Data processing method and system based on nuclear power intelligent monitoring system

Abstract

The invention provides a data processing method and system based on a nuclear power intelligent monitoring system, and the method comprises the steps: extracting original measurement point information from a database of a distributed control system as first measurement point information, and carrying out the preprocessing of the first measurement point information; matching the preprocessed first measuring point information with second measuring point information in a real-time information supervision system, and constructing a point table database according to a matching result; calculating the importance of each measuring point according to the multi-dimensional attribute of each measuring point in the point table database, and dynamically optimizing the sampling frequency according to the importance; and verifying the optimized point table database, and outputting a final point table database after the point table database is verified to be qualified. First measuring point information in a database of the distributed control system is extracted, preprocessed and then matched, so that a point table database can be constructed; system resources can be reasonably distributed through dynamic optimization of the sampling frequency; the verification mechanism further improves the accuracy and reliability of the point table database.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of nuclear power data processing, and particularly relates to a data processing method and system based on a nuclear power intelligent monitoring system. BACKGROUND

[0002] In the field of nuclear power, the naming of the measuring points on the side of the distributed control system (DCS) usually follows the control logic and hardware address, is short, abstract and not unified, and the side of the real-time information monitoring system (KMS) needs clear, standard and rich-semantic measuring point names for monitoring, reporting and analysis, therefore, it is necessary to build a point table database to establish the mapping relationship between the measuring points on the side of the distributed control system and the side of the real-time information monitoring system.

[0003] The prior art has the key technical problems of high artificial dependence, difficulty in guaranteeing multi-dimensional consistency, lack of dynamic maintenance, mapping failure caused by naming rule differences, lack of frequency optimization mechanism and imperfect version tracking system in the point table generation and management process, for example: (1) high artificial dependence: 100,000 measuring point information needs to be manually extracted from the database of the distributed control system, with an error screening rate of about 3% and 5% (nuclear power industry statistical data); (2) difficulty in guaranteeing multi-dimensional consistency: 12 parameters such as measuring point names, data types, units and collection frequencies on the side of the distributed control system and the side of the real-time information monitoring system need to be manually matched, and any mismatch of any parameter will lead to data quality abnormalities; (3) lack of dynamic maintenance: when the measuring points are changed after the upgrade of the distributed control system, manual checking is needed, which takes 40-80 man-days / once (taking a CPR1000 unit as an example). These problems seriously affect the efficiency, accuracy and dynamic maintenance capability of data collection of the intelligent monitoring system of the nuclear power plant.

[0004] Most importantly, the prior art lacks a technology for optimizing the sampling frequency of each measuring point in the point table, and only sets some fixed sampling frequencies for each measuring point during mapping, which is obviously not conducive to the allocation of system resources. SUMMARY

[0005] In view of the defects of the prior art, the present application provides a data processing method and system based on a nuclear power intelligent monitoring system to solve the technical problem that the existing nuclear power data processing cannot dynamically optimize the collection frequency of measuring points, resulting in unreasonable allocation of system resources.

[0006] To achieve the above object and other related objects, the present application provides a data processing method based on a nuclear power intelligent monitoring system, comprising: extracting original measuring point information from a database of a distributed control system as first measuring point information, and preprocessing the first measuring point information; matching the preprocessed first measuring point information with second measuring point information in a real-time information supervision system, and constructing a point table database according to a matching result; calculating an importance degree of each measuring point in the point table database according to multi-dimensional attributes of the measuring point, and dynamically optimizing a sampling frequency of the measuring point according to the importance degree; verifying the optimized point table database, and outputting a final point table database after verification.

[0007] In an embodiment of the present application, preprocessing the first measuring point information comprises: standardizing measuring point names in the first measuring point information according to a preset first rule; and filtering abnormal values in the first measuring point information according to a preset second rule.

[0008] In an embodiment of the present application, matching the preprocessed first measuring point information with the second measuring point information in the real-time information supervision system comprises: matching the preprocessed first measuring point information with the second measuring point information in the real-time information supervision system using a hybrid matching algorithm, the hybrid matching algorithm comprising at least two of rule-based matching, deep learning model-based matching, semantic similarity-based matching, and manual confirmation.

[0009] In an embodiment of the present application, matching the preprocessed first measuring point information with the second measuring point information in the real-time information supervision system comprises: performing word segmentation on measuring point descriptions in the first measuring point information to obtain a word segmentation result; judging whether the word segmentation result matches a preset regular expression: if yes, determining a corresponding second measuring point according to a matching result; if no, querying a nuclear power field knowledge graph according to the word segmentation result to determine a second measuring point candidate set to be matched; calculating cosine similarity between a semantic vector of the first measuring point information and semantic vectors of second measuring point information in the second measuring point candidate set; if a maximum cosine similarity exceeds a preset threshold, matching the first measuring point with a second measuring point having the maximum cosine similarity; and if the preset threshold is not exceeded, generating a recommended mapping relationship for manual confirmation.

