KKS code-based power plant equipment two-dimensional code standardized generation and analysis management system

Abstract

The application discloses a power plant equipment two-dimensional code standardization generation and analysis management system based on KKS coding, comprising a server and a mobile terminal. The data preprocessing module of the server carries out semantic vectorization processing on the KKS coding and generates a dynamic hash index; the two-dimensional code coding module adopts a non-uniform redundancy distribution mode, distributes the core semantic section to the two-dimensional code center high-redundancy protection area, and distributes a lower redundancy factor or a predictive error correction bit to the edge area; the data feedback module returns adaptive display content based on an identity tag. The mobile terminal analysis module preferentially decodes the center area to extract the system attribution vector in an offline state, combines a local process logic library and auxiliary positioning information to restore the equipment information through probability matching recursion; meanwhile, the dynamic hash index and the identity tag can be sent to the server to obtain online display content. The application significantly improves the recognition rate under oil stain shielding, realizes offline analysis, reduces the label reprint cost, and supports role differentiation information display.

Description

Technical Field

[0001] This invention relates to the field of electronic information technology, specifically to a standardized generation and parsing management system for QR codes of power plant equipment based on KKS encoding. Background Technology

[0002] KKS coding (Kraftwerk-Kennzeichen System, Power Plant Identification System) is an internationally recognized standard for power plant equipment coding, developed by the German VGB (Very Power Group). It is widely used throughout the entire lifecycle of power plants, including thermal, nuclear, and wind power plants, encompassing design, construction, operation, and asset management. KKS coding uses a hierarchical and structured approach to uniquely identify systems, equipment, components, and installation locations within a power plant. For example, "10GBA10CE001" can be interpreted as: Unit 10, GBA feedwater pump system, Equipment 10, and CE001 electric valve. Based on KKS coding, power plants can achieve unified association and sharing of equipment information across different systems (such as ERP, EAM, and SCADA).

[0003] Currently, QR code technology has been initially applied in power plant equipment management. The typical approach is to directly encode the equipment's KKS code or a Uniform Resource Locator (URL) containing the KKS code into a standard QR code (such as a QR code), print it as a label, and affix it to the equipment surface. Inspection personnel use mobile devices to scan the QR code, retrieving equipment information by being redirected to a pre-set webpage or querying a backend database. However, this traditional solution has the following significant technical problems in actual power plant environments:

[0004] (i) Information Silos and Non-Standardized Management. In existing systems, QR codes typically store static URLs or fixed KKS strings, which point to specific resources on the backend server. Once the backend server address changes, equipment information is updated, or the system is upgraded, all printed QR code labels become invalid, requiring large-scale reprinting and replacement of labels, resulting in high management costs and maintenance burdens. At the same time, different manufacturers and power plants use QR codes in different ways, lacking a unified standardized solution, making it difficult to share equipment information across systems and power plants.

[0005] (II) Poor robustness of QR codes in harsh power plant environments. Power plant sites generally experience severe conditions such as high temperatures, oil contamination, dust, vibration, and electromagnetic interference. Ordinary standard QR codes (such as QR codes) use uniform error correction coding (such as Reed-Solomon codes), and although their error correction levels can be selected between L (7%), M (15%), Q (25%), and H (30%), they are all globally uniformly distributed. In actual working conditions, oil contamination often spreads from the edge of the label and gradually erodes inward. Experiments show that ordinary QR codes are very prone to failure when the edge is obscured or physically damaged by more than 25%, resulting in the inability to decode. For critical equipment, this low robustness may delay inspections or operations, and even cause safety hazards.

[0006] (III) Access Bottlenecks in Weak or No-Network Environments. Many areas within power plants suffer from severe signal shielding, such as boiler rooms, cable trays, deep within turbine halls, and underground pipe corridors. The traditional "scan-connect-to-webpage" model heavily relies on real-time network connectivity. In the absence of or with weak network access, mobile terminals cannot query the backend database, and inspection personnel are completely unable to obtain any process information about the equipment (such as equipment name, system affiliation, and safety precautions). This forces inspection work to be interrupted, or requires returning to areas with signal coverage to consult paper diagrams, severely reducing maintenance efficiency.

[0007] (iv) Lack of granular access control and limited information display. In existing QR code solutions, all scanners see the same information page. However, the needs of different personnel in a power plant (operators, maintenance personnel, and safety managers) for equipment information are vastly different: operators focus on real-time operating parameters (pressure, temperature, current), maintenance personnel focus on equipment maintenance records, spare parts, and disassembly drawings, and safety managers focus on the status of hot work permits and the distribution of hazards. Existing technology cannot automatically switch the displayed content based on the scanner's identity, leading to either information overload (too much irrelevant information for certain personnel) or information insufficiency (missing key information), and also makes it difficult to achieve secure isolation of production data.

[0008] To address the aforementioned issues, some research has attempted improvements, such as using QR codes with higher error correction levels, employing specialized anti-fouling label materials, or configuring multiple data interfaces in the background. However, these solutions either only address a single problem (e.g., merely improving damage resistance) or introduce new drawbacks (e.g., higher error correction levels significantly increase QR code size, making them unsuitable for space-constrained device labels). Therefore, a comprehensive technical solution is urgently needed that can simultaneously resolve information silos, environmental robustness, weak network access, and differentiated permission display. Summary of the Invention

[0009] To address the aforementioned technical problems, this invention provides a standardized generation and parsing management system for QR codes of power plant equipment based on KKS encoding.

[0010] This invention is achieved through the following technical solution:

[0011] The present invention provides a standardized generation and parsing management system for QR codes of power plant equipment based on KKS encoding, including a server and a mobile terminal that communicates with the server via a wireless network or a wired connection.

[0012] The aforementioned server-side components include:

[0013] The data preprocessing module is configured to obtain the original KKS code of the equipment from the power plant equipment management system, and perform semantic vectorization processing on the original KKS code to generate the corresponding feature vector and dynamic hash index, so as to achieve data compression and logical decoupling. The dynamic hash index is the hash value obtained by XORing the original KKS code with a dynamic offset and then performing hash calculation. The dynamic offset is data containing timestamp and equipment serial number.

[0014] The QR code encoding module, connected to the aforementioned data preprocessing module, is configured to receive the feature vector and dynamic hash index output by the aforementioned data preprocessing module, and generate a QR code image based on the aforementioned feature vector and dynamic hash index; wherein, when generating the error correction encoding matrix of the aforementioned QR code image, a non-uniform redundancy allocation method is adopted, the data stream corresponding to the core semantic segment in KKS encoding, which includes at least the functional identifier segment and the system affiliation segment, is allocated to the high redundancy protection zone in the center of the aforementioned QR code image, while a lower redundancy factor is allocated to the edge region of the aforementioned QR code image or predictive error correction bits generated based on KKS logical topology are introduced;

[0015] The data feedback module is connected to the QR code encoding module and is configured to receive the parsing request sent by the mobile terminal. The parsing request carries the identity tag of the scanner and the dynamic hash index obtained from the QR code image decoding. Based on the identity tag and the dynamic hash index, the data feedback module matches and returns the display content adapted to the identity tag from the multidimensional dataset to the mobile terminal.

[0016] The aforementioned mobile terminal includes a mobile terminal parsing module, configured as follows:

[0017] In offline mode, the QR code image obtained by scanning is parsed, and the central high redundancy protection zone is decoded first to extract the system attribution vector. Based on the system attribution vector, the local process logic library built based on KKS encoding rules and auxiliary positioning information pre-stored on the mobile terminal, the complete equipment information and process association data are recursively restored through probabilistic matching based on KKS hierarchical topology. When the edge area of ​​the QR code image is damaged and the equipment number cannot be directly decoded, the recursive restoration uses the decoded system attribution vector and the auxiliary positioning information to lock the unique matching equipment node from the local process logic library.

[0018] Furthermore, the aforementioned mobile terminal parsing module is also configured to send the decoded dynamic hash index along with the scanner's identity tag to the aforementioned server's data feedback module, and receive the display content returned by the aforementioned data feedback module.

[0019] This invention discloses a standardized generation and parsing management system for QR codes of power plant equipment based on KKS encoding. Its core design concept lies in the deep coupling of the power plant identification system with QR code technology. KKS encoding is an internationally recognized power plant identification system that employs a hierarchical structure to uniquely identify equipment from multiple dimensions, including function, installation location, and equipment type. In this system, the server and mobile terminal communicate via wireless or wired connections. The server is responsible for QR code generation, data storage, and access control, while the mobile terminal is responsible for QR code scanning, offline parsing, and information display.

[0020] The core function of the data preprocessing module is to transform lengthy KKS strings into a compact form suitable for QR code storage. The original KKS encoding typically contains multiple levels of segments; for example, "10GBA10CE001" can be broken down into "10" for the unit number, "GBA" for the water pump system, "10" for the serial number, and "CE001" for the electric valve number. Semantic vectorization mapping maps these meaningful fields to binary feature vectors, preserving the semantic information of the KKS while significantly compressing the data volume. The introduction of dynamic hash indexes decouples physical tags from logical data: the QR code does not store the original KKS, but instead stores a dynamically generated hash value. This hash value is obtained by XORing the original KKS with a dynamic offset and then performing a hash calculation. The dynamic offset includes a timestamp and the device serial number; therefore, the hash value of QR codes generated by the same device at different times will be different. This prevents the tags from being maliciously copied or forged. Furthermore, when device information is updated in the background, there is no need to reprint the tags; only the hash mapping table needs to be updated on the server side.

[0021] The QR code encoding module employs a non-uniform redundancy allocation method, specifically designed for the harsh environment of power plants. Standard QR codes use the Reed-Solomon error correction algorithm, which achieves globally uniform error correction levels. This scheme tilts the limited codeword space towards the core semantic segments: the data stream containing at least the functional identifier segment and system affiliation segment in the KKS encoding is concentrated in the high-redundancy protection zone at the center of the QR code, while a lower redundancy factor is allocated to the edge areas. This way, even if the QR code edges are obscured by oil or physically worn, the system affiliation information of the equipment can be extracted as long as the central area is readable. Furthermore, predictive error correction bits are introduced for the edge areas. These error correction bits are not universal error correction codes but are pre-generated based on the KKS logical topology. For example, if a device belongs to a water pump system, its equipment number range is pre-calculated and encoded as a check bit to assist decoding.

[0022] The data feedback module handles online requests from mobile terminals. When the mobile terminal is in a network environment, it sends the dynamic hash index decoded from the QR code, along with the scanner's identity tag, to the server. The server then matches the corresponding display content from the multidimensional dataset based on the identity tag. Operators see the equipment's real-time operating parameters, maintenance personnel see the equipment's historical defect records and standard operating procedure cards, and safety managers see the hot work permit status and hazard distribution. This multi-functional QR code mechanism avoids the cumbersome process of printing different QR codes for different roles.

[0023] The mobile terminal parsing module is the core of the system's offline capability. In signal dead zones, where the mobile terminal cannot connect to the network, it can still scan QR codes and perform local parsing. It prioritizes decoding the high-redundancy protection zone in the center of the QR code to extract the system attribution vector. Then, combining this with the mobile terminal's pre-stored local process logic library and auxiliary positioning information, it recursively reconstructs complete device information using a probabilistic matching algorithm based on KKS hierarchical topology. Specifically, when the edge area of ​​the QR code is damaged, making the device number unreadable, the system uses the decoded system attribution vector and the current location to search for a matching device node in the local process logic library. Since the number of devices within the same system is limited, combining location information allows for unique device identification. This offline recursive parsing mechanism ensures that even in network-free environments, inspection personnel can obtain the basic process role and safety precautions for the equipment.

