Nuclear energy operation and maintenance optimization method based on graph driving

Abstract

The invention relates to the technical field of intelligent operation and maintenance, in particular to a nuclear energy operation and maintenance optimization method based on atlas driving, which comprises the following steps: acquiring equipment data and maintenance records, marking state matching, screening a maintenance strategy and combining semantic reasoning to extract a fault path, and classifying deviations to generate adaptability labels. Uninterfered tasks are screened, work order mapping distribution is extracted, execution imbalance is judged, resource and task ratio recording period differences are calculated, and operation and maintenance progress monitoring structure indexes are generated. According to the method, the mapping relation between equipment and working conditions is constructed through operation data and maintenance records, accurate evaluation of the health state is achieved, the fault path and deviation degree are recognized in combination with semantic reasoning, the adaptability of a maintenance strategy and the working conditions is improved, task interference risks are avoided, load fluctuation is recognized by means of work order execution data, and resource and task quantity comparison is combined. And a resource investment benefit evaluation mechanism is constructed, so that the fault prediction accuracy, the task execution coordination and the resource configuration rationality are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and maintenance technology, and in particular to a graph-driven method for optimizing nuclear energy operation and maintenance. Background Technology

[0002] The field of intelligent operation and maintenance (O&M) technology mainly involves the intelligent management and optimization of industrial equipment, production processes, and the entire production environment using modern information technology, automation technology, and artificial intelligence. Its core aspects include equipment health management, predictive maintenance, fault diagnosis and repair, and production efficiency improvement. Intelligent O&M relies on technologies such as big data, cloud computing, and the Internet of Things (IoT) to provide intelligent decision support for manufacturing enterprises through real-time monitoring, data collection, and analysis, thereby achieving automation and self-optimization of production systems. Intelligent O&M technology not only reduces equipment failure rates and extends equipment lifespan but also promptly identifies problems during production and makes predictive adjustments, thus improving the overall operational efficiency and resource utilization of enterprises. Traditional nuclear energy O&M optimization methods refer to the operation and maintenance management of nuclear energy facilities based on existing equipment's routine maintenance processes. Traditional methods rely on manual inspections, periodic maintenance, and equipment status monitoring to ensure the safety and stability of nuclear energy facilities. However, traditional O&M methods have limitations. They rely on experience-based judgment, making it difficult to accurately predict equipment failure times and causes. Furthermore, maintenance and repair work has strong time-sensitive requirements, making it impossible to achieve optimal resource allocation and optimization.

[0003] Existing technologies rely on manual inspections and periodic maintenance, making it difficult to promptly detect abnormal signs when nuclear equipment experiences sudden changes in operating status or frequent external interference. Equipment status identification is largely based on experience-based judgments, lacking quantitative assessment methods for health status. This makes it difficult for maintenance strategies to accurately match actual operating conditions. Data collected by sensors fails to form an effective correlation with specific task segments, and resource duplication or execution imbalances are prone to occur during operation and maintenance task scheduling. For example, when abnormal operating data of some equipment is not detected in time, related maintenance tasks are still executed as usual, which can easily lead to task overlap, resource waste, and schedule deviations. Overall, this limits the operation and maintenance system's ability to respond to complex status changes and its resource utilization efficiency. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a graph-driven nuclear energy operation and maintenance optimization method.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a graph-driven nuclear energy operation and maintenance optimization method, comprising the following steps:

[0006] S1: Obtain nuclear energy equipment operation data and original maintenance records, extract equipment nodes and their associations, construct an initial knowledge graph, mark the matching between equipment status and operating conditions, and generate an equipment health status assessment list;

[0007] S2: Based on the equipment health status assessment list, select maintenance strategies that are suitable for the current working conditions, extract fault mode paths by combining semantic reasoning algorithms, and perform deviation calculation and classification processing with the benchmark working conditions to obtain maintenance strategy adaptability labels.

[0008] S3: Call the maintenance strategy adaptability tag, extract the fault mode number, identify the range of sensor data fluctuation within the section, compare the radiation sensor monitoring period with the fault mode section, record the number of overlapping time periods, and generate a list of radiation sensor abnormal impact sections.

[0009] S4: Based on the list of abnormally affected sections of the radiation sensor, filter out maintenance task nodes that are not disturbed, extract work order segment records and operation data, map task allocation and operation time distribution, determine whether there is an execution imbalance of tasks, and obtain work order task execution fluctuation groups.

[0010] As a further aspect of the present invention, the equipment health status assessment list includes equipment number, status label, operating condition deviation, and maintenance classification; the maintenance strategy adaptability label includes adaptability level, fault type, benchmark comparison result, and task association number; the radiation sensor abnormality impact zone list includes equipment type, impact time zone, number of overlapping time periods, and affected task number; and the work order task execution fluctuation group includes task distribution unevenness number, operation duration record, task completion deviation, and operation matching degree.

[0011] As a further aspect of the present invention, the steps for obtaining the equipment health status assessment list are as follows:

[0012] S111: Acquire nuclear energy equipment operation data and original maintenance records, extract equipment nodes and their associations, match equipment status acquisition time with operating condition requirement time, compare time range with status requirements, and generate partition node operation record time periods;

[0013] S112: Based on the time period of the partition node operation record, extract the overlapping time period between the equipment status and the working condition demand time interval, calculate the ratio of the overlapping time to the total working condition demand time, calculate the overlap ratio value, filter nodes with an overlap ratio lower than the benchmark value, and obtain the partition node status coverage deviation rate according to the number of equipment status labels.

[0014] S113: Based on the deviation rate of the partition node status coverage, determine the deviation status of the node number, identify the node number whose deviation rate exceeds the node synchronization threshold, integrate the node number, status coverage information and deviation rate value, and generate a device health status assessment list.

