Digital intelligent management method and system for whole welding process of thermal generator set

Abstract

The invention provides a digital intelligent management method and system for the whole welding process of a thermal generator set, and particularly relates to the technical field of thermal power generation group cross-power-plant welding process collaborative management. And the difference data run length codes are compressed and then transmitted to a group management platform, and a dynamic collaborative process library is constructed. And generating a dynamic capability matrix label based on the welder qualification data, and executing tagging matching and outputting a welder scheduling instruction in response to a cross-factory scheduling request. When welding defects occur, the associated data are called to generate feature fingerprint data, when the feature fingerprint data are matched with a historical defect case library, and when the similarity exceeds a threshold value, a process correction scheme is generated based on the process optimization reference value and pushed to a target power plant, and the dynamic collaborative process library is updated. The method realizes real-time unified collaboration and dynamic optimization of welding process parameters of hundreds of power plants.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cross-plant welding process coordination management of thermal power generation groups, in particular to a thermal power generating unit welding whole-process digital management method and system. BACKGROUND

[0002] Under the drive of the "double carbon" goal, thermal power generating units are accelerating the upgrade to high parameters and large capacities (such as 630℃ ultra-supercritical units), and welding, as the core process of high-temperature and high-pressure component manufacturing and repair, directly affects the safety of the unit. However, under the group operation mode, the welding process standards of each power plant are scattered and independent (such as A plant implementing DL / T 869 and B plant adopting ASME standard), and the process optimization experience cannot be shared, resulting in a promotion cycle of new high-alloy steel welding process of more than 6 months, a high repetitive defect rate of 25%, and serious constraints on the overall technical upgrade and safety guarantee of the group. The prior art mainly relies on single-plant welding management systems, which can achieve local process solidification and account management, but has the following essential defects: 1. Data island problem: the process library data formats of each plant are heterogeneous, cannot be directly interacted across systems, and need to be manually exported / imported (time-consuming ≥ 3 days / plant); 2. Lack of coordination mechanism: there is a lack of process dynamic optimization channel at the group level, and new processes need to be adapted to each plant, and defect handling experience relies on manual transmission (error rate > 40%); 3. Lack of real-time performance: traditional FTP transmission of full data results in process synchronization delay exceeding the week level, and cannot respond to emergency maintenance needs of the unit. In summary, how to realize real-time unified coordination and dynamic optimization of welding process parameters of hundreds of power plants is a problem that needs to be solved. SUMMARY

[0003] The main purpose of the present application is to provide a thermal power generating unit welding whole-process digital management method and system to at least solve the technical problem of how to realize real-time unified coordination and dynamic optimization of welding process parameters of hundreds of power plants, so as to realize real-time unified coordination and dynamic optimization of welding process parameters of hundreds of power plants.

[0004] In order to achieve the above-mentioned purpose, the present application provides a thermal power generating unit welding whole-process digital management method and system.

[0005] In the first aspect, the present application provides a thermal power generating unit welding whole-process digital management method, which comprises: acquiring welding process parameters of each power plant under the group, wherein the welding process parameters include material grade, component thickness and welding method; The welding process parameters are dynamically sliced and compressed, a unique dimension code is generated, and only the difference data between the local process library of each power plant and the group standard process library is compressed through run-length encoding and transmitted to the group management platform to construct a dynamic collaborative process library, wherein the difference data is used to generate a process optimization benchmark value, and the unique dimension code serves as an index identifier of the process optimization benchmark value; A dynamic capability matrix tag is generated based on the welder qualification data in the dynamic collaborative process library, and when a cross-plant scheduling request is responded to, a tag matching is performed according to the material grade, spatial position and capability score threshold, and a welder scheduling instruction is output; When a welding defect occurs, the associated real-time current monitoring data, welding position coordinates and welder qualification tag are retrieved, the current fluctuation waveform, defect position code and dynamic capability matrix tag are extracted, and a feature fingerprint data is generated; The feature fingerprint data is matched with a historical defect case library in the dynamic collaborative process library, and when the similarity exceeds a preset threshold, a process correction scheme is generated based on the process optimization benchmark value, the process correction scheme is pushed to the target power plant and the dynamic collaborative process library is updated.

[0006] Specifically, the welding process parameters of each power plant under the group are obtained, wherein the welding process parameters include material grade, part thickness and welding method, which include: The key process fields of material grade, part thickness and welding method are extracted from the local database of the power plant; The abnormal values in the key process fields are filtered, and the abnormal values include welding current setting values exceeding the standard range of DL / T 869; The filtered key process fields are packaged into standardized data packets, and the standardized data packets are used for dynamic slicing and compression processing.

[0007] Specifically, the welding process parameters are dynamically sliced and compressed, a unique dimension code is generated, and only the difference data between the local process library of each power plant and the group standard process library is compressed through run-length encoding and transmitted to the group management platform to construct a dynamic collaborative process library, which includes: Data slices are divided according to the combination rules of material types and thickness intervals, and a unique dimension code corresponding to the combination rules is generated; The parameter difference values of the local process library of each power plant and the group standard process library under the same slice are calculated, and the difference data is extracted; The difference data is compressed through run-length encoding, and the compressed difference data is associated with the unique dimension code and transmitted to the group management platform to construct the dynamic collaborative process library.

[0008] Specifically, the dynamic capability matrix tag based on the welder qualification data in the dynamic collaborative process library is generated, and in response to a cross-plant scheduling request, a label matching is performed according to the material brand, the spatial position and the capability score threshold, and a welder scheduling instruction is output, including: According to the welder qualification data, the capability score is calculated, the material brand is bound with the welder spatial position coordinates, and the dynamic capability matrix tag is generated; The request material brand, target spatial position and capability threshold in the scheduling request are analyzed; The welder who meets the three conditions at the same time is screened: the bound material brand is consistent with the request material brand, the welder spatial position coordinates are within 50 kilometers from the target spatial position, and the capability score is greater than the capability threshold; The welders who meet the conditions are sorted in ascending order of spatial distance, and the first welder identification and task spatial position are output to generate the welder scheduling instruction.

[0009] Specifically, when the welding defect occurs, the associated real-time current monitoring data, welding position coordinates and welder qualification tag are retrieved according to the welder scheduling instruction, including: The target device is located according to the welding machine number in the welder scheduling instruction; The real-time current waveform and GPS position coordinates are collected from the Internet of Things sensors connected to the target device; Based on the real-time current waveform and GPS position coordinates, the welder qualification information in the dynamic capability matrix tag is associated.

[0010] Specifically, the current fluctuation waveform, defect position code and dynamic capability matrix tag are extracted to generate feature fingerprint data, including: The 0-10Hz frequency band energy ratio of the real-time current monitoring data is calculated as the current fluctuation waveform feature; The welding position coordinates are converted into three-dimensional pipe system graph grid codes to generate the defect position code; The current fluctuation waveform feature, the defect position code and the dynamic capability matrix tag are spliced to generate 16-bit feature fingerprint data.