[0010] In an embodiment of the present application, the multi-dimensional attributes comprise a safety level, a historical fluctuation rate, and a correlation logic diagram; and the importance degree of each measuring point in the point table database is calculated according to the multi-dimensional attributes of the measuring point, comprising: obtaining a fluctuation coefficient of each measuring point according to the historical fluctuation rate of the measuring point in the point table database; obtaining a correlation system number of each measuring point according to the correlation logic diagram of the measuring point in the point table database; and obtaining an importance degree of each measuring point according to the safety level, the fluctuation coefficient, and the correlation system number of the measuring point.

[0011] In an embodiment of the present application, the step of dynamically optimizing the sampling frequency according to the importance degree further comprises: setting a lower limit of the sampling frequency for the measuring points of a preset safety level according to the safety level of each measuring point; switching the measuring points with a fluctuation coefficient greater than a preset value to a sliding window sampling mode; and optimizing the combination of the collection frequencies of all measuring points by using a Monte Carlo simulation optimization algorithm, or a genetic algorithm, or a simulated annealing algorithm.

[0012] In an embodiment of the present application, the combination of the collection frequencies of all measuring points is optimized by using a Monte Carlo simulation optimization algorithm, and the optimization target is to minimize the communication load on the premise of ensuring that the data timeliness is greater than or equal to a preset value.

[0013] In an embodiment of the present application, after the point table database is constructed, the method further comprises: encoding each measuring point in the point table database according to a preset multi-level composite encoding rule, the encoding comprising one or more of a plant code, a unit number, a system code, a system serial number, a device type, a device serial number, a data type, and a check code, to ensure the uniqueness of the measuring point encoding.

[0014] In an embodiment of the present application, the optimized point table database is verified, comprising: verifying the static attributes of each measuring point in the point table database in the distributed control system and the real-time information monitoring system, the static attributes comprising a data type and an engineering unit, and if the engineering units are inconsistent but can be converted, a unit conversion coefficient is automatically injected.

[0015] In an embodiment of the present application, the optimized point table database is verified, comprising: performing data collection based on the point table database in a virtual collection sandbox, and dynamically verifying and tracking the quality bit state of the collected data, if the quality bit is abnormal, error tracing is performed by associating historical configuration records.

[0016] In an embodiment of the present application, the optimized point table database is verified, comprising: triggering a bidirectional verification protocol between the distributed control system and the real-time information monitoring system to verify the consistency of the cross-system data reading and writing functions and data, and if the protocol is found to be mismatched or the data is found to be inconsistent, an error is reported.

[0017] To achieve the above object and other related objects, the present application further provides a data processing system based on a nuclear power intelligent monitoring system, comprising: a data acquisition and preprocessing module, configured to extract original measuring point information from a database of a distributed control system as first measuring point information, and to preprocess the first measuring point information; a point table database construction module, configured to match the preprocessed first measuring point information with second measuring point information in a real-time information supervision system, and to construct a point table database according to a matching result; a sampling frequency optimization module, configured to calculate an importance of each measuring point in the point table database according to multi-dimensional attributes of the measuring point, and to dynamically optimize a sampling frequency of the measuring point according to the importance; and a verification module, configured to verify the optimized point table database, and to output a final point table database after verification.

[0018] In an embodiment of the present application, the sampling frequency optimization module comprises: an importance calculation unit, configured to calculate an importance of each measuring point in the point table database according to multi-dimensional attributes of the measuring point; a sampling mode adjustment unit, configured to switch a measuring point with a fluctuation coefficient greater than a preset value to a sliding window sampling mode; a first frequency optimization unit, configured to dynamically optimize a sampling frequency of each measuring point according to the importance; a second frequency optimization unit, configured to set a lower limit of the sampling frequency for a measuring point with a preset safety level according to the safety level of the measuring point; and a third frequency optimization unit, configured to optimize a collection frequency combination of all measuring points by using a Monte Carlo simulation optimization algorithm, or a genetic algorithm, or a simulated annealing algorithm.

[0019] In an embodiment of the present application, the verification module comprises: a first verification unit, configured to verify static attributes of each measuring point in the point table database in the distributed control system and the real-time information supervision system; a second verification unit, configured to perform data collection based on the point table database in a virtual collection sandbox, and to dynamically verify and track a data quality bit state of collected data; and a third verification unit, configured to trigger a bidirectional verification protocol between the distributed control system and the real-time information supervision system to verify a cross-system data read-write function and data consistency.

[0020] The present application has the following beneficial effects: the data processing method and system based on the nuclear power intelligent monitoring system can conveniently construct a point table database by extracting first measuring point information from a database of a distributed control system and preprocessing the first measuring point information, and then matching the preprocessed first measuring point information with second measuring point information in a real-time information supervision system, and can dynamically optimize a sampling frequency of each measuring point according to an importance of the measuring point after constructing the point table database, so as to reasonably allocate system resources; meanwhile, by introducing a verification mechanism, the accuracy and reliability of the point table database are further improved. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. The drawings herein are incorporated into the description and form a part of the description, show the embodiments consistent with the present application, and are used to explain the principles of the present application. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0022] Figure 1 The flow chart of the data processing method provided by an embodiment of the present application is shown in FIG. 1. Figure 2 The pre-processing flow chart of the first measuring point information provided by an embodiment of the present application is shown in FIG. 2. Figure 3 The matching flow chart of the measuring point provided by an embodiment of the present application is shown in FIG. 3. Figure 4 The flow chart of calculating the importance of the measuring point provided by an embodiment of the present application is shown in FIG. 4. Figure 5 The flow chart of optimizing the sampling frequency and mode provided by an embodiment of the present application is shown in FIG. 5. Figure 6 The schematic diagram of the data processing system provided by an embodiment of the present application is shown in FIG. 6. Figure 7 The schematic diagram of the sampling frequency optimization module provided by an embodiment of the present application is shown in FIG. 7. Figure 8 The schematic diagram of the verification module provided by an embodiment of the present application is shown in FIG. 8.