[0015] As a further aspect of the present invention, the step of obtaining the maintenance strategy adaptability tag specifically includes:

[0016] S211: Based on the equipment health status assessment list, identify maintenance strategies and fault mode paths that are adapted to the current working conditions, extract the start and end fluctuation ranges of the real-time sensors, calculate the start and end difference of the fluctuation range, compare it with the benchmark fluctuation range, and obtain the fluctuation deviation value.

[0017] S212: Call the fluctuation deviation value, combine the segment distribution, fluctuation trend and adjustment frequency, uniformly collect the maintenance strategy deviation data, identify and calculate the adaptability deviation degree according to the segment number, determine the fluctuation direction according to the adjustment frequency, and obtain the maintenance strategy adaptability label.

[0018] As a further aspect of the present invention, the step of obtaining the list of abnormally affected areas of the radiation sensor specifically includes:

[0019] S311: Call the maintenance strategy adaptability tag, filter the fluctuation deviation task segment number, extract the fluctuation time period according to the node association table, process the segment fluctuation time period according to the time dimension, identify the fluctuation time period index table, and obtain the fluctuation task time period set.

[0020] S312: Based on the set of fluctuating task periods, collect the fluctuation range of sensor data in the same period, identify the radiation sensor impact table, determine the daily sensor interference based on the interference threshold, match it with the fluctuating task periods, determine whether there is a fluctuating abnormal correlation, and generate a list of radiation sensor abnormally affected areas.

[0021] As a further aspect of the present invention, the step of obtaining the work order task execution fluctuation group specifically includes:

[0022] S411: Based on the list of abnormally affected sections of the radiation sensor, filter and maintain task nodes, extract the task list, and obtain the task set of nodes not affected by the radiation sensor.

[0023] S412: Call the task set of nodes not affected by radiation sensor interference, match the daily segmented records and operation data of work orders, extract the planned task volume according to the task number, count the number of operators and the real-time operation time period, and generate a work order cooperation execution status matching dataset.

[0024] S413: Based on the work order cooperation execution status matching dataset, evaluate the matching degree between task execution efficiency and operation distribution, identify efficiency fluctuation nodes, mark tasks with fluctuations exceeding the benchmark value as abnormal nodes, analyze work order task execution matching deviation, and obtain work order task execution fluctuation groups.

[0025] As a further aspect of the present invention, the method further includes step S5:

[0026] S5: Call the work order task execution fluctuation group, extract the abnormal fluctuation task group, calculate the ratio of resource input to task volume in the operation and maintenance model, record the difference distribution between resource deployment cycle and task execution cycle, and generate operation and maintenance progress monitoring structure indicators.

[0027] The operational progress monitoring structure indicators include resource allocation ratio, task intensity level, execution cycle difference, and operational efficiency indicators.

[0028] As a further aspect of the present invention, the steps for obtaining the operation and maintenance progress monitoring structure indicators are as follows:

[0029] S511: Call the work order task execution fluctuation group, filter nodes that exceed the threshold, record the time interval and task volume change range, and obtain the progress fluctuation abnormal identifier set;

[0030] S512: Based on the operation and maintenance model resources corresponding to the nodes in the progress fluctuation anomaly identifier set, identify the node resource input ratio sequence, extract the abnormal distribution interval and compare the critical coefficient, record the ratio offset direction and node number, and form an operation and maintenance resource matching offset index group.

[0031] S513: Based on the operation and maintenance resource matching offset index group, extract the time period of operation and maintenance model resource deployment and task execution, identify the difference between the resource deployment cycle and the operation cycle, and sort and label them according to the progress benchmark to generate operation and maintenance progress monitoring structure indexes.

[0032] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0033] In this invention, by analyzing the operating data and maintenance records of nuclear energy equipment, a mapping relationship between equipment nodes and operating conditions is formed, thereby achieving a refined characterization of the equipment's health status. By combining semantic reasoning to extract fault paths and quantify the degree of deviation from the baseline state, the accuracy of the adaptation between maintenance strategies and actual needs can be improved. By relying on sensor data fluctuations to identify abnormally affected sections, potential risk points can be accurately located in unstable areas, thereby avoiding interference factors in task allocation. By using work order execution records to construct a dynamic distribution model of task execution, the trend of task load fluctuations can be effectively identified. Furthermore, by combining the dynamic comparison of resources and task volume, a structured evaluation of the benefits of operation and maintenance resource investment can be achieved, thereby promoting the overall improvement of fault prediction accuracy, task coordination efficiency, and the rationality of resource allocation. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the main steps of the present invention;

[0035] Figure 2 This is a flowchart illustrating the process of obtaining the equipment health status assessment list in this invention.

[0036] Figure 3 This is a flowchart illustrating the process of obtaining the maintenance strategy adaptability tag in this invention;

[0037] Figure 4 This is a flowchart illustrating the process of obtaining the list of radiation sensor anomaly-affected zones in this invention.

[0038] Figure 5 This is a flowchart illustrating the process of obtaining the work order task execution fluctuation group in this invention.

[0039] Figure 6 This is a flowchart illustrating the acquisition of structural indicators for operation and maintenance progress monitoring in this invention. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0041] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0042] Example 1

[0043] Please see Figure 1 This invention provides a technical solution: a graph-driven nuclear energy operation and maintenance optimization method, comprising the following steps:

[0044] S1: Obtain nuclear energy equipment operation data and original maintenance records, extract equipment nodes and their associations, construct an initial knowledge graph, mark the matching between equipment status and operating conditions, and generate an equipment health status assessment list;

[0045] S2: Based on the equipment health status assessment list, select maintenance strategies that are suitable for the current working conditions, extract fault mode paths by combining semantic reasoning algorithms, and perform deviation calculation and classification with the baseline working conditions to obtain maintenance strategy adaptability labels.