[0024] Furthermore, the aforementioned data preprocessing module includes:

[0025] The system comprises a KKS encoding and parsing unit, a semantic vectorization mapping unit, a dynamic hash index generation unit, and a data caching unit connected to the aforementioned KKS encoding and parsing unit, semantic vectorization mapping unit, and dynamic hash index generation unit, respectively.

[0026] The input terminal of the aforementioned KKS code parsing unit is connected to the aforementioned power plant equipment management system and is configured to receive the aforementioned original KKS code and, according to the power plant KKS identification standard, segment the aforementioned original KKS code into a functional identification segment, an installation location identification segment, and an equipment identification segment. The aforementioned functional identification segment is used to describe the function of the process system to which the equipment belongs, the aforementioned installation location identification segment is used to describe the spatial location of the equipment in the power plant, and the aforementioned equipment identification segment is used to describe the specific type and number of the equipment.

[0027] The input of the semantic vectorization mapping unit is connected to the output of the KKS encoding and parsing unit, and is configured to receive the function identifier segment, installation location identifier segment, and equipment identifier segment, and perform binary feature vector mapping on each segment. Specifically, for the function identifier segment, the semantic vectorization mapping unit has a built-in system dictionary table, where each function code corresponds to a fixed-length binary feature code. For the installation location identifier segment, the semantic vectorization mapping unit uses a spatial grid-based encoding method, dividing the power plant space into multiple grid regions, each corresponding to a fixed-length binary spatial code. For the equipment identifier segment, the semantic vectorization mapping unit converts the equipment type and number into a binary sequence. The semantic vectorization mapping unit also merges the three binary feature codes into a single overall feature vector through bit concatenation.

[0028] The input of the aforementioned dynamic hash index generation unit is connected to the output of the aforementioned KKS encoding parsing unit, and is configured to receive the aforementioned original KKS encoding and obtain the current system timestamp and the device serial number of the aforementioned device in the power plant management system; the aforementioned dynamic hash index generation unit performs an XOR operation on the aforementioned original KKS encoding and a dynamic offset, the aforementioned dynamic offset being a value formed by arranging the binary bits of the aforementioned current system timestamp and the binary bits of the aforementioned device serial number in an alternating order; the result of the XOR operation is then input into a hash function to calculate a hash value of fixed length, which serves as the dynamic hash index;

[0029] The input of the aforementioned data caching unit is connected to the output of the aforementioned semantic vectorization mapping unit and the output of the aforementioned dynamic hash index generation unit, respectively, and is configured to temporarily store the aforementioned total feature vector and the aforementioned dynamic hash index, and provide them to the aforementioned QR code encoding module for reading.

[0030] The data preprocessing module is further divided into a KKS encoding and parsing unit, a semantic vectorization mapping unit, a dynamic hash index generation unit, and a data caching unit. These units work together to form a pipeline for converting raw KKS data into QR code encoded data.

[0031] The KKS encoding parsing unit is responsible for parsing the flattened KKS strings stored in the power plant equipment management system according to the internationally recognized KKS identification standard. The structure of a KKS code typically consists of several segments, including a function identifier segment, an installation location identifier segment, and an equipment identifier segment. The parsing unit separates these segments one by one, providing structured input for subsequent vectorized mapping. This separation is necessary because different segments have different semantic weights in subsequent processing: the function identifier segment and the system affiliation segment belong to core semantics and require higher redundancy protection, while the equipment identifier segment can be repaired through reasoning in the event of damage.

[0032] The semantic vectorization mapping unit performs binary feature vector mapping. For the function identifier segment, this unit has a built-in system dictionary table that maps each function code to a fixed-length binary feature code. This lookup encoding is both fast and ensures the Hamming distance between different function codes, facilitating subsequent error correction and matching. For the installation location identifier segment, this unit adopts a spatial grid-based encoding method: the entire power plant space is divided into multiple grid areas, each corresponding to a unique binary spatial code. In this way, the location information of the equipment is quantized into a binary vector, which facilitates matching with the location information of mobile terminals. For the equipment identifier segment, this unit converts the equipment type and number into a binary sequence. Finally, the semantic vectorization mapping unit merges these three binary feature codes into a total feature vector through a bit concatenation operation. Bit concatenation refers to connecting the high-order part, the middle part, and the low-order part in sequence into a longer binary string. For example, the function identifier segment is 10 bits, the installation location identifier segment is 12 bits, and the equipment identifier segment is 12 bits, resulting in a 34-bit total feature vector. Compared to directly storing the original KKS string, the amount of data after compression is significantly reduced, and the freed-up space can be used to deploy a highly redundant error correction matrix.

[0033] The dynamic hash index generation unit is responsible for generating the dynamic hash index. This unit receives the original KKS code and obtains the current system timestamp and the device serial number in the power plant management system. The dynamic offset is constructed by alternating the binary bits of the timestamp and the binary bits of the device serial number. For example, if the timestamp binary is t1 t2 t3… and the serial number binary is s1 s2 s3…, then the offset is t1 s1 t2 s2 t3 s3… This alternating arrangement increases the randomness of the offset, making the hash result more difficult to predict. Then, the original KKS code and the dynamic offset are XORed bitwise. The result of the XOR is then input into a standard hash function to obtain a fixed-length hash value as the dynamic hash index. This index is stored in a QR code, and the server maintains a mapping table to record the correspondence between the hash value and the original KKS. When the device information is updated, only the mapping relationship needs to be updated on the server; there is no need to reprint the QR code label. In addition, the timestamp in the dynamic offset is also used to prevent replay attacks and manage tag expiration.

[0034] The data caching unit, acting as temporary storage, receives the total feature vector from the semantic vectorization mapping unit and the dynamic hash index from the dynamic hash index generation unit, and provides it to the QR code encoding module for reading. This caching unit employs a first-in-first-out queue or register array to ensure that data is not lost or corrupted during transmission.

[0035] Furthermore, the aforementioned QR code encoding module includes:

[0036] The data stream segmentation unit, the redundancy allocation decision unit, the error correction coding unit, the QR code matrix generation unit, and the matrix region mapping unit connecting the error correction coding unit and the QR code matrix generation unit.

[0037] The input end of the data stream segmentation unit is connected to the output end of the data preprocessing module, and is configured to receive the feature vector and the dynamic hash index, and further split the feature vector into core semantic segment data stream and edge semantic segment data stream; the core semantic segment data stream includes at least the part of the feature vector corresponding to the function identifier segment and the system affiliation segment, and the edge semantic segment data stream includes the part of the feature vector corresponding to the device number and auxiliary parameters.

[0038] The aforementioned redundancy allocation decision unit is connected to the aforementioned data stream segmentation unit and is configured to allocate different error correction redundancy levels to the aforementioned core semantic segment data stream and the aforementioned edge semantic segment data stream according to a preset non-uniform redundancy allocation strategy; wherein, a first redundancy level is allocated to the aforementioned core semantic segment data stream, and the amount of redundant data corresponding to the aforementioned first redundancy level is not less than 40% of the original data amount of the aforementioned core semantic segment data stream; a second redundancy level is allocated to the aforementioned edge semantic segment data stream, and the amount of redundant data corresponding to the aforementioned second redundancy level is not more than 20% of the original data amount of the aforementioned edge semantic segment data stream.

[0039] The aforementioned error correction coding unit is connected to the aforementioned redundancy allocation decision unit and is configured to receive the aforementioned core semantic segment data stream, edge semantic segment data stream and their respective redundancy levels, and generate corresponding error correction codewords using the Reed-Solomon coding algorithm.

[0040] The aforementioned matrix region mapping unit is connected to the aforementioned error correction coding unit and is configured to map the aforementioned core semantic segment data stream and its error correction codeword to the central region of the QR code matrix, and to map the aforementioned edge semantic segment data stream and its error correction codeword to the edge region of the QR code matrix; the aforementioned central region is defined as a circular or rectangular region centered on the geometric center of the QR code and occupying 30% to 50% of the total area of ​​the QR code; the aforementioned edge region is defined as the surrounding area of ​​the QR code matrix excluding the central region;

[0041] The aforementioned QR code matrix generation unit is connected to the aforementioned matrix region mapping unit and is configured to generate the final QR code image based on the mapped data bits and error correction code bits. In addition, for the insufficient data protection that may occur in the aforementioned edge region due to the allocation of a lower redundancy level, the aforementioned QR code matrix generation unit also introduces a predictive error correction bit generated based on KKS logical topology. The aforementioned predictive error correction bit is a range of possible values ​​of the equipment number pre-calculated from the process topology library pre-stored on the server based on the system affiliation vector in the aforementioned core semantic segment, and this range is encoded as a check bit and filled into the reserved position in the edge region.

[0042] The QR code encoding module receives feature vectors and dynamic hash indices from the data preprocessing module. Through a series of steps including data stream segmentation, redundancy allocation, error correction coding, region mapping, and matrix generation, it finally outputs a printable QR code image. This module's design breaks through the uniform error correction framework of standard QR codes and is specifically optimized for the KKS semantic features of power plant equipment.

[0043] The data flow segmentation unit first splits the feature vector into core semantic segment data streams and edge semantic segment data streams. The core semantic segment includes at least the portions of the feature vector corresponding to the function identifier segment and the system affiliation segment, as this information determines the device's technological role and security level; even with severely damaged QR codes, readability must be prioritized. The edge semantic segment includes the device number and auxiliary parameters. This segmentation is not a simple truncation, but an intelligent division based on the semantic importance of KKS encoding.

[0044] The redundancy allocation decision unit assigns different error correction redundancy levels to the core semantic segment and the edge semantic segment according to a preset non-uniform redundancy allocation strategy. While the error correction level of the standard QR code is global, this scheme breaks this uniformity: the core semantic segment is assigned the first redundancy level, with a corresponding redundant data volume of no less than 40% of the original data volume of the core semantic segment. This means that the effective error correction capability of the core semantic segment far exceeds that of the standard H level; the edge semantic segment is assigned the second redundancy level, with a redundant data volume of no more than 20% of the original data volume of the edge semantic segment. This polarized design, combined with subsequent logical blind filling, is feasible because even if the device number of the edge segment is completely lost, it can still be inferred from the system attribution vector and location information of the core segment.

[0045] The error-correcting coding unit uses the Reed-Solomon coding algorithm to generate error-correcting codewords. Reed-Solomon coding is a widely used error-correcting code in QR codes. Its basic idea is to treat data as a polynomial and construct an error-correcting polynomial by adding redundant symbols. This unit generates error-correcting codewords of appropriate strength for each of the two data streams based on different redundancy levels given by the redundancy allocation decision unit. The core semantic segment uses a higher-order generator polynomial, generating more redundant symbols.

[0046] The matrix region mapping unit is responsible for mapping data and error correction codewords to specific areas of the QR code matrix. This unit maps the core semantic segment data stream and its error correction codewords to the central area of ​​the QR code, defined as a circular or rectangular area centered on the geometric center of the QR code, occupying 30% to 50% of the total area. This is because in actual scanning, the camera is usually aimed at the center of the QR code, and oil stains tend to spread from the edges, making the central area relatively safer. The edge area is defined as the area surrounding the central area in the QR code matrix, used to carry the edge semantic segment data stream and a smaller number of error correction codewords.

[0047] The QR code matrix generation unit generates the final QR code image based on the mapped data bits and error correction bits. Specifically, to address potential data protection deficiencies in edge areas due to lower redundancy levels, this unit introduces predictive error correction bits generated based on KKS logical topology. This predictive error correction bit generation method involves retrieving all possible device number sets from the server-pre-stored process topology library based on the system affiliation vector in the core semantic segment. This device number set is then encoded into a series of check bits, using either bitmap encoding or interval encoding. The data length of the check bits is determined by the size of the device number set. The QR code matrix generation unit fills these check bits into reserved positions in the edge area, which were originally intended to store error correction bits for the edge semantic segment data stream. When the mobile terminal decodes, if the device number in the edge area is unreadable, it reads the check bits from the reserved positions and, combined with the decoded system affiliation segment, reverse-engineers the correct device number from the local process logic library.