[0023] Legend: 601, data acquisition and pre-processing module; 602, point table database construction module; 603, sampling frequency optimization module; 604, verification module; 701, importance calculation unit; 702, sampling mode adjustment unit; 703, first frequency optimization unit; 704, second frequency optimization unit; 705, third frequency optimization unit; 801, first verification unit; 802, second verification unit; 803, third verification unit. DETAILED DESCRIPTION

[0024] The present application will be described in more detail by the following specific examples. Other advantages and embodiments of the present application will be more clearly understood from this description. The following examples and features in the examples can be combined, if not incompatible, to provide further embodiments of the application. Except for specific methods, devices, materials used in the examples, any methods, devices and materials similar or equivalent to those used in the examples of the present application can be used to implement the present application according to the knowledge of the skilled in the art and the description of the present application.

[0025] It should be understood that the terminology used herein is for the purpose of describing particular embodiments of the application only and is not intended to limit the scope of the present application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art.

[0026] In the following description, numerous specific details are discussed so as to provide a thorough understanding of embodiments of the application. However, it will be apparent to one of ordinary skill in the art that embodiments of the application can be practiced without these specific details. In some instances, well-known structures and devices are shown in block diagram form, rather than in detail, in order to avoid obscuring the embodiments of the application.

[0027] The flow diagrams and block diagrams in the drawings are schematic representations for purposes of illustration and description. They are not necessarily drawn to scale, and some of the representations have been simplified for clarity. It should be noted that in some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently or the blocks may be executed in the reverse order, depending upon the functionality involved.

[0028] See Figure 1 , Figure 1 A data processing method based on a nuclear power intelligent monitoring system according to an embodiment of the present application includes steps S101-S104.

[0029] Step S101, extract original measurement point information from the database of the distributed control system as first measurement point information, and pre-process the first measurement point information. The distributed control system is abbreviated as DCS. In this step, the original measurement point information is first extracted for subsequent processing.

[0030] See Figure 2 In an embodiment of the present application, the pre-processing of the first measurement point information includes steps S201 and S202.

[0031] Step S201, according to a preset first rule, standardize the measurement point name in the first measurement point information. The measurement point naming on the DCS side usually follows the control logic and hardware address, is short, abstract and not unified, for example, a pressure sensor may be named as 10PT001A (No. 10 cabinet, pressure transmitter, No. 001, etc.), and the engineer may be used to use psi (pound per square inch) as the pressure unit, in order to ensure more accurate matching later, a first rule can be preset according to the naming habit of the measurement point of the distributed control system, so as to standardize the first measurement point name, for example, convert "DCS_PT101" to the standard format "PRESSURE_TANK101".

[0032] Step S202, according to a preset second rule, filter the abnormal values in the first measurement point information. The abnormal values mainly include two types: one is semantic anomaly, for example, the engineering unit of the measurement point is not configured (the unit field is "NULL", "NA" or null), or the unit is obviously not in line with the physical specification (such as the pressure unit is incorrectly marked as "℃"); the second is configuration anomaly, for example, the upper and lower limits of the measurement point range are set incorrectly (the upper limit is less than the lower limit), or the communication address format is illegal. If such abnormal data is not removed, it will seriously damage the data quality and cause the matching process to fail or produce systematic errors.

[0033] Step S102, match the pre-processed first measurement point information with the second measurement point information in the real-time information supervision system, and construct a point table database according to the matching result.

[0034] In an embodiment of the present application, the pre-processed first measurement point information and the second measurement point information in the real-time information supervision system are matched, including: the pre-processed first measurement point information and the second measurement point information in the real-time information supervision system are matched by using a hybrid matching algorithm, and the hybrid matching algorithm includes at least two of rule-based matching, deep learning model-based matching, semantic similarity-based matching and manual confirmation.

[0035] In the hybrid matching algorithm, first, automatic matching is performed based on some rules or models, and then manual confirmation is introduced for the points that cannot be matched, which can greatly reduce the workload of manual work and avoid the many shortcomings of pure manual matching and checking mentioned in the background art.

[0036] See Figure 3 In an embodiment of the present application, the first point information after preprocessing and the second point information in the real-time information monitoring system are matched, including steps S301-S304.

[0037] Step S301, the point description in the first point information is processed by word segmentation to obtain a word segmentation result. The point description in the first point information can include multiple words, so in this step, it is first processed by word segmentation, for example, a certain point description is: “RESSURE_MAIN_STEAM”, after word segmentation, the three words “PRESSURE”, “MAIN” and “STEAM” can be obtained.