[0046] S3: Call the maintenance strategy adaptability tag, extract the fault mode number, identify the range of sensor data fluctuation within the section, compare the radiation sensor monitoring period with the fault mode section, record the number of overlapping time periods, and generate a list of radiation sensor abnormal impact sections.

[0047] S4: Based on the list of abnormally affected sections of the radiation sensor, filter out the maintenance task nodes that are not disturbed, extract the work order segment records and operation data, map the task allocation and operation time distribution, determine whether there is an execution imbalance of the task, and obtain the work order task execution fluctuation group.

[0048] S5: Call the work order task execution fluctuation group, extract the abnormal fluctuation task group, calculate the ratio of resource input to task volume in the operation and maintenance model, record the difference distribution between resource deployment cycle and task execution cycle, and generate operation and maintenance progress monitoring structure indicators.

[0049] The equipment health status assessment checklist includes equipment number, status label, operating condition deviation, and maintenance category. The maintenance strategy adaptability label includes adaptability level, fault type, benchmark comparison result, and task association number. The radiation sensor anomaly impact area list includes equipment type, impact time period, number of overlapping time periods, and affected task number. The work order task execution fluctuation group includes task distribution unevenness number, operation duration record, task completion deviation, and operation matching degree. The operation and maintenance progress monitoring structure indicators include resource allocation ratio, task intensity level, execution cycle difference, and operation and maintenance efficiency indicators.

[0050] Please see Figure 2 The specific steps for obtaining the equipment health status assessment list are as follows:

[0051] S111: Acquire nuclear energy equipment operation data and original maintenance records, extract equipment nodes and their associations, match equipment status acquisition time with operating condition requirement time, compare time range with status requirements, and generate partition node operation record time periods;

[0052] Acquiring operational data and original maintenance records for nuclear energy equipment involves collecting key operating parameters, status parameters, and historical maintenance activity data. For example, this includes real-time vibration sensor data, bearing temperature data, and current data for a reactor coolant pump (number: RCP001), as well as original maintenance records such as regular inspection records, fault repair records, and component replacement records. From the acquired data, the individual functional units or monitoring points within the equipment, i.e., equipment nodes, are accurately identified. Simultaneously, the interdependencies and logical connections between these nodes, i.e., their correlations, are clarified. For example, the RCP001 pump body is the main node, while its internal motor, bearings, and impeller are child nodes, and there is a direct correlation between bearing temperature and motor current. In nuclear energy equipment management, the timestamps of the collected equipment status data are strictly matched with the standard operating condition timestamps required by the nuclear power unit or specific equipment unit. For example, RCP001... The sensor data collection period for pump 01 on October 26th was from 08:00 to 17:00, while the pump's specified operating period for that day was from 07:00 to 18:00. The matched time range was then compared one by one with the preset equipment status requirements. For example, it was checked whether the data from 08:00 to 17:00 fully covered the key operating status information from 07:00 to 18:00. If any data was missing or redundant, it was trimmed or supplemented accordingly. Based on this, the aforementioned nuclear equipment operating data, original maintenance records, and the results of time matching and comparison were integrated to generate a record for each partition node containing all its key operating statuses and related events within that time period. For example, the recording period for pump RCP001 on October 26th was from 08:00 to 17:00, including vibration data, temperature data, and corresponding maintenance event records, to generate the partition node operating record period.

[0053] S112: Based on the partition node's operation record time period, extract the overlapping time periods between the equipment status and the operating condition demand time interval, and calculate the ratio of the overlapping time to the total operating condition demand duration using the formula:

[0054] ;

[0055] Calculate the overlap ratio, filter nodes with an overlap ratio lower than the baseline value, and obtain the partition node status coverage deviation rate based on the number of device status labels.

[0056] in, This represents the overlap ratio. This represents the duration of overlap between the equipment status and operational requirements within the i-th recording period. The normalized adjustment factor representing the number of state labels in the i-th recording period. This represents the duration of the device status within the i-th recording period. This represents the total duration of the operating condition within the i-th recording period. This indicates the total number of record periods involved in the calculation;

[0057] Based on the operating time periods of the partitioned nodes of nuclear energy equipment under specific operating conditions, the overlapping time periods of the equipment status data collection time interval and the corresponding operating condition demand time interval for each equipment node are accurately identified and extracted. For example, for nuclear energy equipment node RCP001, its equipment status data collection time is from 08:00 to 17:00 on the same day, and the operating condition demand time is from 07:00 to 18:00 on the same day. Then, the overlapping time period is from 08:00 to 17:00, which is 9 hours. Next, the ratio of the overlapping time period to the total operating condition demand time is calculated using the following formula, which is the overlap ratio value: ,in, This represents the overlap ratio, which indicates the extent to which the equipment status data covers the operating requirements. Its absolute value ensures that the result is positive. Representing the The time interval between the equipment status and operating requirements within each recording period overlaps, in hours. This measure indicates the duration of actual effective data coverage. Representing the The normalized adjustment coefficient for the number of status labels within a recording period is used to measure the richness of human or labeled information in the device status data. It is calculated by comparing the number of status labels within that period to a preset maximum number of labels. For example, if there are 25 status labels in a period, and the maximum number that period can accommodate is 50, then... ; Representing the The device status duration within each recording period, i.e. the actual duration for which the device outputs data, is expressed in hours. This item represents the actual running time of data acquisition. Representing the The total operating time required within a recording period, that is, the total operating time that the equipment should theoretically meet within that period, in hours. This item represents the theoretical working time of the equipment under specific operating conditions. This represents the total number of recording periods involved in the calculation; it is the summation symbol in the formula. The upper limit represents the number of data samples included in the calculation, and the formula is obtained through... The difference between the calculated state duration and the required duration is multiplied by an adjustment factor. The duration of overlap Add them together to form the numerator, which aims to comprehensively consider data coverage, data volume, and demand satisfaction. The denominator is... The sum of squares of the total operating time required under each condition is considered, and the square root of the sum is added by 1. This is used to normalize the numerator, preventing the ratio value from becoming abnormally large due to an excessively small denominator. Simultaneously, it provides a stable scale for the ratio value, thereby ensuring the overlap ratio value... It can objectively reflect the comprehensive coverage of equipment status data to operating conditions. For example, data from three consecutive recording periods of the RCP001 pump are used for calculation, as shown in the table below:

[0058] Table 1: Statistical Table of Operating Data for Nuclear Equipment Node RCP001

[0059]

[0060] Assuming a maximum preset number of state labels of 50, the normalized adjustment coefficient for each period is... They are respectively: , , ;

[0061] Substitute the above data into the formula to calculate. value:

[0062] ;

[0063] The overlap ratio value This indicates that the nuclear energy equipment has good data coverage and status recording within the selected period, and the overlap ratio value is [value missing]. The data coverage of a device node is compared with a preset benchmark value. For example, if the benchmark value is set to 0.8, and the calculated overlap ratio is lower than this benchmark value, the data coverage of that device node is considered insufficient and requires further analysis. For example, if the calculated overlap ratio of a node is 0.75, which is lower than the benchmark value of 0.8, it is filtered out. Based on the number of device status labels corresponding to the filtered device nodes and their operating requirements, the device status coverage deviation rate of each node is calculated. The device status coverage deviation rate is calculated as: 1 - (actual number of status labels / expected number of status labels). The expected number of status labels is set according to the device type, operating conditions, and historical data quality requirements. For example, for the RCP001 pump, if its overlap ratio is lower than 0.8 and it is filtered out, and its actual number of status labels in a recording period is 10, while the expected number of status labels is 40, then its status coverage deviation rate is... Get the partition node status coverage deviation rate.

[0064] S113: Based on the partition node status coverage deviation rate, determine the deviation status of the node number, identify the node number whose deviation rate exceeds the node synchronization threshold, integrate the node number, status coverage information and deviation rate value, and generate a device health status assessment list.

[0065] Based on the partition node status coverage deviation rate, for example, for device RCP001, its status coverage deviation rate is 0.75. This deviation rate is compared with a pre-set node synchronization threshold, for example, 0.50. This threshold is determined based on long-term experience data and risk assessment of nuclear power equipment operation and represents an acceptable level of data coverage deviation. Any deviation rate exceeding this threshold is considered a status anomaly requiring attention. In this determination process, if the deviation rate of a node exceeds the threshold, for example, 0.75 is significantly higher than 0.50, then that node number is identified as a node number with a deviation rate exceeding the node synchronization threshold. This means that the data of that node... If there are issues with coverage quality or status synchronization, the identified node numbers, their corresponding status coverage information (e.g., overlap ratio of 1.2378 and actual number of status labels of 10), and specific deviation rate values ​​(e.g., 0.75) are then integrated to form a structured dataset. For example, an entry might be a record with the following structure: {Node Number: RCP001, Status Coverage Information: {Overlap Ratio: 1.2378, Actual Number of Labels: 10}, Deviation Rate: 0.75}. The integrated information is then organized into a comprehensive equipment health status assessment list. This list provides the foundational data for subsequent maintenance strategy development and failure mode analysis, generating the equipment health status assessment list.

[0066] Please see Figure 3 The specific steps for obtaining the maintenance strategy adaptability tag are as follows:

[0067] S211: Based on the equipment health status assessment checklist, identify maintenance strategies and failure mode paths that meet the current operating conditions, extract the start and end fluctuation ranges of real-time sensors, and use the formula:

[0068] ;

[0069] Calculate the start and end difference of the fluctuation range and compare it with the benchmark fluctuation range to obtain the fluctuation deviation value;

[0070] in, Represents the characteristic value of fluctuation difference. The value of the termination sensor represents the j-th fluctuation segment. The initial sensor value represents the j-th fluctuation segment. The normalization coefficient represents the number of fault mode paths corresponding to the sensor within the j-th fluctuation segment. This represents the expected fluctuation value in the equipment health status assessment list that matches the current operating conditions within the j-th fluctuation range. Indicates the number of fluctuation segments involved in the calculation;