[0048] Furthermore, the aforementioned data feedback module includes:

[0049] The system includes a request receiving unit, an identity authentication unit, a dynamic hash parsing unit, a role-data mapping library, a data retrieval unit, and a response sending unit.

[0050] The input end of the aforementioned request receiving unit is connected to the wireless communication module of the aforementioned mobile terminal and is configured to receive the parsing request sent by the aforementioned mobile terminal; the data packet format of the aforementioned parsing request includes: a request header, an identity tag field, a dynamic hash index field, a timestamp field, and a verification field; the value of the aforementioned verification field is a verification code obtained by hashing the aforementioned identity tag field, dynamic hash index field, and timestamp field.

[0051] The input terminal of the aforementioned identity authentication unit is connected to the aforementioned request receiving unit and is configured to extract the identity tag field from the aforementioned parsing request and compare it with the pre-stored list of authorized identities in the aforementioned request receiving unit; if the aforementioned identity tag is not in the list of authorized identities, the aforementioned identity authentication unit outputs an authentication failure signal to the aforementioned response sending unit;

[0052] The input of the dynamic hash parsing unit is connected to the request receiving unit and configured to extract the dynamic hash index field from the parsing request and perform a reverse match between the dynamic hash index and the dynamic hash index generated by the data preprocessing module. The dynamic hash parsing unit internally maintains a dynamic hash index cache table, which stores the mapping relationship between the dynamic hash index generated by the data preprocessing module and its corresponding original KKS code. The dynamic hash parsing unit queries the cache table based on the received dynamic hash index to restore the corresponding original KKS code.

[0053] The aforementioned role-data mapping library is a relational database table stored in the server-side storage medium. The aforementioned relational database table contains at least the following fields: role identifier, data source address, and data display template; wherein, the aforementioned role identifier includes operator identifier, maintenance personnel identifier, and safety management personnel identifier; the aforementioned data source address points to a specific dataset in the power plant's real-time database, maintenance management system, or safety management system;

[0054] The input terminals of the aforementioned data retrieval unit are respectively connected to the output terminals of the aforementioned identity authentication unit, the aforementioned dynamic hash parsing unit, and the aforementioned role-data mapping library. The configuration is as follows: upon receiving an authentication pass signal from the aforementioned identity authentication unit, based on the parsed identity tag and original KKS code, the corresponding data source address is searched in the aforementioned role-data mapping library. Then, based on the data source address, the display content adapted to the identity tag is retrieved from the corresponding database. For operator identifiers, real-time operating parameters and alarm thresholds are retrieved. For maintenance personnel identifiers, equipment maintenance history records, standard operating cards, and spare parts inventory information are retrieved. For safety management personnel identifiers, the status of hot work permits, hazard distribution, and safety hazard investigation forms are retrieved.

[0055] The aforementioned response sending unit is connected to the output end of the aforementioned data retrieval unit and the aforementioned identity authentication unit, respectively, and is configured to: if an authentication failure signal is received from the aforementioned identity authentication unit, return an authentication failure response to the mobile terminal; if the aforementioned display content is received from the aforementioned data retrieval unit, encapsulate the aforementioned display content into a response data packet and send it to the aforementioned mobile terminal via the wireless network.

[0056] The data feedback module is the core component of the server-side processing of online requests from mobile terminals. It receives parsing requests from mobile terminals, completes identity authentication, reverse hash matching, data retrieval, and response sending. The design of this module takes into account the security isolation requirements of power plant production data and the different information needs of different roles.

[0057] The request receiving unit establishes a connection with the mobile terminal's wireless communication module to receive and parse the request. The data packet format for parsing the request includes a request header, an identity tag field, a dynamic hash index field, a timestamp field, and a verification field. The verification field is a checksum calculated by hashing the concatenated identity tag, dynamic hash index, and timestamp, used to prevent request tampering or replay attacks. This design is similar to a message authentication code in network protocols, ensuring the integrity and authenticity of the request.

[0058] The authentication unit extracts the identity tag field from the request and compares it with a pre-stored list of authorized identities. The identity tag can be a user's employee ID, role code, or an encrypted digital certificate. If the identity tag is not in the authorized list, the authentication unit immediately outputs an authentication failure signal to the response sending unit, rejecting further processing. This ensures that only authorized personnel can access device information, preventing the leakage of sensitive data.

[0059] The dynamic hash parsing unit extracts the dynamic hash index field from the request and performs a reverse match. Since the QR code stores a hash value rather than the original KKS, the server needs to maintain a dynamic hash index cache table. This table stores the mapping relationship between each pair of dynamic hash indices generated by the data preprocessing module and the original KKS encoding. This unit queries the cache table based on the received hash value to reconstruct the corresponding original KKS encoding. Because the hash function is one-way, the original KKS cannot be directly calculated from the hash value; therefore, the cache table is essential. The cache table is stored in high-speed memory to speed up the query. The timestamp portion of the dynamic hash index is also used to determine if the request has timed out. If the timestamp differs significantly from the current time, the request is rejected to prevent the use of expired QR codes.

[0060] The role-data mapping library is a relational database table stored on the server-side storage medium. This table contains at least three fields: role identifier, data source address, and data display template. The role identifier includes identifiers for operators, maintenance personnel, and safety management personnel. The data source address can be a database connection string or an application programming interface endpoint, pointing to a specific dataset in the power plant's real-time database, maintenance management system, or safety management system. The data display template defines the format of the returned content; for operators, it displays real-time values ​​of measuring points such as temperature, pressure, and vibration in tabular form; for maintenance personnel, it displays disassembly diagrams and work procedures in graphic and textual form.

[0061] After receiving the authentication pass signal from the identity authentication unit, the data retrieval unit searches for the corresponding data source address in the role-data mapping library based on the parsed identity tag and original KKS code, and then retrieves the appropriate display content from the corresponding database. For operator identifiers, it retrieves real-time operating parameters and alarm thresholds, and also queries the historical trend data of the equipment. For maintenance personnel identifiers, it retrieves equipment maintenance history records, standard operating procedures, and spare parts inventory information. For safety management personnel identifiers, it retrieves the hot work permit status, hazard distribution, and safety hazard investigation form. This role-based data routing mechanism achieves secure isolation and efficient flow of production data.

[0062] The response sending unit generates a response based on the authentication result and the data retrieval result. If authentication fails, it returns an authentication failure response containing an error code; if the retrieval is successful, it encapsulates the content to be displayed into a response data packet and sends it back to the mobile terminal via the wireless network. The response data packet also contains caching control instructions, instructing the mobile terminal to cache certain content locally for offline use.

[0063] Furthermore, the aforementioned mobile terminal parsing module includes:

[0064] Image acquisition unit, image preprocessing unit, region decoding unit, local logic library storage unit, probability matching inference unit, offline output unit, online request unit and online receiving unit;

[0065] The image acquisition unit is a camera built into the mobile terminal, configured to scan the QR code image attached to the power plant equipment and output the acquired raw image data to the image preprocessing unit.

[0066] The input end of the image preprocessing unit is connected to the image acquisition unit and is configured to perform grayscale conversion, binarization, noise reduction and distortion correction processing on the original image data in sequence. In particular, for the oil pollution environment of the power plant, the image preprocessing unit has a built-in adaptive binarization algorithm. The adaptive binarization algorithm dynamically adjusts the binarization threshold according to the pixel brightness distribution of the local area of ​​the image to enhance the edge features of the QR code in the oil pollution occlusion area.

[0067] The input of the aforementioned region decoding unit is connected to the output of the aforementioned image preprocessing unit, and is configured to perform region segmentation on the preprocessed QR code image to identify the central region and edge regions; the aforementioned region decoding unit first decodes the aforementioned central region and extracts the system attribution vector stored therein; if the aforementioned central region is successfully decoded, the aforementioned region decoding unit outputs the aforementioned system attribution vector to the aforementioned probability matching inference unit; if the aforementioned central region is not decoded, the aforementioned region decoding unit outputs a decoding failure signal to the aforementioned probability matching inference unit;

[0068] The aforementioned local logic library storage unit is a non-volatile memory, which pre-stores a local process logic library constructed based on KKS encoding rules. The data structure of the aforementioned local process logic library is a compressed prefix tree. Each node in the compressed prefix tree corresponds to a level segment of KKS encoding, with the root node corresponding to the function identifier segment and the child nodes corresponding to the installation location identifier segment and the equipment identifier segment, respectively. Multiple KKS codes with the same prefix share nodes on the prefix path. Each node in the aforementioned local process logic library is also associated with and stores the attribute vector of that node. The attribute vector includes: equipment type, security level, maintenance cycle, and standard operating procedure index.

[0069] The input of the aforementioned probabilistic matching inference unit is connected to the output of the aforementioned region decoding unit, the output of the aforementioned local logic library storage unit, and the positioning module of the mobile terminal, respectively. The probabilistic matching inference unit is configured to: when receiving the aforementioned system attribution vector, locate the corresponding subtree in the aforementioned local logic library storage unit using the aforementioned system attribution vector as an index; then obtain auxiliary positioning information provided by the aforementioned positioning module, the auxiliary positioning information including GPS coordinates or Bluetooth beacon signal strength; the probabilistic matching inference unit calculates the spatial matching probability between each device node in the subtree and the current auxiliary positioning information, and selects the device node with the highest probability as a candidate node; if the aforementioned region... The domain decoding unit simultaneously outputs partial edge region decoding results. The aforementioned probability matching inference unit also calculates the edit distance between the device number string of the candidate node and the partial decoded string output by the aforementioned region decoding unit, and performs a weighted summation of the aforementioned spatial matching probability and the aforementioned edit distance similarity. The weight coefficient of the spatial matching probability is automatically increased when there is no network signal. Based on the weighted summation result, the uniquely matching device node and its complete attribute vector are output. If the aforementioned region decoding unit outputs a decoding failure signal, the aforementioned probability matching inference unit searches for the device node with the closest spatial location in the entire local logic library based only on the aforementioned auxiliary positioning information, and uses that node as the inference result.

[0070] The input end of the aforementioned offline output unit is connected to the output end of the aforementioned probability matching inference unit, and is configured to format the information in the aforementioned complete attribute vector into readable text or a graphical interface, and display it on the mobile terminal screen; the aforementioned readable text includes at least the device name, the system to which it belongs, and security precautions;

[0071] The input end of the above-mentioned online request unit is connected to the output end of the above-mentioned area decoding unit, and is configured to obtain the dynamic hash index decoded by the above-mentioned area decoding unit from the QR code image, and obtain the identity tag of the currently logged-in user of the mobile terminal, encapsulate the above-mentioned dynamic hash index, identity tag and current timestamp into a parsing request data packet according to a preset format, and send it to the data feedback module of the above-mentioned server through the wireless communication module of the mobile terminal.

[0072] The input end of the aforementioned online receiving unit is connected to the wireless communication module of the mobile terminal, and the output end is connected to the aforementioned offline output unit. It is configured to receive the display content returned by the aforementioned data feedback module and transmit the display content to the aforementioned offline output unit for display. When the aforementioned online request unit successfully sends a request and receives the display content, the aforementioned online receiving unit also caches the display content in the aforementioned local logic library storage unit for quick access in subsequent offline states.

[0073] The mobile terminal parsing module is the front-end execution unit of the entire system. It integrates multiple functions such as image acquisition, image preprocessing, region decoding, local logic library storage, probabilistic matching inference, offline output, online request, and online reception. The design goal of this module is to be able to reconstruct device information using local data and inference algorithms even if the QR code is partially damaged in offline mode; and to obtain richer real-time data through online requests when the network is available.

