment frequency and conditions, and generate a basic adjustment plan;

[0036] S603: Based on the basic adjustment scheme, simulate the adjustment effect of the fault detection threshold, test the effect of the new threshold under simulated operating conditions, analyze the accuracy and responsiveness of fault prediction after the threshold adjustment, optimize the fault detection threshold, and generate a fault threshold adjustment scheme.

[0037] A system for predicting equipment failures in a radiation monitoring station based on a small sample training model, wherein the system for predicting equipment failures in a radiation monitoring station based on a small sample training model is used to execute the method for predicting equipment failures in a radiation monitoring station based on a small sample training model, and the system comprises:

[0038] The data collection module captures the reading records of radiation detectors based on the operating environment of the radiation monitoring station, combines the equipment maintenance logs and failure time records, integrates the data, and generates historical data records;

[0039] The model training module extracts key influencing factors based on the historical data records, performs small sample learning framework training, trains a fault classification prediction model that matches the target environment of the radiation monitoring station, and obtains an equipment fault prediction model;

[0040] The data analysis module analyzes the monitoring data in real time based on the equipment failure prediction model, analyzes the matching degree between the data and the known failure mode, and generates the predicted failure analysis result;

[0041] The risk assessment module evaluates the fault risk based on the predicted fault analysis results and the current environmental conditions, calculates the probability of the predicted fault, identifies the probability values ​​of multiple types of predicted faults, sets the risk level, and generates a risk level assessment result;

[0042] The threshold adjustment module formulates matching maintenance measures based on the risk level assessment results, instantly updates the fault management equipment of the radiation monitoring station, analyzes the performance changes and environmental adaptability of the radiation monitoring equipment, adjusts the fault detection threshold, matches the current radiation monitoring equipment status, and obtains a fault threshold adjustment plan.

[0043] Compared with the prior art, the advantages and positive effects of the present invention are:

[0044] In the present invention, limited data resources are effectively utilized through a small sample learning model to build a fault prediction model, thereby improving the accuracy of the prediction. Through continuous monitoring of radiation intensity and equipment operating status, data patterns matching known fault patterns can be identified in a timely manner, and warnings can be issued in a timely manner before the problem becomes serious. Through quantitative evaluation of the probability of failure, real-time setting and adjustment of risk levels are performed, making maintenance more targeted, effectively extending the service life of the equipment and reducing operating costs. By dynamically adjusting the fault detection threshold to match the current operating status and environment, the response speed and sensitivity are optimized. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0046] Figure 2 This is a detailed flow chart of S1 of the present invention;

[0047] Figure 3 This is a detailed flow chart of S2 of the present invention;

[0048] Figure 4 This is a detailed flow chart of S3 of the present invention;

[0049] Figure 5 This is a detailed flow chart of S4 of the present invention;

[0050] Figure 6 This is a detailed flow chart of S5 of the present invention;

[0051] Figure 7 This is a detailed flow chart of S6 of the present invention;

[0052] Figure 8 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0054] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0055] Embodiment 1

[0056] See also Figure 1 The present invention provides a technical solution: a method for predicting equipment failures of radiation monitoring stations based on a small sample training model, comprising the following steps:

[0057] S1: Based on the operating environment of the radiation monitoring station, historical data of the radiation detectors are collected, including radiation reading frequency, equipment maintenance logs and failure time records, capturing the historical operation records of the radiation monitoring equipment to obtain the historical archives of the radiation monitoring equipment;

[0058] S2: Based on the historical archives of radiation monitoring equipment, a small sample learning framework is trained. By analyzing the radiation threshold and radiation safety standard associated data, a fault type classification prediction model matching the radiation monitoring station equipment is constructed to obtain an equipment failure prediction model.

[0059] S3: Based on the equipment failure prediction model, combined with the real-time monitored radiation intensity and the status and operation data of the radiation monitoring equipment, data comparison and analysis are performed. The data of radiation intensity and the operating temperature of the radiation monitoring equipment are analyzed to determine the matching degree between the analysis data and the known failure modes, and the predicted failure matching results are obtained.