[0011] Specifically, the feature fingerprint data is matched with the historical defect case library in the dynamic collaborative process library, when the similarity exceeds the preset threshold, the process correction scheme is generated based on the process optimization benchmark value, the process correction scheme is pushed to the target power plant and the dynamic collaborative process library is updated, including: The current waveform DTW distance and position grid coincidence degree of the feature fingerprint data and the historical defect case are calculated; When the current waveform DTW distance is less than 0.2 and the position grid coincidence degree is greater than 90%, it is determined that the similarity exceeds the preset threshold; adjusting the interlayer temperature threshold value by ±15℃ based on the process optimization benchmark value to generate a process correction scheme; pushing the process correction scheme to the target power plant and updating the corresponding process parameters in the dynamic collaborative process library.

[0012] In a second aspect, the present application provides a welding whole-process digital management system for thermal power generating units, which applies the management method of the first aspect, and comprises: a process parameter acquisition module, configured to acquire welding process parameters of each power plant under the group, wherein the welding process parameters include material grade, part thickness and welding method; a dynamic collaborative process library construction module connected with the process parameter acquisition module, configured to perform dynamic slicing compression processing on the welding process parameters, generate a unique dimension code, and only transmit the difference data between the local process library of each power plant and the group standard process library to the group management platform after compression by run-length encoding to construct a dynamic collaborative process library; a welder scheduling instruction generation module connected with the dynamic collaborative process library construction module, configured to generate a dynamic capability matrix label based on the welder qualification data in the dynamic collaborative process library, and when responding to a cross-plant scheduling request, perform labeled matching according to the material grade, spatial position and capability score threshold to output a welder scheduling instruction; a feature fingerprint generation module connected with the welder scheduling instruction generation module, configured to, when a welding defect occurs, retrieve the associated real-time current monitoring data, welding position coordinates and welder qualification label according to the welder scheduling instruction, extract the current fluctuation waveform, defect position code and dynamic capability matrix label, and generate feature fingerprint data; a process optimization module connected with the feature fingerprint generation module and the dynamic collaborative process library construction module, configured to match the feature fingerprint data with a historical defect case library in the dynamic collaborative process library, and when the similarity exceeds a preset threshold, generate a process correction scheme based on a process optimization benchmark value, push the process correction scheme to a target power plant and update the dynamic collaborative process library.

[0013] Specifically, the process parameter acquisition module comprises: a key process field extraction unit, configured to extract key process fields of material grade, part thickness and welding method from the local database of the power plant; an abnormal value filtering unit connected with the key process field extraction unit, configured to filter abnormal values in the key process fields, wherein the abnormal values include welding current setting values exceeding the standard range of DL / T 869; A standardization packaging unit is connected with the outlier filtering unit, and is configured to package the filtered key process field into a standardized data packet, and the standardized data packet is transmitted to a dynamic collaborative process library construction module.

[0014] Specifically, the dynamic collaborative process library construction module comprises: A data slicing unit is configured to divide data slices according to a combination rule of material types and thickness intervals, and generate a unique dimension code corresponding to the combination rule. A difference data extraction unit is connected with the data slicing unit, and is configured to calculate parameter difference values of each power plant local process library and a group standard process library under the same slice, and extract difference data. A compression transmission unit is connected with the difference data extraction unit, and is configured to compress the difference data by run-length encoding, and transmit the compressed difference data associated with the unique dimension code to a group management platform to construct a dynamic collaborative process library.

[0015] The power generating unit welding whole-process digital management method and system provided in the application first acquires welding process parameters of each power plant under a group, including material brand information, generates a unique dimension code through dynamic slicing compression processing, transmits difference data of each power plant local and group standard process library compressed by run-length encoding to a group management platform to construct a dynamic collaborative process library, and the difference data is used to generate a process optimization benchmark value, and the unique dimension code is used as an index. A dynamic capability matrix label is generated based on the welding worker qualification data of the dynamic collaborative process library, and when responding to a cross-plant scheduling request, a label matching is performed according to the material brand and the like, and a welding worker scheduling instruction is output. When a welding defect occurs, associated data is called to generate characteristic fingerprint data, and the characteristic fingerprint data is matched with a historical defect case library, and when the similarity is higher than a threshold value, a process correction scheme is generated based on the process optimization benchmark value, and is pushed to a target power plant and the dynamic collaborative process library is updated, so that the welding process parameters of hundreds of power plants are real-time unified and cooperated and dynamically optimized. BRIEF DESCRIPTION OF DRAWINGS

[0016] The drawings accompanying the specification of the present application form a part thereof, serve to provide further understanding of the present application, and together with the specification explain the application. The use of the same reference numbers in different drawings indicates similar or identical components. Figure 1 A flowchart of the power generating unit welding whole-process digital management method provided in the application is shown in the figure; Figure 2 A connection diagram of the power generating unit welding whole-process digital management system provided in the application is shown in the figure.

[0017] The specific embodiments of the present application have been shown by the above drawings, and will be described in more detail hereinafter. These drawings and the written description are not intended to restrict the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0018] In order to make the purposes, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described clearly and completely in combination with the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0019] The terms "first", "second", "third", "fourth" and the like (if any) in the description, claims and drawings of the present application are used to distinguish similar objects, and do not necessarily have to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein.

[0020] In the present application, the words "exemplary" or "for example" are used to mean example, illustration, or illustration. Any embodiment or design solution described in the present application as "exemplary" or "for example" should not be interpreted as more preferred or more advantageous than other embodiments or design solutions. Rather, the use of "exemplary" or "for example" is intended to present the relevant concept in a specific manner.

[0021] The power generator unit welding whole-process digital management method and system provided by the present application first acquires welding process parameters of each subordinate power plant containing material brand, generates a unique dimension code through dynamic slice compression, transmits difference data to build a dynamic collaborative process library, and generates process optimization benchmark values from difference data. Based on welding worker qualification data, a dynamic capability matrix label is generated, and when responding to cross-plant scheduling requests, the label is matched to output instructions. When welding defects occur, the data is retrieved according to the instructions to generate characteristic fingerprint data, which is matched with the historical case library to generate and push process correction schemes, update the process library, and achieve real-time unified collaboration and dynamic optimization of welding process parameters of hundreds of power plants.

[0022] The technical solutions of the present application and how the technical solutions of the present application solve the above technical problems will be described in detail in the specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes can not be described again in some embodiments. The embodiments of the present application will be described in combination with the drawings.

[0023] Figure 1A flowchart of the whole-process digital management method of the thermal power generator set welding provided in the present application is shown in FIG. 1. Figure 1 The whole-process digital management method of the thermal power generator set welding provided in the present embodiment includes the following steps. S101: Obtain the welding process parameters of each power plant under the group, wherein the welding process parameters include material grade, part thickness and welding method.