[0074] The image acquisition unit is the camera built into the mobile terminal. In power plant inspection scenarios, staff use a personal digital assistant or smartphone to point at the QR code label on the device, and the camera automatically focuses and acquires raw image data. To improve the success rate of acquisition, the application provides a viewfinder and automatic trigger function, automatically taking a picture when the QR code is centered in the viewfinder.

[0075] The image preprocessing unit sequentially performs grayscale conversion, binarization, noise reduction, and distortion correction on the raw image data. Grayscale conversion converts the color image to a grayscale image, reducing computational load. Binarization converts the grayscale image to a black-and-white binary image, distinguishing between black and white modules. Noise reduction uses median filtering or Gaussian filtering to remove salt-and-pepper noise. Distortion correction utilizes the positioning pattern of the QR code to correct perspective distortion. For the oil-contaminated environment of a power plant, this image preprocessing unit incorporates an adaptive binarization algorithm. Traditional global binarization uses a fixed threshold, which can easily lead to black-and-white inversion or blurring in oil-covered areas. Adaptive binarization dynamically adjusts the threshold based on the pixel brightness distribution of local image regions, dividing the image into multiple small windows, each with its own independently calculated threshold. This ensures that even if oil contamination causes uneven local brightness, the QR code's module pattern can be correctly restored.

[0076] The region decoding unit performs region segmentation on the preprocessed QR code image. It first uses three position probes to locate the QR code's boundaries and orientation, then identifies the central and edge regions based on preset geometric relationships. This unit prioritizes decoding the central region because it carries the core semantic data stream and its highly redundant error-correcting code, exhibiting strong robustness. If the central region decoding is successful, the system attribution vector is extracted and output to the probability matching inference unit. If the central region decoding also fails, a decoding failure signal is output to the probability matching inference unit; in this case, the system can only rely on auxiliary positioning information for pure positional inference.

[0077] The local logic library storage unit is a non-volatile memory that pre-stores a local process logic library built based on KKS encoding rules. The data structure of this logic library is a compressed prefix tree. In the compressed prefix tree, each node corresponds to a level segment of the KKS encoding, with the root node corresponding to the function identifier segment, and the child nodes corresponding to the installation location identifier segment and the device identifier segment, respectively. For example, if multiple devices share the same prefix, the path is stored only once in the tree, and subsequent devices become branch nodes. This structure significantly reduces storage space. Each node is also associated with a stored attribute vector, which includes device type, security level, maintenance cycle, and standard operating procedure index. The local logic library is periodically updated from the server through the synchronization mechanism described in claim 9.

[0078] The input of the probabilistic matching inference unit is connected to the output of the region decoding unit, the output of the local logic library storage unit, and the positioning module of the mobile terminal, respectively. This unit is configured to: upon receiving a system attribution vector, locate the corresponding subtree in the local logic library storage unit using this vector as an index. Then, it acquires auxiliary positioning information provided by the positioning module, including GPS coordinates or Bluetooth beacon signal strength. The probabilistic matching inference unit calculates the spatial matching probability of each device node in the subtree with the current auxiliary positioning information, selecting the device node with the highest probability as a candidate node. If the region decoding unit simultaneously outputs partial edge region decoding results, the probabilistic matching inference unit also calculates the edit distance between the device number string of the candidate node and the partial decoded string output by the region decoding unit, and performs a weighted sum of the spatial matching probability and the edit distance similarity. The weight coefficient of the spatial matching probability is automatically increased when there is no network signal. Based on the weighted sum, it outputs the uniquely matching device node and its complete attribute vector. If the region decoding unit outputs a decoding failure signal, the probabilistic matching inference unit searches for the device node with the closest spatial location in the entire local logic library based solely on the auxiliary positioning information, and uses this node as the inference result. While this reasoning has some uncertainty, in actual power plant scenarios, inspection personnel usually stand directly in front of the equipment to be inspected, and the positioning information can be accurate to 1-2 meters. Combined with the density of equipment within the system, the accuracy rate is very high.

[0079] The input of the offline output unit is connected to the output of the probability matching inference unit, and is configured to format the information in the complete attribute vector into readable text or a graphical interface, and display it on the mobile terminal screen. The readable text includes at least the device name, the system it belongs to, and security precautions. This information can be obtained even without a network, meeting the basic inspection needs in weak network environments.

[0080] The input end of the online request unit is connected to the output end of the regional decoding unit. It is configured to obtain the dynamic hash index decoded by the regional decoding unit from the QR code image, obtain the identity tag of the currently logged-in user of the mobile terminal, encapsulate the dynamic hash index, identity tag and current timestamp into a parsing request data packet according to a preset format, and send it to the data feedback module of the server through the wireless communication module of the mobile terminal.

[0081] The input of the online receiving unit is connected to the wireless communication module of the mobile terminal, and the output is connected to the offline output unit. It is configured to receive the display content returned by the data feedback module and transmit the display content to the offline output unit for display. When the online request unit successfully sends a request and receives the display content, the online receiving unit also caches the display content in its local logic library storage unit for quick access in subsequent offline states. This hybrid mode, rich in online content and basic offline content, maximizes the user experience.

[0082] Furthermore, the aforementioned semantic vectorization mapping unit is also configured to perform bit-map encoding based on character set compression;

[0083] For the characters in the aforementioned function identifier segment, installation location identifier segment, and device identifier segment, the aforementioned semantic vectorization mapping unit uses 5-bit binary encoding instead of the standard 8-bit ASCII encoding. The 5-bit binary code corresponding to each character is obtained by querying a predefined character mapping table. The aforementioned character mapping table covers all uppercase letters A to Z and numbers 0 to 9 that are allowed to appear in KKS encoding.

[0084] The semantic vectorization mapping unit converts each character in the above-mentioned function identifier segment, installation location identifier segment and equipment identifier segment into the corresponding 5-bit binary code in sequence, and then concatenates the binary codes of each character in order to form the binary feature vectors of the above-mentioned function identifier segment, installation location identifier segment and equipment identifier segment respectively.

[0085] The above bit concatenation operation is as follows: the binary feature vector of the above function identifier segment is used as the high-order part, the binary feature vector of the above installation location identifier segment is used as the middle part, and the binary feature vector of the above device identifier segment is used as the low-order part. The two bits are concatenated in order from high-order to low-order to obtain the above total feature vector. The length of the above total feature vector does not exceed 40 bits.

[0086] In the semantic vectorization mapping unit, bit-mapped encoding based on character set compression is used to further compress the data volume. The KKS encoding character set contains only uppercase Latin letters A to Z (26 characters) and Arabic numerals 0 to 9 (10 characters), totaling 36 characters. Standard ASCII encoding uses 8 bits to represent a character, which is redundant for these 36 characters. This system uses 5-bit binary encoding instead of standard 8-bit ASCII encoding. The 5-bit binary code corresponding to each character is obtained from a predefined character mapping table. The character mapping table covers all uppercase letters A to Z and numerals 0 to 9 allowed in KKS encoding. Since 5 bits can represent a maximum of 32 different values, and the character set contains 36 characters, the character mapping table adopts a variable-length encoding strategy: the 32 most frequently occurring characters are directly mapped to 32 5-bit codewords, and the remaining 4 rare characters are encoded using 6 bits, with a special 5-bit escape code as a prefix. In practical KKS encoding, the character distribution exhibits a clear bias—the numbers 0 to 9 and common letters (such as A, B, C, E, G, H, K, M, N, P, S, T, U, V, W) appear far more frequently than other letters, resulting in an average encoding length close to 5 bits. Statistical verification shows that after adopting this compression scheme, the average encoding length per character does not exceed 5.2 bits. The semantic vectorization mapping unit converts each character in the function identifier segment, installation location identifier segment, and device identifier segment into its corresponding binary code, and then concatenates the binary codes of each character in sequence to form the binary feature vector of each segment. The specific rules for the bit concatenation operation are as follows: the binary feature vector of the function identifier segment is used as the high-order part, the binary feature vector of the installation location identifier segment is used as the middle part, and the binary feature vector of the device identifier segment is used as the low-order part, and they are concatenated sequentially from high-order to low-order to obtain the total feature vector. After the compression described above, the typical KKS encoding "10GBA10CE001" consists of 11 characters. Using 8-bit ASCII would require 88 bits, but with this compression scheme, it requires approximately 57 bits. Combined with storage optimization of the shared prefix in the prefix tree, the total feature vector length for each device in actual storage does not exceed 40 bits. This saves valuable codeword space for QR codes, allowing for higher levels of error correction redundancy without increasing the number of QR code versions.

[0087] Furthermore, the aforementioned redundancy allocation decision unit is also configured to dynamically adjust the specific values ​​of the first redundancy level and the second redundancy level according to the device type corresponding to the system affiliation segment in the aforementioned core semantic segment data stream.

[0088] For devices with a security level higher than the preset threshold in the aforementioned core semantic segment data stream, the amount of redundant data corresponding to the first redundancy level is increased to more than 50% of the original data amount of the aforementioned core semantic segment data stream; for devices with a security level not higher than the aforementioned preset threshold, the amount of redundant data corresponding to the first redundancy level is set to between 30% and 40% of the original data amount of the aforementioned core semantic segment data stream.

[0089] The process of generating the aforementioned predictive error correction bits is as follows: The aforementioned QR code matrix generation unit extracts the system affiliation segment from the aforementioned core semantic segment data stream, uses the system affiliation segment as the query key, and retrieves all possible equipment number sets in the system to which the equipment belongs from the process topology library pre-stored on the server; the equipment number set is encoded into a series of check bits, and the data length of the aforementioned check bits is determined according to the size of the equipment number set;

[0090] The aforementioned QR code matrix generation unit fills the aforementioned check bit into the reserved position in the aforementioned edge region. The aforementioned reserved position is part of the position in the aforementioned edge region that was originally used to store the error correction code bit of the edge semantic segment data stream. When the aforementioned mobile terminal parsing module finds that the device number data in the aforementioned edge region is unreadable during decoding, it reads the check bit in the aforementioned reserved position and, in combination with the decoded system affiliation segment, reverse-calculates the correct device number from the aforementioned local process logic library.

[0091] The redundancy allocation decision unit can not only assign fixed redundancy levels to core and edge semantic segments, but also dynamically adjust the specific values ​​of the redundancy levels according to the equipment type. Different equipment in a power plant has different levels of criticality. For example, main feedwater pumps and main steam isolation valves are directly related to the safe operation of the unit, while auxiliary equipment such as general lighting control switches and ventilation valves have a smaller impact on safety. The system queries the safety level of the equipment from a predefined safety level table based on the equipment type corresponding to the system affiliation segment in the core semantic segment data stream. For equipment with a safety level higher than the preset threshold (such as nuclear safety-related equipment or important process equipment), the amount of redundant data corresponding to the first redundancy level is increased to more than 50% of the original data volume of the core semantic segment. This means that for such equipment, the proportion of error correction codes in the central area of ​​the QR code exceeds that of the original data, thus greatly enhancing its resilience. For ordinary equipment with a safety level not higher than the preset threshold, the amount of redundant data corresponding to the first redundancy level is set between 30% and 40% of the original data volume of the core semantic segment, ensuring basic reliability while leaving more space for the edge area.