[0060] S4: Based on the predicted fault matching results, combined with the radiation monitoring equipment aging, historical maintenance quality and current radiation environment status, the probability of the predicted fault under the current operating conditions is evaluated to obtain the monitoring equipment failure probability information;

[0061] S5: Based on the failure probability information of the monitoring equipment, risk level is set, matching maintenance measures are formulated, and sent to the fault management equipment of the radiation monitoring station, the fault management equipment status of the radiation monitoring station is updated in real time, and the radiation risk management update results are generated;

[0062] S6: Based on the updated results of radiation risk management, the fault detection threshold of the radiation monitoring equipment is adjusted. By using historical data analysis and the monitoring results of the performance changes of the radiation monitoring equipment, the detection threshold is adjusted to match the current operating status and radiation environment of the radiation monitoring equipment to obtain a fault threshold adjustment plan.

[0063] The historical archives of radiation monitoring equipment include the usage time, maintenance times and fault records of the radiation monitoring equipment. The equipment failure prediction model includes the fault categories of abnormal radiation intensity, equipment temperature fluctuation and power supply interruption. The predicted fault matching results are specifically the matching scores between the current equipment status and the various fault types in the prediction model. The monitoring equipment failure probability information includes the probability value of the fault type, potential impact assessment and emergency response level. The radiation risk management update results include the risk level, matched maintenance measures and scheduled equipment inspection time points. The fault threshold adjustment plan includes the adjusted detection threshold parameters and the detection threshold adjustment frequency.

[0064] See also Figure 2 Based on the operating environment of the radiation monitoring station, the historical data of the radiation detector is collected, including the radiation reading frequency, equipment maintenance log and failure time record, and the historical operation record of the radiation monitoring equipment is captured. The specific steps for obtaining the historical archive of the radiation monitoring equipment are as follows:

[0065] S101: Based on the operating environment of the radiation monitoring station, obtain the reading record of the radiation detector, record the date, time and reading result of each monitoring, and obtain the specific process of the reading frequency record as follows;

[0066] S101: Based on the operating environment of the radiation monitoring station, the radiation detector is configured in a timing mode to automatically record data once every hour. The detector transmits the data to the central database after each monitoring, and each record in the database will contain a timestamp, reading, and device identification. In order to ensure the accuracy of the data, the preliminary screening formula "effective reading = actual reading - average error" is applied, where the average error is calculated based on the standard deviation of the readings at the same time point in the past week to obtain the reading frequency record.

[0067] S102: Based on the reading frequency record, analyze the maintenance log of the radiation monitoring equipment, record the date, operator and operation content of each maintenance, integrate the data to evaluate the frequency and type of maintenance, and obtain the specific process of the maintenance log analysis result;

[0068] S102: Based on the reading frequency record, analyze the maintenance log of the radiation monitoring equipment, record the date, operator and operation content of each maintenance, integrate data to evaluate the frequency and type of maintenance, and extract all maintenance records in the equipment maintenance database, including maintenance date, maintenance type and maintenance person in charge. The impact evaluation of each maintenance is calculated by the formula "maintenance effect = equipment effectiveness after maintenance / equipment effectiveness before maintenance", where equipment effectiveness is based on the average effective reading of the equipment within one week before and after maintenance, and the maintenance log analysis results are obtained.

[0069] S103: Based on the maintenance log analysis results and the reading frequency records, the specific process of summarizing the fault time records of the radiation monitoring equipment, including the date and time of the fault, the post-fault processing, analyzing the trend and frequency of the fault, and obtaining the historical archives of the radiation monitoring equipment is as follows;

[0070] S103: Based on the maintenance log analysis results and reading frequency records, summarize the fault time records of the radiation monitoring equipment, including the date and time of the fault and the post-fault processing, analyze the trend and frequency of faults, and calculate the equipment fault interval time through the formula "fault interval time = current fault date - last fault date" to analyze the periodicity of faults. Each fault event records in detail the specific time of the fault, the affected equipment part and the repair measures taken, and obtains the historical archive of the radiation monitoring equipment.

