Task matching method and system for wind power industry

Abstract

The invention discloses a wind power industry task matching method and system which are applied to the technical field of intelligent employment, and the method comprises the steps: carrying out the splitting and marking of a task exclusive demand, calculating the multi-dimensional similarity based on a deployed task matching model, combining a cosine distance, a Manhattan distance and a Jaccard distance, and dynamically adjusting the weight, and outputting the optimal technical talents corresponding to each subtask, matching the confidence, replacing talent sorting and subtask parallel execution suggestions, and realizing accurate matching of task requirements and the technical talents. Compared with the prior art, specific wind power scene attributes are given to sub-task splitting for the problem of matching of intelligent employment and task talents of a wind power enterprise, a method for accurately matching the sub-tasks and technical talents is established, intelligent employment is modeled as the problem of matching of a demand pool and a talent pool, simple system design is achieved, and the intelligent employment and task talent matching efficiency is improved. System expansion and function updating are facilitated, and flexible application of the intelligent employment method in wind power enterprises is greatly promoted.

Description

Technical Field

[0001] This invention relates to the field of intelligent workforce management, and in particular to a task matching method and system for the wind power industry. Background Technology

[0002] Enterprises can flexibly hire and manage talent according to their own staffing needs, supplementing their full-time workforce. This approach is more flexible in form and simpler in process, effectively supporting business expansion, cost savings, and maximizing the value of talent. It is an important employment model for rapidly developing enterprises today. (Existing technology)

[0003] Existing technologies cannot achieve a precise match between task requirements and technical personnel. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a task matching method and system for the wind power industry, which solves the technical problem that tasks and personnel cannot be accurately matched in the prior art.

[0005] To address the aforementioned technical problems, this invention provides a task matching method for the wind power industry, comprising: S1. Based on task requirements and wind power-specific information, a wind power-specific requirement pool is established. The task requirements are collected according to categories such as operation and maintenance, technical upgrades, emergency repairs, and pre-research. The wind power-specific information corresponding to each task is recorded. The wind power-specific information includes wind turbine model, operating environment, urgency level, equipment operating years, and fault codes. Each task in the wind power-specific requirement pool is configured with a status monitoring module and a task quality detection module. The status monitoring module is used to detect whether the task has been matched with technical personnel, and the task quality detection module is used to inspect and evaluate the task completion status according to the wind power task acceptance standards. S2. Based on the physical space dimension, technical difficulty dimension, and timeliness requirement dimension of the wind power task, the task is broken down into sub-tasks. The dependencies between sub-tasks are verified and resource conflicts are avoided. According to the preset updatable and correctable tag library, each sub-task is labeled with a category tag, a technical requirement tag, and a wind power-specific tag. The wind power-specific tag includes a wind turbine model tag, a working environment tag, a certification requirement tag, and a task association tag. The tag weights are dynamically adjusted based on historical matching data using a reinforcement learning algorithm. S3, acquire talent information of enterprise technical personnel and social technical personnel, the talent information includes identity information, education information, technical ability information, work experience information, wind power qualification certification information, turbine experience information and operation restriction information; configure a status monitoring module for each technical personnel, the status monitoring module is used to update the idle status or task matching status of the personnel in real time, and synchronously update the temporary unavailable status of the personnel in combination with the operation scenario; S4, statistically analyze the talent information mentioned in S3, mark the talent's technical ability, work experience, wind power-specific skills and qualification certification, establish a technical talent profile including general skill vectors and wind power-specific skill vectors, and update the parameters of the technical talent profile according to the talent's skill learning and project performance changes at a preset time period; S5, taking the work experience tags and wind power-specific skill tags from step S4 as input data, and the technical ability tags as label data, a neural network with a dual-branch attention mechanism is used. Through transfer learning, the pre-trained model of general employment data is adjusted using small sample data from the wind power industry until the model loss no longer decreases. The trained task matching model is then deployed to the terminal device. The loss function of the neural network is a predefined loss function, which includes a smooth L1 loss term and a wind power task fault tolerance penalty term. S6: Obtain the task-specific requirements from the wind power-specific requirement pool in S1. After being split and labeled in S2, the task-specific requirements are used to calculate multi-dimensional similarity based on the task matching model deployed in S5, combined with cosine distance, Manhattan distance and Jaccard distance, and dynamically adjust the weights. Output the optimal technical talent, matching confidence, alternative talent ranking and parallel execution suggestions for each sub-task.

[0006] Optionally, the dependencies of the verification subtasks in S2 include: verifying the logical order of execution of the subtasks.

[0007] Optionally, the task association tag in S2 includes the dependencies between subtasks and other tasks, as well as information on required tools and equipment.

[0008] Optionally, the status monitoring module described in S3 further includes: using a talent skill decay model, reducing the corresponding matching weight of talents who have not performed tasks related to the corresponding model or skill for more than a preset period, requiring them to be re-certified and then restored.

[0009] Optionally, the neural network with the dual-branch attention mechanism in S5 includes processing general skill matching and enhancing wind power-specific skill matching, and dynamically assigning the priority of the two branches through attention weights.

[0010] Optionally, the formula for calculating the smooth L1 loss in S5 is as follows: ,in, Let represent the model's predicted value and the true value for the i-th input sample, respectively. The L1 loss represents the smooth L1 loss, and N is the number of training samples; the fault tolerance penalty term for the wind power task is... , Let be the fault tolerance coefficient for the i-th subtask. This is the penalty coefficient, with a value of 2.0.

[0011] Optionally, the weight adjustment rules for multi-dimensional similarity in S6 are as follows: the cosine distance weight for technology-intensive sub-tasks is set to 0.6, and the Manhattan distance weight for experience-intensive sub-tasks is set to 0.7.