[0092] The generation process of predictive error correction bits leverages the hierarchical constraint characteristics of KKS coding. Within the same KKS system (e.g., a feedwater pump system), equipment numbers typically fall within a continuous range or follow a certain numbering rule. After extracting the system affiliation segment from the core semantic segment data stream, the QR code matrix generation unit uses this system affiliation segment as the query key to retrieve all possible equipment number sets belonging to the system from the pre-stored process topology database on the server. The size of this set is typically much smaller than the total number of equipment in the entire power plant. This set of equipment numbers is encoded into a series of check bits, using either bitmap encoding or interval encoding. The data length of the check bits is determined by the size of the equipment number set; for example, a system containing 50 devices requires 50 check bits. The QR code matrix generation unit fills these check bits into reserved positions in the edge region, positions originally intended for storing error correction bits from the edge semantic segment data stream. When the mobile terminal parsing module finds that the device number data in the edge area is unreadable due to oil stains or damage during decoding, it reads the check bit in the reserved position, combines it with the decoded system affiliation segment, obtains the list of device numbers under the system from the local process logic library, filters it using the check bit, and then combines it with auxiliary positioning information to reverse-calculate the correct device number.

[0093] Furthermore, the spatial matching probability in the above-mentioned probability matching inference unit is calculated as follows:

[0094] For each device node in the above subtree, obtain the preset installation coordinates of the device node in the power plant spatial coordinate system; obtain the current terminal coordinates output by the mobile terminal positioning module; calculate the Euclidean distance between the preset installation coordinates and the current terminal coordinates; input the Euclidean distance into a Gaussian probability density function to obtain the spatial matching probability of the device node; set the mean of the Gaussian probability density function to 0, and precalibrate the variance according to the spatial distribution density of the power plant equipment.

[0095] The above-mentioned edit distance similarity is calculated as follows: the device number string of the candidate node is aligned with the partially decoded string output by the above-mentioned region decoding unit, and the minimum number of single-character edit operations required to convert the partially decoded string into the device number string of the candidate node is calculated. The above-mentioned single-character edit operations include insertion, deletion and replacement; the above-mentioned minimum number of single-character edit operations is divided by the length of the device number string of the candidate node to obtain the normalized edit distance; the above-mentioned edit distance similarity is equal to 1 minus the above-mentioned normalized edit distance.

[0096] The specific method of the above weighted summation is as follows: the comprehensive score is equal to the first weight coefficient multiplied by the spatial matching probability plus one minus the first weight coefficient multiplied by the edit distance similarity, where the first weight coefficient is the weight coefficient of the spatial matching probability; the default value of the first weight coefficient is 0.6; when the mobile terminal detects no network signal, the first weight coefficient is adjusted to 0.8 or higher.

[0097] The spatial matching probability calculation method in the probabilistic matching inference unit is based on a Gaussian distribution model. In the three-dimensional spatial coordinate system of the power plant, each equipment node stores its preset installation coordinates in its local process logic library. These coordinates are determined during the power plant design phase. The positioning module of the mobile terminal outputs the real-time coordinates of the current terminal. The positioning module can be a GPS receiver or a Bluetooth beacon positioning system. The Euclidean distance between the preset installation coordinates and the current terminal coordinates is calculated, which is the straight-line distance in three-dimensional space. Due to measurement errors and human positioning deviations, the terminal coordinates follow a Gaussian distribution with the equipment coordinates as the mean. The Euclidean distance is input into a Gaussian distribution probability density function with a mean of 0 to obtain the spatial matching probability of the equipment node. The form of the Gaussian distribution probability density function is f(d)=1 / (σ√(2π))·exp(-d² / (2σ²)), where d is the Euclidean distance and σ is the standard deviation. The mean of the Gaussian distribution probability density function is set to 0, and the variance is pre-calibrated according to the spatial distribution density of the power plant equipment: the variance is set to be smaller in densely populated equipment areas to distinguish adjacent equipment, and the variance is set to be larger in sparsely populated equipment areas to tolerate greater positioning errors.

[0098] Edit distance similarity is calculated to handle cases where edge regions are partially readable. Edit distance, also known as Levinstein distance, refers to the minimum number of single-character edit operations required to convert one string into another. Single-character edit operations include insertion, deletion, and replacement. The device ID string of the candidate node is aligned with the partially decoded string output by the region decoding unit, and the minimum number of single-character edit operations required to convert the partially decoded string into the device ID string of the candidate node is calculated. Dividing this minimum number of single-character edit operations by the length of the device ID string of the candidate node yields the normalized edit distance, which ranges from 0 to 1. Edit distance similarity equals 1 minus the normalized edit distance; a perfect match results in a similarity of 1, while a complete mismatch results in a similarity close to 0.

[0099] The weighted summation is performed as follows: the overall score equals the first weight coefficient multiplied by the spatial matching probability, plus one, minus the first weight coefficient multiplied by the edit distance similarity, where the first weight coefficient is the weighting factor for the spatial matching probability. The default value of the first weight coefficient is 0.6, indicating that by default, spatial information contributes slightly more to the inference result than string similarity. When the mobile terminal detects no network signal, the first weight coefficient is adjusted to 0.8 or higher. This means that when offline and unable to request server assistance, the system relies almost entirely on spatial location to determine the device, with string similarity only used for fine-tuning.

[0100] Furthermore, the aforementioned server also includes a logic library generation and synchronization module, and the aforementioned mobile terminal also includes a logic library synchronization unit;

[0101] The input of the aforementioned logic library generation and synchronization module is connected to the aforementioned data preprocessing module, and is configured to: obtain the feature vectors of all devices and their corresponding original KKS codes after processing by the aforementioned data preprocessing module; construct a compressed prefix tree for the KKS codes of all devices according to the hierarchical structure of the KKS codes; each node of the compressed prefix tree stores the binary feature code of the KKS segment corresponding to the node and the attribute vector of the node; the aforementioned logic library generation and synchronization module also serializes the aforementioned compressed prefix tree into a binary data stream and sends it to the aforementioned mobile terminal via a wireless network or wired connection;

[0102] The input end of the logic library synchronization unit of the mobile terminal is connected to the wireless communication module of the mobile terminal, and the output end is connected to the local logic library storage unit. It is configured to receive the binary data stream sent by the server, deserialize the binary data stream into a compressed prefix tree structure, and store it in the local logic library storage unit to update or replace the original local process logic library.

[0103] The aforementioned logic library synchronization unit is also configured to automatically initiate a synchronization request to the aforementioned server according to a preset synchronization cycle, or to check whether the aforementioned server has an updated version of the process logic library after each successful network scan by the aforementioned mobile terminal. If it does, incremental synchronization is triggered. The aforementioned incremental synchronization only transmits the device nodes and their attribute vectors that have changed since the last synchronization, so as to reduce the amount of data transmitted over the network.

[0104] In addition to data preprocessing, QR code encoding, and data feedback modules, the server-side also includes a logic library generation and synchronization module. The mobile terminal further includes a logic library synchronization unit. The input of the logic library generation and synchronization module is connected to the data preprocessing module to obtain the feature vectors of all devices processed by the data preprocessing module and their corresponding original KKS codes. The logic library generation and synchronization module constructs a compressed prefix tree of all devices' KKS codes according to the hierarchical structure of the KKS codes. The construction process is as follows: First, a root node is created; then, the KKS codes of all devices are traversed, and each segment is inserted into the tree in the hierarchical order of function identifier segment, installation location identifier segment, and device identifier segment. If a path for a certain segment already exists, the existing node is reused; otherwise, a new node is created. Each node stores the binary feature code of the corresponding KKS segment and the node's attribute vector. The attribute vector includes information such as device type, security level, maintenance cycle, and standard operating procedure index. For devices sharing a prefix, only one copy of the nodes on their common path is stored, thus significantly reducing storage space. After the build is complete, the logic library generation and synchronization module serializes the compressed prefix tree into a binary data stream and sends it to the mobile terminal via wireless network or wired connection.

[0105] The input of the mobile terminal's logic library synchronization unit is connected to the mobile terminal's wireless communication module, and the output is connected to the local logic / library storage unit. The logic library synchronization unit receives binary data streams sent by the server, deserializes the binary data streams into a compressed prefix tree structure, and stores it in the local logic library storage unit to update or replace the existing local process logic library. The logic library synchronization unit automatically initiates a synchronization request to the server according to a preset synchronization cycle, carrying the current version number or timestamp of the local logic library in the synchronization request. After comparing the version numbers, if the server finds an update, it calculates the device nodes that have changed since the last synchronization, including newly added, modified, and deleted devices, and only packages and sends the paths and attribute vectors of these changed nodes. The logic library synchronization unit is also configured to check whether the server has an updated process logic library version after each successful network scan by the mobile terminal; if so, incremental synchronization is triggered. Incremental synchronization only transmits the device nodes and their attribute vectors that have changed since the last synchronization, thereby reducing the amount of data transmitted over the network.

[0106] Furthermore, the QR code encoding module is also configured to embed a self-verification bit in the QR code image; the self-verification bit includes the hash value of the QR code generation timestamp and the QR code version verification code.

[0107] The mobile terminal parsing module is further configured to: after parsing the QR code image, extract the self-check bit, calculate the difference between the timestamp of the QR code generation obtained by decoding and the current system time; if the difference exceeds the preset tag lifespan threshold, the mobile terminal parsing module adds the information that the tag has expired and prompts for replacement to the displayed content.

[0108] The aforementioned mobile terminal parsing module also records the recognition time for each scan. When the recognition time exceeds the preset time threshold for three consecutive scans, the aforementioned mobile terminal parsing module automatically generates a tag replacement work order and sends the tag replacement work order to the server via the wireless network.

[0109] The aforementioned server also includes a tag lifecycle management module, configured to receive and store the aforementioned tag replacement work orders, generate tag replacement tasks based on the equipment identification and QR code image information in the aforementioned tag replacement work orders, and push them to the power plant maintenance management system.

[0110] The aforementioned server is also configured to automatically mark the label status of the device as pending replacement when the cumulative number of reported recognition times exceeding the threshold for the QR code image of the same device reaches a preset number, and add the label replacement reminder to the next synchronization data generated by the above logic library and synchronization module.

[0111] The QR code encoding module embeds a self-checking bit into the QR code image. This self-checking bit includes the hash value of the QR code generation timestamp and a QR code version checksum. The self-checking bit is written to a fixed, pre-reserved location within the QR code, such as a specific module area in the upper right corner. After parsing the QR code image, the mobile terminal parsing module extracts the self-checking bit and calculates the difference between the currently decoded QR code generation timestamp and the system's current time. If this difference exceeds a preset tag lifespan threshold, it indicates that the QR code tag has been used for an extended period and may be worn or faded. The mobile terminal parsing module then adds an expired tag notification to the displayed content, prompting the user to replace the tag.

[0112] The mobile terminal parsing module also records the recognition time for each scan, defined as the time elapsed from image acquisition to successful decoding. This metric reflects the readability of the QR code: when the QR code is slightly worn or stained with oil, the error correction algorithm requires more time to correct errors, leading to increased recognition time. When the recognition time exceeds the preset threshold for three consecutive scans, the mobile terminal parsing module automatically generates a tag replacement work order and sends it to the server via the wireless network. The tag replacement work order includes the device identifier, QR code image information, and specific data on the recognition time.

[0113] The server also includes a tag lifecycle management module. This module receives and stores tag replacement work orders, generates tag replacement tasks based on the equipment identifier and QR code image information in the work order, and pushes them to the power plant maintenance management system. The tag replacement task includes a list of equipment to be replaced, a suggested replacement time, and printing parameters for the new tags. When the cumulative number of times the QR code image of the same device is reported to have exceeded a threshold for recognition time reaches a preset number, the server automatically marks the tag status of that device as pending replacement and adds a tag replacement reminder to the next synchronization data generated by the logic library and synchronization module. After subsequent synchronization, the mobile terminal records this reminder in its local logic library and proactively prompts the inspection personnel that the tag is about to expire when scanning the device. Through this mechanism, the system achieves proactive management of the entire lifecycle of QR code tags.