[0071] See also Figure 3 Based on the historical archives of radiation monitoring equipment, a small sample learning framework is trained. By analyzing the radiation threshold and radiation safety standard associated data, a fault type classification prediction model matching the radiation monitoring station equipment is constructed. The specific steps to obtain the equipment fault prediction model are as follows:

[0072] S201: Based on the historical archives of radiation monitoring equipment, extract key influencing factors, which include failure time exceeding preset thresholds and maintenance activity records, and combine with abnormal radiation reading data to screen key parameters that cause failures. The specific process for obtaining the influencing factor screening results is as follows;

[0073] S201: Based on the historical archives of radiation monitoring equipment, extract key influencing factors, including failure time and maintenance activity records exceeding the preset threshold, combined with abnormal radiation reading data, screen the key parameters that cause failures. The formula "screening factor = abnormal reading frequency / standard reading frequency" is applied during the screening process, where the standard reading frequency is the average reading frequency based on long-term monitoring data. By comparing the maintenance activities in the maintenance records with the equipment failure records, determine the key maintenance operations that affect the performance of the equipment. All failure events that exceed the preset threshold are recorded and analyzed to identify the most common failure types and related influencing factors, and obtain the influencing factor screening results.

[0074] S202: Based on the influencing factor screening results, the influencing factor training small sample learning framework is used to set the training cycle and parameter adjustment strategy, and the model is trained and adjusted to match the fault prediction under the target environment of the radiation monitoring station. The specific process of obtaining the training optimization model is as follows;

[0075] S202: Based on the results of the influencing factor screening, the influencing factor training small sample learning framework is used to set the training cycle and parameter adjustment strategy, and the model is trained and adjusted to match the fault prediction under the target environment of the radiation monitoring station. During the model training process, the calculation method of "model error = actual fault data - predicted fault data" is used to evaluate the model performance. After each iteration, the learning rate is adjusted according to the model error to optimize the prediction ability of the model. The training cycle is set to dynamically adjust according to the frequency of changes in historical data to ensure that the model can adapt to environmental changes and obtain a trained optimization model.

[0076] S203: Based on the training optimization model, the parameter optimization model is verified through the historical archives of the radiation monitoring equipment to verify the accuracy of the training optimization model in predicting the equipment fault classification, and the model parameters are adjusted to optimize the accuracy of the model in predicting the fault classification. The specific process of generating the equipment fault prediction model is as follows;

[0077] S203: Based on the training optimization model, the parameter optimization model is verified through the historical archives of the radiation monitoring equipment to verify the accuracy of the training optimization model in predicting the equipment fault classification, and the model parameters are adjusted to optimize the accuracy of the model in predicting the fault classification. In the verification phase, the formula "accuracy = number of correctly predicted fault categories / total number of fault categories" is applied to evaluate the classification accuracy of the model. According to the accuracy results, the weights of the influencing factors and the structural parameters of the model, such as the number of nodes in the hidden layer and the type of activation function, are adjusted to improve the adaptability and prediction accuracy of the model to complex fault environments, and to generate an equipment fault prediction model.

[0078] See also Figure 4 ,Based on the equipment fault prediction model, combined with the real-time monitored radiation intensity and radiation monitoring equipment status and operation data, data comparison and analysis are carried out, the data of radiation intensity and radiation monitoring equipment operating temperature are analyzed, and the matching degree between the analysis data and the known fault mode is analyzed. The specific steps for obtaining the predicted fault matching results are as follows:

[0079] S301: Based on the equipment fault prediction model, the radiation intensity recorded in the real-time monitoring data is traversed, compared with the historical intensity data recorded in the fault prediction model, and the difference and abnormal points of the radiation intensity are calculated. The specific process of generating the radiation intensity difference analysis result is as follows;

[0080] S301: Based on the equipment fault prediction model, traverse the radiation intensity recorded in the real-time monitoring data, compare it with the historical intensity data recorded in the fault prediction model, calculate the difference and abnormal points of radiation intensity, compare the radiation intensity of each real-time monitoring point with the historical average intensity at the corresponding time point, use the formula "intensity difference = real-time intensity - historical average intensity", apply the threshold analysis method according to the difference result, where the threshold is set to twice the standard deviation of the historical intensity, so as to screen out abnormal points that exceed the normal fluctuation range, and the abnormal points are marked and recorded in the difference analysis report to generate the radiation intensity difference analysis results.