[0038] Step S302, it is judged whether the word segmentation result matches the preset regular expression: if yes, the corresponding second point is determined according to the matching result; if no, the word segmentation result is queried in the nuclear power field knowledge graph to determine the second point candidate set to be matched. In this step, the regular expression is a rule-based matching, and the first matching is realized by constructing the regular expression. The rule-based matching is placed in the first step because this matching method is fast and accurate, and the disadvantage is that it is not flexible enough.

[0039] In this step, if the rule matching is not successful, the intelligent matching process based on semantics is started. This process first takes the key terms (such as “pressure”, “main steam”) in the word segmentation result as a query input to search the nuclear power field knowledge graph constructed in advance. This knowledge graph integrates system architecture, device coding, functional logic and other field knowledge, and can understand the semantic relationship between concepts (such as “main steam” belongs to “secondary circuit system”, and “pressure” is a kind of attribute of “main steam”). Through graph reasoning, the KMS standard point set related to the semantics of the DCS point, i.e. the “second point candidate set”, can be accurately located, so that the large-scale global matching problem is transformed into a small-range accurate screening problem, laying a foundation for the subsequent similarity calculation.

[0040] Step S303, cosine similarity between the semantic vector of the first measuring point information and the semantic vector of each second measuring point information in the second measuring point candidate set is calculated. In this step, the DCS measuring point description and the KMS candidate measuring point description are first mapped into semantic vectors in a high-dimensional space by using a pre-trained natural language processing model, and the vectors can capture the deep semantic features of the words in the measuring point name and their context association. Then, the cosine similarity algorithm is used to calculate the cosine value of the included angle between the vectors, so as to quantify the semantic association strength. The advantages of this scheme are as follows: firstly, it can effectively overcome the sensitivity of traditional keyword matching to wording variants and abbreviation word sequences, and realize intelligent matching of "synonyms with different words" (such as "PRESSURE_MAIN_STEAM" and "main steam pressure"); secondly, the similarity calculation is performed based on the candidate set filtered by the knowledge graph, which has both flexibility of semantic matching and constraint of domain knowledge, greatly improves the matching range and accuracy, and significantly reduces the computational complexity.

[0041] Step S304, if the maximum cosine similarity exceeds a preset threshold, the first measuring point is matched with the second measuring point with the maximum cosine similarity; if the preset threshold is not exceeded, a recommended mapping relationship is generated for manual confirmation. In this step, the preset threshold can be set to 0.85, for example. When the cosine similarity between the semantic vector of a certain first measuring point information and the semantic vector of one or more second measuring point information exceeds the threshold, it indicates that there is a matching relationship, but since the measuring points are one-to-one, the first measuring point is matched with the second measuring point with the maximum cosine similarity. If all the cosine similarities are less than the preset threshold, it means that there is no matching relationship, and manual confirmation is needed at this time.

[0042] When manual confirmation is performed, in order to reduce the artificial burden, multiple groups of recommended mapping relationships can be generated for the unmatched items for the final confirmation and modification by the artificial. Since the first matching has been performed by using the regular expression and the second matching has been performed by using the semantic vector cosine similarity based on the knowledge graph, the remaining unmatched items are very few, and even if the artificial confirmation is introduced, it will not increase too much workload, but can improve the accuracy of the matching.

[0043] Further, after the manual confirmation, the knowledge graph can be updated according to the result of the manual confirmation, so as to improve the self-updating ability of the knowledge graph and lay a foundation for subsequent accurate and automatic matching.

[0044] Step S103, the importance of each measuring point in the point table database is calculated according to the multi-dimensional attributes of the measuring point, and the sampling frequency is dynamically optimized according to the importance.

[0045] Please refer to Figure 4In a specific embodiment of the present application, the multi-dimensional attributes include a security level, a historical fluctuation rate, and a correlation logic diagram, and can also include a measurement point ID, a physical quantity type, a subsystem to which the measurement point belongs, and the like.

[0046] With the above multi-dimensional attributes, the importance of each measurement point in the point table database can be calculated according to the multi-dimensional attributes, and the calculation specifically includes steps S401-S403.

[0047] In step S401, a fluctuation coefficient of each measurement point is obtained according to a historical fluctuation rate of each measurement point in the point table database. In this step, time series data of each measurement point in a certain period is first extracted from a power plant historical database, and the historical fluctuation rate is quantified by calculating a standard deviation or a variance to objectively reflect the stability of the measurement point value. Then, in order to eliminate the influence of different physical quantity dimensions and magnitudes, the absolute fluctuation rate is converted into a dimensionless relative index between 0 and 1, i.e., the fluctuation coefficient, by using a minimum-maximum normalization method or the like. The advantages of this scheme are as follows: first, the absolute fluctuation information which is difficult to compare directly is converted into a standardized characteristic parameter, which provides a scientific basis for subsequent unified importance evaluation across measurement points; and second, the system can automatically identify and distinguish between slowly varying parameters and rapidly varying parameters, which lays a data-driven foundation for subsequent dynamic optimization of sampling frequency and avoids the blindness of resource allocation.