[0012] Optionally, it also includes: subtask collaborative scheduling steps: based on the constraint satisfaction algorithm of the wind power operation process, scheduling personnel, equipment and time resources for multiple subtasks.

[0013] Optionally, it also includes: wind power knowledge base linkage steps: integrating wind power knowledge base, which includes wind power equipment fault database, maintenance manuals and industry standards, automatically associating with wind power knowledge base during sub-task matching, and giving priority to matching technical personnel who have consulted the corresponding wind power knowledge base and have relevant experience.

[0014] This invention also provides a task matching system for the wind power industry, comprising: The demand pool construction module is used to execute S1, establish a wind power-specific demand pool based on task requirements and wind power-specific information, collect the task requirements according to categories such as operation and maintenance, technical upgrades, emergency repairs, and pre-research, and record the wind power-specific information corresponding to each task. The wind power-specific information includes wind turbine model, operating environment, urgency level, equipment operating years, and fault codes. Each task in the wind power-specific demand pool is configured with a status monitoring module and a task quality detection module. The status monitoring module is used to detect whether the task has been matched with technical personnel, and the task quality detection module is used to inspect and evaluate the task completion status according to the wind power task acceptance standards. The subtask splitting module is used to execute S2. It splits the wind power task into subtasks based on the physical space dimension, technical difficulty dimension, and timeliness requirement dimension. It verifies the dependencies between subtasks and avoids resource conflicts. According to the preset updatable and correctable tag library, it marks each subtask with category tags, technical requirement tags, and wind power-specific tags. The wind power-specific tags include wind turbine model tags, working environment tags, certification requirement tags, and task association tags. The tag weights are dynamically adjusted based on historical matching data through reinforcement learning algorithms. The talent pool construction module is used to execute S3 to obtain talent information of enterprise technical talents and social technical talents. The talent information includes identity information, education information, technical ability information, work experience information, wind power qualification certification information, turbine experience information, and operation restriction information. A status monitoring module is configured for each technical talent. The status monitoring module is used to update the talent's idle status or task matching status in real time, and to update the talent's temporary unavailable status in conjunction with the operation scenario. The talent profile marking module is used to execute S4, collect the talent information in S3, mark the talent's technical ability, work experience, wind power-specific skills and qualification certification, establish a technical talent profile containing general skill vectors and wind power-specific skill vectors, and update the parameters of the technical talent profile according to the talent's skill learning and project performance changes at a preset time period. The talent demand modeling module is used to execute S5. It takes the work experience tags and wind power-specific skill tags from step S4 as input data, and the technical ability tags as label data. It adopts a neural network with a dual-branch attention mechanism and uses a transfer learning mode to adjust the pre-trained model of general employment data using small sample data from the wind power industry until the model loss no longer decreases. The trained task matching model is then deployed to the terminal device. The loss function of the neural network is a set loss function, which includes a smooth L1 loss term and a wind power task fault tolerance penalty term. The matching execution module is used to execute S6, obtain the task-specific requirements in the wind power-specific demand pool in S1. After the task-specific requirements are split and labeled in S2, the multi-dimensional similarity is calculated based on the task matching model deployed in S5, combined with cosine distance, Manhattan distance and Jaccard distance, and the weights are dynamically adjusted. The module outputs the optimal technical talent, matching confidence, alternative talent ranking and parallel execution suggestions for each sub-task.

[0015] The present invention also provides a task matching device for the wind power industry, comprising: Memory, used to store computer programs; A processor is used to execute the computer program to implement the steps of the wind power industry task matching method described above.

[0016] The present invention also provides a readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the above-described wind power industry task matching method.

[0017] The present invention also provides a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of a wind power industry task matching method.

[0018] It is evident that the beneficial effects of this invention are as follows: ① Combining the basic rules unique to the wind power industry, we endow the wind power scenario with exclusive logic for the sub-task decomposition, and based on this, we have established a demand pool and a talent pool; based on historical matching data, we dynamically adjust the label weights through reinforcement learning algorithms, so that the label system has the ability to dynamically and intelligently upgrade. ② We innovatively designed a neural network model architecture based on the unique characteristics of wind power, adopting a "dual-branch attention mechanism." One branch handles general skill matching (such as mechanical maintenance capabilities), while the other branch enhances wind power-specific skill matching (such as wind turbine pitch system debugging). Priorities are dynamically allocated through attention weights, and a customized loss function is designed: on the basis of the general smooth L1 loss, a "wind power task fault tolerance penalty term" is added; through multi-dimensional matching, it not only outputs the "optimal matching talent" but also the "matching confidence" and "alternative talent ranking." ③ By modeling the intelligent employment problem as a matching problem between the demand pool and the talent pool, a simple system design is achieved, which is conducive to system expansion and function updates.

[0019] In addition, the present invention also provides a task matching system for the wind power industry, which also has the above-mentioned beneficial effects. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0021] Figure 1 A flowchart of a wind power industry task matching method provided in an embodiment of the present invention; Figure 2 A schematic diagram of the structural framework of a precise matching method for intelligent labor sub-tasks specific to the wind power industry provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of task requirement decomposition and marking in a precise matching method for intelligent labor subtasks specific to the wind power industry, provided by an embodiment of the present invention. Figure 4 A schematic diagram illustrating the technical talent profile and labeling in a precise matching method for intelligent labor sub-tasks specific to the wind power industry, provided in an embodiment of the present invention; Figure 5 A work order flow diagram of a smart labor sub-task precision matching system specifically for the wind power industry, provided as an embodiment of the present invention; Figure 6 A schematic diagram illustrating the expanded labor settlement function of a smart labor sub-task precision matching system specifically for the wind power industry, provided as an embodiment of the present invention; Figure 7 This is a schematic diagram of a wind power industry task matching system provided in an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] It should be noted that, unless otherwise specified, the features in the following embodiments and implementation methods can be combined with each other.

