Multi-agent large-model monitoring system and method for industrial electricity safety

Abstract

The invention belongs to the technical field of industrial power utilization monitoring, and particularly relates to a multi-agent large model monitoring system and method for industrial power utilization safety, and the system comprises a cloud analysis unit, a data collection unit executes the deployment of an industrial-grade sensor, collects the operation data of equipment in real time, and transmits the monitored data information to the cloud analysis unit. The multi-agent collaboration unit is connected with the cloud analysis unit and is used for executing and utilizing a reinforcement learning algorithm to enable each agent to continuously learn from historical data and environment feedback and optimize the decision-making process of the agent, so that highly intelligent monitoring and early warning are realized, timely fault diagnosis and response are provided, safe and stable operation of a power system is ensured, and the service life of the power system is prolonged. The automation degree of the power system is improved, the ability of the power system to deal with emergencies is enhanced, and the risk of power failure to industrial production is effectively reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial electricity monitoring, and in particular to a multi-agent large-scale model monitoring system and method for industrial electricity safety. Background Art

[0002] Industrial electricity consumption accounts for a large proportion, and public transformer users are important service targets of power grid companies. However, current industrial electricity safety monitoring has the following pain points:

[0003] 1. Data silo problem: Industrial equipment data is scattered across different subsystems, making global collaborative analysis difficult.

[0004] 2. Lack of real-time performance: Traditional methods rely on manual inspections or single sensors, resulting in high response delays.

[0005] 3. Poor adaptability to complex scenarios: Distributed photovoltaic access and frequent start-up and shutdown of high-power equipment are prone to voltage fluctuations and safety hazards;

[0006] Therefore, there is an urgent need to provide a multi-agent large-scale model monitoring system and method for industrial electricity safety. Summary of the Invention

[0007] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid blurring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0008] Therefore, the purpose of the present invention is to provide a multi-agent large-scale model monitoring system and method for industrial electricity safety, which can realize highly intelligent monitoring and early warning, provide timely fault diagnosis and response, ensure the safe and stable operation of the power system, improve the degree of automation of the power system, and enhance its ability to respond to emergencies, effectively reducing the risks brought by power failures to industrial production.

[0009] To solve the above technical problems, according to one aspect of the present invention, the present invention provides the following technical solutions:

[0010] A multi-agent large-scale model monitoring system and method for industrial power safety, comprising:

[0011] The cloud-based analysis unit, which serves as the central control unit of the system, receives data from subordinate units and uses a multi-agent model, combined with knowledge graphs and self-supervised learning, to mine related equipment failure modes.

[0012] The data acquisition unit is connected to the cloud analysis unit and is used to deploy industrial-grade sensors, collect equipment operation data in real time, and transmit the monitored data information to the cloud analysis unit;

[0013] The execution control unit is connected to the cloud analysis unit, receives control commands from the cloud analysis unit, automatically triggers protection actions through the digital energy controller, and generates maintenance work orders;

[0014] The multi-agent collaboration unit is connected to the cloud analysis unit and is used to execute the reinforcement learning algorithm to enable each agent to continuously learn from historical data and environmental feedback and optimize its decision-making process.

[0015] As an optimal solution of the multi-agent large-scale model monitoring system for industrial electricity safety described in the present invention, the cloud-based analysis unit has a built-in edge computing layer, which is used to execute lightweight agents deployed locally in the factory to achieve data preprocessing, initial screening of anomalies and privacy desensitization.

[0016] As a preferred solution of the multi-agent large-scale model monitoring system for industrial electricity safety described in the present invention, the cloud-based analysis unit is connected to the data transmission unit, and the data transmission unit is used to form a data transmission channel through a wireless network and Internet of Things technology, transmit data to the cloud-based analysis unit, and connect the cloud-based analysis unit with each subordinate unit.

[0017] As a preferred solution of the multi-agent large-scale model monitoring system for industrial power safety described in the present invention, the data acquisition unit includes:

[0018] Current / voltage monitoring: real-time monitoring of current and voltage changes in the power system to promptly detect abnormal fluctuations, overvoltage or undervoltage;

[0019] Equipment health monitoring: monitors the operating status of electrical equipment, including temperature, vibration, corrosion, etc., and predicts equipment failures;

[0020] Load forecasting, based on historical data and current electricity consumption conditions, predicts the electricity load in the future and whether it will be overloaded.

[0021] As a preferred solution of the multi-agent large-scale model monitoring system for industrial electricity safety described in the present invention, the cloud-based analysis unit is connected to the monitoring and early warning unit, which is used to send an alarm to the operator immediately when the system detects abnormal information, provide fault location and repair suggestions, and provide the operator with real-time viewing of the power system operation status, monitoring data charts and alarm information.

[0022] A multi-agent large-scale model monitoring method for industrial power safety includes the following steps:

[0023] S1. Define the main objectives of industrial power safety monitoring and decompose the system into multiple agents based on the objectives, with each agent responsible for a specific monitoring task.