[0048] In step S402, a correlation system number of each measurement point is obtained according to a correlation logic diagram of each measurement point in the point table database. Specifically, by analyzing the relationship of the measurement point and the control logic diagram, the entities such as measurement points, devices, and subsystems and their belonging and connection relationship are structured and stored in a graph database, and a complete power plant topology knowledge network is constructed. Then, for each specific measurement point, the upstream system and the downstream system connected to the device to which the measurement point belongs are traversed by using a graph query algorithm such as breadth-first search, and the total number of independent systems directly or indirectly associated with the measurement point is counted, i.e., the correlation system number of the measurement point. The advantages of this scheme are as follows: first, the static topology hidden in the drawing is converted into a dynamic relationship model that can be calculated and queried, and the quantification of system correlation is realized; and second, key measurement points that bear interfaces or monitoring functions among multiple systems (such as cross-system flow meters and shared heat exchanger temperatures) can be accurately identified, and higher acquisition and monitoring priorities can be given to the key measurement points, so that resource allocation is optimized.

[0049] Step S403, according to the security level, fluctuation coefficient and associated system number of each measuring point, the importance of each measuring point is obtained. Specifically, the following formula can be used for calculation: importance = a x security level + b x fluctuation coefficient + g x associated system number, wherein a, b and g are weight coefficients based on the experience of domain experts or the analytic hierarchy process (AHP) and are used for quantifying the influence of safety, real-time and system association on the collection strategy. The advantages of this scheme are as follows: firstly, through multi-dimensional attribute fusion and weighted comprehensive, the qualitative engineering experience is converted into quantitative decision index, the one-sidedness of single index evaluation is overcome, and the systematization and scientization of collection importance evaluation are realized; secondly, the weight coefficients can be flexibly adjusted for different units or system types, so that the importance model has good scalability and adaptability; thirdly, it provides direct and objective calculation basis for the differentiated dynamic configuration of collection frequency in the subsequent process, so as to ensure the real-time of key data and effectively improve the resource utilization efficiency of the whole data collection system.

[0050] In the above steps, the sampling frequency optimization according to the importance is generally a fixed rule, for example, the measuring points can be divided into multiple levels (high importance, medium importance and low importance) according to the importance, and then each level corresponds to a sampling frequency.

[0051] Please refer to Figure 5 In a specific embodiment of the present application, the step of dynamically optimizing the sampling frequency according to the importance further includes steps S501-S503.

[0052] Step S501, according to the security level of each measuring point, setting the lower limit of the sampling frequency for the measuring points with the preset security level. For example, for the measuring points related to a certain security level, the default sampling interval is improved to 500ms level, that is, the sampling frequency of such measuring points cannot be lower than 2 seconds / time.

[0053] Step S502, switching the measuring points with fluctuation coefficient greater than the preset value to the sliding window sampling mode. This is a special processing rule, which is aimed at the measuring points with abnormal rapid changes, and needs to change the sampling mode (not only the frequency) to capture transient changes.

[0054] Step S503, using a Monte Carlo simulation optimization algorithm, or a genetic algorithm, or a simulated annealing algorithm, to optimize the collection frequency combination of all measurement points. After the processing of the previous steps, a preliminary frequency allocation scheme can be obtained (based on importance ranking and specific rules), but this allocation scheme is based on the independent calculation of each measurement point, and thousands of "local optimal" decisions are superimposed together, which may not constitute a "global optimal" system, and the total communication load may still be high. The role of Monte Carlo simulation (or genetic algorithm, annealing algorithm) is to start from the preliminary scheme and perform global search and fine-tuning.

[0055] In a specific embodiment of the present application, in step S503, the Monte Carlo simulation optimization algorithm is used to optimize the collection frequency combination of all measurement points, and the optimization target is to minimize the communication load under the premise of ensuring that the data timeliness is greater than or equal to a preset value. The Monte Carlo simulation optimization algorithm can intelligently and tentatively reduce the frequency of non-core measurement points in the "importance" under the premise of ensuring global data timeliness (≥99.9%), such as relaxing the collection frequency of a measurement point with low importance from 2 seconds to 5 seconds, and checking whether the performance requirements can still be met. Ultimately, it will find a frequency combination scheme that can meet all performance requirements and minimize the total communication load. This step belongs to global optimization, that is, considering the sampling frequency of all measurement points as a whole.

[0056] In a specific embodiment of the present application, after constructing the point table database, it further includes: encoding each measurement point in the point table database according to a preset multi-level composite encoding rule, the encoding including one or more of a plant area code, a unit number, a system code, a system serial number, a device type, a device serial number, a data type, and a check code, to ensure the uniqueness of the measurement point encoding. The reason for encoding each measurement point in the point table database is that, first, to avoid confusion, that is, there may be measurement points with the same name in different units and different systems, which can be distinguished by encoding; second, by forcibly embedding the plant area code, the unit number, and the system code, it is ensured that the ID of each measurement point in the whole plant is absolutely unique; third, intelligent encoding itself is a data specification, and any engineer or system can understand and interpret the information that it belongs to which plant area and which unit, etc.

[0057] Step S104, verifying the optimized point table database, and outputting the final point table database after passing the verification.

