Search

CN-121981446-A - Welding task distribution method and system based on welder dynamic image and knowledge graph

CN121981446ACN 121981446 ACN121981446 ACN 121981446ACN-121981446-A

Abstract

The invention relates to the field of thermal power station welding process management, and discloses a thermal power station construction welding task distribution system and method based on welder dynamic images and knowledge maps. The system comprises a data acquisition module, a welder dynamic image module, a knowledge graph construction module, a task suitability evaluation module and a task allocation suggestion module, wherein the method comprises five steps of data acquisition, welder dynamic image construction, knowledge graph construction and reasoning, task suitability evaluation and task allocation suggestion generation. According to the invention, dynamic images are constructed by quantifying the technical characteristics and quality trends of welders, knowledge maps are constructed, and the welding task and the technical skills of the welders are intelligently matched and risk early-warned by excavating the association rule of the welders and the defects by combining an Apriori algorithm, so that the problems of subjectivity of task allocation, loss of risk predictability and insufficient data value excavation in the prior art are solved, and the welding quality and the production efficiency of the pipeline of the thermal power station are remarkably improved.

Inventors

  • TAO JIANJUN
  • LI BO
  • QIAO JIAN
  • ZHANG ZHE
  • XU JINGCHONG
  • MEI LIHONG
  • PENG JUN
  • YUAN YIPING
  • XU DEHENG
  • DUAN YONGGANG
  • HUANG YONGXIANG
  • ZHU PAN
  • LIU JINGGANG
  • JIANG TAO

Assignees

  • 中电建湖北电力建设有限公司

Dates

Publication Date
20260505
Application Date
20251231

Claims (10)

  1. 1. A thermal power station construction welding task distribution system based on welder dynamic images and knowledge maps is characterized by comprising: The data acquisition module is used for extracting welder task data from a welding task book, a welder operation record and a repair record; The welder dynamic portrayal module is used for constructing a welder comprehensive evaluation model from the skill dimension and the welding quality trend dimension based on the historical welding task data acquired by the data acquisition module and calculating welder skill comprehensive scores and welding quality trend values; the knowledge graph construction module is used for integrating the welder image data output by the welder dynamic image module and the welder-defect association rule, constructing a dynamic knowledge graph and carrying out association analysis reasoning; The task suitability evaluation module is used for calculating suitability scores of welders and tasks based on welding task requirements and welder portrait data output by the welder dynamic portrait module, and generating a recommended welder list; and the task allocation suggestion module is used for generating task allocation suggestions and welder construction notes according to the suitability score of the task suitability evaluation module and the reasoning result of the knowledge graph construction module.
  2. 2. The system of claim 1, wherein the welder task data extracted by the data collection module includes welder number, welding time, number of craters, crater material, welding location, and defect condition.
  3. 3. The system of claim 1 or 2, wherein the skill dimension score of the welder dynamic representation module is calculated by the formula Calculation of wherein As the weight value of the weight, For the average yield of the historical tasks, For the number of the historical craters, For the duty cycle of the stainless steel crater in the welding task, The welding quality trend dimension is the difficulty coefficient of the welding position, and the welding quality trend dimension is calculated by a formula Calculation of wherein As the reject ratio of the upper cycle, The failure rate of the present cycle is defined.
  4. 4. The system of claim 3, wherein the weld location difficulty factor By the formula The calculation is performed such that, The value range is [0,1].
  5. 5. The system of claim 3, wherein the welding quality trend value The judgment standard of (1) is that when When the content is greater than 1.0, the quality trend is good, and when When the ratio is=1.0, the quality trend is stable, and when the ratio is 0< <1.0, The quality trend is deteriorated, when At=0, the quality trend is severely deteriorated.
  6. 6. A system according to claim 1 or 3, wherein the entities defined by the knowledge graph construction module include welder, welding task type, welding target and defect type, and the entity relationships include welder-execution-task, task-dependency-welding target, task-generation-defect type and welder-quality trend-score.
  7. 7. The system of claim 6, wherein the association analysis reasoning adopts Apriori algorithm, and the welder-defect association rule is mined by setting a minimum support threshold value min_sup and a minimum confidence threshold value min_conf, wherein the minimum support threshold value min_sup is 0.01, and the minimum confidence threshold value min_conf is 0.01.
  8. 8. A system according to claim 1 or 3, wherein in the task suitability assessment module, the suitability matching degree M of welder and task is calculated by the formula Calculating, wherein S is the skill comprehensive score, And the quality trend value is obtained.
  9. 9. A thermal power station construction welding task allocation method based on a welder dynamic image and a knowledge graph is characterized by comprising the following steps: s1, data acquisition, namely extracting welder task data from a welding task book, a welder operation record and a repair record; S2, constructing a welder dynamic image, constructing a comprehensive evaluation model from skill dimension and welding quality trend dimension based on the historical welding task data acquired in the step S1, and calculating welder skill comprehensive score and welding quality trend value; S3, knowledge graph construction and reasoning, namely integrating welder portrait data obtained in the step S2 with welder-defect association rules, constructing a dynamic knowledge graph, and mining potential risk rules through an association analysis algorithm; S4, task suitability assessment, namely calculating suitability scores of welders and tasks based on welding task requirements and welder portrait data obtained in the step S2, and generating a recommended welder list; and S5, generating a task allocation suggestion, and outputting the task allocation suggestion and the welder construction notice according to the suitability score of the step S4 and the reasoning result of the step S3.
  10. 10. The method of claim 9, wherein the risk potential hint at step S5 is generated based on excavated welder-defect association rules, the construction optimization advice including wire and groove surface cleaning requirements, gun angle control criteria, and welding speed adjustment advice.