[0081] S302: Based on the radiation intensity difference analysis results, the operating temperature and pressure data of the radiation monitoring equipment are extracted, and the deviation between the parameters and the typical values ​​in the fault prediction model is evaluated. The specific process of obtaining the operating parameter deviation analysis results is as follows;

[0082] S302: Based on the results of the radiation intensity difference analysis, the operating temperature and pressure data of the radiation monitoring equipment are extracted to evaluate the degree of deviation between the parameters and the typical values ​​in the fault prediction model. The degree of deviation of the operating parameters is calculated by the formula "parameter deviation = (measured temperature - model typical temperature) / model typical temperature", and the deviation of pressure is calculated in the same way. By statistically analyzing the degree of deviation of all operating parameters, possible abnormal trends related to faults can be found. The trends are recorded in detail in the operating parameter deviation report, which explains the specific circumstances and potential impacts of the deviation of each parameter from the typical value, and obtains the results of the operating parameter deviation analysis.

[0083] S303: Based on the operation parameter deviation analysis result, the specific process of comparing the fault mode in the fault prediction model, analyzing the fit between the current data mode and the known fault, identifying the expected fault type, and obtaining the predicted fault matching result is as follows;

[0084] S303: Based on the analysis results of operating parameter deviation, the fault mode in the fault prediction model is compared, the fit between the current data mode and the known faults is analyzed, the expected fault type is identified, and the consistency between the current monitoring data and the known fault mode is evaluated by applying the formula "matching degree = number of parameters that meet the fault mode / total number of detected parameters". The most likely fault type can be effectively identified, and this information can be integrated into the fault prediction report to obtain the predicted fault matching result.

[0085] See also Figure 5 Based on the predicted fault matching results, combined with the radiation monitoring equipment aging, historical maintenance quality and current radiation environment status, the statistical analysis method is used to evaluate the probability of predicted faults under current operating conditions. The specific steps to obtain the monitoring equipment failure probability information are as follows:

[0086] S401: Based on the predicted fault matching results, the aging data and historical maintenance quality records of the radiation monitoring equipment are collected, the potential impact of the aging and maintenance data on the current fault diagnosis is analyzed, and the specific process of generating the aging and maintenance impact assessment results is as follows;

[0087] S401: Based on the predicted fault matching results, collect the aging data and historical maintenance quality records of the radiation monitoring equipment, analyze the potential impact of aging and maintenance data on the current fault diagnosis, and calculate the formula "impact factor = (equipment age factor + maintenance quality factor) / 2", where the equipment age factor is determined by the relationship between the equipment's service life and the frequency of faults, and the maintenance quality factor is evaluated based on the degree of improvement or deterioration of equipment performance reflected in the maintenance records. Through the comprehensive evaluation of these two factors, the degree of impact of aging and maintenance on the current equipment performance can be quantitatively revealed to help more accurately diagnose the cause of the fault and generate aging and maintenance impact assessment results.

[0088] S402: Based on the aging and maintenance impact assessment results, collect the current environmental radiation, temperature and humidity data, evaluate the impact of the current environmental data on the performance of the radiation monitoring equipment, and obtain the specific process of the environmental impact analysis results as follows;

[0089] S402: Based on the results of aging and maintenance impact assessment, collect current environmental radiation, temperature and humidity data, evaluate the impact of current environmental data on the performance of radiation monitoring equipment, and use the formula "environmental impact index = (environmental radiation impact + temperature impact + humidity impact) / 3", where the impact of each environmental factor is determined by correlation analysis with equipment performance parameters. The analysis results will show how environmental conditions affect equipment operation alone or together, provide a scientific basis for equipment maintenance and adjustment, and obtain environmental impact analysis results;

[0090] S403: Based on the environmental impact analysis results, the predicted fault data, environmental data and radiation monitoring equipment status information are integrated, and the predicted fault probability of the radiation monitoring equipment is quantitatively analyzed through the Bayesian network algorithm to evaluate the probability of the predicted fault under the current operating conditions. The specific process of obtaining the monitoring equipment failure probability information is as follows;

[0091] S403: Based on the results of environmental impact analysis, the predicted fault data, environmental data and radiation monitoring equipment status information are integrated, and the predicted fault probability of the radiation monitoring equipment is quantitatively analyzed through the Bayesian network algorithm, and the probability of predicted faults under current operating conditions is evaluated. Through the Bayesian network comprehensive analysis, the specific contribution of the current equipment status, historical fault data and environmental factors to the probability of fault occurrence is quantified. This makes the fault prediction not only rely on historical and status data, but also takes into account the real-time impact of the environment, and obtains the monitoring equipment failure probability information.