[0071] Based on the equipment health status assessment checklist, each equipment node in the checklist is analyzed to identify and match the most suitable maintenance strategy and potential failure mode paths under the current nuclear power operating conditions. For example, for the RCP001 pump, which is marked as having poor health status in the assessment checklist, strategies such as "condition-based predictive maintenance" or "fault diagnosis-based repair" are identified from the preset maintenance strategy library based on its current operating parameters (such as vibration data and bearing temperature) and operating conditions (such as full-load operation and startup phase), and associated with failure mode paths such as "bearing wear" and "impeller imbalance". At the same time, the starting and ending sensor values ​​of the equipment fluctuation range are extracted from real-time sensor data. For example, for the vibration sensor of the RCP001 pump, the starting value of a certain fluctuation range is... The termination value is The starting and ending difference of the fluctuation range is calculated using the following formula and compared with the preset benchmark fluctuation range to obtain the fluctuation deviation value: ,in, This represents the characteristic value of fluctuation difference, which is used to quantify the degree of fluctuation in sensor data and its deviation from the expected value; Representing the The termination sensor value of a fluctuation segment, for example, in vibration sensor monitoring, the vibration amplitude at the end of a certain monitoring period, in mm / s; Representing the The initial sensor value of each fluctuation segment, i.e., the vibration amplitude at the beginning of the monitoring period, in mm / s; Representing the The normalized coefficient for the number of fault mode paths corresponding to a sensor within a fluctuation range is calculated by comparing the number of fault mode paths associated with that fluctuation range to the maximum number of fault mode paths associated with that type of sensor. For example, if a vibration sensor's data fluctuation is associated with 3 fault modes (such as bearing wear, rotor imbalance, pump cavitation), and that type of sensor can be associated with a maximum of 5 typical fault modes, then... ; Representing the The expected fluctuation value in the equipment health status assessment list within each fluctuation range matches the current operating conditions. This value is set based on the equipment's historical health data and standard operating specifications. For example, for the RCP001 pump in a healthy state, its expected vibration fluctuation value is 0.05 mm / s. This represents the number of fluctuation segments involved in the calculation, and is the summation symbol in the formula. The upper limit represents the number of sensor data fluctuation sample segments included in the calculation, as shown in the formula. The term calculates the squared difference within the fluctuation range, reflecting the intensity of the fluctuation itself, while The term combines the difference between the fluctuation termination value and the expected value, and considers the complexity of the failure modes. The average of the two values ​​is then taken, resulting in... The values ​​comprehensively evaluate the fluctuation characteristics of the sensor data and its deviation from the expected values. For example, the values ​​of the RCP001 pump vibration sensor in three consecutive fluctuation ranges are collected and calculated, as shown in Table 2:

[0072] Table 2: Vibration Sensor Fluctuation Data for Nuclear Power Equipment

[0073]

[0074] Assuming the maximum number of associated fault modes for this type of sensor is 5, then the normalization coefficients for each segment are... They are respectively: , , Substitute the above data into the formula to calculate. value:

[0075] ;

[0076] This fluctuation difference characteristic value This indicates that the vibration sensor data of the RCP001 pump exhibits certain fluctuations within the monitored area, with a slight deviation from the expected value. The calculated fluctuation difference characteristic value will be used to further illustrate this. Compare with a preset benchmark fluctuation range, for example, a preset benchmark fluctuation range of 0.01 to 0.05 mm / s. If the value exceeds this range, it is considered to have a fluctuation deviation. For example, 0.05856 exceeds 0.05. This comparison result is taken as the fluctuation deviation value, and the fluctuation deviation value is obtained.

[0077] S212: Call the fluctuation deviation value, combine the segment distribution, fluctuation trend and adjustment frequency, uniformly collect the maintenance strategy deviation data, identify and calculate the adaptability deviation degree according to the segment number, determine the fluctuation direction according to the adjustment frequency, and obtain the maintenance strategy adaptability label.

[0078] The fluctuation deviation value is retrieved and comprehensively considered in conjunction with the distribution of equipment operating sections, the fluctuation trend of sensors, and the frequency of maintenance strategy adjustments. For example, for the RCP001 pump, its fluctuation deviation value is 0.05856. This is combined with the operating section where the deviation value occurs (e.g., frequent occurrence during startup, occasional occurrence during steady-state operation), the trend of the fluctuation deviation value over time (e.g., a slow upward trend in vibration deviation), and the frequency of maintenance strategy adjustments for this pump (e.g., monthly preventative maintenance or quarterly overhaul). Based on this, all relevant data are uniformly collected to form maintenance strategy deviation data. The deviation data is then identified according to the equipment section number and calculated. The adaptability deviation of each segment is calculated. The adaptability deviation measures the effectiveness of the current maintenance strategy for a specific fluctuation segment. The calculation takes into account the magnitude, duration, and frequency of fluctuation deviation. For example, if the vibration deviation of the RCP001 pump is consistently high and the adjustment frequency is low in the start-up segment, its adaptability deviation will be calculated as high. The direction of equipment fluctuation is judged based on the adjustment frequency of the maintenance strategy. If the adjustment frequency is high, for example, the parameters are fine-tuned weekly, it indicates a rapid response to fluctuations and that the fluctuation direction tends to be stable or under control. If the adjustment frequency is low, for example, a major overhaul is only performed once a year, and the fluctuation deviation continues to increase, it indicates that the fluctuation direction tends to worsen or become out of control. Combining the analysis results, the adaptability label of the maintenance strategy is finally obtained.

[0079] Please see Figure 4 The specific steps for obtaining the list of areas affected by radiation sensor anomalies are as follows:

[0080] S311: Call the maintenance strategy adaptability tag, filter the fluctuation deviation task segment number, extract the fluctuation time period according to the node association table, process the segment fluctuation time period by time dimension, identify the fluctuation time period index table, and obtain the fluctuation task time period set.

[0081] The system invokes maintenance strategy adaptability tags to precisely filter out all task segment numbers with significant fluctuation deviations or low maintenance strategy adaptability. For example, if the "start-up" operation segment of RCP001 pump is marked as having low adaptability, then that segment number is selected. Next, based on a pre-established equipment node association table, the system extracts the fluctuation time periods associated with the task segment number. The node association table records details for each equipment node, sensor data type, and corresponding historical fluctuation times. For instance, the system retrieves the time periods within the past month for all vibration data in the RCP001 pump "start-up" segment that exceeded the normal range, such as "October 27th, 07:00-07:30". Subsequently, the extracted segment fluctuation time periods are refined according to the time dimension, including merging overlapping time periods. Segments with excessively short filling time intervals or those filtered according to a preset minimum fluctuation duration are processed. For example, if "07:00-07:10" and "07:15-07:30" are components of the same fluctuation event and the interval between them is less than 10 minutes, they are merged into "07:00-07:30". After completing the time dimension processing, a fluctuation time period index table is identified and generated. This table lists in detail the final fluctuation time period and its unique identifier for each filtered task segment. For example, {Task segment number: RCP001-Start, Fluctuation time period: [10-27 07:00-07:30, 11-03 06:45-07:15]}. All fluctuation time periods in the fluctuation time period index table are collected to obtain the fluctuation task time period set.