[0024] S2. After clarifying the goal of step S1, data collection and preprocessing are performed according to the task, and the collected data is transmitted in real time to the central data platform or the local processing unit of each intelligent agent through wired and wireless communication networks;

[0025] S3, cleaning, denoising, and standardizing the raw data collected in step S2 to ensure data quality;

[0026] S4. Through a message passing mechanism, each agent shares the data-related information obtained in step S3 when making decisions. Based on the data feedback between the agents, they jointly decide whether to trigger an alarm and build a system fault model. Using machine learning and pattern recognition technology, they identify potential faults or abnormal trends in the power system.

[0027] S5. Use reinforcement learning algorithms to allow each agent to continuously learn from historical data and environmental feedback and optimize its decision-making process.

[0028] As a preferred solution of the multi-agent large-scale model monitoring method for industrial power safety described in the present invention, in step S5, the collaborative processing includes the following:

[0029] Condition monitoring agent: uses a time-series graph neural network to analyze the electrical topology of the equipment;

[0030] Abnormal analysis agent: Based on contrastive learning and GAN generative adversarial networks, it simulates abnormal signal characteristics under extreme working conditions;

[0031] Fault prediction agent: Using deep reinforcement learning to optimize equipment maintenance strategies and reduce the risk of unplanned downtime.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] 1. Multimodal data fusion: Integrate multi-source data such as vibration, infrared thermal imaging, and current waveforms, and extract cross-modal features through the Transformer model.

[0034] 2. Self-healing closed-loop mechanism: Automates the entire "prediction-warning-self-healing" process, reducing fault identification time from minutes to milliseconds.

[0035] 3. Privacy protection design: The edge uses a federated learning framework and only uploads desensitized feature parameters to meet industrial data confidentiality requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and detailed embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be derived from these drawings without inventive effort. Among them:

[0037] Figure 1 This is a module block diagram of the monitoring system of the present invention;

[0038] Figure 2 This is a flowchart of the monitoring method steps of the present invention. DETAILED DESCRIPTION

[0039] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0040] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0041] Next, the present invention is described in detail with reference to schematic diagrams. For ease of illustration, cross-sectional views of device structures may be partially enlarged and not to scale when describing the embodiments of the present invention. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of protection of the present invention. Furthermore, in actual production, three-dimensional dimensions, including length, width, and depth, should be included.

[0042] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0043] The present invention provides a multi-agent large-scale model monitoring system and method for industrial power safety. Figure 1-2 ,include:

[0044] The cloud-based analysis unit, which serves as the central control unit of the system, receives data from subordinate units and uses a multi-agent model, combined with knowledge graphs and self-supervised learning, to mine related equipment failure modes.

[0045] The data acquisition unit is connected to the cloud analysis unit and is used to deploy industrial-grade sensors, collect equipment operation data in real time, and transmit the monitored data information to the cloud analysis unit;

[0046] The execution control unit is connected to the cloud analysis unit, receives control commands from the cloud analysis unit, automatically triggers protection actions through the digital energy controller, and generates maintenance work orders;

[0047] The multi-agent collaboration unit is connected to the cloud analysis unit and is used to execute the reinforcement learning algorithm to enable each agent to continuously learn from historical data and environmental feedback and optimize its decision-making process;

[0048] The cloud analysis unit has a built-in edge computing layer, which is used to execute lightweight intelligent agents deployed locally in the factory to achieve data preprocessing, anomaly screening and privacy desensitization;

[0049] The cloud analysis unit is connected to the data transmission unit, which is used to form a data transmission channel through a wireless network and Internet of Things technology, transmit data to the cloud analysis unit, and connect the cloud analysis unit with each subordinate unit;

[0050] The data acquisition unit includes:

[0051] Current / voltage monitoring: real-time monitoring of current and voltage changes in the power system to promptly detect abnormal fluctuations, overvoltage or undervoltage;

[0052] Equipment health monitoring: monitors the operating status of electrical equipment, including temperature, vibration, corrosion, etc., and predicts equipment failures;

[0053] Load forecasting: predicting the power load in the future based on historical data and current power consumption conditions, and predicting whether it will be overloaded;

[0054] The cloud analysis unit is connected to the monitoring and early warning unit. The monitoring and early warning unit is used to send an alarm to the operator immediately when the system detects abnormal information, provide fault location and repair suggestions, and allow the operator to view the power system operation status, monitoring data charts and alarm information in real time;

[0055] The multi-agent large-scale model monitoring method for industrial power safety includes the following steps:

[0056] S1. Define the main objectives of industrial power safety monitoring and decompose the system into multiple agents based on the objectives, with each agent responsible for a specific monitoring task.