[0092] Bayesian network algorithm, according to the formula:

[0093] ;

[0094] Calculate the failure probability value, where, For environmental data, is the device status, To maintain records, For fault prediction, For a given environment data , Device Status and maintenance records Under the condition of The posterior probability of occurrence, To predict failure Occurs, considering the device status Under the condition of environmental data The probability of For existing maintenance records Equipment status under the condition The probability of Consider maintaining records Failure prediction in the case of The probability of occurrence, To maintain the probability of record occurrence, is the marginal probability of the environmental data, is the marginal probability of the device state.

[0095] The specific execution process of the formula is as follows:

[0096] Collecting environmental data , device status , and maintenance records Now, use historical failure data and maintenance logs to calculate and , and According to the frequency analysis of historical records, The data is obtained by analyzing the performance change data after equipment maintenance, reflecting the statistical relationship between maintenance activities and equipment performance. The weight coefficient of each parameter is calculated by the frequency of occurrence in historical data and its importance to system performance, combined with real-time monitoring data update and , and Through real-time data collection and instant analysis, ensure the timeliness of data and synthesize according to Bayesian rules , calculate the failure probability under current conditions.

[0097] See also Figure 6 Based on the failure probability information of the monitoring equipment, the risk level is set, the matching maintenance measures are formulated, and sent to the fault management device of the radiation monitoring station, and the fault management device status of the radiation monitoring station is updated in real time. The specific steps for generating the radiation risk management update results are as follows:

[0098] S501: Based on the monitoring equipment failure probability information, risk level setting is performed according to the occurrence probability and potential impact of each predicted failure category, and low risk, medium risk and high risk levels are set respectively. The specific process of generating a risk level assessment table is as follows;

[0099] S501: Based on the monitoring equipment failure probability information, risk levels are set according to the probability of occurrence and potential impact of each predicted failure category, and low, medium and high risk levels are set respectively. Rating is performed according to the failure probability and the degree of impact caused by the failure, using the formula "risk rating = failure probability × impact severity", where the impact severity is scored according to the potential risks of economic losses, safety accidents and operation interruptions. The rating results will be classified into three risk levels: low, medium and high. Each level lists the corresponding failure type and expected response measures in detail to generate a risk level assessment table.

[0100] S502: Based on the risk level assessment table, a matching maintenance strategy is formulated for each risk level. The maintenance strategy includes regular inspection cycle adjustment, preventive maintenance measures and emergency repair measures. The specific process of obtaining the maintenance strategy solution is as follows;

[0101] S502: Based on the risk level assessment table, a matching maintenance strategy is formulated for each risk level. The maintenance strategy includes regular inspection cycle adjustment, preventive maintenance measures and emergency repair measures. For low risk levels, the formulated maintenance strategy includes a longer inspection cycle and basic preventive maintenance. For medium risk levels, the inspection cycle needs to be shortened, and more frequent maintenance inspections and preventive replacement parts strategies need to be introduced. High risk levels require urgent maintenance measures, including immediate equipment inspections and immediate replacement of key components. A maintenance strategy plan is obtained. The strategy is recorded in detail in the maintenance strategy plan and organized by risk level to ensure that the strategies at each level can be implemented quickly and effectively.

[0102] S503: Based on the maintenance strategy, the data is sent to the fault management device of the radiation monitoring station, and the fault management device automatically updates the fault management and preventive measures settings according to the maintenance strategy to obtain the specific process of the radiation risk management update result;

[0103] S503: Based on the maintenance strategy, the data is sent to the fault management device of the radiation monitoring station. The fault management device automatically updates the fault management and preventive measures according to the maintenance strategy, and the maintenance strategy is digitized and programmed into the fault management device. The fault management device automatically adjusts its operating parameters according to the strategy, such as adjusting the detection frequency and warning threshold. The device can automatically identify the risk level changes according to the real-time data, and adjust the preventive measures to ensure that the equipment operates in the best state, thereby effectively reducing the probability of failure and the resulting losses, and obtaining the updated radiation risk management results.