[0082] S312: Based on the set of fluctuating task periods, collect the fluctuation range of sensor data in the same period, identify the radiation sensor impact table, determine the daily sensor interference based on the interference threshold, match it with the fluctuating task periods, determine whether there is a correlation with fluctuation anomalies, and generate a list of radiation sensor anomaly affected areas.

[0083] Based on the set of fluctuating task periods, the fluctuation range of sensor data for all segments during a specific time period is collected and acquired in real time. For example, for the RCP001 pump in the fluctuating task period set from 07:00 to 07:30 on October 27th, the actual measured values ​​of all sensors (not limited to vibration sensors, but also including temperature, pressure, etc.) during this time period are collected, and their fluctuation amplitude is calculated. For example, the vibration fluctuation range from 0.05 mm / s to 0.15 mm / s is acquired. Simultaneously, a preset radiation sensor influence table is referenced and identified. This table details the known effects of different radiation levels in the nuclear energy environment on the readings of various sensors. For example, the table lists that "when the gamma ray intensity is higher than 100 nGy / h, the readings of some temperature sensors produce an error of ±2%". A preset interference threshold is used to determine whether there is radiation interference affecting sensors daily. For example, the gamma ray interference threshold is set to... If the radiation sensor reading consistently exceeds 150 nGy / h on a given day, radiation interference is considered to exist. This threshold is set based on experimental data and industry standards regarding the equipment's radiation tolerance and is precisely matched to the fluctuation task period. For example, if the RCP001 pump synchronously detects a nearby radiation sensor reading of 200 nGy / h during the fluctuation task period "October 27th, 07:00-07:30," exceeding the interference threshold of 150 nGy / h, further investigation is conducted to determine whether there is an abnormal correlation during this fluctuation task period. That is, whether the fluctuation of the sensor data has a causal relationship or strong correlation with radiation interference. For example, by comparing the synchronicity between the radiation intensity curve and the sensor fluctuation curve, if the two are highly consistent in time and the amplitude changes are consistent, an abnormal correlation is determined. Based on this, all identified sections affected by radiation anomalies are integrated to generate a list of sections affected by radiation sensor anomalies.

[0084] Please see Figure 5 The specific steps for obtaining the work order task execution fluctuation group are as follows:

[0085] S411: Based on the list of sections affected by abnormal radiation sensors, filter maintenance task nodes, extract the task list, and obtain the task set of nodes not affected by radiation sensors.

[0086] Based on the list of radiation sensor anomaly-affected zones, the maintenance task nodes included are first precisely screened. For example, all maintenance tasks related to the RCP001 pump are identified from this list. Simultaneously, a corresponding task list is extracted from the overall nuclear equipment maintenance plan. This list contains information such as all planned maintenance activities, execution time, and expected resources. By comparing the maintenance task nodes with the list of radiation sensor anomaly-affected zones, those node tasks in the list are excluded from the overall task list. For example, if a maintenance task for the RCP001 pump (e.g., bearing inspection) is listed in the list of radiation sensor anomaly-affected zones, that task will be temporarily excluded from the current execution scope to ensure the accuracy of subsequent operations. Finally, a set of node tasks unaffected by radiation sensor interference is obtained.

[0087] S412: Call the task set of nodes not affected by radiation sensor interference, match the daily segmented records and operation data of work orders, extract the planned task volume by task number, count the number of operators and real-time operation time period, and generate a work order cooperation execution matching dataset.

[0088] The system calls upon the task set of nodes unaffected by radiation sensor interference. Each task in this set is precisely matched against the daily segmented records and actual operation data recorded in the daily work order management. For example, for the daily inspection task of pump RCP001 (task number: RCP_DAILY_CHK_001), the system matches its work order record on October 28th, including the actual execution time period and operator information. Then, according to the task number, the planned task quantity of the task is extracted from the matched work order record. For example, the planned task quantity of the inspection task is "inspect 20 measuring points". The system also simultaneously counts the number of operators and the real-time operation time during the actual operation of the task. For example, the actual number of operators is 2, and the actual operation time period is from 09:00 to 10:00. The matching, extraction, and statistical results are integrated, including task number, planned task quantity, actual number of operators, and deviation between the actual operation time period and the plan, to form a work order collaborative execution matching dataset.

[0089] S413: Match the dataset based on the work order cooperation execution status, evaluate the matching degree between task execution efficiency and operation distribution, identify efficiency fluctuation nodes, mark tasks with fluctuations exceeding the benchmark value as abnormal nodes, analyze the work order task execution matching deviation, and obtain the work order task execution fluctuation group.

[0090] Based on the work order collaboration execution data matching dataset, the matching degree between the execution efficiency and operation distribution of each task is evaluated. For example, for the inspection task with task number RCP_DAILY_CHK_001, its planned task volume is "inspect 20 measurement points", the actual number of operators is 2, and the actual operation time is 09:00 to 10:00. If the actual number of measurement points completed is 18, then its execution efficiency is 18 / 20 = 0.9. The matching degree of operation distribution evaluates the reasonableness of the number of operators and task complexity. Based on this evaluation result, efficiency fluctuation nodes are identified, that is, those task nodes where the execution efficiency or operation distribution differs from the expectation. For example, if the execution efficiency of a task is lower than 0.95 or higher than 1.05 (i.e., efficiency deviation), then... If the efficiency deviation exceeds 5%, it is identified as an efficiency fluctuation node. The efficiency deviation rate is calculated as |actual efficiency - 1|. Tasks whose fluctuation exceeds a preset benchmark value are marked as abnormal nodes. For example, if the benchmark value for the efficiency deviation rate is set to 0.10 (i.e., efficiency fluctuations exceeding 10% are considered abnormal), if the efficiency deviation rate of a certain task is 0.12, then the task is marked as an abnormal node. This benchmark value is set based on historical task execution data and best practices in the nuclear energy industry. By further analyzing abnormal nodes, the work order task execution matching deviation is evaluated in depth. For example, it is analyzed whether the abnormal deviation rate of 0.12 is due to insufficient personnel, non-standard operating procedures, or external environmental interference. After evaluation and analysis, the tasks are classified and integrated to obtain the work order task execution fluctuation group.