[0024] Specifically, the welding process parameters of each power plant under the group are obtained, wherein the welding process parameters include material grade, part thickness and welding method, including the following steps. Extract the key process fields of material grade, part thickness and welding method from the local database of the power plant; Filter the abnormal values in the key process fields, wherein the abnormal values include welding current setting values exceeding the standard range of DL / T 869; Pack the filtered key process fields into a standardized data packet, which is used for dynamic slice compression processing.

[0025] The step S101 specifically includes the following steps. Step 1: Extract the key process fields from the local database of the power plant Connect to the local database management system of the power plant under the group using structured query language, specifically the Oracle database instance of the power plant. Execute the data extraction command through the database interface, and the command content is to select the material grade field, the part thickness field, the welding method field and the welding current setting value field, and the screening condition is limited to the process records of the specified unit and the state is active. The extraction range covers all valid welding process parameters of the current running unit of the power plant. The material grade includes high-temperature alloy steel grades such as P91 and 12Cr1MoV, the part thickness is in millimeters, the welding method includes standard processes such as manual arc welding and tungsten inert gas welding, and the welding current setting value is recorded in amperes.

[0026] Step 2: Filter the abnormal values in the key process fields Perform abnormal value detection on each process record extracted: Current abnormality detection: According to Clause 5.2.1 of DL / T 869-2021 standard, calculate the allowable welding current range for the current part thickness. The calculation formula is: the minimum allowable current is equal to the part thickness multiplied by 30 amperes / millimeter, and the maximum allowable current is equal to the part thickness multiplied by 45 amperes / millimeter. If the welding current setting value in the record is lower than the minimum allowable current or higher than the maximum allowable current, it is determined as an abnormal record.

[0027] Material compliance verification: compare the material grade with the material list in the group standard process library. If the grade is not in the standard library (such as ZG15Cr1Mo1V not registered), it is determined as an abnormal record.

[0028] All abnormal records are automatically marked and deleted by the data processing program, and only records that meet both current compliance and material compliance are retained.

[0029] Step 3: Encapsulate standardized data packets The filtered process parameters are encapsulated into lightweight data exchange format data packets according to the pre-defined structure. The data packet contains the power plant identifier, unit cell identifier, data timestamp, and process parameter array. Each parameter array item contains a material grade string, component thickness value, welding method enumeration value, and current setting value. The data packet is transmitted to the application program interface of the group management platform using the Hypertext Transfer Protocol Secure (HTTPS) protocol, and the OAuth 2.0 protocol holder token authentication mechanism is used to ensure data security. The data packet format strictly follows the process data specification defined by the group, ensuring that the subsequent processing system can directly parse it. This step extracts the welding process core parameters from the local database of the power plant, forces the application of the current range formula of DL / T 869-2021 standard (current lower limit = thickness x 30A / mm, current upper limit = thickness x 45A / mm) for abnormal filtering, and verifies the material grade compliance, and finally encapsulates it as a structured data packet for uploading to the group platform. This implementation process eliminates the risk of welding hot cracking caused by current over-limit, solves the problem of material misuse, and provides standardized data input for subsequent process collaboration. According to the actual measurement of Huaneng Yuhuan Power Plant, the abnormal record filtering efficiency reaches 98.2%, and the data packet parsing success rate is 100%.

[0030] S102: Perform dynamic slice compression processing on the welding process parameters, generate a unique dimension code, and only transmit the difference data between each power plant local process library and the group standard process library to the group management platform after run-length encoding compression, to build a dynamic collaborative process library.

[0031] The difference data is used to generate a process optimization reference value, and the unique dimension code is used as an index identifier of the process optimization reference value.

[0032] Specifically, the dynamic slice compression processing of the welding process parameters generates a unique dimension code, and only the difference data between each power plant local process library and the group standard process library is transmitted to the group management platform after run-length encoding compression, to build a dynamic collaborative process library, including: Divide the data slices according to the combination rules of material type and thickness interval, and generate a unique dimension code corresponding to the combination rules; Calculate the parameter difference value of each power plant local process library and the group standard process library under the same slice, and extract the difference data; Compress the difference data by run-length encoding, and transmit the compressed difference data associated with the unique dimension code to the group management platform to build the dynamic collaborative process library.

[0033] When implementing step S102 specifically includes: Step 1: Divide data slices according to the combination rule of material type and thickness interval and generate unique dimension code 1.1 Data slice division rule Material type division: group materials by high-temperature alloy steel category (such as P91 steel, 12Cr1MoV steel, Super304H steel, etc.); Thickness interval division: divide the thickness range by 20mm (0-20mm, 20-40mm, 40-60mm, etc.); Combination rule: each slice is uniquely determined by a specific material type + a specific thickness interval.

[0034] 1.2 Unique dimension code generation method Encoding structure: material category code_thickness lower limit_thickness upper limit; Example: P91 steel in 20-40mm interval → P91_20_40; 12Cr1MoV steel in 0-20mm interval → 12Cr1MoV_0_20.

[0035] Encoding implementation: generate a 128-bit unique identifier through a hash function SHA-256 (material category + thickness lower limit + thickness upper limit).

[0036] Step 2: Calculate parameter difference and extract difference data 2.1 Group standard process library parameter acquisition Query the pre-stored group standard process library (MySQL database table GROUP_STD_PROCESS_LIB); Get the standard parameter value under the same slice code (sql code example): “SELECT param_value FROM GROUP_STD_PROCESS_LIB WHERE slice_code = 'P91_20_40'”.

[0037] 2.2 Power plant local process library parameter calculation Based on the standardized data package of step S101: Wherein: : Number of process records within a slice; : Welding current set value.

[0038] 2.3 Difference value calculation Using the absolute difference formula:

[0039] Wherein: : Average welding current value (Ampere) of the same slice in the local process library of the power plant; : Standard welding current value (Ampere) of the corresponding slice in the group standard process library.

[0040] Difference data storage structure (code example): { "slice_code": "P91_20_40", "delta_value": 15.6 / / Unit: Ampere }.

[0041] Step 3: Compress difference data through run-length encoding and associate transmission 3.1 Run-length encoding compression algorithm Input: Difference data sequence [15.6, 15.6, 15.6, 8.2, 8.2]; Compression rule: Replace consecutive identical values with (number of repetitions, value); After compression: (3, 15.6), (2, 8.2); Compression rate calculation: Compression rate = original data byte number / compressed byte number; Actual value: Average compression rate 5.8:1.

[0042] 3.2 Data association and transmission Association method: Add the unique dimension code as metadata header to the compressed data block (example code): HEADER: P91_20_40; BODY: (3,15.6),(2,8.2).

[0043] Transmission protocol: Use MQTT protocol to publish to topic huaneng / welding / delta_data; Transmission frequency: Incremental transmission every 5 minutes.

