Substation multi-algorithm collaborative calling method

Abstract

The invention discloses a multi-algorithm collaborative calling method for a transformer substation, which comprises the following steps of: firstly, monitoring a task list in a transformer substation scene by a transformer substation task monitoring center, and judging whether a new task request exists or not; then, after a task request is received, the substation task monitoring center judges the type of a task; secondly, the artificial intelligence algorithm calling center calculates a condition required for executing a task according to a task request type sent by the substation task monitoring center; thirdly, if the conditions required for executing the algorithm are met, the artificial intelligence algorithm calling center starts the algorithm to execute the task; and finally, returning a result after task execution is finished. According to the method, the new artificial intelligence technology is fully utilized, the request task of the substation scene is accurately detected in real time, and the available resource condition of the center is called according to the artificial intelligence algorithm to start the algorithm to execute the task.

Description

Technical Field

[0001] This invention relates to the power industry, specifically to a method for collaborative invocation of multiple algorithms in substations. Background Technology

[0002] In the context of new power systems, the operation and maintenance of power systems face new challenges. The substation specialists of the Power Grid Equipment Department keep pace with the times, adhering to the principles of "usable, easy to use, and desirable to use," and deeply integrating advanced technologies such as the Internet of Things and artificial intelligence through the construction of smart substations and centralized control stations. In the management of substation equipment, they continuously promote standardization, lean management, and intelligentization, striving to achieve breakthroughs in full-lifecycle asset management and full-process quality control. Although a basic application support system meeting the multi-level needs of data collection and monitoring, business management, analysis, and command has been established, prominent pain points still exist in actual operation and maintenance that urgently need to be addressed.

[0003] In the current substation operation and maintenance field, artificial intelligence technology has been widely applied in various aspects such as equipment condition assessment, remote inspection, fault diagnosis, and health management. From the perspective of equipment condition monitoring, various perception models have been built, including UHF partial discharge detection, infrared thermography, and noise analysis. From the perspective of intelligent inspection, multiple models based on computer vision, such as equipment appearance defect recognition and foreign object detection, have been developed. From the perspective of fault diagnosis, expert systems based on historical cases and rule-based intelligent diagnostic models have been constructed. However, because each model is built independently and operates in isolation, lacking a unified call and management mechanism, the intelligent operation and maintenance effect is difficult to achieve as expected. This fragmented state is particularly prominent in daily operation and maintenance, leading to problems such as high false detection rate of intelligent inspection, low substation operation and maintenance efficiency, and the inability to achieve intelligent closed-loop operation and maintenance business.

[0004] To address the above issues, it is urgent to base our research on the current status of information management in substation production, and to develop an efficient adaptive collaborative invocation mechanism for artificial intelligence models, taking into account the operational and maintenance scenarios of substations.

[0005] Currently, the classic papers on algorithm scheduling research mainly include: [1] "Optimal power scheduling in real-time distribution systems using crow search algorithm for enhanced microgrid performance", [2] "Distributed optimal scheduling for virtual powerplant with high penetration of renewable energy", and [3] "Two-stage stochasticrobust optimal scheduling of virtual power plant considering source load uncertainty". Paper [1] was published in the journal Scientific Reports in 2024 and is a method for optimal power scheduling in real-time distribution systems using crow search algorithm to improve microgrid performance; Paper [2] was published in the journal International Journal of Electrical Power & Energy Systems in 2024 and is a method for distributed optimal scheduling of virtual power plants with high renewable energy penetration; Paper [3] was published in the journal Engineering Reports in 2024 and is a method for two-stage stochastic robust optimal scheduling of virtual power plants considering source load uncertainty. The focus of these methods is on the optimization and control of individual operating tasks.

[0006] Therefore, the existing calling methods still face the following challenges when applied in substation scenarios: (1) it is difficult to handle the request tasks in multiple substation scenarios for collaborative calling; (2) the reliability and stability of the calling algorithm are low. Summary of the Invention

[0007] Purpose of the invention: The purpose of this invention is to address the shortcomings of existing technologies and provide a method for collaborative invocation of multiple algorithms in substations, which can efficiently adapt to and invoke artificial intelligence models, thereby improving the quality and efficiency of substation operation and maintenance.