[0104] See also Figure 7 Based on the updated results of radiation risk management, the fault detection threshold of the radiation monitoring equipment is adjusted. The detection threshold is adjusted to match the current operating status and radiation environment of the radiation monitoring equipment by using historical data analysis and the monitoring results of the performance changes of the radiation monitoring equipment. The specific steps for obtaining the fault threshold adjustment plan are as follows:

[0105] S601: Based on the radiation risk management update results, the performance monitoring results and historical radiation data of the radiation monitoring equipment are sorted and analyzed, and the adaptability of the current state of the radiation monitoring equipment to the environmental conditions is evaluated. The specific process of generating the equipment status analysis results is as follows;

[0106] S601: Based on the updated results of radiation risk management, the performance monitoring results and historical radiation data of the radiation monitoring equipment are collated and analyzed to evaluate the compatibility of the current status of the radiation monitoring equipment with the environmental conditions. The equipment status analysis results are generated by calculating the formula "Compatibility score = (equipment stability score + environmental compatibility score) / 2", where the equipment stability score is based on the failure frequency and maintenance response speed of the equipment during operation, and the environmental compatibility score is calculated based on the performance of the equipment's historical data under the current environmental conditions.

[0107] S602: Based on the equipment status analysis result, identify the fault detection threshold that needs to be adjusted in the current operating environment, update the fault detection threshold, set the matching adjustment frequency and conditions, and generate the specific process of the basic adjustment plan;

[0108] S602: Based on the equipment status analysis results, identify the fault detection threshold that needs to be adjusted in the current operating environment, update the fault detection threshold, set the matching adjustment frequency and conditions, and apply the formula "adjusted threshold = current threshold + Δ adjustment", where Δ adjustment is determined by the equipment adaptability score and fault response record to ensure that the threshold can more sensitively reflect the actual operating status of the equipment. Set the adjustment frequency and conditions, generate a basic adjustment plan, and ensure that the threshold can be re-evaluated and adjusted when the environment or equipment operating status changes.

[0109] S603: Based on the basic adjustment plan, simulate the adjustment effect of the fault detection threshold, test the effect of the new threshold under the simulated operating conditions, analyze the accuracy and response ability of fault prediction after the threshold adjustment, and optimize the fault detection threshold. The specific process of generating the fault threshold adjustment plan is as follows;

[0110] S603: Based on the basic adjustment plan, simulate the adjustment effect of the fault detection threshold, test the effect of the new threshold under simulated operating conditions, analyze the accuracy and responsiveness of fault prediction after the threshold adjustment, and optimize the fault detection threshold. During the simulation test, the effect of the new threshold is evaluated by the formula "test effect score = (simulated fault detection success times / total simulation times) × 100%). According to the test results, optimize the fault detection threshold to ensure that the threshold can effectively identify expected faults while reducing false alarms, and generate a fault threshold adjustment plan.

[0111] See also Figure 8 , a system for predicting equipment failures of radiation monitoring stations based on a small sample training model, a system for predicting equipment failures of radiation monitoring stations based on a small sample training model is used to execute the above-mentioned method for predicting equipment failures of radiation monitoring stations based on a small sample training model, and the system includes:

[0112] The data collection module captures the reading records of radiation detectors based on the operating environment of the radiation monitoring station, including the monitoring date, time and reading, and integrates the data with the equipment maintenance log and failure time record to generate historical data records;

[0113] The model training module extracts key influencing factors based on historical data records, conducts small sample learning framework training, trains the fault classification prediction model that matches the target environment of the radiation monitoring station, and obtains the equipment fault prediction model;

[0114] The data analysis module analyzes the monitoring data in real time based on the equipment failure prediction model, analyzes the matching degree between the data and the known failure modes, and generates the predicted failure analysis results;

[0115] The risk assessment module collects and analyzes the aging data and maintenance quality of the equipment based on the predicted failure analysis results, evaluates the failure risk in combination with the current environmental conditions, calculates the probability of predicted failures, identifies the probability values ​​of multiple types of predicted failures, sets the risk level, and generates the risk level assessment results;

[0116] Based on the risk level assessment results, the threshold adjustment module formulates matching maintenance measures, instantly updates the fault management equipment of the radiation monitoring station, analyzes the performance changes and environmental adaptability of the radiation monitoring equipment, adjusts the fault detection threshold, matches the current status of the radiation monitoring equipment, and obtains the fault threshold adjustment plan.