[0091] Please see Figure 6 The specific steps for obtaining the structural indicators of operation and maintenance progress monitoring are as follows:

[0092] S511: Call the work order task execution fluctuation group, filter nodes that exceed the threshold, record the time interval and task volume change range, and obtain the progress fluctuation anomaly identifier set;

[0093] The work order task execution fluctuation group is invoked. First, all task nodes within it are precisely screened to identify and extract those nodes whose execution efficiency or operation distribution fluctuations exceed a preset threshold. For example, from the work order task execution fluctuation group, a maintenance task for pump RCP001 is screened out, with an efficiency fluctuation rate of 0.12, exceeding the set fluctuation threshold of 0.10. For the screened nodes, the time interval of their task execution is precisely recorded, for example, from 09:00 to 10:00 on October 28th. The specific change in the task volume within this time interval is also recorded in detail. For example, if 20 measuring points are planned to be inspected, but 18 measuring points are actually completed, the change in the task volume is 2 measuring points or 10% of the uncompleted amount. The recorded time interval and task volume change information are organized and labeled to form a progress fluctuation anomaly label set.

[0094] S512: Based on the operation and maintenance model resources corresponding to the nodes in the progress fluctuation anomaly identifier set, identify the node resource input ratio sequence, extract the abnormal distribution interval and compare the critical coefficient, record the ratio offset direction and node number, and form an operation and maintenance resource matching offset index group.

[0095] Based on the nodes identified by the schedule fluctuation anomaly markers, the corresponding operation and maintenance model resources are first retrieved. These resources include standard maintenance procedures, required manpower, materials, and time configurations for that type of equipment. For example, for the RCP001 pump maintenance task identified as abnormal, the operation and maintenance model specifies a standard resource input of 2 senior engineers and a time consumption of 2 hours. The resource input ratio sequence for this node is identified and calculated, which is the ratio of actual resource input to the operation and maintenance model resource input. For instance, if the task actually involved 3 senior engineers and took 2.5 hours, the manpower ratio is 3 / 2 = 1.5, and the time ratio is 2.5 / 2 = 1.25. Further, the abnormal distribution interval is extracted from the resource input ratio sequence. For example, if the normal ratio range is between 0.8 and 1.2, then 1.5 and 1.25 both fall within the abnormal distribution range and are compared with a preset critical coefficient. The critical coefficient is a threshold used to judge the degree of abnormality of the resource input ratio. For example, the critical coefficient is set to 1.3, which means that if the ratio exceeds 1.3, the resource input is considered to be significantly deviated. This coefficient is set based on historical operation and maintenance data and risk assessment. After comparison, the direction of each abnormal ratio deviation is recorded (for example, a deviation higher than 1.3 is a positive deviation, and a deviation lower than 0.7 is a negative deviation) and the corresponding node number. For example, the manpower ratio of 1.5 for this task of RCP001 pump is a positive deviation. The information including the deviation direction and node number is integrated to form the operation and maintenance resource matching deviation index group.

[0096] S513: Based on the offset index group of operation and maintenance resources, extract the time period of resource deployment and task execution in the operation and maintenance model, identify the difference between the resource deployment cycle and the operation cycle, and sort and label them according to the progress benchmark to generate operation and maintenance progress monitoring structure indicators.

[0097] Based on the operation and maintenance resource matching offset index group, the operation and maintenance model resource deployment information (e.g., planned resource deployment amount and time) and the actual execution time period of the task are accurately extracted for each task. For example, the operation and maintenance model resource deployment plan for a certain task is 4 hours, and the actual execution time period is from 08:00 to 13:00 (i.e., 5 hours). The difference between the resource deployment cycle and the operation cycle is further identified, that is, the deviation between the plan and the actual time. For example, the difference between the planned 4 hours and the actual 5 hours is 1 hour. At the same time, the difference information is sorted and labeled with the preset progress benchmark. The progress benchmark is the standard for measuring whether the task completion progress meets the expectations. For example, the progress benchmark can be set as "on time (difference ≤ 0)", "minor delay (0 < difference ≤ 1 hour)", "moderate delay (1 hour < difference ≤ 3 hours)" or "serious delay (difference > 3 hours)". Based on the benchmark, the progress of each task is sorted and marked. For example, the 1-hour difference of the above task will be marked as "minor delay". The information including the difference between resource deployment and operation cycle and the progress benchmark sorting and labeling is organized to generate the operation and maintenance progress monitoring structure index.

[0098] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A graph-driven nuclear energy operation and maintenance optimization method, characterized in that, Includes the following steps: S1: Obtain nuclear energy equipment operation data and original maintenance records, extract equipment nodes and their associations, construct an initial knowledge graph, mark the matching between equipment status and operating conditions, and generate an equipment health status assessment list; S2: Based on the equipment health status assessment list, select maintenance strategies that are suitable for the current working conditions, extract fault mode paths by combining semantic reasoning algorithms, and perform deviation calculation and classification processing with the benchmark working conditions to obtain maintenance strategy adaptability labels. S3: Call the maintenance strategy adaptability tag, extract the fault mode number, identify the range of sensor data fluctuation within the section, compare the radiation sensor monitoring period with the fault mode section, record the number of overlapping time periods, and generate a list of radiation sensor abnormal impact sections. S4: Based on the list of abnormally affected sections of the radiation sensor, filter out maintenance task nodes that are not disturbed, extract work order segment records and operation data, map task allocation and operation time distribution, determine whether there is an execution imbalance of tasks, and obtain work order task execution fluctuation groups.