Description

Welding task distribution method and system based on welder dynamic image and knowledge graph Technical Field The invention relates to the field of thermal power station welding process management, in particular to a pipeline welding task distribution system and method in thermal power station construction based on welder dynamic images and knowledge maps. Background In the construction process of the thermal power plant, the pipeline welding of pressure-bearing key equipment such as a boiler, a water-cooled wall, a superheater, a reheater and the like is dependent on manual work, and the final quality of pipeline welding is directly determined by the matching degree of the welder's personal capacity and the specific welding task requirement. The welding task management mode commonly adopted in the industry at present has obvious defects: the subjective decision dependence is serious, the task allocation mainly depends on personal experience and intuitive judgment of management staff such as a team leader, and the situation that skills are not matched with tasks easily occurs due to unstable decision quality caused by lack of unified and objective quantification standards; The risk predictability is lost, namely, the quality risk possibly occurring in future tasks cannot be systematically estimated based on welder history defect data, and preventive measures are difficult to take in advance; The data value mining is insufficient, namely, although a large amount of welding process record and quality detection data are accumulated in the industry, the data exist in isolation, an associated network which can be used for predictive analysis is not formed, and the data value is not fully exerted. In the prior art, the welder archive management system only realizes the digital storage of basic information and lacks the deep analysis capability of data, and a welder rating method based on statistics has single evaluation dimension and can not construct a model which comprehensively reflects the comprehensive capability of a welder, so that the requirements of thermal power station construction on the allocation accuracy and the risk controllability of welding tasks are difficult to meet. Disclosure of Invention The invention aims to overcome the defects of the prior art, and provides a scientific and accurate thermal power station construction welding task distribution system and method based on dynamic images and knowledge maps of welders, which realize intelligent matching and risk early warning of welding tasks and welder skills by quantitatively analyzing welder skills and quality trends and mining data association rules, provide reliable decision support for dispatching the welder tasks and improve welding quality and production efficiency of thermal power station pipelines. In order to achieve the above purpose, the invention provides a thermal power station construction welding task distribution system based on welder dynamic images and knowledge maps, comprising: The data acquisition module is used for extracting welder task data from a welding task book, a welder operation record and a repair record; The welder dynamic portrayal module is used for constructing a welder comprehensive evaluation model from the skill dimension and the welding quality trend dimension based on the historical welding task data acquired by the data acquisition module and calculating welder skill comprehensive scores and welding quality trend values; the knowledge graph construction module is used for integrating the welder image data output by the welder dynamic image module and the welder-defect association rule, constructing a dynamic knowledge graph and carrying out association analysis reasoning; The task suitability evaluation module is used for calculating suitability scores of welders and tasks based on welding task requirements and welder portrait data output by the welder dynamic portrait module, and generating a recommended welder list; and the task allocation suggestion module is used for generating task allocation suggestions and welder construction notes according to the suitability score of the task suitability evaluation module and the reasoning result of the knowledge graph construction module. Further, the welder task data extracted by the data acquisition module comprises welder numbers, welding time, welding number, welding material, welding position and defect condition. Further, the skill dimension score of the welder dynamic image module is calculated by a formulaCalculation of whereinAs the weight value of the weight,For the average yield of the historical tasks,For the number of the historical craters,For the duty cycle of the stainless steel crater in the welding task,The welding quality trend dimension is the difficulty coefficient of the welding position, and the welding quality trend dimension is calculated by a formulaCalculation of whereinAs the reject ratio of the upper cycle,The failure rate of th