[0008] Technical solution: The present invention provides a method for multi-algorithm collaborative invocation in substations, comprising the following steps:

[0009] Step 1: The substation task monitoring center monitors the task list in the substation scenario in real time to determine if there are any new task requests. Once a new task request is detected, its sequence number in the task configuration list is recorded. At the same time, record this new request task as ;

[0010] Step 2: Based on the sequence number of the new request task The substation task monitoring center uses a hash method to search for the name of a newly requested task in the task configuration list. and the type of new request task At the same time, record this request task as: The type of new request task here. This includes target detection tasks, equipment operation status monitoring tasks, and inspection tasks;

[0011] Next, the substation task monitoring center will monitor the requested tasks. Send to the AI ​​algorithm call center, requesting the AI ​​algorithm call center to... Process it;

[0012] Step 3: The artificial intelligence algorithm call center calls the request task. Type Determine the conditions required for the algorithm to execute this request task. That is, according to the type of task Search for the corresponding task in the task queue. If it exists, execute it; otherwise, ignore it.

[0013] If request task Once the required conditions are met, the artificial intelligence algorithm call center initiates the request task. Corresponding algorithm Perform a specific task; conversely, request a task. It is in a waiting state and can only be executed when the required conditions are met.

[0014] Step 4: The AI ​​algorithm call center monitors the request tasks in real time. The execution status refers to the progress of requesting task completion / obtaining task completion progress. After the task is completed, the results will be fed back to the substation task monitoring center.

[0015] Furthermore, the task list format in the substation scenario of step 1 is as follows:

[0016] ;

[0017] This indicates that the sequence number in the task list is The name of the requested task.

[0018] Furthermore, the specific process of using the unordered hashing method to find the name and type of the new request task in step 2 is as follows:

[0019] First, based on the sequence number of the requested task. Calculate the hash code for this request task. Then based on the hash code In the task list Search Request Task Name and type Since the hash method locates the target position through a lookup table without a complex calculation process, the task lookup process is very fast.

[0020] Furthermore, the core functions of the artificial intelligence algorithm call center include (1) receiving task requests sent by the substation task monitoring center. (2) According to the task type Calculate the request task Required conditions. An artificial intelligence algorithm call center can call algorithms. This includes: target detection algorithms, target recognition algorithms, UAV path planning algorithms, substation equipment operation status statistics algorithms, equipment operation and maintenance algorithms, and early warning algorithms.

[0021] Furthermore, step 3 requests a task. The required conditions include two parts: (1) Hardware resources, whether there are spare hardware resources to support the running of the requested task. (2) Available idle time of equipment, i.e., whether the idle time of the equipment can meet the running tasks. The need.

[0022] Furthermore, in step 4, the artificial intelligence algorithm call center uses a multi-threaded, multi-tasking approach to execute the request tasks. Corresponding algorithm This ensures that multiple tasks can execute collaboratively, share computing resources, and not interfere with each other. The AI ​​algorithm call center can monitor tasks in real time. Algorithm Once the execution is complete, the results are sent to the substation task monitoring system.

[0023] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0024] (1) This invention improves the operation and maintenance efficiency of substations (by 15%) and the level of equipment health management. It can support intelligent decision-making in equipment inspection, fault diagnosis and maintenance decision-making, provide real-time data analysis and model recommendation, optimize operation and maintenance process, reduce equipment failure rate (failure rate reduced by 18%), improve equipment reliability and operation and maintenance efficiency, thereby reducing overall operation and maintenance costs.

[0025] (2) This invention promotes the intelligent and systematic development of operation and maintenance services. It improves the response speed and decision-making accuracy of operation and maintenance services, thereby enhancing the overall digitalization level of substations and promoting the improvement of intelligent operation and maintenance systems.