[0058] In a specific embodiment of the present application, the optimized point table database is verified, including: verifying the static attributes of each measuring point in the distributed control system and the real-time information monitoring system in the point table database, the static attributes including data type and engineering unit, if the engineering units are inconsistent but can be converted (for example, the DCS side unit is "psi" and the KMS standard unit is "MPa"), then the unit conversion coefficient (for example, the proportion factor 0.00689476) is automatically injected. The advantages of this scheme are: first, it realizes the automatic compatible processing of unit system differences, avoids data analysis errors or monitoring deviations caused by inconsistent units, and significantly improves the standardization and usability of data; second, by dynamically injecting conversion rules instead of directly modifying the original data, the original characteristics of the data source are preserved, and the requirements of the upper system for data consistency are met, embodying the intelligent principle of "source specification, process controllable" in the data governance process.

[0059] In a specific embodiment of the present application, the optimized point table database is verified, including: in the virtual acquisition sandbox, based on the point table database, data acquisition is performed, and the quality bit state of the acquired data is dynamically verified and tracked, if the quality bit is abnormal (such as communication interruption, sensor failure, etc.), then the historical configuration record is associated to trace the error. This embodiment can actively find intermittent and configuration-type hidden problems before deployment by simulating the real acquisition environment and monitoring the data quality bit, effectively preventing abnormal data from flowing into the upper application system; it can also intelligently trace the historical configuration record, quickly locate the fault root cause (such as communication address modification error, acquisition card failure or signal line interference), greatly improve the diagnosis efficiency of operation and maintenance personnel, shorten the system debugging cycle, and embody the advanced management concept from passive alarm to proactive predictive maintenance.

[0060] In a specific embodiment of the present application, the optimized point table database is checked, including: triggering a two-way checking protocol between the distributed control system and the real-time information monitoring system to verify the consistency of the cross-system data reading and writing functions and data, specifically, by simulating in a virtual collection sandbox that the KMS system sends a control instruction or a set value modification request to the DCS system, and real-time acquires the corresponding feedback data on the DCS side, and compares the consistency of the instruction value and the actual feedback value in value, timing and state bit. If it is found that the protocol does not match (such as register address resolution error) or the data is inconsistent (such as the deviation of the written value and the read-back value is out of limit), an error is reported or a detailed report containing error type, positioning information and repair suggestion is automatically generated. By actively triggering cross-system two-way interaction, the integrity of the data link and the compatibility of the communication protocol under the point table configuration are deeply verified, which can discover deep integration defects that cannot be covered by one-way monitoring in advance; the traditional manual point-by-point verification is upgraded to automated closed-loop testing, which greatly improves the checking efficiency and coverage, especially suitable for acceptance testing of large-scale point tables in high-reliability scenarios such as nuclear power plants, significantly reducing the operation risk after the system is put into operation.

[0061] It should be noted that the step division of the above methods is only for the purpose of clear description, and when implemented, one step can be combined or some steps can be split and decomposed into multiple steps, as long as the same logical relationship is included, all within the protection scope of the present application; adding insignificant modifications or introducing insignificant designs in the algorithm or process, but not changing the core design of the algorithm and process, are within the protection scope of the patent.

[0062] Please refer to Figure 6 , Figure 6 is a data processing system based on a nuclear power intelligent monitoring system provided by an embodiment of the present application, including a data acquisition and preprocessing module 601, a point table database construction module 602, a sampling frequency optimization module 603, and a checking module 604. The data acquisition and preprocessing module 601 is used to extract original measurement point information from the database of the distributed control system as first measurement point information, and pre-process the first measurement point information; the point table database construction module 602 is used to match the pre-processed first measurement point information with second measurement point information in the real-time information monitoring system, and construct a point table database according to the matching result; the sampling frequency optimization module 603 is used to calculate the importance of each measurement point in the point table database according to its multi-dimensional attributes, and dynamically optimize its sampling frequency according to the importance; the checking module 604 is used to check the optimized point table database, and output the final point table database after passing the check.

[0063] It should be noted that the data processing system of the embodiment corresponds to the data processing method described above, and the functional modules in the data processing system correspond to the respective steps in the data processing method. The data processing system of the embodiment can be implemented in cooperation with the data processing method, that is, in the case of no conflict, the technical details mentioned in the data processing method of the above embodiment can also be applied to the data processing system of the embodiment.

[0064] Referring to Figure 7 In an embodiment of the present application, the sampling frequency optimization module 603 includes an importance calculation unit 701, a sampling mode adjustment unit 702, a first frequency optimization unit 703, a second frequency optimization unit 704, and a third frequency optimization unit 705. The importance calculation unit 701 is configured to calculate the importance of each measurement point in the point table database according to the multi-dimensional attributes of the measurement point. The sampling mode adjustment unit 702 is configured to switch the measurement point with a fluctuation coefficient greater than a preset value to a sliding window sampling mode. The first frequency optimization unit 703 is configured to dynamically optimize the sampling frequency of each measurement point according to the importance. The second frequency optimization unit 704 is configured to set a lower limit of the sampling frequency for the measurement point with a preset safety level according to the safety level of each measurement point. The third frequency optimization unit 705 is configured to optimize the combination of the collection frequencies of all measurement points by using a Monte Carlo simulation optimization algorithm, a genetic algorithm, or a simulated annealing algorithm. The functions of these units have been described in detail in the description of the method, and will not be repeated here.