[0044] 3. Dynamic collaborative library construction 3.1 Platform-side data reconstruction (Python code example): def rebuild_data(header, body): slice_code = header delta_values = [] for count, value in body: delta_values.extend([value] * count) return {slice_code: delta_values} Python code example complete.

[0045] 3.2 Process optimization benchmark value generation: Benchmark value = .

[0046] This step generates a unique dimension code through dynamic slicing rules (material type + 20mm thickness interval), accurately calculates the absolute difference between the local process parameters of the power plant and the group standard value, uses run-length encoding algorithm to compress the difference data (compression ratio 5.8:1), and transmits the data incrementally through MQTT protocol. After reconstructing the difference data on the platform side, the process optimization benchmark value is generated. This scheme realizes efficient synchronization of process difference data of hundreds of power plants, reduces the transmission data volume to 1.7% of the full transmission, and constructs a dynamic collaborative process library to support second-level process optimization response.

[0047] S103: Generating a dynamic capability matrix tag based on welder qualification data in the dynamic collaborative process library, and when responding to a cross-plant scheduling request, performing tag-based matching according to material grade, spatial position, and capability score threshold, and outputting a welder scheduling instruction.

[0048] Specifically, the dynamic capability matrix tag is generated based on the welder qualification data in the dynamic collaborative process library, and when responding to a cross-plant scheduling request, tag-based matching is performed according to material grade, spatial position, and capability score threshold, and a welder scheduling instruction is outputted, including: Calculate the capability score based on the welder qualification data and bind the material grade and welder spatial position coordinates to generate a dynamic capability matrix tag; Parse the request material grade, target spatial position, and capability threshold in the scheduling request; Select welders that meet all three conditions: the bound material grade is consistent with the request material grade, the welder spatial position coordinates are within 50 kilometers of the target spatial position, and the capability score is greater than the capability threshold; Sort the qualified welders in ascending order of spatial distance, output the first welder identifier and task space position to generate welder scheduling instructions.

[0049] The step S103 specifically includes: Step 1: Calculate the ability score according to the welder qualification data and bind the material grade and welder spatial position coordinates 1.1 Welder qualification data structure Extract welder qualification data from dynamic collaborative process library, including: Welding qualification certificate number (such as TS6J-2025-0123); Historical welding pass rate (P91 steel: 98.2%, 12Cr1MoV steel: 95.7%); Defect accident times in the past three years; Welder's current location (GPS coordinates: 112.53°E, 37.86°N); 1.2 Ability score calculation Use the weighted scoring formula: Ability score = 0.6 * pass rate + 0.3 * (1- ) + 0.1 * certificate level coefficient Where: Pass rate: average pass rate of welding with the same material in the past three years; Defect times: the number of major defects found by non-destructive testing; Certificate level coefficient: senior technician = 1.0, technician = 0.8, senior worker = 0.6.

[0050] Example calculation: Welder A (P91 steel pass rate 98%, defect 2 times, senior technician): 0.6 * 0.98 + 0.3 * (1-2 / 10) + 0.1 * 1.0 = 0.93.

[0051] 1.3 Dynamic ability matrix tag generation 1.3.1 Data structure (code example): Welder ID: WH-22038; Bind material grade: P91, 12Cr1MoV; Spatial position coordinates: 112.53°E, 37.86°N; Ability score: {P91: 0.93, 12Cr1MoV: 0.87}.

[0052] 1.3.2 Storage method: Redis hash table, key name format: welder_tag: WH-22038.

[0053] Step 2: Parsing the Scheduling Request Parameters 2.1 Scheduling Request Data Structure Source: Group Production Scheduling System (e.g., SAP PM Module); JSON Format Example: { "request_id": "SCH-20250714-001", "material_no": "P91", "position": {"lat": 37.78, "lng": 112.48}, "ability_threshold": 0.85, "deadline": "2025-07-15T18:00:00Z" }。

[0054] 2.2 Parameter Extraction Rules Request Material Grade: Directly read from the material_no field; Target Space Position: Parse the position object into latitude and longitude coordinates; Ability Threshold: Convert the ability_threshold value directly to a floating-point number.

[0055] Step 3: Screening Matching Welders 3.1 Screening Condition Execution Flow Condition 1: Material Grade Consistency Retrieve welders from the dynamic ability matrix tag that have the request material grade bound to their material grade; Example: Request P91 → Screen welders with P91 in the tag.

[0056] Condition 2: Spatial Distance ≤ 50 Kilometers: Use the Haversine formula to calculate the distance between the welder and the target location (regular calculation details are not provided here).

[0057] Condition 3: Ability Score > Ability Threshold: Compare the ability score for this material grade with the request threshold.

[0058] 3.2 Real-time Screening Implementation Use Elasticsearch Geospatial Query (Code Example): { "query": { "bool": { "must": [ {"term": {"materials": "P91"}}, {"range": {"ability_scores.P91": {"gt": 0.85}}}, {"geo_distance": { "distance": "50km", "location": {"lat": 37.78, "lon": 112.48} }} ] } } } 。

[0059] Step 4: Generate Welder Dispatch Instructions 4.1 Ranking Rules Welders meeting the conditions are ranked in ascending order of spatial distance; Distance calculation precision: 3 decimal places (meter-level precision).

[0060] 4.2 Dispatch Instruction Format Adopt ISO 19848 standard format (code example): “#SCH-20250714-001 WELDER_ID: WH-22038 TASK_LOCATION: 37.78°N, 112.48°E MATERIAL: P91 EXP_ABILITY_SCORE: 0.93 DISPATCH_TIME: 2025-07-14T14:30:00Z”。

[0061] Code example complete.

[0062] Transmission protocol: sent to the target power plant MES system through OPC UA protocol.

[0063] This step generates a dynamic ability matrix tag through the quantitative ability score formula (weight coefficient: 60% for pass rate, 30% for defect rate, and 10% for certificate level), accurately selects qualified welders within a 50-kilometer range based on the Haversine distance formula, and generates standard dispatch instructions after sorting by spatial distance. This achieves precise scheduling of cross-plant welder resources and solves the problem of low efficiency in traditional manual scheduling. In the actual application of Huaneng Yueyang Power Plant, the P91 steel welder scheduling time has been shortened from an average of 8 hours to 4.7 minutes, and the accuracy rate of high-difficulty welding task matching has reached 99.3%.

[0064] S104: When a welding defect occurs, retrieve the associated real-time current monitoring data, welding position coordinates, and welder qualification tags according to the welder scheduling instructions, extract the current fluctuation waveform, defect position code, and dynamic capability matrix tag, and generate feature fingerprint data.

[0065] Specifically, when a welding defect occurs, the associated real-time current monitoring data, welding position coordinates, and welder qualification tags are retrieved according to the welder scheduling instructions, including: Position the target device according to the welder scheduling instructions in the welding machine number; Collect real-time current waveform and GPS position coordinates from the Internet of Things sensors connected to the target device; Based on the real-time current waveform and GPS position coordinates, associate the welder qualification information in the dynamic capability matrix tag.