[0024] Example 1: Please refer to Figure 1 , Figure 1 A flowchart illustrating a task matching method for the wind power industry provided by an embodiment of the present invention. The method may include: S1. Based on task requirements and wind power-specific information, a wind power-specific requirement pool is established. Task requirements are collected in categories such as operation and maintenance, technical upgrades, emergency repairs, and pre-research. Wind power-specific information corresponding to each task is recorded, including wind turbine model, operating environment, urgency level, equipment operating years, and fault codes. Each task in the wind power-specific requirement pool is equipped with a status monitoring module and a task quality detection module. The status monitoring module is used to detect whether the task has been matched with technical personnel, and the task quality detection module is used to inspect and evaluate the task completion status according to the wind power task acceptance standards.

[0025] Each step in this embodiment can be executed by a designated electronic device, which can be a server, a portable terminal, or other forms. For example, consider the task requirements issued during the implementation of a project, such as monthly, semi-annual, and annual inspection tasks for a wind power company. Please refer to the structural diagram of this invention's intelligent labor sub-task precise matching method specifically for the wind power industry. Figure 2 , Figure 2 This invention provides a structural framework diagram of a precise matching method for intelligent labor sub-tasks specific to the wind power industry. Taking the release of task requirements from a project-specific demand pool as an example, the method includes the following steps: Establishing a wind power-specific demand pool: Collecting task requirements that need to be addressed during project implementation, categorized by operation and maintenance, technical upgrades, emergency repairs, and pre-research. For example, monthly, semi-annual, and annual inspection tasks for wind turbines across the country by wind power companies. Recording wind power-specific information such as the turbine model, operating environment, urgency level, equipment operating years, and fault codes corresponding to the tasks; Each task in the demand pool is configured with a status monitoring module and a task quality detection module. The status monitoring module is used to detect whether the task has been matched with technical personnel, and the task quality detection module is used to inspect and evaluate the task completion status according to the wind power task acceptance standards.

[0026] S2 breaks down the wind power task into subtasks based on the physical space dimension, technical difficulty dimension, and timeliness requirement dimension. It verifies the dependencies between subtasks and avoids resource conflicts. Based on a pre-set updatable and correctable tag library, it labels each subtask with category tags, technical requirement tags, and wind power-specific tags. The wind power-specific tags include wind turbine model tags, working environment tags, certification requirement tags, and task association tags. The tag weights are dynamically adjusted based on historical matching data using a reinforcement learning algorithm.

[0027] This embodiment is used to perform subtask decomposition and dynamic tagging: Subtasks are decomposed based on the physical space dimension (e.g., turbine nacelle, tower base, blades), technical difficulty dimension (basic maintenance, fault diagnosis, technical upgrades), and timeliness requirement dimension (emergency repair, routine inspection, planned maintenance) of wind power tasks. Subtask dependencies are verified and resource conflicts are avoided (e.g., blade inspection must be completed before blade repair, and the same technician cannot simultaneously undertake multiple parallel subtasks on the same turbine). Based on a preset, updatable, and correctable tagging library, each subtask is tagged with category tags, technical requirement tags, and wind power-specific tags. Wind power-specific tags include turbine model tags, working environment tags, certification requirement tags, and task association tags. The tagging system supports dynamic adjustment of tag weights based on historical matching data using reinforcement learning algorithms. See the attached diagram for a schematic of task decomposition and tagging. Figure 3 , Figure 3 This is a schematic diagram illustrating the task requirement decomposition and marking in a precise matching method for intelligent labor subtasks specific to the wind power industry, provided as an embodiment of the present invention.

[0028] It should be further explained that, in order to improve the accuracy of subtask dependency verification, the subtask dependency verification in step S2 above includes: verifying the logical order of subtask execution to ensure that the subtasks that need to be completed in advance are executed before the subsequent related subtasks are started.

[0029] It should be further explained that, in order to improve the accuracy of determining task relationships, the task association tag of the wind power-specific tag in step S2 above includes the dependency relationship between subtasks and other tasks, as well as the required tool and equipment information.

[0030] S3 acquires talent information from enterprise technical personnel and social technical personnel. The talent information includes identity information, education information, technical ability information, work experience information, wind power qualification certification information, turbine experience information, and operation restriction information. A status monitoring module is configured for each technical personnel. The status monitoring module is used to update the idle status or task matching status of the personnel in real time, and to update the temporary unavailable status of the personnel in combination with the operation scenario.

[0031] This embodiment is used to establish a wind power talent pool: collecting information on internal technical personnel and individual technical personnel from the community, including identity information, education information, technical ability information, work experience information, wind power qualification certification information, turbine model experience information, and operation restriction information; configuring a status monitoring module for each technical personnel to update the idle status or task matching status of the personnel in real time, and supporting the synchronous update of the temporary unavailability status of the personnel in combination with the operation scenario.

[0032] It should be further explained that, in order to improve the accuracy of talent status updates, step S3 also includes: using a talent skill decay model, reducing the corresponding matching weight of talents who have not engaged in corresponding model or skill-related tasks for more than a preset period, and requiring re-certification to restore their status.