[0114] The beneficial effects of this invention are as follows:

[0115] (1) Significantly improved the recognition success rate of QR codes in the harsh environment of power plants. By allocating the core semantic segments (functional identifier segment and system ownership segment) of KKS encoding to the high redundancy protection zone in the center of the QR code, and using a lower redundancy factor or predictive error correction bits for the edge areas, even when oil stains cover or the physical damage to the edges reaches 50%, the system ownership information can still be extracted from the central area first. Combined with offline recursive parsing, the complete equipment information can be restored, and the recognition success rate is improved by more than 40% compared with the standard QR code.

[0116] (2) It completely solves the problem of information acquisition in weak network / no network environments. By pre-storing a local process logic library (compressed prefix tree structure) based on KKS encoding rules on the mobile terminal, and using a probabilistic matching recursive parsing algorithm based on system affiliation vector and auxiliary positioning information, inspection personnel can obtain basic process information of equipment (equipment name, system to which it belongs, safety level, maintenance cycle, operating procedure index, etc.) in milliseconds without network access in signal blind areas, reducing the information acquisition delay from minutes to milliseconds.

[0117] (3) Significantly reduced label reprinting costs when equipment is replaced or information changes. A dynamic hash index (XORing the original KKS code with a dynamic offset containing a timestamp and device serial number before hashing) is used to decouple the physical QR code from the backend logical data. The QR code label does not store the original KKS or a fixed URL, only the hash value. When device information is updated or the backend address changes, only the hash mapping table needs to be updated on the server side; there is no need to reprint or replace the labels, reducing the label reprint rate by more than 80%.

[0118] (4) Role-based “one code, multiple effects” differentiated information display was realized. By parsing the identity tag (operation, maintenance, security) of the scanner automatically carried in the request, the server dynamically matches and returns the appropriate display content from the multidimensional dataset (real-time database, maintenance management system, security management system) according to the identity tag, realizing the secure isolation and efficient flow of production data, and avoiding information overload and missing key information.

[0119] (5) A proactive management mechanism for the entire lifecycle of QR code labels has been established. By embedding self-verification bits (generating timestamp hash and version verification code) in the QR code and monitoring the scanning and recognition time, the system can automatically determine the aging degree of the label, generate a replacement work order before the label expires and push it to the maintenance system, thus realizing closed-loop management from label generation, use, aging to replacement.

[0120] (6) The storage space utilization and query efficiency of mobile terminals have been optimized. A compressed prefix tree based on KKS hierarchical topology is used to store the local process logic library. KKS codes with the same prefix share storage nodes. Combined with 5-bit binary character set compression encoding, the storage occupation of the terminal is greatly reduced. At the same time, it supports fast subtree positioning based on system belonging vector, and the retrieval efficiency is significantly improved. Attached Figure Description

[0121] Figure 1 : System architecture diagram of the present invention. Detailed Implementation

[0122] The present invention will be further described below with reference to the accompanying drawings and specific embodiments:

[0123] Example: Figure 1 As shown, a standardized generation and parsing management system for QR codes of power plant equipment based on KKS encoding is presented.

[0124] I. System Overall Architecture

[0125] This embodiment provides a standardized generation and parsing management system for QR codes of power plant equipment based on KKS encoding. The system includes a server and mobile terminals (such as inspection PDAs or smartphones) that communicate with the server via wireless networks (e.g., 4G, 5G, Wi-Fi) or wired connections. The server is deployed in the power plant information center and is responsible for QR code generation, data storage, access control, and the maintenance and distribution of the local process logic library. The mobile terminals are carried by inspection personnel and are responsible for QR code scanning, offline parsing, information display, and online requests.

[0126] like Figure 1As shown, the server includes a data preprocessing module, a QR code encoding module, a data feedback module, a logic library generation and synchronization module, and a tag lifecycle management module. The mobile terminal includes a mobile terminal parsing module, which further includes an image acquisition unit, an image preprocessing unit, a region decoding unit, a local logic library storage unit, a probability matching inference unit, an offline output unit, an online request unit, an online receiving unit, and a logic library synchronization unit.

[0127] II. QR Code Generation Process

[0128] 1. Data Preprocessing

[0129] The data preprocessing module obtains the original KKS code of the equipment from the power plant's equipment management system (such as EAM or ERP). Taking the electric valve at the outlet of the feedwater pump of Unit 1 in a power plant as an example, its KKS code is "10GBA10CE001". The data preprocessing module first performs semantic vectorization processing on this code.

[0130] Specifically, the KKS encoding and parsing unit within the data preprocessing module, according to the power plant's KKS identification standard, breaks down "10GBA10CE001" into three segments: the functional identifier segment "GBA" (feedwater pump system), the installation location identifier segment "10" (10-meter layer), and the equipment identifier segment "CE001" (electric valve No. 1). The semantic vectorization mapping unit maps these three segments into binary feature vectors: for the functional identifier segment, "GBA" is mapped to a 10-bit binary code (e.g., 0001010010) using the built-in system dictionary; for the installation location identifier segment, a spatial grid-based encoding method is used, dividing the power plant space into a 10m × 10m grid, and determining its grid code as a 12-bit binary code based on the actual coordinates of the equipment; for the equipment identifier segment, the equipment type letters in "CE001" are converted to 5-bit binary codes, and the numerical part is converted to 7-bit binary codes, for a total of 12 bits. Then, a bit-concatenation operation merges these three parts into a single overall feature vector (e.g., 34 bits). Simultaneously, the dynamic hash index generation unit obtains the current system timestamp (e.g., 2025-03-15 10:30:00) and the device serial number (e.g., SN123456). It then arranges the binary bits of the timestamp and the binary bits of the device serial number in alternating order to form a dynamic offset. Next, it performs an XOR operation between the original KKS encoding and this dynamic offset. Finally, it performs a hash calculation on the XOR result (e.g., taking the first 64 bits of SHA-256) to obtain the dynamic hash index. The data caching unit temporarily stores the total feature vector and the dynamic hash index.

[0131] 2. QR code encoding

[0132] The QR code encoding module receives the total feature vector and dynamic hash index, and performs non-uniform redundancy distribution encoding.

[0133] The data flow segmentation unit splits the total feature vector into core semantic segment data streams (containing at least the binary bits corresponding to the function identifier segment and the system affiliation segment) and edge semantic segment data streams (containing device numbers and auxiliary parameters). The redundancy allocation decision unit assigns a first redundancy level to the core semantic segment according to a preset strategy, ensuring that its redundant data volume is no less than 40% of the original core semantic segment data volume; and assigns a second redundancy level to the edge semantic segment, ensuring that its redundant data volume is no more than 20% of the original edge semantic segment data volume. For devices with higher security levels (such as nuclear safety-related devices), the system automatically increases the first redundancy level to over 50%.

[0134] The error correction coding unit uses the Reed-Solomon coding algorithm to generate error correction codewords. The matrix region mapping unit maps the core semantic segment data stream and its error correction codewords to the central region of the QR code (defined as a circular area centered on the geometric center of the QR code, occupying 30% to 50% of the total area of ​​the QR code), and maps the edge semantic segment data stream and its error correction codewords to the edge region of the QR code (the surrounding area outside the central region). In addition, the QR code matrix generation unit introduces predictive error correction bits: based on the system affiliation segment in the core semantic segment, it retrieves all possible equipment number sets (e.g., CE001 to CE050) from the pre-stored process topology library on the server, encodes this set as a check bit (e.g., a 50-bit bitmap), and fills it into the reserved positions in the edge region. Finally, the QR code matrix generation unit embeds the self-check bit (containing the hash value of the QR code generation timestamp and the version check code) into the fixed reserved module of the QR code to generate the final QR code image. This QR code image is printed as a physical label and affixed to the power plant equipment.

[0135] III. Construction and Synchronization of Local Process Logic Library for Mobile Terminals

[0136] When a mobile terminal is used for the first time or during routine maintenance, the server-side logic library generation and synchronization module constructs a compressed prefix tree from the KKS codes of all devices. The construction process is as follows: a root node is created; the KKS codes of all devices are traversed, and each segment is inserted into the tree in hierarchical order of function identifier segment, installation location identifier segment, and device identifier segment, with shared prefix paths stored only once. Each node stores the corresponding binary feature code and attribute vector (including device type, security level, maintenance cycle, and standard operating procedure index). After completion, the compressed prefix tree is serialized into a binary data stream and sent to the mobile terminal via the wireless network.

[0137] The mobile terminal's logic library synchronization unit receives binary data streams, deserializes them into compressed prefix trees, and stores them in the local logic library storage unit (non-volatile memory). The logic library synchronization unit also automatically initiates synchronization requests to the server according to a preset period (e.g., 2 AM daily), or checks for updates after each network scan. The server compares version numbers and only transmits incremental changes (added, modified, or deleted device nodes), and the mobile terminal updates its local tree structure accordingly.

[0138] IV. Offline parsing workflow

[0139] In signal dead zones such as power plant boiler rooms and cable tunnels, mobile terminals cannot connect to the network and rely on offline parsing capabilities to function.

[0140] 1. Image Acquisition and Preprocessing

[0141] Inspection personnel use the built-in camera of their mobile terminals to point at the QR code labels on the equipment, and the image acquisition unit captures the raw images. The image preprocessing unit sequentially performs grayscale conversion, binarization, noise reduction, and distortion correction. For oily environments, an adaptive binarization algorithm is used: the image is divided into multiple small windows, and the binarization threshold for each window is calculated independently, thereby overcoming local brightness unevenness.

[0142] 2. Region Decoding and Prioritized Extraction of Core Semantics

[0143] The regional decoding unit uses three locations on the QR code to detect the graphic's boundaries and orientation, segmenting it into a central region and edge regions. This unit prioritizes decoding the central high-redundancy protected area. Because the central region stores the core semantic segment data stream and has high redundancy error correction capabilities, it can still be successfully decoded even if the QR code edges are obscured by oil or physically worn, extracting the system attribution vector (e.g., the binary codes of the function identifier segment and the system attribution segment). If the central region is successfully decoded, the unit continues to attempt to decode the edge regions; if the central region decoding fails, a decoding failure signal is output to the probability matching inference unit.

[0144] 3. Probability matching recursive restoration

[0145] After receiving the system attribution vector, the probabilistic matching inference unit uses this vector as an index to locate the corresponding subtree (e.g., the water pump system subtree) in the local compressed prefix tree. Then, it acquires the auxiliary positioning information of the mobile terminal (GPS coordinates or Bluetooth beacon signal strength). For each device node in the subtree, it calculates the Euclidean distance between its preset installation coordinates and the current terminal coordinates. This distance is then input into a Gaussian probability density function (mean 0, variance pre-calibrated based on the device spatial distribution density) to obtain the spatial matching probability. The node with the highest probability is selected as the candidate node.

[0146] If the edge region is partially readable, and the region decoding unit outputs a partially decoded string (e.g., "CE0"), the probabilistic matching inference unit calculates the edit distance (minimum number of insertion, deletion, and replacement operations) between the candidate node's device ID string (e.g., "CE001") and this partially decoded string. This edit distance similarity is then normalized and summed. A weighted summation is then performed: Overall Score = α × Spatial Matching Probability + (1-α) × Edit Distance Similarity, where α defaults to 0.6. When the mobile terminal detects no network signal, α is automatically adjusted to 0.8 or higher, making the inference more reliant on location information. Based on the overall score, the uniquely matching device node and its complete attribute vector are output.

[0147] If the regional decoding unit outputs a decoding failure signal (the central region is also completely unreadable), the probability matching inference unit searches for the nearest device node in the entire local logic library based solely on the auxiliary positioning information, and uses that node as the inference result. For example, if an inspector is standing next to water pump No. 1 and the Bluetooth beacon positioning is displayed in the "water pump area," the system will directly match the only water pump device in that area.