[0117] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A method for predicting equipment failures in radiation monitoring stations based on a small sample training model, characterized in that: The following steps are involved: Based on the operating environment of the radiation monitoring station, historical data of radiation detectors are collected to obtain historical archives of radiation monitoring equipment; Based on the historical archives of the radiation monitoring equipment, a small sample learning framework is trained, and by analyzing the radiation threshold and radiation safety standard associated data, a fault type classification prediction model matching the radiation monitoring station equipment is constructed to obtain an equipment fault prediction model; Based on the equipment fault prediction model, combined with the real-time monitored radiation intensity and the state and operation data of the radiation monitoring equipment, the matching degree between the data and the known fault mode is analyzed to obtain the predicted fault matching result; Based on the predicted fault matching result, the probability of the predicted fault under the current operating conditions is evaluated to obtain the monitoring equipment failure probability information; Based on the predicted fault matching result, the probability of the predicted fault under the current operating conditions is evaluated to obtain the monitoring equipment failure probability information in the following steps: Based on the predicted fault matching results, aging data and historical maintenance quality records of the radiation monitoring equipment are collected, the potential impact of the aging and maintenance data on the current fault diagnosis is analyzed, and aging and maintenance impact assessment results are generated; Based on the aging and maintenance impact assessment results, current environmental radiation, temperature and humidity data are collected to assess the impact of current environmental data on the performance of radiation monitoring equipment and obtain environmental impact analysis results; Based on the environmental impact analysis results, the predicted fault data, environmental data and radiation monitoring equipment status information are integrated, and the predicted fault probability of the radiation monitoring equipment is quantitatively analyzed through the Bayesian network algorithm to evaluate the probability of the predicted fault under the current operating conditions and obtain the monitoring equipment failure probability information; The Bayesian network algorithm is based on the formula: ； Calculate the failure probability value, where, For environmental data, is the device status, To maintain records, For fault prediction, For a given environment data , Device Status and maintenance records Under the condition of The posterior probability of occurrence, To predict failure Occurs, considering the device status Under the condition of environmental data The probability of For existing maintenance records Equipment status under the condition The probability of Consider maintaining records Failure prediction in the case of The probability of occurrence, To maintain the probability of record occurrence, is the marginal probability of environmental data, is the marginal probability of the device state; Based on the failure probability information of the monitoring equipment, risk level is set and sent to the fault management device of the radiation monitoring station to generate radiation risk management update results; Based on the radiation risk management update result, the detection threshold is adjusted to match the current operating state and radiation environment of the radiation monitoring equipment to obtain a fault threshold adjustment plan.

2. The method for predicting equipment failures of radiation monitoring stations based on a small sample training model according to claim 1 is characterized in that: The radiation monitoring equipment history archive includes the usage time, maintenance times and fault records of the radiation monitoring equipment. The equipment fault prediction model includes radiation intensity abnormality, equipment temperature fluctuation and power supply interruption fault categories. The predicted fault matching result is specifically a matching score between the current equipment status and multiple fault types in the prediction model. The monitoring equipment failure probability information includes the probability value of the fault type, potential impact assessment and emergency response level. The radiation risk management update result includes the risk level, matched maintenance measures and scheduled equipment inspection time point. The fault threshold adjustment plan includes adjusted detection threshold parameters and detection threshold adjustment frequency.

3. The method for predicting equipment failures of radiation monitoring stations based on a small sample training model according to claim 1 is characterized in that: Based on the historical archives of the radiation monitoring equipment, a small sample learning framework training is performed, and by analyzing the radiation threshold and radiation safety standard associated data, a fault type classification prediction model matching the radiation monitoring station equipment is constructed. The specific steps of obtaining the equipment fault prediction model are as follows: Extract key influencing factors based on the historical archives of the radiation monitoring equipment, the key influencing factors including the failure time exceeding a preset threshold and the maintenance activity records, and screen the key parameters causing the failure in combination with the abnormal radiation reading data to obtain the influencing factor screening results; Based on the influencing factor screening results, a small sample learning framework is trained through influencing factors, a training cycle and a parameter adjustment strategy are set, and the model is trained and adjusted to match the fault prediction under the target environment of the radiation monitoring station, so as to obtain a training optimization model; Based on the training optimization model, the parameter optimization model is verified through the historical archives of the radiation monitoring equipment to verify the accuracy of the training optimization model in predicting equipment fault classification, and the model parameters are adjusted to optimize the accuracy of the model in predicting fault classification, thereby generating an equipment fault prediction model.