2. The graph-driven nuclear energy operation and maintenance optimization method according to claim 1, characterized in that, The equipment health status assessment list includes equipment number, status label, operating condition deviation, and maintenance category. The maintenance strategy adaptability label includes adaptability level, fault type, benchmark comparison result, and task association number. The radiation sensor abnormality impact section list includes equipment type, impact time section, number of overlapping time periods, and affected task number. The work order task execution fluctuation group includes task distribution unevenness number, operation duration record, task completion deviation, and operation matching degree.

3. The graph-driven nuclear energy operation and maintenance optimization method according to claim 1, characterized in that, The specific steps for obtaining the equipment health status assessment list are as follows: S111: Acquire nuclear energy equipment operation data and original maintenance records, extract equipment nodes and their associations, match equipment status acquisition time with operating condition requirement time, compare time range with status requirements, and generate partition node operation record time periods; S112: Based on the time period of the partition node operation record, extract the overlapping time period between the equipment status and the working condition demand time interval, calculate the ratio of the overlapping time to the total working condition demand time, calculate the overlap ratio value, filter nodes with an overlap ratio lower than the benchmark value, and obtain the partition node status coverage deviation rate according to the number of equipment status labels. S113: Based on the deviation rate of the partition node status coverage, determine the deviation status of the node number, identify the node number whose deviation rate exceeds the node synchronization threshold, integrate the node number, status coverage information and deviation rate value, and generate a device health status assessment list.

4. The graph-driven nuclear energy operation and maintenance optimization method according to claim 3, characterized in that, The specific steps for obtaining the maintenance strategy adaptability tag are as follows: S211: Based on the equipment health status assessment list, identify maintenance strategies and fault mode paths that are adapted to the current working conditions, extract the start and end fluctuation ranges of the real-time sensors, calculate the start and end difference of the fluctuation range, compare it with the benchmark fluctuation range, and obtain the fluctuation deviation value. S212: Call the fluctuation deviation value, combine the segment distribution, fluctuation trend and adjustment frequency, uniformly collect the maintenance strategy deviation data, identify and calculate the adaptability deviation degree according to the segment number, determine the fluctuation direction according to the adjustment frequency, and obtain the maintenance strategy adaptability label.

5. The graph-driven nuclear energy operation and maintenance optimization method according to claim 4, characterized in that, The specific steps for obtaining the list of abnormally affected areas by the radiation sensor are as follows: S311: Call the maintenance strategy adaptability tag, filter the fluctuation deviation task segment number, extract the fluctuation time period according to the node association table, process the segment fluctuation time period according to the time dimension, identify the fluctuation time period index table, and obtain the fluctuation task time period set. S312: Based on the set of fluctuating task periods, collect the fluctuation range of sensor data in the same period, identify the radiation sensor impact table, determine the daily sensor interference based on the interference threshold, match it with the fluctuating task periods, determine whether there is a fluctuating abnormal correlation, and generate a list of radiation sensor abnormally affected areas.

6. The graph-driven nuclear energy operation and maintenance optimization method according to claim 5, characterized in that, The specific steps for obtaining the work order task execution fluctuation group are as follows: S411: Based on the list of abnormally affected sections of the radiation sensor, filter and maintain task nodes, extract the task list, and obtain the task set of nodes not affected by the radiation sensor. S412: Call the task set of nodes not affected by radiation sensor interference, match the daily segmented records and operation data of work orders, extract the planned task volume according to the task number, count the number of operators and the real-time operation time period, and generate a work order cooperation execution status matching dataset. S413: Based on the work order cooperation execution status matching dataset, evaluate the matching degree between task execution efficiency and operation distribution, identify efficiency fluctuation nodes, mark tasks with fluctuations exceeding the benchmark value as abnormal nodes, analyze work order task execution matching deviation, and obtain work order task execution fluctuation groups.

7. The graph-driven nuclear energy operation and maintenance optimization method according to claim 1, characterized in that, The method also includes step S5: S5: Call the work order task execution fluctuation group, extract the abnormal fluctuation task group, calculate the ratio of resource input to task volume in the operation and maintenance model, record the difference distribution between resource deployment cycle and task execution cycle, and generate operation and maintenance progress monitoring structure indicators. The operational progress monitoring structure indicators include resource allocation ratio, task intensity level, execution cycle difference, and operational efficiency indicators.

8. The graph-driven nuclear energy operation and maintenance optimization method according to claim 7, characterized in that, The specific steps for obtaining the operation and maintenance progress monitoring structure indicators are as follows: S511: Call the work order task execution fluctuation group, filter nodes that exceed the threshold, record the time interval and task volume change range, and obtain the progress fluctuation abnormal identifier set; S512: Based on the operation and maintenance model resources corresponding to the nodes in the progress fluctuation anomaly identifier set, identify the node resource input ratio sequence, extract the abnormal distribution interval and compare the critical coefficient, record the ratio offset direction and node number, and form an operation and maintenance resource matching offset index group. S513: Based on the operation and maintenance resource matching offset index group, extract the time period of operation and maintenance model resource deployment and task execution, identify the difference between the resource deployment cycle and the operation cycle, and sort and label them according to the progress benchmark to generate operation and maintenance progress monitoring structure indexes.

Images