[0033] S4 statistically analyzes the talent information in S3, marks the talent's technical capabilities, work experience, wind power-specific skills, and qualification certifications, and establishes a technical talent profile that includes general skill vectors and wind power-specific skill vectors. The parameters of the technical talent profile are updated according to the talent's skill learning and project performance changes at a preset time period.

[0034] This embodiment is used to perform talent profiling and labeling: It statistically analyzes the talent information in step S3, labels the talent's technical capabilities, work experience, wind power-specific skills, and qualification certifications, and establishes a technical talent profile containing general skill vectors and wind power-specific skill vectors. The profile parameters are periodically updated based on changes in talent skill learning and project performance. See the diagram for a schematic of the technical talent profile and labeling. Figure 4 , Figure 4 This is a schematic diagram illustrating the technical talent profile and labeling in a precise matching method for intelligent labor sub-tasks specific to the wind power industry, provided as an embodiment of the present invention.

[0035] S5. Using the work experience tags and wind power-specific skill tags from step S4 as input data, and the technical ability tags as label data, a neural network with a dual-branch attention mechanism is used. Through transfer learning, the model is pre-trained on general employment data using small sample data from the wind power industry and adjusted until the model loss no longer decreases. The trained task matching model is then deployed to the terminal device. The loss function of the neural network is a predefined loss function, which includes a smooth L1 loss term and a wind power task fault tolerance penalty term.

[0036] This embodiment is used for talent demand modeling: Work experience tags and wind power-specific skill tags from step S4 are used as input data, and technical ability tags are used as label data. A neural network with a dual-branch attention mechanism is used for training. The loss function of the neural network is a customized loss function, including a smooth L1 (smoothed average absolute error) loss term and a wind power task tolerance penalty term. Through transfer learning, after pre-training the model based on general employment data, it is fine-tuned using small sample data from the wind power industry until the model loss no longer decreases. The trained model is then deployed to terminal devices. In this embodiment, the dual-branch attention mechanism neural network is the initial architecture (based on general data pre-training) to obtain a general pre-trained model. The obtained general pre-trained model is then fine-tuned based on wind power data to obtain a trained task matching model.

[0037] It should be further explained that, in order to improve the accuracy of the neural network setup, the neural network with the dual-branch attention mechanism in step S5 has one branch for general skill matching and the other for wind power-specific skill matching, and the priority of the two types of matching is dynamically allocated through attention weights.

[0038] It should be further explained that, in order to improve the accuracy of the loss function determination, the formula for calculating the smooth L1 loss in step S5 is as follows: ,in Let represent the model's predicted value and the true value for the i-th input sample, respectively. The loss term represents the smooth L1 loss, where N is the number of training samples; the wind power task fault tolerance penalty term is... ,in Let be the fault tolerance coefficient for the i-th subtask, with a value of 0.8 for emergency repair tasks and 0.2 for routine maintenance tasks. The penalty coefficient is 2.0. In this embodiment, the emergency repair task is one that needs to be completed within a set minimum time.

[0039] S6 retrieves the task-specific requirements from the wind power-specific requirement pool in S1. After being split and labeled by S2, the task-specific requirements are used to calculate multi-dimensional similarity based on the task matching model deployed in S5, combined with cosine distance, Manhattan distance and Jaccard distance, and dynamically adjust the weights. The output includes the optimal technical talent, matching confidence, alternative talent ranking and parallel execution suggestions for each subtask.

[0040] This embodiment is used to perform multi-dimensional matching of talent needs: The task publisher publishes the task requirements to the demand pool in step S1. After being split and labeled in step S2, based on the model deployed in step S5, multi-dimensional similarity is calculated by combining cosine distance, Manhattan distance and Jaccard distance and the weights are dynamically adjusted. The optimal technical talent, matching confidence, alternative talent ranking and parallel execution suggestions for each sub-task are output to achieve accurate matching between task requirements and technical talent.

[0041] It should be further explained that, in order to improve the accuracy of multi-dimensional similarity determination, the multi-dimensional similarity weight adjustment rule in step S6 is as follows: the cosine distance weight for technology-intensive sub-tasks is set to 0.6, and the Manhattan distance weight for experience-intensive sub-tasks is set to 0.7.

[0042] The beneficial effects of the wind power industry task matching method provided by this invention are: ① Combining the basic rules unique to the wind power industry, we endow the wind power scenario with exclusive logic for the sub-task decomposition, and based on this, we have established a demand pool and a talent pool; based on historical matching data, we dynamically adjust the label weights through reinforcement learning algorithms, so that the label system has the ability to dynamically and intelligently upgrade. ② We innovatively designed a neural network model architecture based on the unique characteristics of wind power, adopting a "dual-branch attention mechanism." One branch handles general skill matching (such as mechanical maintenance capabilities), while the other branch enhances wind power-specific skill matching (such as wind turbine pitch system debugging). Priorities are dynamically allocated through attention weights, and a customized loss function is designed: on the basis of the general smooth L1 loss, a "wind power task fault tolerance penalty term" is added; through multi-dimensional matching, it not only outputs the "optimal matching talent" but also the "matching confidence" and "alternative talent ranking." ③ By modeling the intelligent employment problem as a matching problem between the demand pool and the talent pool, a simple system design is achieved, which is conducive to system expansion and function updates.

[0043] Example 2: Taking a wind power base encountering a sudden malfunction and urgently needing technical personnel to resolve it as an example, this invention provides a precise matching method for intelligent labor sub-tasks specific to the wind power industry. Taking the release of task requirements from a project-specific demand pool as an example, the method includes the following steps: Step 1: Establish a wind power-specific demand pool: Collect task requirements related to sudden failures and aggregate these requirements into the demand pool; record wind power-specific information such as the wind turbine model, operating environment, urgency level, equipment operating years, and fault codes corresponding to the tasks; configure a status monitoring module and a task quality inspection module for each task in the demand pool. The status monitoring module is used to detect whether the task has been matched with technical personnel, and the task quality inspection module is used to inspect and evaluate the task completion status according to the wind power task acceptance standards.