[0065] Referring to Figure 8 In an embodiment of the present application, the verification module 604 includes a first verification unit 801, a second verification unit 802, and a third verification unit 803. The first verification unit 801 is configured to verify the static attributes of each measurement point in the point table database in the distributed control system and the real-time information monitoring system. The second verification unit 802 is configured to perform data collection based on the point table database in the virtual collection sandbox, and dynamically verify and track the data quality bit state. The third verification unit 803 is configured to trigger a bidirectional verification protocol between the distributed control system and the real-time information monitoring system to verify the consistency of the cross-system data reading and writing function and the data. The functions of these units have been described in detail in the description of the method, and will not be repeated here.

[0066] In summary, the present application solves the key technical problems in the prior art, such as high dependence on manual operation, difficulty in ensuring multi-dimensional consistency, lack of dynamic maintenance, mapping failure caused by naming rule differences, lack of frequency optimization mechanism, and imperfect version tracking system, through the dynamic metadata matching, dynamic sampling frequency optimization, and point table point name intelligent verification and self-repair system. The specific technical effects are as follows.

[0067] Significantly reduce artificial dependence and improve data quality. Through a hybrid matching algorithm based on regular expressions and semantic similarity, combined with a nuclear power field knowledge graph, intelligent matching of DCS and KMS measurement points is achieved. This algorithm can automatically extract and match measurement point information, significantly reducing the need for manual intervention and improving data quality, reducing the manual screening error rate from 3%-5% to nearly 0%.

[0068] Ensure multi-dimensional consistency and reduce data quality anomalies. By establishing a three-level verification mechanism (static verification, dynamic verification, and closed-loop verification), data type and unit consistency are ensured, quality position anomalies are tracked and associated with historical configuration records to locate error sources, and cross-system data consistency verification is carried out, fundamentally solving the problem of multi-dimensional consistency and reducing the occurrence of data quality anomalies.

[0069] Improve dynamic maintenance efficiency and achieve intelligent tracking and repair. By developing a point table point name intelligent verification and self-repair system, through virtual acquisition sandbox and historical data playback testing, real-time monitoring and self-repair of point table health are achieved. At the same time, a perfect point table version change tracking system is established, significantly improving the dynamic maintenance efficiency, reducing the maintenance time from 40-80 person-days to 10-20 person-days.

[0070] Solve naming rule differences and improve automation level. Through a hybrid matching algorithm based on regular expressions and semantic similarity, combined with a nuclear power field knowledge graph, the problem of automatic mapping failure caused by differences in DCS and KMS metadata naming rules is effectively solved, improving the automation level of data acquisition.

[0071] Achieve dynamic optimization of acquisition frequency and optimize resource utilization. An adaptive frequency adjustment mechanism based on measurement point importance classification is established, combining Monte Carlo simulation to optimize acquisition frequency combination, ensuring data timeliness (≥99.9%) while minimizing communication load, achieving optimal resource allocation, reducing resource waste, and improving data acquisition efficiency.

[0072] Perfect version tracking system and improve management efficiency. By developing a closed-loop verification system, supporting historical data playback testing and stress testing, and automatically generating a "point table health report", intelligent tracking and management of point table version changes are achieved, avoiding the problems of point table management confusion and version tracing difficulty.

[0073] The above examples only exemplarily illustrate the principles and effects of the present application, and are not intended to limit the present application. Any person skilled in the art can modify or change the above examples without departing from the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by those skilled in the art without departing from the spirit and technical thought disclosed by the present application shall be covered by the claims of the present application.

Claims

1. A data processing method based on a nuclear power intelligent monitoring system, characterized in that, The method comprises the following steps: extracting raw point information from a database of a distributed control system as first point information, and preprocessing the first point information; matching the preprocessed first point information with second point information in a real-time information supervision system, and constructing a point table database according to a matching result; calculating an importance degree of each point in the point table database according to multi-dimensional attributes of the point, and dynamically optimizing a sampling frequency of the point according to the importance degree; verifying the optimized point table database, and outputting a final point table database after passing the verification.

2. The data processing method based on the nuclear power intelligent monitoring system according to claim 1, characterized in that, The preprocessing of the first point information comprises the following steps: standardizing a point name in the first point information according to a preset first rule; filtering an abnormal value in the first point information according to a preset second rule.

3. The data processing method based on the nuclear power intelligent monitoring system according to claim 1, characterized in that, The matching of the preprocessed first point information with the second point information in the real-time information supervision system comprises the following steps: matching the preprocessed first point information with the second point information in the real-time information supervision system by using a hybrid matching algorithm, the hybrid matching algorithm comprising at least two of rule-based matching, deep learning model-based matching, semantic similarity-based matching, and manual confirmation.