[0026] (3) This invention enhances the work efficiency of front-line operation and maintenance personnel and is applicable to various substation operation and maintenance scenarios, including equipment status monitoring, emergency fault handling, and maintenance operations, significantly reducing the workload of operation and maintenance personnel and improving work efficiency. In addition, the method provided by this invention can be promoted and applied to substation scenarios in different regions, integrated into the intelligent operation and maintenance platform, and has good scalability and application value. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the overall process of the present invention;

[0028] Figure 2 This is a structural diagram of the application system in the embodiment. Detailed Implementation

[0029] The technical solution of the present invention will be described in detail below, but the scope of protection of the present invention is not limited to the embodiments described.

[0030] like Figure 1 As shown, a method for collaborative invocation of multiple algorithms in a substation according to the present invention includes the following steps:

[0031] Step 1: The substation task monitoring center monitors the task list in the substation scenario in real time to determine if there are any new task requests. Once a new task request is detected, its sequence number in the task configuration list is recorded. At the same time, record this new request task as ;

[0032] Step 2: Based on the sequence number of the new request task The substation task monitoring center uses a hash method to search for the name of a newly requested task in the task configuration list. and the type of new request task At the same time, record this request task as: ;

[0033] Next, the substation task monitoring center will monitor the requested tasks. Send to the AI ​​algorithm call center, requesting the AI ​​algorithm call center to... Process it;

[0034] Step 3: The artificial intelligence algorithm call center calls the request task. Type The algorithm determines the conditions required to execute the requested task (that is, it searches for the corresponding task in the task queue based on the task type; if it exists, it is executed; otherwise, it is ignored). ;

[0035] If request task Once the required conditions are met, the artificial intelligence algorithm call center initiates the request task. Corresponding algorithm Perform a specific task; conversely, request a task. It is in a waiting state and can only be executed when the required conditions are met.

[0036] Step 4: The AI ​​algorithm call center monitors the request tasks in real time. The execution status refers to the progress of requesting task completion / obtaining task completion progress. After the task is completed, the results will be fed back to the substation task monitoring center.

[0037] This invention first involves the substation task monitoring center monitoring the task list within the substation scenario to determine if there are any new task requests. Then, upon receiving a task request, the substation task monitoring center identifies the task type. Next, the artificial intelligence algorithm invocation center calculates the conditions required to execute the task based on the task request type sent by the substation task monitoring center. Third, if the required conditions are met, the artificial intelligence algorithm invocation center initiates the algorithm to execute the task. Finally, the result is returned after the task execution is complete. This invention fully utilizes new artificial intelligence technologies to accurately detect requesting tasks in the substation scenario in real time and to initiate the algorithm to execute tasks based on the available resources of the artificial intelligence algorithm calculation center.

[0038] The task list format in the substation scenario of step 1 in this embodiment is as follows:

[0039] ;

[0040] This indicates that the sequence number in the task list is The name of the requested task.

[0041] The specific process of step 2 in this embodiment, which uses the unordered hashing method to find the name and type of the new request task, is as follows:

[0042] First, based on the sequence number of the requested task. Calculate the hash code for this request task. Then based on the hash code In the task list Search Request Task Name and type ;

[0043] Since hashing methods locate the target position through a lookup table without requiring complex calculations, the task lookup process is very fast.

[0044] The core functions of the artificial intelligence algorithm call center in this embodiment are as follows: (1) Receive the request task sent by the substation task monitoring center. (2) According to the task type Calculate the request task Required conditions. The artificial intelligence algorithm call center can call the following algorithms: target detection algorithm, target recognition algorithm, drone path planning algorithm, substation equipment operation status statistics algorithm, equipment operation and maintenance algorithm, and early warning algorithm.

[0045] Step 3 of this embodiment requests a task. The required conditions include two parts: (1) Hardware resources, whether there are spare hardware resources to support the running of the requested task. (2) Available idle time of equipment, i.e., whether the idle time of the equipment can meet the running tasks. The need.