[0148] 4. Offline output

[0149] The offline output unit formats the attribute vectors of the matched device nodes into readable text or a graphical interface, displaying them on the mobile terminal screen. The content includes the device name, system to which it belongs, safety precautions, maintenance schedule, etc. This information can be obtained even without a network connection, ensuring the continuity of inspection work.

[0150] V. Online Requests and Dynamic Data Display

[0151] When the mobile terminal is in a network environment, in addition to performing the above offline parsing, it also makes online requests at the same time.

[0152] The online request unit obtains the dynamic hash index (usually stored in the central area of ​​the QR code, still readable even when damaged) from the regional decoding unit, and also obtains the identity tag of the currently logged-in user on the mobile terminal (such as operators, maintenance personnel, or security managers). The dynamic hash index, identity tag, and current timestamp are then encapsulated into a parsing request data packet according to a preset format and sent to the server's data feedback module via the wireless communication module.

[0153] After receiving a request, the server-side request receiving unit extracts the identity tag and compares it with the pre-stored authorization list. If authentication fails, a failure response is returned; if successful, the dynamic hash parsing unit queries the cache table based on the dynamic hash index to restore the original KKS code. The data retrieval unit uses the identity tag and the original KKS code to find the corresponding data source address in the role-data mapping library, and then retrieves the appropriate display content from the power plant's real-time database, maintenance management system, or safety management system. For example:

[0154] Operators: Real-time operating parameters (pressure, temperature, current), alarm thresholds, historical trend curves;

[0155] Maintenance personnel: Equipment maintenance history records, standard operating procedures (SOPs), spare parts inventory, 3D disassembly diagrams;

[0156] Safety management personnel: hot work permit status, hazard source distribution map, safety hazard investigation form.

[0157] The response sending unit encapsulates the displayed content into a response data packet and returns it to the mobile terminal. After receiving the packet, the online receiving unit transmits the content to the offline output unit for display, and also caches it in the local logic library storage unit for subsequent offline quick access.

[0158] VI. Tag Lifecycle Management

[0159] QR code physical tags will gradually age in the power plant environment. This system achieves proactive management through self-verification and recognition time monitoring.

[0160] When generating a QR code, the QR code encoding module embeds a self-verification bit in a fixed reserved location, which contains the hash value of the QR code generation timestamp. After each scan, the mobile terminal parsing module extracts the self-verification bit and calculates the difference between the generation timestamp and the current system time. If the difference exceeds a preset tag lifespan threshold (e.g., 2 years), a message "Tag expired, please replace" is added to the displayed content.

[0161] Meanwhile, the mobile terminal's parsing module records the recognition time for each scan (the time from image acquisition to successful decoding). If the recognition time for the same device's QR code exceeds a preset threshold (e.g., 800 milliseconds) three times consecutively, it indicates a severe degradation in tag quality. In this case, the mobile terminal automatically generates a tag replacement work order, including the device identifier, QR code image information, and time data, and sends it to the server via the wireless network.

[0162] The server-side tag lifecycle management module receives work orders, generates tag replacement tasks (including a list of devices to be replaced, suggested replacement time, and new tag printing parameters), and pushes them to the power plant maintenance management system. When the cumulative number of reported identification times for the same device exceeds a preset threshold (e.g., 5 times), the server automatically marks the tag status of that device as "pending replacement" and adds this reminder to the next logic database synchronization. After synchronization, the mobile terminal will proactively prompt "Tag is about to expire, please contact maintenance personnel for replacement" when scanning the device. Through this closed loop, full lifecycle management of QR code tags is achieved.

[0163] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A standardized generation and parsing management system for QR codes of power plant equipment based on KKS encoding, characterized in that, This includes a server and a mobile terminal that communicates with the server via a wireless network or a wired connection. The server includes: The data preprocessing module is configured to obtain the original KKS code of the equipment from the power plant equipment management system, and perform semantic vectorization processing on the original KKS code to generate corresponding feature vectors and dynamic hash indexes, so as to achieve data compression and logical decoupling; wherein, the dynamic hash index is: the hash value obtained by XORing the original KKS code with a dynamic offset and then performing hash calculation; the dynamic offset is data containing timestamp and equipment serial number; A QR code encoding module, connected to the data preprocessing module, is configured to receive the feature vector and dynamic hash index output by the data preprocessing module, and generate a QR code image based on the feature vector and dynamic hash index. When generating the error correction encoding matrix of the QR code image, a non-uniform redundancy allocation method is used. The data stream corresponding to the core semantic segment in KKS encoding, which includes at least a functional identifier segment and a system affiliation segment, is allocated to the high redundancy protection zone in the center of the QR code image, while a lower redundancy factor is allocated to the edge regions of the QR code image, or predictive error correction bits generated based on KKS logical topology are introduced. The data feedback module is connected to the QR code encoding module and is configured to receive a parsing request sent by the mobile terminal. The parsing request carries the identity tag of the scanner and a dynamic hash index obtained from the QR code image. Based on the identity tag and the dynamic hash index, the data feedback module matches and returns display content adapted to the identity tag from the multidimensional dataset to the mobile terminal. The mobile terminal includes a mobile terminal parsing module, configured as follows: In offline mode, the scanned QR code image is parsed, and the central high-redundancy protection zone is decoded first to extract the system attribution vector. Based on the system attribution vector, the local process logic library built on KKS encoding rules and auxiliary positioning information pre-stored on the mobile terminal, the complete equipment information and process association data are recursively restored through probabilistic matching based on KKS hierarchical topology. When the edge area of ​​the QR code image is damaged, making it impossible to directly decode the equipment number, the recursive restoration uses the decoded system attribution vector and the auxiliary positioning information to lock the unique matching equipment node from the local process logic library. Furthermore, the mobile terminal parsing module is also configured to send the decoded dynamic hash index along with the scanner's identity tag to the server's data feedback module, and receive the display content returned by the data feedback module.

2. The system according to claim 1, characterized in that, The data preprocessing module includes: The system comprises a KKS encoding and parsing unit, a semantic vectorization mapping unit, a dynamic hash index generation unit, and a data caching unit connected to the KKS encoding and parsing unit, the semantic vectorization mapping unit, and the dynamic hash index generation unit, respectively. The input terminal of the KKS encoding parsing unit is connected to the power plant equipment management system and is configured to receive the original KKS encoding and, according to the power plant KKS identification standard, segment the original KKS encoding into a functional identifier segment, an installation location identifier segment, and an equipment identifier segment. The functional identifier segment describes the function of the process system to which the equipment belongs, the installation location identifier segment describes the spatial location of the equipment in the power plant, and the equipment identifier segment describes the specific type and number of the equipment. The input of the semantic vectorization mapping unit is connected to the output of the KKS encoding and parsing unit, and is configured to receive the function identifier segment, installation location identifier segment, and equipment identifier segment, and perform binary feature vector mapping on each segment. Specifically, for the function identifier segment, the semantic vectorization mapping unit has a built-in system dictionary table, where each function code corresponds to a fixed-length binary feature code. For the installation location identifier segment, the semantic vectorization mapping unit uses a spatial grid-based encoding method, dividing the power plant space into multiple grid regions, each corresponding to a fixed-length binary spatial code. For the equipment identifier segment, the semantic vectorization mapping unit converts the equipment type and number into a binary sequence. The semantic vectorization mapping unit also merges the three binary feature codes into a single overall feature vector through bit concatenation. The input of the dynamic hash index generation unit is connected to the output of the KKS encoding parsing unit, and is configured to receive the original KKS encoding and obtain the current system timestamp and the device serial number of the device in the power plant management system; the dynamic hash index generation unit performs an XOR operation on the original KKS encoding and a dynamic offset, the dynamic offset being a value formed by arranging the binary bits of the current system timestamp and the binary bits of the device serial number in alternating order; the XOR operation result is then input into a hash function to calculate a hash value of fixed length, which serves as the dynamic hash index; The input of the data caching unit is connected to the output of the semantic vectorization mapping unit and the output of the dynamic hash index generation unit, respectively, and is configured to temporarily store the total feature vector and the dynamic hash index, and provide them to the QR code encoding module for reading.

3. The system according to claim 1, characterized in that, The QR code encoding module includes: The system includes a data stream segmentation unit, a redundancy allocation decision unit, an error correction coding unit, a QR code matrix generation unit, and a matrix region mapping unit connecting the error correction coding unit and the QR code matrix generation unit. The input end of the data stream segmentation unit is connected to the output end of the data preprocessing module, and is configured to receive the feature vector and the dynamic hash index, and further split the feature vector into core semantic segment data stream and edge semantic segment data stream; the core semantic segment data stream includes at least the parts of the feature vector corresponding to the function identifier segment and the system affiliation segment, and the edge semantic segment data stream includes the parts of the feature vector corresponding to the device number and auxiliary parameters; The redundancy allocation decision unit is connected to the data stream segmentation unit and is configured to allocate different error correction redundancy levels to the core semantic segment data stream and the edge semantic segment data stream according to a preset non-uniform redundancy allocation strategy; wherein, a first redundancy level is allocated to the core semantic segment data stream, and the amount of redundant data corresponding to the first redundancy level is not less than 40% of the original data volume of the core semantic segment data stream; a second redundancy level is allocated to the edge semantic segment data stream, and the amount of redundant data corresponding to the second redundancy level is not more than 20% of the original data volume of the edge semantic segment data stream. The error correction coding unit is connected to the redundancy allocation decision unit and is configured to receive the core semantic segment data stream, the edge semantic segment data stream and their respective redundancy levels, and generate corresponding error correction codewords using the Reed-Solomon coding algorithm. The matrix region mapping unit is connected to the error correction coding unit and is configured to map the core semantic segment data stream and its error correction codewords to the central region of the QR code matrix, and to map the edge semantic segment data stream and its error correction codewords to the edge region of the QR code matrix; the central region is defined as a circular or rectangular region centered on the geometric center of the QR code and occupying 30% to 50% of the total area of ​​the QR code; the edge region is defined as the surrounding area of ​​the QR code matrix excluding the central region. The QR code matrix generation unit is connected to the matrix region mapping unit and is configured to generate the final QR code image based on the mapped data bits and error correction code bits. In addition, for the insufficient data protection that may occur in the edge region due to the allocation of a lower redundancy level, the QR code matrix generation unit also introduces a predictive error correction bit generated based on KKS logical topology. The predictive error correction bit is a range of possible values ​​of the equipment number pre-calculated from the process topology library pre-stored on the server based on the system affiliation vector in the core semantic segment, and this range is encoded as a check bit and filled into the reserved position in the edge region.