[0044] Step 2, Subtask Decomposition and Dynamic Tagging: Subtasks are decomposed based on the physical space dimension (e.g., turbine nacelle, tower base, blades), technical difficulty dimension (basic operation and maintenance, fault diagnosis, technical upgrades), and timeliness requirement dimension (emergency repair, routine inspection, planned maintenance) of wind power tasks. Subtask dependencies are verified and resource conflicts are avoided (e.g., blade inspection must be completed before blade repair can be carried out, and the same technician cannot undertake multiple parallel subtasks of the same turbine at the same time). According to the preset updatable and correctable tag library, each subtask is labeled with category tags, technical requirement tags, and wind power-specific tags. The wind power-specific tags include turbine model tags, working environment tags, certification requirement tags, and task association tags. The tag system supports dynamic adjustment of tag weights based on historical matching data through reinforcement learning algorithms.

[0045] In this embodiment, the subtask dependency verification in step 2 includes: verifying the logical order of subtask execution to ensure that subsequent associated subtasks are started only after the subtasks that need to be completed in advance have been completed; the task association tag of the wind power exclusive tag in step 2 includes the dependency relationship between the subtask and other tasks, and the required tool and equipment information.

[0046] Step 3, establish a wind power talent pool: collect information on internal technical personnel and individual technical personnel from the community, including identity information, education information, technical ability information, work experience information, wind power qualification certification information, turbine model experience information, and operation restriction information; configure a status monitoring module for each technical personnel to update the idle status or task matching status of the personnel in real time, and support the synchronous update of the temporary unavailability status of the personnel in combination with the operation scenario.

[0047] In this embodiment, step 3 of the talent status update also includes: using a talent skill decay model, reducing the corresponding matching weight of talents who have not engaged in corresponding model or skill-related tasks for more than a preset period, and requiring re-certification to restore their status.

[0048] Step 4, Talent Profile Tagging: Compile the talent information from Step 3, tag the talent's technical capabilities, work experience, wind power-specific skills, and qualification certifications, and establish a technical talent profile that includes general skill vectors and wind power-specific skill vectors. Regularly update the profile parameters based on changes in talent skill learning and project performance. Schematic diagram of technical talent profile and tagging.

[0049] Step 5, Talent Demand Modeling: Using the work experience tags and wind power-specific skill tags from Step 4 as input data, and the technical ability tags as label data, a neural network with a dual-branch attention mechanism is used for training. The loss function of the neural network is a customized loss function, which includes a smooth L1 loss term and a wind power task fault tolerance penalty term. After pre-training the model based on general employment data through transfer learning, it is fine-tuned using small sample data from the wind power industry until the model loss no longer decreases. The trained model is then deployed to the terminal device.

[0050] In this embodiment, the neural network with a dual-branch attention mechanism in step 5 has one branch handling general skill matching and the other reinforcing wind power-specific skill matching. The priority of the two types of matching is dynamically allocated through attention weights. The smoothL1 loss calculation formula in step 5 is as follows: ,in Let represent the model's predicted value and the true value for the i-th input sample, respectively. The L1 loss represents the smooth L1 loss, and N is the number of training samples; the fault tolerance penalty term for the wind power task is... ,in Let be the fault tolerance coefficient for the i-th subtask, with a value of 0.8 for emergency repair tasks and 0.2 for routine maintenance tasks. This is the penalty coefficient, with a value of 2.0.

[0051] Step 6, Multi-dimensional matching of talent needs: The task publisher publishes the task requirements to the requirement pool in Step 1. After being split and labeled in Step 2, based on the model deployed in Step 5, multi-dimensional similarity is calculated by combining cosine distance, Manhattan distance and Jaccard distance and dynamically adjusting the weights. The optimal technical talent, matching confidence, alternative talent ranking and parallel execution suggestions for each sub-task are output, so as to achieve accurate matching between task requirements and technical talent.

[0052] In this embodiment, the multi-dimensional similarity weight adjustment rule in step 6 is as follows: the cosine distance weight for technology-intensive sub-tasks is set to 0.6, and the Manhattan distance weight for experience-intensive sub-tasks is set to 0.7.

[0053] Example 3: To facilitate understanding of this invention, taking the monitoring of work order flow records and the expansion of labor settlement functions as an example, this invention provides a method and system for precise matching of intelligent labor sub-tasks specific to the wind power industry. (See attached document.) Figure 5 , Figure 5 The work order flow diagram of a smart labor sub-task precision matching system specifically for the wind power industry provided in this embodiment of the invention includes the following steps: S501, Establish a dedicated wind power demand pool: Collect task requirements that need to be addressed during project implementation, categorized by operation and maintenance, technical upgrades, emergency repairs, and pre-research. For example, wind power companies' monthly, semi-annual, and annual inspection tasks for wind turbines across the country, recording wind power-specific information such as turbine model, operating environment, urgency level, equipment operating years, and fault codes. Each task in the demand pool is equipped with a status monitoring module and a task quality inspection module. The status monitoring module is used to detect whether the task has been matched with technical personnel, and the task quality inspection module is used to inspect and evaluate the task completion status according to wind power task acceptance standards.