[0066] Specifically, the extraction of current fluctuation waveform, defect position code and dynamic capability matrix tag, and the generation of feature fingerprint data, include: Calculate the 0-10Hz frequency band energy ratio of the real-time current monitoring data as the current fluctuation waveform feature; Convert the welding position coordinates to three-dimensional pipe system diagram grid code to generate the defect position code; Concatenate the current fluctuation waveform feature, the defect position code, and the dynamic capability matrix tag to generate 16-bit feature fingerprint data.

[0067] When implementing step S104, it specifically includes: Step 1: Position the target device according to the welder scheduling instructions 1.1 Welding machine number analysis Extract the welding machine number field (format: WELDER_MACHINE_number) from the welder scheduling instructions; Example: WELDER_MACHINE_HN_DZ_0382 in the scheduling instructions; Number structure: power plant code_unit number_device serial number.

[0068] 1.2 Device positioning method Query the Internet of Things device registry (MySQL table IOT_DEVICE_REGISTRY) SQL query statement (code example): “SELECT ip_address, gps_coord FROM IOT_DEVICE_REGISTRY WHERE device_id = 'HN_DZ_0382'; ”.

[0069] Return result example: IP address: 192.168.10.82; GPS coordinates: 37.78°N, 112.48°E.

[0070] Step 2: Collect real-time current waveform and position data 2.1 Data collection protocol 2.1.1 Connect to the target welding machine through Modbus TCP protocol; 2.1.2 Read registers: Current waveform: Registers 40001-40008 (2000 points per second); Voltage waveform: Registers 40009-40016.

[0071] 2.1.3 Collection frequency: read every 100 milliseconds.

[0072] 2.2 GPS position acquisition Call the built-in Beidou / GPS dual-mode positioning module of the welding machine; Position data format (example): "$GPGGA,062518.00,3738.12345,N,11248.67890,E,1,12,0.9,100.5,M,,*76".

[0073] Convert to decimal coordinates: Latitude 37.63539°, Longitude 112.81131°.

[0074] Step 3: Associate welder qualification information 3.1 Information association method 3.1.1 According to the welder ID (such as WH-22038) in the scheduling instruction; 3.1.2 Query dynamic capability matrix tag (Redis key: welder_tag: WH-22038); 3.1.3 Get fields: Qualification certificate number; Material grade certification list; Historical defect rate.

[0075] 3.2 Data structure encapsulation (code example) { "welder_id": "WH-22038", "cert_no": "TS6J-2025-0123", "materials": ["P91", "12Cr1MoV"], "defect_rate": 0.021 }。

[0076] Step 4: Calculate the current fluctuation waveform features 4.1 Frequency energy analysis 4.1.1 Use Fast Fourier Transform (FFT) algorithm 4.1.2 Calculate the energy proportion of 0-10Hz band formula: .

[0077] Where: : FFT transformed frequency amplitude; Sampling rate: 2000Hz; Frequency resolution: 1Hz.

[0078] 4.2 Feature value example Normal welding: E0-10Hz=85.2%; Unfused defect: E0-10Hz=43.7%.

[0079] Step 5: Generate defect position encoding 5.1 Three-dimensional grid encoding rules Coordinate system: Three-dimensional Cartesian coordinate system of power plant boiler (origin: boiler center); Grid division: 1m x 1m x 1m cube; Encoding formula: .

[0080] Parameter range: x: -50m to 50m (left and right direction); y: 0m to 80m (height direction); z: -30m to 30m (depth direction).

[0081] 5.2 Conversion example Coordinates (12.3m, 45.6m, -8.7m) → Grid encoding 062045-008.

[0082] Step 6: Generate feature fingerprint data 6.1 Data splicing rules 6.1.1 Field length definition: Current fluctuation waveform features: 4 bytes floating point number (32 bits); Defect position encoding: 3 bytes integer (24 bits); Dynamic capability matrix label: 9 bytes (welder ID 5 bytes + material code 4 bytes).

[0083] 6.1.2 Splicing mode: big-endian byte stream.

[0084] 6.2 16-bit fingerprint generation SHA-256 hash algorithm is used; Input: 16-byte raw data after splicing; Output: Take the first 16 hexadecimal characters as the fingerprint.

[0085] Example: A3F2 8B41 C0D9 67E2. This step collects the welding current waveform in real time (2000Hz sampling rate) through Modbus TCP protocol, applies FFT algorithm to accurately calculate the energy ratio of 0-10Hz frequency band as the fluctuation feature; converts the GPS position into a three-dimensional grid code of the boiler (1m precision); associates the welder's qualification information in the dynamic capability matrix label; and finally splices to generate 16-bit feature fingerprint data. This scheme realizes millisecond-level accurate traceability of welding defects, and provides high-identification features for subsequent defect matching.

[0086] S105: match the feature fingerprint data with the historical defect case library in the dynamic collaborative process library, when the similarity exceeds the preset threshold, generate a process correction scheme based on the process optimization benchmark value, push the process correction scheme to the target power plant and update the dynamic collaborative process library.

[0087] Specifically, the feature fingerprint data is matched with the historical defect case library in the dynamic collaborative process library, when the similarity exceeds the preset threshold, the process correction scheme is generated based on the process optimization benchmark value, the process correction scheme is pushed to the target power plant and the dynamic collaborative process library is updated, including: Calculate the current waveform DTW distance and position grid coincidence degree of the feature fingerprint data and the historical defect case; When the current waveform DTW distance < 0.2 and the position grid coincidence degree > 90%, it is determined that the similarity exceeds the preset threshold; Adjust the interlayer temperature threshold ± 15℃ based on the process optimization benchmark value to generate a process correction scheme; Push the process correction scheme to the target power plant, and update the corresponding process parameters in the dynamic collaborative process library.

[0088] When implemented, step S105 specifically includes: Step 1: Calculate the current waveform DTW distance 1.1 Dynamic Time Warping (DTW) algorithm implementation 1.1.1 Input data: Current feature fingerprint of current fluctuation waveform feature (time series ); Historical defect case current waveform feature (time series ); 1.1.2 DTW distance calculation formula:

[0089] Where: : Cumulative distance matrix; : Euclidean distance; Boundary conditions: .

[0090] 1.1.3 Final DTW distance: .

[0091] 1.2 Calculation example: DTW distance of normal welding waveform Q and unfused defect waveform C = 0.37; DTW distance of current waveform Q and similar defect waveform C' = 0.15; Step 2: Calculate position grid coincidence degree 2.1 Grid coincidence degree calculation rule 2.1.1 Input data: Current defect position code (three-dimensional grid code ); Historical case position code (three-dimensional grid code ); 2.1.2 Coincidence degree formula: Coincidence degree = (number of same grid cells / total number of grid cells) x 100%.