4. The data processing method based on the nuclear power intelligent monitoring system according to claim 3, characterized in that, The matching of the preprocessed first point information with the second point information in the real-time information supervision system comprises the following steps: performing word segmentation processing on a point description in the first point information to obtain a word segmentation result; judging whether the word segmentation result matches a preset regular expression: if yes, determining a corresponding second point according to a matching result; if no, querying a nuclear power field knowledge graph according to the word segmentation result to determine a second point candidate set to be matched; calculating a cosine similarity between a semantic vector of the first point information and semantic vectors of second point information in the second point candidate set; if the maximum cosine similarity exceeds a preset threshold, matching the first point with a second point having the maximum cosine similarity; if the preset threshold is not exceeded, generating a recommended mapping relationship for manual confirmation.

5. The data processing method based on the nuclear power intelligent monitoring system according to claim 1, characterized in that, The multi-dimensional attributes comprise a safety level, a historical fluctuation rate, and a correlation logic diagram; The calculation of the importance degree of each point in the point table database according to the multi-dimensional attributes of the point comprises the following steps: obtaining a fluctuation coefficient of each point according to the historical fluctuation rate of each point in the point table database; obtaining a correlation system number of each point according to the correlation logic diagram of each point in the point table database; obtaining an importance degree of each point according to the safety level, the fluctuation coefficient, and the correlation system number of each point.

6. The data processing method based on the nuclear power intelligent monitoring system according to claim 5, characterized in that, The step of dynamically optimizing the sampling frequency of each point according to the importance degree further comprises the following steps: setting a sampling frequency lower limit for a point of a preset safety level according to the safety level of each point; switching a point whose fluctuation coefficient is greater than a preset value to a sliding window sampling mode; optimizing a collection frequency combination of all points by using a Monte Carlo simulation optimization algorithm, a genetic algorithm, or a simulated annealing algorithm.

7. The data processing method based on the nuclear power intelligent monitoring system according to claim 6, characterized in that, The optimization of the collection frequency combination of all points by using the Monte Carlo simulation optimization algorithm has an optimization target of minimizing communication load on the premise of ensuring that data timeliness is greater than or equal to a preset value.

8. The data processing method based on the nuclear power intelligent monitoring system according to claim 1, characterized in that, The point table database construction further comprises: According to a preset multi-level composite coding rule, each measurement point in the point table database is coded, and the coding includes one or more of a plant code, a unit number, a system code, a system serial number, a device type, a device serial number, a data type, and a check code, to ensure the uniqueness of the measurement point coding.

9. The data processing method based on the nuclear power intelligent monitoring system according to claim 1, characterized in that, The optimized point table database is checked, including: The static attributes of each measurement point in the point table database in the distributed control system and the real-time information supervision system are checked, and the static attributes include data type and engineering unit. If the engineering units are inconsistent but can be converted, the unit conversion coefficient is automatically injected.

10. The data processing method based on the nuclear power intelligent monitoring system according to claim 1, characterized in that, The optimized point table database is checked, including: In a virtual acquisition sandbox, data acquisition is performed based on the point table database, and dynamic checking and tracking of the quality bit state of the acquired data are performed. If the quality bit is abnormal, error tracing is performed based on historical configuration records.

11. The data processing method based on the nuclear power intelligent monitoring system according to claim 1, characterized in that, The optimized point table database is checked, including: A bidirectional checking protocol between the distributed control system and the real-time information supervision system is triggered to verify the cross-system data read-write function and data consistency. If the protocol does not match or the data is inconsistent, an error is reported.

12. A data processing system based on a nuclear power plant intelligent monitoring system, characterized by, It comprises: a data acquisition and preprocessing module for extracting raw measurement point information from the database of the distributed control system as first measurement point information and preprocessing the first measurement point information; a point table database construction module for matching the preprocessed first measurement point information with second measurement point information in the real-time information supervision system and constructing a point table database according to the matching result; a sampling frequency optimization module for calculating the importance of each measurement point in the point table database according to its multi-dimensional attributes and dynamically optimizing its sampling frequency according to the importance; and a checking module for checking the optimized point table database and outputting the final point table database after passing the check.

13. The data processing system based on nuclear power intelligent monitoring system according to claim 12, characterized in that, The sampling frequency optimization module comprises: an importance calculation unit for calculating the importance of each measurement point in the point table database according to its multi-dimensional attributes; a sampling mode adjustment unit for switching the measurement points with a fluctuation coefficient greater than a preset value to a sliding window sampling mode; a first frequency optimization unit for dynamically optimizing the sampling frequency of each measurement point according to the importance; a second frequency optimization unit for setting a lower limit of the sampling frequency for measurement points of a preset safety level according to the safety level of each measurement point; and a third frequency optimization unit for optimizing the combination of the acquisition frequencies of all measurement points using a Monte Carlo simulation optimization algorithm, a genetic algorithm, or a simulated annealing algorithm.

14. The data processing system based on nuclear power intelligent monitoring system according to claim 12, characterized in that, The checking module comprises: a first checking unit for checking the static attributes of each measurement point in the point table database in the distributed control system and the real-time information supervision system; a second checking unit for performing data acquisition in a virtual acquisition sandbox based on the point table database and dynamically checking and tracking the quality bit state of the acquired data; The third checking unit is configured to trigger a bidirectional checking protocol between the distributed control system and the real-time information monitoring system to verify consistency of cross-system data reading and writing functions and data.

Images