[0054] S502, Subtask Decomposition and Dynamic Tagging: Subtasks are decomposed based on the physical space dimension (e.g., turbine nacelle, tower base, blades), technical difficulty dimension (basic maintenance, fault diagnosis, technical upgrades), and timeliness requirement dimension (emergency repair, routine inspection, planned maintenance) of wind power tasks. Subtask dependencies are verified and resource conflicts are avoided (e.g., blade inspection must be completed before blade repair; the same technician cannot simultaneously undertake multiple parallel subtasks on the same turbine). Based on a pre-set, updatable, and correctable tagging library, each subtask is tagged with category tags, technical requirement tags, and wind power-specific tags. The wind power-specific tags include turbine model tags, working environment tags, certification requirement tags, and task association tags. The tagging system supports dynamic adjustment of tag weights based on historical matching data using reinforcement learning algorithms. See the task decomposition and tagging diagram for details. Figure 3 .

[0055] In this embodiment, the subtask dependency verification in step S502 includes: verifying the logical order of subtask execution to ensure that the subtasks that need to be completed in advance are completed before the subsequent associated subtasks are started; the task association tag of the wind power exclusive tag in step S502 includes the dependency relationship between the subtask and other tasks, and the required tool and equipment information.

[0056] S503, Establish a wind power talent pool: Collect information on internal technical personnel and individual technical personnel from the community. The information includes identity information, education information, technical ability information, work experience information, wind power qualification certification information, turbine experience information, and operation restriction information; Configure a status monitoring module for each technical personnel to update the idle status or task matching status of the personnel in real time, and support the synchronous update of the temporary unavailable status of the personnel in combination with the operation scenario.

[0057] In this embodiment, the talent status update in step S503 also includes: using a talent skill decay model, reducing the corresponding matching weight of talents who have not engaged in corresponding model or skill-related tasks for more than a preset period, and requiring re-certification to restore their status.

[0058] S504, Talent Profile Tagging: This step involves statistically analyzing the talent information from step S503, tagging the talent's technical capabilities, work experience, wind power-specific skills, and certifications to create a technical talent profile that includes both general skill vectors and wind power-specific skill vectors. The profile parameters are updated periodically based on changes in talent skill learning and project performance. See the diagram for a technical talent profile and tagging illustration. Figure 4 .

[0059] S505, Talent Demand Modeling: Using the work experience tags and wind power-specific skill tags from step S504 as input data, and the technical ability tags as label data, a neural network with a dual-branch attention mechanism is used for training; the loss function of the neural network is a customized loss function, including a smooth L1 loss term and a wind power task fault tolerance penalty term; through transfer learning, after pre-training the model based on general employment data, it is fine-tuned using small sample data from the wind power industry until the model loss no longer decreases, and the trained model is deployed to the terminal device.

[0060] In this embodiment, the neural network with the dual-branch attention mechanism in step S505 has one branch for general skill matching and the other for wind power-specific skill matching, and dynamically allocates the priority of the two types of matching through attention weights.

[0061] In this embodiment, the formula for calculating the smooth L1 loss in step S505 is as follows: ,in Let represent the model's predicted value and the true value for the i-th input sample, respectively. The L1 loss represents the smooth L1 loss, and N is the number of training samples; the fault tolerance penalty term for the wind power task is... ,in Let be the fault tolerance coefficient for the i-th subtask, with a value of 0.8 for emergency repair tasks and 0.2 for routine maintenance tasks. This is the penalty coefficient, with a value of 2.0.

[0062] S506, Multi-dimensional matching of talent needs: The task publisher releases the task requirements to the requirement pool in step S501. After being split and labeled in step S502, based on the model deployed in step S505, multi-dimensional similarity is calculated by combining cosine distance, Manhattan distance and Jaccard distance and the weights are dynamically adjusted. The optimal technical talent, matching confidence, alternative talent ranking and parallel execution suggestions for each sub-task are output, so as to achieve accurate matching between task requirements and technical talent.

[0063] In this embodiment, the multi-dimensional similarity weight adjustment rule in step S506 is as follows: the cosine distance weight for technology-intensive sub-tasks is set to 0.6, and the Manhattan distance weight for experience-intensive sub-tasks is set to 0.7.

[0064] This embodiment utilizes a wind power industry-specific intelligent labor sub-task precision matching system based on a task matching method. This system can connect to other modules, such as a labor settlement module, by retrieving work orders. (See also...) Figure 6 , Figure 6 This is a schematic diagram illustrating the expanded labor settlement function of a smart labor sub-task precision matching system specifically for the wind power industry, provided as an embodiment of the present invention.

[0065] This invention employs a "dual-branch attention mechanism," with one branch handling general skill matching (such as mechanical maintenance capabilities) and the other reinforcing wind power-specific skill matching (such as wind turbine pitch system debugging). Priorities are dynamically allocated through attention weights, and a customized loss function is designed: based on the general smooth L1 loss, a "wind power task fault tolerance penalty term" is added; through multi-dimensional matching, not only is the "optimal matching talent" output, but also the "matching confidence" and "alternative talent ranking," thereby improving the accuracy of task and talent matching.

[0066] The wind power industry task matching system provided in the embodiments of the present invention will be described below. The wind power industry task matching system described below and the wind power industry task matching method described above can be referred to in correspondence.