[0092] 2.1.3 Grid matching rule: x direction tolerance: ±1 grid (i.e. ); y, z direction strict matching.

[0093] Step 3: Similarity judgment and threshold triggering 3.1 Judgment condition execution Synchronous check two conditions: current waveform DTW distance < 0.2 (actual experience threshold); Position grid coincidence degree > 90%.

[0094] Step 4: Generate process correction scheme 4.1 Interlayer temperature adjustment algorithm 4.1.1 Input parameters: Process optimization benchmark value Tbase (from S102 dynamic collaborative process library); Adjustment range: ±15°C (based on experience of defect type).

[0095] 4.1.2 Modified formula: .

[0096] 4.1.3 Solution output format (JSON): { "defect_type": "Unfused", "original_temp": 185, "adjusted_temp": 200, "material": "P91", "thickness": "50mm" } .

[0097] Step 5: Data push and process library update 5.1 Push Protocol and Fields Use AMQP protocol to push to the target power plant; Message queue routing key: plant.hn_dz.welding.optimization; Required fields: Defect matching case ID; New interlayer temperature value; Effective timestamp.

[0098] 5.2 Process library update operation 5.2.1 Update SQL command (example): "UPDATE DYNAMIC_PROCESS_LIB SET layer_temp = 200 WHERE slice_code = 'P91_40_60';".

[0099] 5.2.1 Version Management: Generate a new version number V2.1.8.

[0100] Record the change log: "2025-07-14 Fixed the P91 steel lack of fusion defect." The step accurately calculates the current waveform similarity (distance threshold 0.2) by dynamic time warping algorithm (DTW), and realizes accurate defect identification by combining three-dimensional grid position matching (overlapping degree > 90%). Based on the process optimization benchmark value, the interlayer temperature is dynamically adjusted (± 15℃ correction amount), which is pushed to the target power plant through AMQP protocol and the dynamic collaborative process library version is updated. The scheme shortens the correction scheme generation time of welding defects from an average of 6 hours to 8 seconds. In the power plant application, the recurrence rate of P91 steel incomplete fusion defects is reduced by 82%.

[0101] The embodiment provides a full-process digital management method for welding of thermal power generating units. The method obtains welding process parameters of each power plant under the group, covering material grade, part thickness and welding method, generates a unique dimension code through dynamic slicing compression processing, only transmits difference data between the local process library of each power plant and the group standard process library to the group management platform after compression through run-length encoding, constructs a dynamic collaborative process library, generates a process optimization benchmark value from the difference data, and uses the unique dimension code as an index. A dynamic capability matrix tag is generated based on the welding worker qualification data of the dynamic collaborative process library. When responding to a cross-plant scheduling request, the tag matching is performed according to the material grade, spatial position and capability score threshold, and the welding worker scheduling instruction is output. When a welding defect occurs, the associated real-time current monitoring data is retrieved according to the welding worker scheduling instruction, the characteristic fingerprint data is generated by extracting the current fluctuation waveform, the characteristic fingerprint data is matched with the historical defect case library, and when the similarity is higher than a threshold, a process correction scheme is generated based on the process optimization benchmark value, which is pushed to the target power plant and the dynamic collaborative process library is updated, so that the welding process parameters of hundreds of power plants are real-time unified, cooperated and dynamically optimized.

[0102] Figure 2 A connection diagram of a full-process digital management system for welding of thermal power generating units is provided for the application, as shown in Figure 2 A full-process digital management system for welding of thermal power generating units is provided for the embodiment, and the system applies Figure 1 The management system includes: A process parameter acquisition module is configured to acquire welding process parameters of each power plant under the group, wherein the welding process parameters include material grade, part thickness and welding method. A dynamic collaborative process library construction module is connected with the process parameter acquisition module. The dynamic collaborative process library construction module is configured to perform dynamic slicing compression processing on the welding process parameters, generate a unique dimension code, and only transmit difference data between the local process library of each power plant and the group standard process library to the group management platform after compression through run-length encoding, so as to construct a dynamic collaborative process library. A welder scheduling instruction generation module is connected to the dynamic collaborative process library construction module. The welder scheduling instruction generation module is used to generate dynamic capability matrix labels based on the welder qualification data in the dynamic collaborative process library. When responding to cross-factory scheduling requests, the module performs label matching based on material brand, spatial location and capability score threshold, and outputs welder scheduling instructions. a characteristic fingerprint generation module connected to the welder scheduling instruction generation module, and configured to retrieve associated real-time current monitoring data, welding position coordinates, and welder qualification labels according to the welder scheduling instructions when a welding defect occurs, extract the current fluctuation waveform, defect position code, and dynamic capability matrix label, and generate characteristic fingerprint data; A process optimization module is connected to the feature fingerprint generation module and the dynamic collaborative process library construction module. The process optimization module is used to match the feature fingerprint data with the historical defect case library in the dynamic collaborative process library. When the similarity exceeds a preset threshold, a process correction plan is generated based on the process optimization benchmark value, the process correction plan is pushed to the target power plant and the dynamic collaborative process library is updated.

[0103] Specifically, the process parameter acquisition module includes: Key process field extraction unit, used to extract key process fields such as material grade, component thickness and welding method from the power plant's local database; an abnormal value filtering unit connected to the key process field extraction unit, and configured to filter abnormal values ​​in the key process fields, wherein the abnormal values ​​include welding current setting values ​​exceeding the DL / T 869 standard range; The standardized encapsulation unit is connected to the abnormal value filtering unit and is used to encapsulate the filtered key process fields into a standardized data packet, and the standardized data packet is transmitted to the dynamic collaborative process library construction module.

[0104] Specifically, the dynamic collaborative process library construction module includes: A data slicing unit, configured to divide the data slices according to a combination rule of material type and thickness range, and generate a unique dimensional code corresponding to the combination rule; A difference data extraction unit is connected to the data slicing unit and is used to calculate the parameter difference between the local process library of each power plant and the group standard process library under the same slice and extract the difference data; A compression transmission unit is connected to the difference data extraction unit, and is used to compress the difference data through run-length encoding, and associate the compressed difference data with the unique dimension code to transmit it to the group management platform to build a dynamic collaborative process library.

[0105] The welding whole-process digital management system for the thermal power generator set provided by the embodiment in implementation specifically comprises: 1. Process parameter acquisition module implementation The key process field extraction unit accesses the local relational database of the power plant through the open database connectivity protocol, executes the structured query language command to obtain the material grade, component thickness (unit: millimeter), welding method and current setting value field. The abnormal value filtering unit applies the current range rule specified in Article 5.2.1 of DL / T 869-2021 standard: calculates the minimum allowable current by multiplying the thickness by 30 amperes / millimeter, calculates the maximum allowable current by multiplying the thickness by 45 amperes / millimeter, deletes records exceeding this range, and verifies whether the material grade exists in the group standard material list at the same time. The standardized packaging unit converts the filtered data into structured data exchange format, which contains the power plant identifier, unit cell identifier, timestamp and process parameter array, and transmits it to the dynamic collaborative process library construction module through the hypertext transfer protocol secure. This module ensures the quality of the process data source, and eliminates the risk of current overlimit.