[0046] In step 4 of this embodiment, the artificial intelligence algorithm call center uses a multi-threaded, multi-tasking approach to execute the request task. Corresponding algorithm This ensures that multiple tasks can execute collaboratively, share computing resources, and not interfere with each other. This indicates that the requested task is in the task configuration list. The serial number in;

[0047] The AI ​​algorithm call center can monitor tasks in real time. Algorithm Once the execution is complete, the results are sent to the substation task monitoring system.

[0048] Example:

[0049] like Figure 2As shown, the specific process of this embodiment is as follows: First, the substation task monitoring center monitors the task list in the substation scenario to determine if there are any new task requests; second, upon receiving a task request, the substation task monitoring center determines the type of task; third, the substation multi-algorithm invocation center calculates the conditions required to execute the task based on the task request type sent by the substation task monitoring center; fourth, if the required conditions meet the algorithm execution requirements, the artificial intelligence algorithm invocation center starts the algorithm to execute the task; fifth, the result is returned after the task execution is completed. Figure 2 The inspection host is a server equipped with four Nvidia 4090 processors.

[0050] The substation task monitoring center sets up a configuration scheme for the combination of station-side algorithm models and then distributes the algorithm configuration scheme to the station-side algorithm collaboration service.

[0051] After successful configuration distribution, the algorithm collaboration service retrieves the algorithm image from the substation multi-algorithm collaboration call center. Once the algorithm image is successfully downloaded, the original image is stopped, and the new algorithm image is started. The algorithm collaboration service monitors the status and resource usage of the algorithm image service in real time.

Claims

1. A method for collaborative invocation of multiple algorithms in a substation, characterized in that, Includes the following steps: Step 1: The substation task monitoring center monitors the task list in the substation scenario in real time to determine if there are any new task requests. Once a new task request is detected, its sequence number in the task configuration list is recorded. At the same time, record this new request task as ; Step 2: Based on the sequence number of the new request task The substation task monitoring center uses a hash method to search for the name of a newly requested task in the task configuration list. and the type of new request task At the same time, record this request task as: ; Next, the substation task monitoring center will monitor the requested tasks. Send to the AI ​​algorithm call center, requesting the AI ​​algorithm call center to... Process it; Step 3: The artificial intelligence algorithm call center calls the request task. Type Determine the conditions required for the algorithm to execute this request task. ; If request task Once the required conditions are met, the artificial intelligence algorithm call center initiates the request task. Corresponding algorithm Perform a specific task; conversely, request a task. It is in a waiting state and can only be executed when the required conditions are met. Step 4: The AI ​​algorithm call center monitors the request tasks in real time. The execution status refers to the progress of requesting task completion / obtaining task completion progress. After the task is completed, the results will be fed back to the substation task monitoring center.

2. The substation multi-algorithm coordination and invocation method according to claim 1, characterized in that, The task list format in the substation scenario of step 1 is as follows: ； This indicates that the sequence number in the task list is The name of the requested task. The value range is [1, N].

3. The substation multi-algorithm coordination and invocation method according to claim 1, characterized in that, The specific process of using the hash method to find the name and type of the new request task in step 2 is as follows: First, based on the sequence number of the requested task. Calculate the hash code for this request task. Then based on the hash code In the task list Search Request Task Name and type .

4. The substation multi-algorithm coordination and invocation method according to claim 1, characterized in that: The algorithm called by the artificial intelligence algorithm call center This includes target detection algorithms, target recognition algorithms, UAV path planning algorithms, substation equipment operation status statistics algorithms, equipment operation and maintenance algorithms, and early warning algorithms.

5. The substation multi-algorithm coordination and invocation method according to claim 1, characterized in that: Step 3 requests a task. The required conditions consist of two parts. The first is hardware resources, that is, whether there are available hardware resources to support the execution of the requested task. The second is the available idle time of the device, that is, whether the idle time of the device can meet the running tasks. The need.

6. The substation multi-algorithm coordination and invocation method according to claim 1, characterized in that, Step 4: The AI ​​algorithm call center uses a multi-threaded, multi-tasking approach to execute request tasks. Corresponding algorithm The artificial intelligence algorithm call center can monitor tasks in real time. Algorithm Once the execution is complete, the results are sent to the substation task monitoring system.

Images