4. The method for predicting equipment failures of radiation monitoring stations based on a small sample training model according to claim 1, characterized in that: Based on the equipment fault prediction model, combined with the real-time monitored radiation intensity and the state and operation data of the radiation monitoring equipment, the matching degree between the data and the known fault mode is analyzed, and the steps of obtaining the predicted fault matching result are specifically as follows: Based on the equipment fault prediction model, traverse the radiation intensity recorded in the real-time monitoring data, compare it with the historical intensity data recorded in the fault prediction model, calculate the difference and abnormal points of the radiation intensity, and generate the radiation intensity difference analysis result; Based on the radiation intensity difference analysis results, extract the operating temperature and pressure data of the radiation monitoring equipment, evaluate the degree of deviation between the parameters and the typical values ​​in the fault prediction model, and obtain the operating parameter deviation analysis results; Based on the operation parameter deviation analysis result, the fault mode in the fault prediction model is compared, the fit between the current data mode and the known fault is analyzed, the expected fault type is identified, and the predicted fault matching result is obtained.

5. The method for predicting equipment failures of radiation monitoring stations based on a small sample training model according to claim 1, characterized in that: Based on the monitoring equipment failure probability information, risk level setting is performed and sent to the fault management device of the radiation monitoring station. The steps of generating radiation risk management update results are specifically as follows: Based on the failure probability information of the monitoring equipment, risk levels are set according to the probability of occurrence and potential impact of each predicted failure category, and low risk, medium risk and high risk levels are set respectively to generate a risk level assessment table; Based on the risk level assessment table, a matching maintenance strategy is formulated for each risk level, the maintenance strategy including regular inspection cycle adjustment, preventive maintenance measures and emergency repair measures, to obtain a maintenance strategy solution; Based on the maintenance strategy solution, it is sent to the fault management device of the radiation monitoring station, and the fault management and preventive measures settings are automatically updated by the fault management device according to the maintenance strategy solution to obtain the radiation risk management update result.

6. The method for predicting equipment failures of radiation monitoring stations based on a small sample training model according to claim 1, characterized in that: Based on the radiation risk management update result, the detection threshold is adjusted to match the current operating state and radiation environment of the radiation monitoring equipment, and the steps of obtaining the fault threshold adjustment scheme are specifically as follows: Based on the radiation risk management update results, organize and analyze the performance monitoring results and historical radiation data of the radiation monitoring equipment, evaluate the adaptability of the current state of the radiation monitoring equipment to the environmental conditions, and generate equipment state analysis results; Based on the equipment status analysis result, identifying the fault detection threshold that needs to be adjusted in the current operating environment, updating the fault detection threshold, setting a matching adjustment frequency and condition, and generating a basic adjustment plan; Based on the basic adjustment scheme, the adjustment effect of the fault detection threshold is simulated, the effect of the new threshold under simulated operating conditions is tested, the accuracy and responsiveness of fault prediction after the threshold adjustment are analyzed, and the fault detection threshold is optimized to generate a fault threshold adjustment scheme.

7. A system for predicting equipment failures in radiation monitoring stations based on a small sample training model, characterized in that, according to the method for predicting equipment failures in radiation monitoring stations based on a small sample training model according to any one of claims 1 to 6, the system comprises: The data collection module captures the reading records of radiation detectors based on the operating environment of the radiation monitoring station, combines the equipment maintenance logs and failure time records, integrates the data, and generates historical data records; The model training module extracts key influencing factors based on the historical data records, performs small sample learning framework training, trains a fault classification prediction model that matches the target environment of the radiation monitoring station, and obtains an equipment fault prediction model; The data analysis module analyzes the monitoring data in real time based on the equipment failure prediction model, analyzes the matching degree between the data and the known failure mode, and generates the predicted failure analysis result; The risk assessment module evaluates the fault risk based on the predicted fault analysis results and the current environmental conditions, calculates the probability of the predicted fault, identifies the probability values ​​of multiple types of predicted faults, sets the risk level, and generates a risk level assessment result; The threshold adjustment module formulates matching maintenance measures based on the risk level assessment results, instantly updates the fault management equipment of the radiation monitoring station, analyzes the performance changes and environmental adaptability of the radiation monitoring equipment, adjusts the fault detection threshold, matches the current radiation monitoring equipment status, and obtains a fault threshold adjustment plan.

Images