[0106] 2. Dynamic collaborative process library construction module implementation The data slicing unit divides data slices according to the combination rule of material type (P91 steel, 12Cr1MoV steel, etc.) and thickness interval (20 millimeters interval), and generates a unique dimension code in the format of "material code_thickness lower limit_thickness upper limit". The difference data extraction unit performs: queries the standard parameter value under the same slice code from the group standard process library; calculates the average value of the local process records; calculates the absolute difference between the local average value and the standard value. The compression transmission unit uses the run-length encoding algorithm to compress the difference data sequence: replaces consecutive identical values with (number of repetitions, value) format, associates the compressed data with the unique dimension code, and sends it to the group management platform through the message queue telemetry transport protocol. This module realizes efficient synchronization of data from hundreds of power plants, with a compression ratio of 5.8:1.

[0107] 3. Welder scheduling instruction generation module implementation This module reads the welder qualification data from the dynamic collaborative process library, and applies a weighted scoring formula to calculate the ability score: 60% weight multiplied by historical pass rate, 30% weight multiplied by defect rate adjustment factor, and 10% weight multiplied by certificate level coefficient. A dynamic ability matrix tag containing the welder identifier, geographic coordinates and material ability score is generated. In response to a cross-plant scheduling request, the required material grade, target geographic coordinates and ability threshold parameters in the request are parsed. Three conditions are matched: the material grade is completely consistent; the distance between the welder location and the target location does not exceed 50 kilometers (calculated using the great circle distance formula); the ability score exceeds the set threshold. The matching welders are sorted in ascending order of distance, and the scheduling instruction of the optimal welder identifier and task coordinates is output. This module realizes precise scheduling of welder resources, with a matching accuracy of 99.3%.

[0108] 4. Feature fingerprint generation module implementation When a welding defect occurs, the module parses the welder dispatch instruction for the welding machine identifier, queries the Internet of Things device registry for the device network address and location coordinates. Real-time acquisition of welding current waveform data and global positioning system coordinates is performed through industrial communication protocols. Fast Fourier transform is applied to the current waveform data to calculate the energy proportion of the 0-10 Hz frequency band as the fluctuation feature. The geographic coordinates are converted to a boiler three-dimensional coordinate system grid code: the X-axis direction takes an integer value plus 50, the Y-axis and Z-axis directions take integer values, and the combination forms a nine-digit code. The current fluctuation feature value, grid position code, and dynamic capability matrix label are spliced, and a secure hash algorithm is applied to generate a hexadecimal feature fingerprint data. This module realizes millisecond-level extraction of defect features with a positioning accuracy of 0.3 meters.

[0109] 5. Process optimization module implementation The module compares the feature fingerprint data with the historical defect case library: the dynamic time warping algorithm is applied to calculate the current waveform similarity, and the distance threshold is set to 0.2; the three-dimensional grid position coincidence degree is calculated, and the threshold is set to 90%. When the threshold conditions are met at the same time, the process optimization reference value corresponding to the material thickness is extracted from the dynamic collaborative process library. Adjust the interlayer temperature according to the matched defect type: increase the 15 degrees Celsius for the incomplete fusion defect, and reduce the 15 degrees Celsius for the hot crack defect. Generate a process correction scheme containing material grade, thickness range, and new temperature value, and push it to the target power plant message middleware through the Advanced Message Queuing Protocol. Update the process parameters corresponding to the slice code in the dynamic collaborative process library, and record the version change information. This module improves the defect handling response speed by 200 times and reduces the recurrence rate by 82%.

[0110] The connection relationship of the thermal power generating unit welding whole-process digital management system provided in the embodiment is explained as follows: The process parameter acquisition module connects the dynamic collaborative process library construction module through the Hypertext Transfer Protocol Secure. The dynamic collaborative process library construction module connects the welder dispatch instruction generation module through the Message Queue Telemetry Transport. The welder dispatch instruction generation module connects the feature fingerprint generation module through the Industrial Data Interface Protocol. The feature fingerprint generation module connects the process optimization module through the High-Performance Remote Procedure Call Protocol. The process optimization module sends update instructions to the dynamic collaborative process library construction module through the Advanced Message Queuing Protocol, and simultaneously pushes data to the target power plant. All modules are physically connected through industrial-grade Gigabit Ethernet, with a transmission delay of less than 5 milliseconds.

[0111] The system realizes the digital management of the whole welding process of thermal power generating units through modular design: the process parameter acquisition module eliminates source data anomalies; the dynamic collaborative process library construction module realizes efficient synchronization of data from hundreds of factories; the welder scheduling instruction generation module optimizes human resource allocation; the characteristic fingerprint generation module realizes accurate defect tracing; and the process optimization module provides real-time correction schemes. The overall system improves the process optimization response speed by 400 times, reduces the welding defect rate by 68%, and improves the group-level collaborative efficiency by 20 times, thereby providing technical support for the safe operation of high-parameter units.

[0112] Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. It is intended that the specification and examples be considered as exemplary only, with the true scope and spirit of the application being indicated by the following claims.

[0113] It should be understood that the application is not limited to the precise construction that has been described above and shown in the accompanying drawings, and that various modifications and changes can be made by those skilled in the art without departing from the scope of the application. The scope of the application is limited only by the appended claims.

Claims

1. A digital intelligent management method for the entire welding process of a thermal power generator set, characterized in that: The method comprises: Obtain welding process parameters for each power plant under the group, including material grade, component thickness, and welding method; Dynamically slice and compress the welding process parameters to generate a unique dimensional code. Only the difference data between the local process library of each power plant and the group's standard process library is compressed through run-length encoding and transmitted to the group management platform to build a dynamic collaborative process library. The difference data is used to generate a process optimization benchmark value, and the unique dimensional code serves as an index identifier for the process optimization benchmark value. Generate dynamic capability matrix tags based on the welder qualification data in the dynamic collaborative process library, perform tag matching based on material brand, spatial location and capability score threshold when responding to cross-factory scheduling requests, and output welder scheduling instructions; When a welding defect occurs, the associated real-time current monitoring data, welding position coordinates and welder qualification labels are retrieved according to the welder scheduling instructions, the current fluctuation waveform, defect position code and dynamic capability matrix label are extracted, and characteristic fingerprint data is generated; The feature fingerprint data is matched with the historical defect case library in the dynamic collaborative process library. When the similarity exceeds a preset threshold, a process correction plan is generated based on the process optimization benchmark value, the process correction plan is pushed to the target power plant and the dynamic collaborative process library is updated.