[0067] Please refer to the details. Figure 7 , Figure 7 A schematic diagram of a wind power industry task matching system provided in an embodiment of the present invention may include: The demand pool construction module 100 is used to execute S1, establish a wind power-specific demand pool based on task requirements and wind power-specific information, collect the task requirements according to categories such as operation and maintenance, technical upgrades, emergency repairs, and pre-research, and record the wind power-specific information corresponding to each task. The wind power-specific information includes wind turbine model, operating environment, urgency level, equipment operating years, and fault codes. Each task in the wind power-specific demand pool is configured with a status monitoring module and a task quality detection module. The status monitoring module is used to detect whether the task has been matched with technical personnel, and the task quality detection module is used to inspect and evaluate the task completion status according to the wind power task acceptance standards. The subtask splitting module 200 is used to execute S2. It splits the wind power task into subtasks based on the physical space dimension, technical difficulty dimension, and timeliness requirement dimension. It verifies the dependencies between subtasks and avoids resource conflicts. According to the preset updatable and correctable tag library, it marks each subtask with category tags, technical requirement tags, and wind power-specific tags. The wind power-specific tags include wind turbine model tags, working environment tags, certification requirement tags, and task association tags. The tag weights are dynamically adjusted based on historical matching data through reinforcement learning algorithms. The talent pool construction module 300 is used to execute S3 to obtain talent information of enterprise technical talents and social technical talents. The talent information includes identity information, education information, technical ability information, work experience information, wind power qualification certification information, turbine experience information, and operation restriction information. A status monitoring module is configured for each technical talent. The status monitoring module is used to update the talent's idle status or task matching status in real time, and to update the talent's temporary unavailable status in conjunction with the operation scenario. The talent profile marking module 400 is used to execute S4, count the talent information in S3, mark the talent's technical ability, work experience, wind power-specific skills and qualification certification, establish a technical talent profile containing general skill vectors and wind power-specific skill vectors, and update the parameters of the technical talent profile according to the talent's skill learning and project performance changes at a preset time period. The talent demand modeling module 500 is used to execute S5. It takes the work experience tags and wind power-specific skill tags from step S4 as input data, and the technical ability tags as label data. It adopts a neural network with a dual-branch attention mechanism and uses a transfer learning mode to adjust the pre-trained model of general employment data using small sample data from the wind power industry until the model loss no longer decreases. The trained task matching model is then deployed to the terminal device. The loss function of the neural network is a set loss function, which includes a smooth L1 loss term and a wind power task fault tolerance penalty term. The matching execution module 600 is used to execute S6, obtain the task-specific requirements in the wind power-specific demand pool in S1. After the task-specific requirements are split and labeled in S2, the multi-dimensional similarity is calculated based on the task matching model deployed in S5, and the weights are dynamically adjusted by combining cosine distance, Manhattan distance and Jaccard distance. The module outputs the optimal technical talent, matching confidence, alternative talent ranking and parallel execution suggestions for each sub-task.

[0068] It should be noted that the order of modules and units in the above-mentioned wind power industry task matching system can be changed without affecting the logic.

[0069] This invention provides a wind power industry task matching system that employs a "dual-branch attention mechanism." One branch handles general skill matching (such as mechanical maintenance capabilities), while the other branch enhances wind power-specific skill matching (such as wind turbine pitch system debugging). Priorities are dynamically allocated through attention weights, and a customized loss function is designed: based on the general smooth L1 loss, a "wind power task fault tolerance penalty term" is added. Through multi-dimensional matching, the system not only outputs the "optimal matching talent" but also the "matching confidence" and "alternative talent ranking," thereby improving the accuracy of task and talent matching.

[0070] The following describes the computer-readable storage medium provided in the embodiments of the present invention. The computer-readable storage medium described below can be referred to in correspondence with the wind power industry task matching method described above.

[0071] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the wind power industry task matching method described above.

[0072] The computer-readable storage medium may include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0073] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0074] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0075] Finally, it should be noted that in this document, relationships such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0076] The above provides a detailed description of a wind power industry task matching method and system provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A task matching method for the wind power industry, characterized in that, include: S1. Based on task requirements and wind power-specific information, a wind power-specific requirement pool is established. The task requirements are collected according to categories such as operation and maintenance, technical upgrades, emergency repairs, and pre-research. The wind power-specific information corresponding to each task is recorded. The wind power-specific information includes wind turbine model, operating environment, urgency level, equipment operating years, and fault codes. Each task in the wind power-specific requirement pool is configured with a status monitoring module and a task quality detection module. The status monitoring module is used to detect whether the task has been matched with technical personnel, and the task quality detection module is used to inspect and evaluate the task completion status according to the wind power task acceptance standards. S2. Based on the physical space dimension, technical difficulty dimension, and timeliness requirement dimension of the wind power task, the task is broken down into sub-tasks. The dependencies between sub-tasks are verified and resource conflicts are avoided. According to the preset updatable and correctable tag library, each sub-task is labeled with a category tag, a technical requirement tag, and a wind power-specific tag. The wind power-specific tag includes a wind turbine model tag, a working environment tag, a certification requirement tag, and a task association tag. The tag weights are dynamically adjusted based on historical matching data using a reinforcement learning algorithm. S3, acquire talent information of enterprise technical personnel and social technical personnel, the talent information includes identity information, education information, technical ability information, work experience information, wind power qualification certification information, turbine experience information and operation restriction information; configure a status monitoring module for each technical personnel, the status monitoring module is used to update the idle status or task matching status of the personnel in real time, and synchronously update the temporary unavailable status of the personnel in combination with the operation scenario; S4, statistically analyze the talent information mentioned in S3, mark the talent's technical ability, work experience, wind power-specific skills and qualification certification, establish a technical talent profile including general skill vectors and wind power-specific skill vectors, and update the parameters of the technical talent profile according to the talent's skill learning and project performance changes at a preset time period; S5, taking the work experience tags and wind power-specific skill tags from step S4 as input data, and the technical ability tags as label data, a neural network with a dual-branch attention mechanism is used. Through transfer learning, the pre-trained model of general employment data is adjusted using small sample data from the wind power industry until the model loss no longer decreases. The trained task matching model is then deployed to the terminal device. The loss function of the neural network is a predefined loss function, which includes a smooth L1 loss term and a wind power task fault tolerance penalty term. S6: Obtain the task-specific requirements from the wind power-specific requirement pool in S1. After being split and labeled in S2, the task-specific requirements are used to calculate multi-dimensional similarity based on the task matching model deployed in S5, combined with cosine distance, Manhattan distance and Jaccard distance, and dynamically adjust the weights. Output the optimal technical talent, matching confidence, alternative talent ranking and parallel execution suggestions for each sub-task.