[0057] S2. After clarifying the goal of step S1, data collection and preprocessing are performed according to the task, and the collected data is transmitted in real time to the central data platform or the local processing unit of each intelligent agent through wired and wireless communication networks;

[0058] S3, cleaning, denoising, and standardizing the raw data collected in step S2 to ensure data quality;

[0059] S4. Through a message passing mechanism, each agent shares the data-related information obtained in step S3 when making decisions. Based on the data feedback between the agents, they jointly decide whether to trigger an alarm and build a system fault model. Using machine learning and pattern recognition technology, they identify potential faults or abnormal trends in the power system.

[0060] S5. Use reinforcement learning algorithms to enable each agent to continuously learn from historical data and environmental feedback and optimize its decision-making process;

[0061] In step S5, the collaborative processing includes the following:

[0062] Condition monitoring agent: uses a time-series graph neural network to analyze the electrical topology of the equipment;

[0063] Abnormal analysis agent: Based on contrastive learning and GAN generative adversarial networks, it simulates abnormal signal characteristics under extreme working conditions;

[0064] Fault prediction agent: Using deep reinforcement learning to optimize equipment maintenance strategies and reduce the risk of unplanned downtime.

[0065] Although the present invention has been described above with reference to embodiments, various modifications may be made thereto and equivalent components may be substituted without departing from the scope of the present invention. In particular, as long as there are no structural conflicts, the various features of the embodiments disclosed herein may be combined with each other in any manner, and the omission of an exhaustive description of such combinations in this specification is solely for the sake of space and resource conservation. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A multi-agent large-scale model monitoring system for industrial power safety, characterized by: include: The cloud-based analysis unit, which serves as the central control unit of the system, receives data from subordinate units and uses a multi-agent model, combined with knowledge graphs and self-supervised learning, to mine related equipment failure modes. The data acquisition unit is connected to the cloud analysis unit and is used to deploy industrial-grade sensors, collect equipment operation data in real time, and transmit the monitored data information to the cloud analysis unit; The execution control unit is connected to the cloud analysis unit, receives control commands from the cloud analysis unit, automatically triggers protection actions through the digital energy controller, and generates maintenance work orders; The multi-agent collaboration unit is connected to the cloud analysis unit and is used to execute the reinforcement learning algorithm to enable each agent to continuously learn from historical data and environmental feedback and optimize its decision-making process.

2. The multi-agent large-scale model monitoring system for industrial power safety according to claim 1 is characterized in that: The cloud analysis unit has a built-in edge computing layer, which is used to execute lightweight intelligent agents deployed locally in the factory to achieve data preprocessing, initial screening of anomalies and privacy desensitization.

3. The multi-agent large-scale model monitoring system for industrial power safety according to claim 1 is characterized in that: The cloud analysis unit is connected to the data transmission unit, which is used to form a data transmission channel through a wireless network and Internet of Things technology, transmit data to the cloud analysis unit, and connect the cloud analysis unit with each subordinate unit.

4. The multi-agent large-scale model monitoring system for industrial power safety according to claim 1 is characterized in that: The data acquisition unit includes: Current / voltage monitoring: real-time monitoring of current and voltage changes in the power system to promptly detect abnormal fluctuations, overvoltage or undervoltage; Equipment health monitoring: monitors the operating status of electrical equipment, including temperature, vibration, corrosion, etc., and predicts equipment failures; Load forecasting, based on historical data and current electricity consumption conditions, predicts the electricity load in the future and whether it will be overloaded.

5. The multi-agent large-scale model monitoring system for industrial power safety according to claim 1 is characterized in that: The cloud analysis unit is connected to the monitoring and early warning unit, which is used to send an alarm to the operator immediately when the system detects abnormal information, provide fault location and repair suggestions, and provide the operator with real-time viewing of the power system operation status, monitoring data charts and alarm information.

6. A multi-agent large-scale model monitoring method for industrial power safety, characterized by: The steps include: S1. Define the main objectives of industrial power safety monitoring and decompose the system into multiple agents based on the objectives, with each agent responsible for a specific monitoring task. S2. After clarifying the goal of step S1, data collection and preprocessing are performed according to the task, and the collected data is transmitted in real time to the central data platform or the local processing unit of each intelligent agent through wired and wireless communication networks; S3, cleaning, denoising, and standardizing the raw data collected in step S2 to ensure data quality; S4. Through a message passing mechanism, each agent shares the data-related information obtained in step S3 when making decisions. Based on the data feedback between the agents, they jointly decide whether to trigger an alarm and build a system fault model. Using machine learning and pattern recognition technology, they identify potential faults or abnormal trends in the power system. S5. Use reinforcement learning algorithms to allow each agent to continuously learn from historical data and environmental feedback and optimize its decision-making process.

7. The multi-agent large-scale model monitoring method for industrial power safety according to claim 6 is characterized in that: In step S5, the collaborative processing includes the following: Condition monitoring agent: uses a time-series graph neural network to analyze the electrical topology of the equipment; Abnormal analysis agent: Based on contrastive learning and GAN generative adversarial networks, it simulates abnormal signal characteristics under extreme working conditions; Fault prediction agent: Using deep reinforcement learning to optimize equipment maintenance strategies and reduce the risk of unplanned downtime.

Images