4. The system according to claim 1, characterized in that, The data feedback module includes: The system includes a request receiving unit, an identity authentication unit, a dynamic hash parsing unit, a role-data mapping library, a data retrieval unit, and a response sending unit. The input terminal of the request receiving unit is connected to the wireless communication module of the mobile terminal and configured to receive the parsing request sent by the mobile terminal. The data packet format of the parsing request includes: a request header, an identity tag field, a dynamic hash index field, a timestamp field, and a verification field. The value of the verification field is a checksum obtained by hashing the concatenation of the identity tag field, the dynamic hash index field, and the timestamp field. The input terminal of the identity authentication unit is connected to the request receiving unit and is configured to extract the identity tag field from the parsed request and compare it with the pre-stored list of authorized identities in the request receiving unit; if the identity tag is not in the list of authorized identities, the identity authentication unit outputs an authentication failure signal to the response sending unit. The input of the dynamic hash parsing unit is connected to the request receiving unit and is configured to extract the dynamic hash index field from the parsing request and perform a reverse match between the dynamic hash index and the dynamic hash index generated by the data preprocessing module. The dynamic hash parsing unit internally maintains a dynamic hash index cache table, which stores the mapping relationship between the dynamic hash index generated by the data preprocessing module and its corresponding original KKS code. The dynamic hash parsing unit queries the cache table based on the received dynamic hash index to restore the corresponding original KKS code. The role-data mapping library is a relational database table stored in the server-side storage medium. The relational database table contains at least the following fields: role identifier, data source address, and data display template. The role identifier includes operator identifier, maintenance personnel identifier, and safety management personnel identifier. The data source address points to a specific dataset in the power plant's real-time database, maintenance management system, or safety management system. The input terminal of the data retrieval unit is connected to the output terminal of the identity authentication unit, the output terminal of the dynamic hash parsing unit, and the role-data mapping library, respectively. The configuration is as follows: upon receiving an authentication pass signal from the identity authentication unit, based on the parsed identity tag and original KKS code, the corresponding data source address is searched in the role-data mapping library, and then the display content adapted to the identity tag is retrieved from the corresponding database based on the data source address; for operator identifiers, real-time operating parameters and alarm thresholds are retrieved; for maintenance personnel identifiers, equipment maintenance history records, standard operating cards, and spare parts inventory information are retrieved; for safety management personnel identifiers, hot work permit status, hazard distribution, and safety hazard investigation forms are retrieved. The response sending unit is connected to the output of the data retrieval unit and the identity authentication unit respectively, and is configured to: if an authentication failure signal is received from the identity authentication unit, return an authentication failure response to the mobile terminal; if the display content is received from the data retrieval unit, encapsulate the display content into a response data packet and send it to the mobile terminal via the wireless network.

5. The system according to claim 1, characterized in that, The mobile terminal parsing module includes: Image acquisition unit, image preprocessing unit, region decoding unit, local logic library storage unit, probability matching inference unit, offline output unit, online request unit and online receiving unit; The image acquisition unit is a camera built into the mobile terminal, configured to scan the QR code image attached to the power plant equipment and output the acquired raw image data to the image preprocessing unit. The input end of the image preprocessing unit is connected to the image acquisition unit and is configured to sequentially perform grayscale conversion, binarization, noise reduction and distortion correction on the original image data; wherein, for the oil pollution environment of the power plant, the image preprocessing unit has a built-in adaptive binarization algorithm, which dynamically adjusts the binarization threshold according to the pixel brightness distribution of the local area of ​​the image to enhance the edge features of the QR code in the oil pollution occlusion area; The input of the region decoding unit is connected to the output of the image preprocessing unit, and is configured to perform region segmentation on the preprocessed QR code image to identify the central region and edge regions. The region decoding unit first decodes the central region and extracts the system attribution vector stored therein. If the central region is successfully decoded, the region decoding unit outputs the system attribution vector to the probability matching inference unit. If the central region is not decoded, the region decoding unit outputs a decoding failure signal to the probability matching inference unit. The local logic library storage unit is a non-volatile memory, which pre-stores a local process logic library constructed based on KKS encoding rules. The data structure of the local process logic library is a compressed prefix tree. Each node in the compressed prefix tree corresponds to a level segment of KKS encoding. The root node corresponds to the function identifier segment, and the child nodes correspond to the installation location identifier segment and the equipment identifier segment, respectively. Multiple KKS codes with the same prefix share nodes on the prefix path. Each node in the local process logic library is also associated with and stored as an attribute vector, which includes: equipment type, security level, maintenance cycle, and standard operating procedure index. The input of the probabilistic matching inference unit is connected to the output of the region decoding unit, the output of the local logic library storage unit, and the positioning module of the mobile terminal, respectively. The probabilistic matching inference unit is configured to: when the system attribution vector is received, locate the corresponding subtree in the local logic library storage unit using the system attribution vector as an index; then obtain auxiliary positioning information provided by the positioning module, including GPS coordinates or Bluetooth beacon signal strength; the probabilistic matching inference unit calculates the spatial matching probability between each device node in the subtree and the current auxiliary positioning information, and selects the device node with the highest probability as a candidate node; if the region... The domain decoding unit simultaneously outputs partial edge region decoding results. The probabilistic matching inference unit also calculates the edit distance between the device number string of the candidate node and the partial decoded string output by the region decoding unit, and performs a weighted summation of the spatial matching probability and the edit distance similarity. The weight coefficient of the spatial matching probability is automatically increased when there is no network signal. Based on the weighted summation result, the uniquely matching device node and its complete attribute vector are output. If the region decoding unit outputs a decoding failure signal, the probabilistic matching inference unit searches for the device node with the closest spatial location in the entire local logic library based only on the auxiliary positioning information, and uses that node as the inference result. The input terminal of the offline output unit is connected to the output terminal of the probability matching inference unit, and is configured to format the information in the complete attribute vector into readable text or a graphical interface, and display it on the mobile terminal screen; the readable text includes at least the device name, the system to which it belongs, and security precautions; The input end of the online request unit is connected to the output end of the regional decoding unit, and is configured to obtain the dynamic hash index decoded by the regional decoding unit from the QR code image, and obtain the identity tag of the currently logged-in user of the mobile terminal. The dynamic hash index, identity tag and current timestamp are encapsulated into a parsing request data packet according to a preset format and sent to the data feedback module of the server through the wireless communication module of the mobile terminal. The input end of the online receiving unit is connected to the wireless communication module of the mobile terminal, and the output end is connected to the offline output unit. It is configured to receive the display content returned by the data feedback module and transmit the display content to the offline output unit for display. When the online request unit successfully sends a request and receives the display content, the online receiving unit also caches the display content in the local logic library storage unit for quick access in the subsequent offline state.

6. The system according to claim 2, characterized in that, The semantic vectorization mapping unit is also configured to perform bit-map encoding based on character set compression; Specifically, for the characters in the function identifier segment, installation location identifier segment, and device identifier segment, the semantic vectorization mapping unit uses 5-bit binary encoding instead of the standard 8-bit ASCII encoding. The 5-bit binary code corresponding to each character is obtained by querying a predefined character mapping table. The character mapping table covers all uppercase letters A to Z and numbers 0 to 9 that are allowed to appear in KKS encoding. The semantic vectorization mapping unit converts each character in the function identifier segment, installation location identifier segment, and equipment identifier segment into a corresponding 5-bit binary code in sequence, and then concatenates the binary codes of each character in order to form the binary feature vectors of the function identifier segment, installation location identifier segment, and equipment identifier segment respectively. The bit concatenation operation specifically involves: taking the binary feature vector of the function identifier segment as the high-order part, the binary feature vector of the installation location identifier segment as the middle part, and the binary feature vector of the device identifier segment as the low-order part, and concatenating them sequentially from high-order to low-order to obtain the total feature vector; the length of the total feature vector does not exceed 40 bits.

7. The system according to claim 3, characterized in that, The redundancy allocation decision unit is further configured to dynamically adjust the specific values ​​of the first redundancy level and the second redundancy level according to the device type corresponding to the system affiliation segment in the core semantic segment data stream. For devices with a security level higher than a preset threshold in the core semantic segment data stream, the amount of redundant data corresponding to the first redundancy level is increased to more than 50% of the original data amount of the core semantic segment data stream; for devices with a security level not higher than the preset threshold, the amount of redundant data corresponding to the first redundancy level is set to between 30% and 40% of the original data amount of the core semantic segment data stream. The process of generating the predictive error correction bits is as follows: The QR code matrix generation unit extracts the system affiliation segment from the core semantic segment data stream, uses the system affiliation segment as the query key, and retrieves all possible equipment number sets in the system to which the equipment belongs from the pre-stored process topology library on the server; the equipment number set is encoded into a series of check bits, and the data length of the check bits is determined according to the size of the equipment number set. The QR code matrix generation unit fills the check bits into the reserved positions in the edge region, where the reserved positions are part of the error correction code bits originally used to store the edge semantic segment data stream in the edge region; When the mobile terminal parsing module finds that the device number data in the edge area is unreadable during decoding, it reads the check bit in the reserved position and, in combination with the decoded system affiliation segment, reverse-engineers the correct device number from the local process logic library.

8. The system according to claim 5, characterized in that, The spatial matching probability calculation method in the probability matching inference unit is as follows: For each device node in the subtree, obtain the preset installation coordinates of the device node in the power plant spatial coordinate system; obtain the current terminal coordinates output by the mobile terminal positioning module; calculate the Euclidean distance between the preset installation coordinates and the current terminal coordinates; input the Euclidean distance into a Gaussian probability density function to obtain the spatial matching probability of the device node; the mean of the Gaussian probability density function is set to 0, and the variance is pre-calibrated according to the spatial distribution density of the power plant equipment; The edit distance similarity is calculated as follows: align the device ID string of the candidate node with the partially decoded string output by the region decoding unit; calculate the minimum number of single-character edit operations required to convert the partially decoded string into the device ID string of the candidate node, where the single-character edit operations include insertion, deletion, and replacement; divide the minimum number of single-character edit operations by the length of the device ID string of the candidate node to obtain the normalized edit distance; the edit distance similarity is equal to 1 minus the normalized edit distance. The specific method of weighted summation is as follows: the comprehensive score is equal to the first weight coefficient multiplied by the spatial matching probability plus one minus the first weight coefficient multiplied by the edit distance similarity, wherein the first weight coefficient is the weight coefficient of the spatial matching probability; the default value of the first weight coefficient is 0.6; when the mobile terminal detects no network signal, the first weight coefficient is adjusted to 0.8 or higher.

9. The system according to claim 1, characterized in that, The server also includes a logic library generation and synchronization module, and the mobile terminal also includes a logic library synchronization unit. The input terminal of the logic library generation and synchronization module is connected to the data preprocessing module and configured to: acquire the feature vectors of all devices processed by the data preprocessing module and their corresponding original KKS codes; construct a compressed prefix tree of the KKS codes of all devices according to the hierarchical structure of the KKS codes, wherein each node of the compressed prefix tree stores the binary feature code of the KKS segment corresponding to the node and the attribute vector of the node; the logic library generation and synchronization module also serializes the compressed prefix tree into a binary data stream and sends it to the mobile terminal via a wireless network or wired connection; The input end of the logic library synchronization unit of the mobile terminal is connected to the wireless communication module of the mobile terminal, and the output end is connected to the local logic library storage unit. It is configured to receive the binary data stream sent by the server, deserialize the binary data stream into a compressed prefix tree structure, and store it in the local logic library storage unit to update or replace the original local process logic library. The logic library synchronization unit is also configured to automatically initiate a synchronization request to the server according to a preset synchronization cycle, or to check whether the server has an updated process logic library version after each successful network scan by the mobile terminal, and if so, to trigger incremental synchronization; the incremental synchronization only transmits the device nodes and their attribute vectors that have changed since the last synchronization, so as to reduce the amount of data transmitted over the network.

10. The system according to claim 9, characterized in that, The QR code encoding module is also configured to embed a self-verification bit in the QR code image; the self-verification bit includes the hash value of the QR code generation timestamp and the QR code version verification code; The mobile terminal parsing module is further configured to: after parsing the QR code image, extract the self-verification bit and calculate the difference between the QR code generation timestamp obtained by decoding and the current system time; If the difference exceeds the preset tag lifespan threshold, the mobile terminal parsing module will add an expiration tag and prompt for replacement to the displayed content; The mobile terminal parsing module also records the recognition time for each scan. When the recognition time exceeds the preset time threshold for three consecutive scans, the mobile terminal parsing module automatically generates a tag replacement work order and sends the tag replacement work order to the server via the wireless network. The server also includes a tag lifecycle management module, configured to receive and store the tag replacement work order, generate a tag replacement task based on the equipment identifier and QR code image information in the tag replacement work order, and push it to the power plant maintenance management system. The server is also configured to automatically mark the label status of the device as pending replacement when the cumulative number of times the QR code image of the same device is reported to exceed the threshold for recognition time reaches a preset number, and add the label replacement reminder to the next synchronization data generated by the logic library and the synchronization module.

Images