2. The wind power industry task matching method according to claim 1, characterized in that, The dependencies of the verification subtasks in S2 include the logical order in which the verification subtasks are executed.

3. The wind power industry task matching method according to claim 1, characterized in that, In S2, the task association tag contains the dependencies between subtasks and other tasks, as well as information on the required tools and equipment.

4. The wind power industry task matching method according to claim 1, characterized in that, The status monitoring module described in S3 also includes: using a talent skill decay model, reducing the corresponding matching weight of talents who have not engaged in corresponding model or skill-related tasks for more than a preset period, and requiring re-certification to restore their matching weight.

5. The wind power industry task matching method according to claim 1, characterized in that, The neural network with the dual-branch attention mechanism in S5 includes processing general skill matching and enhancing wind power-specific skill matching, and dynamically assigns the priority of the two branches through attention weights.

6. The wind power industry task matching method according to claim 1, characterized in that, The formula for calculating smooth L1 loss in S5 is: ,in, Let represent the model's predicted value and the true value for the i-th input sample, respectively. The L1 loss represents the smooth L1 loss, and N is the number of training samples; the fault tolerance penalty term for the wind power task is... , Let be the fault tolerance coefficient for the i-th subtask. This is the penalty coefficient, with a value of 2.

0.

7. The wind power industry task matching method according to claim 1, characterized in that, The weight adjustment rules for multi-dimensional similarity in S6 are as follows: the cosine distance weight for technology-intensive sub-tasks is set to 0.6, and the Manhattan distance weight for experience-intensive sub-tasks is set to 0.

7.

8. The wind power industry task matching method according to claim 1, characterized in that, Also includes: Subtask collaborative scheduling steps: Based on the constraint satisfaction algorithm of wind power operation process, schedule personnel, equipment and time resources for multiple subtasks.

9. The wind power industry task matching method according to claim 1, characterized in that, Also includes: Wind power knowledge base linkage steps: Integrate the wind power knowledge base, which includes a wind power equipment fault database, maintenance manuals and industry standards. When matching sub-tasks, automatically associate the wind power knowledge base, and give priority to matching technical personnel who have consulted the corresponding wind power knowledge base and have relevant experience.

10. A task matching system for the wind power industry, characterized in that, include: The demand pool construction module is used to execute S1, establish a wind power-specific demand pool based on task requirements and wind power-specific information, collect the task requirements according to categories such as operation and maintenance, technical upgrades, emergency repairs, and pre-research, and record the wind power-specific information corresponding to each task. The wind power-specific information includes wind turbine model, operating environment, urgency level, equipment operating years, and fault codes. Each task in the wind power-specific demand pool is configured with a status monitoring module and a task quality detection module. The status monitoring module is used to detect whether the task has been matched with technical personnel, and the task quality detection module is used to inspect and evaluate the task completion status according to the wind power task acceptance standards. The subtask splitting module is used to execute S2. It splits the wind power task into subtasks based on the physical space dimension, technical difficulty dimension, and timeliness requirement dimension. It verifies the dependencies between subtasks and avoids resource conflicts. According to the preset updatable and correctable tag library, it marks each subtask with category tags, technical requirement tags, and wind power-specific tags. The wind power-specific tags include wind turbine model tags, working environment tags, certification requirement tags, and task association tags. The tag weights are dynamically adjusted based on historical matching data through reinforcement learning algorithms. The talent pool construction module is used to execute S3 to obtain talent information of enterprise technical talents and social technical talents. The talent information includes identity information, education information, technical ability information, work experience information, wind power qualification certification information, turbine experience information, and operation restriction information. A status monitoring module is configured for each technical talent. The status monitoring module is used to update the talent's idle status or task matching status in real time, and to update the talent's temporary unavailable status in conjunction with the operation scenario. The talent profile marking module is used to execute S4, collect the talent information in S3, mark the talent's technical ability, work experience, wind power-specific skills and qualification certification, establish a technical talent profile containing general skill vectors and wind power-specific skill vectors, and update the parameters of the technical talent profile according to the talent's skill learning and project performance changes at a preset time period. The talent demand modeling module is used to execute S5. It takes the work experience tags and wind power-specific skill tags from step S4 as input data, and the technical ability tags as label data. It adopts a neural network with a dual-branch attention mechanism and uses a transfer learning mode to adjust the pre-trained model of general employment data using small sample data from the wind power industry until the model loss no longer decreases. The trained task matching model is then deployed to the terminal device. The loss function of the neural network is a set loss function, which includes a smooth L1 loss term and a wind power task fault tolerance penalty term. The matching execution module is used to execute S6, obtain the task-specific requirements in the wind power-specific demand pool in S1. After the task-specific requirements are split and labeled in S2, the multi-dimensional similarity is calculated based on the task matching model deployed in S5, combined with cosine distance, Manhattan distance and Jaccard distance, and the weights are dynamically adjusted. The module outputs the optimal technical talent, matching confidence, alternative talent ranking and parallel execution suggestions for each sub-task.

Images