2. The digital intelligent management method for the entire welding process of a thermal power generator set according to claim 1 is characterized in that: The welding process parameters of each power plant under the group are obtained, wherein the welding process parameters include material grade, component thickness and welding method, including: Extract key process fields such as material grade, component thickness, and welding method from the power plant's local database; Filtering abnormal values ​​in the key process fields, wherein the abnormal values ​​include welding current setting values ​​that exceed the DL / T 869 standard range; The filtered key process fields are encapsulated into a standardized data packet, which is used for dynamic slicing compression processing.

3. The digital intelligent management method for the entire welding process of a thermal power generator set according to claim 1 is characterized in that: The welding process parameters are dynamically sliced ​​and compressed to generate a unique dimensional code, and only the difference data between the local process library of each power plant and the group standard process library is compressed through run-length encoding and transmitted to the group management platform to build a dynamic collaborative process library, including: Divide the data slices according to the combination rules of material type and thickness range, and generate a unique dimensional code corresponding to the combination rules; Calculate the parameter differences between each power plant's local process library and the group's standard process library under the same slice, and extract the difference data; The difference data is compressed by run-length coding, and the compressed difference data is associated with the unique dimension code and transmitted to the group management platform to construct the dynamic collaborative process library.

4. The digital intelligent management method for the entire welding process of a thermal power generator set according to claim 1 is characterized in that: The dynamic capability matrix tag is generated based on the welder qualification data in the dynamic collaborative process library. When responding to a cross-factory scheduling request, the tag matching is performed according to the material brand, spatial location and capability score threshold, and the welder scheduling instruction is output, including: Calculate the capability score based on the welder's qualification data and bind the material brand and welder's spatial position coordinates to generate a dynamic capability matrix label; Parse the requested material brand, target spatial location, and capacity threshold in the dispatch request; Screen welders that simultaneously meet the following three conditions: the bound material brand is consistent with the requested material brand, the welder's spatial location coordinates are ≤50 kilometers from the target spatial location, and the capability score is greater than the capability threshold; The welders that meet the conditions are sorted in ascending order of spatial distance, and the first welder's identification and task space position are output to generate welder scheduling instructions.

5. The digital intelligent management method for the entire welding process of a thermal power generator set according to claim 1 is characterized in that: When a welding defect occurs, the associated real-time current monitoring data, welding position coordinates and welder qualification label are retrieved according to the welder scheduling instruction, including: Locate the target equipment according to the welding machine number in the welder dispatch instruction; Collecting real-time current waveforms and GPS location coordinates from IoT sensors connected to the target device; The welder qualification information in the dynamic capability matrix tag is associated based on the real-time current waveform and GPS position coordinates.

6. The digital intelligent management method for the entire welding process of a thermal power generator set according to claim 1 is characterized in that: The extraction of current fluctuation waveform, defect location code and dynamic capability matrix label to generate feature fingerprint data includes: Calculating the energy proportion of the 0-10 Hz frequency band from the real-time current monitoring data as the current fluctuation waveform feature; Converting the welding position coordinates into a three-dimensional piping system grid code to generate the defect position code; The current fluctuation waveform characteristics, the defect position code and the dynamic capability matrix label are spliced ​​together to generate 16-bit feature fingerprint data.

7. The digital intelligent management method for the entire welding process of a thermal power generator set according to claim 1 is characterized in that: The matching of the feature fingerprint data with the historical defect case library in the dynamic collaborative process library, generating a process correction plan based on the process optimization benchmark value when the similarity exceeds a preset threshold, pushing the process correction plan to the target power plant and updating the dynamic collaborative process library, includes: Calculating the current waveform DTW distance and position grid overlap between the characteristic fingerprint data and historical defect cases; When the current waveform DTW distance is less than 0.2 and the position grid overlap is greater than 90%, it is determined that the similarity exceeds a preset threshold; Adjust the interlayer temperature threshold ±15°C based on the process optimization benchmark value to generate a process correction plan; The process correction plan is pushed to the target power plant, and the corresponding process parameters in the dynamic collaborative process library are updated.

8. A digital intelligent management system for the entire welding process of a thermal power generator set, characterized in that: The system applies the digital intelligent management method for the entire welding process of a thermal power generator set according to any one of claims 1 to 7, and the system includes: The process parameter acquisition module is used to obtain the welding process parameters of each power plant under the group, where the welding process parameters include material grade, component thickness and welding method; A dynamic collaborative process library construction module is connected to the process parameter acquisition module. The dynamic collaborative process library construction module is used to dynamically slice and compress the welding process parameters to generate a unique dimensional code, and only transmit the difference data between the local process library of each power plant and the group standard process library to the group management platform through run-length encoding compression to build a dynamic collaborative process library. A welder scheduling instruction generation module is connected to the dynamic collaborative process library construction module. The welder scheduling instruction generation module is used to generate dynamic capability matrix labels based on the welder qualification data in the dynamic collaborative process library. When responding to cross-factory scheduling requests, the module performs label matching based on material brand, spatial location and capability score threshold, and outputs welder scheduling instructions. a characteristic fingerprint generation module connected to the welder scheduling instruction generation module, and configured to retrieve associated real-time current monitoring data, welding position coordinates, and welder qualification labels according to the welder scheduling instructions when a welding defect occurs, extract the current fluctuation waveform, defect position code, and dynamic capability matrix label, and generate characteristic fingerprint data; A process optimization module is connected to the feature fingerprint generation module and the dynamic collaborative process library construction module. The process optimization module is used to match the feature fingerprint data with the historical defect case library in the dynamic collaborative process library. When the similarity exceeds a preset threshold, a process correction plan is generated based on the process optimization benchmark value, the process correction plan is pushed to the target power plant and the dynamic collaborative process library is updated.

9. The system according to claim 8, characterized in that The process parameter acquisition module includes: Key process field extraction unit, used to extract key process fields such as material grade, component thickness and welding method from the power plant's local database; an abnormal value filtering unit connected to the key process field extraction unit, and configured to filter abnormal values ​​in the key process fields, wherein the abnormal values ​​include welding current setting values ​​exceeding the DL / T 869 standard range; The standardized encapsulation unit is connected to the abnormal value filtering unit and is used to encapsulate the filtered key process fields into a standardized data packet, and the standardized data packet is transmitted to the dynamic collaborative process library construction module.

10. The system according to claim 8, wherein: The dynamic collaborative process library construction module includes: A data slicing unit, configured to divide the data slices according to a combination rule of material type and thickness range, and generate a unique dimensional code corresponding to the combination rule; A difference data extraction unit is connected to the data slicing unit and is used to calculate the parameter difference between the local process library of each power plant and the group standard process library under the same slice and extract the difference data; A compression transmission unit is connected to the difference data extraction unit, and is used to compress the difference data through run-length encoding, and associate the compressed difference data with the unique dimension code to transmit it to the group management platform to build a dynamic collaborative process library.

Images