GIS installation intelligent integrated environment control method and system

Abstract

The application provides a GIS installation intelligent integrated environment control method and system. The GIS installation intelligent integrated environment control method comprises the following steps: acquiring multi-modal real-time data; acquiring an installation process stage based on the multi-modal real-time data; inputting the multi-modal real-time data and the installation process stage into an environment prediction and decision model to generate an optimized regulation and control instruction, wherein the environment prediction and decision model is obtained by training based on data of a historical database; and controlling an execution module according to the optimized regulation and control instruction to dynamically adjust the installation environment. The technical problem solved by the application is that the prior art cannot perceive and adapt to the differentiated and dynamic requirements of different installation process stages on environmental conditions, the regulation and control are lagged and not accurate enough, thereby increasing the risk of equipment quality defects and even operation failure caused by substandard environment in the installation process.

Description

Technical Field

[0001] This invention relates to the field of electrical equipment technology, and more specifically, to a method and system for intelligent integrated environmental control of GIS installation. Background Technology

[0002] Gas-insulated metal-enclosed switchgear (GIS) is a key piece of equipment in modern power systems, and its installation quality directly affects the long-term safe and stable operation of substations. The installation process of GIS has extremely stringent requirements for parameters such as the cleanliness, temperature, humidity, and slight positive pressure of the assembly environment. Any slight dust or moisture intrusion may lead to a decrease in internal insulation performance, causing partial discharge or even equipment failure, resulting in significant economic losses and safety risks.

[0003] Currently, environmental control at GIS installation sites mainly relies on traditional fixed or semi-fixed environmental control equipment, combined with manually set static parameter thresholds.

[0004] However, the relevant technologies have at least one of the following problems: existing technologies are difficult to perceive and adapt to the differentiated and dynamic requirements of environmental conditions at different stages of installation processes, and the control is lagging and not precise enough, thereby increasing the risk of equipment quality defects or even operational failures due to substandard environment during installation. Summary of the Invention

[0005] The technical problem solved by this invention is that existing technologies are unable to perceive and adapt to the differentiated and dynamic requirements of environmental conditions at different stages of installation processes. The control is lagging and not precise enough, thereby increasing the risk of equipment quality defects or even operational failures due to substandard environment during installation.

[0006] To address the aforementioned problems, this invention provides an intelligent integrated environmental control method for GIS installation, comprising: acquiring multimodal real-time data including video image information, personnel positioning information, and environmental parameters; acquiring the installation process stage based on the multimodal real-time data; determining whether a linkage trigger event has occurred based on the video image information and personnel positioning information; if so, executing a preset environmental control action bound to the linkage trigger event; if not, inputting the multimodal real-time data and the installation process stage into an environmental prediction and decision model to generate optimized control instructions; generating optimized control instructions includes: the installation process stage includes a vacuuming and gas injection stage; when the equipment is in the vacuuming and gas injection stage, a feedforward temperature control instruction is generated based on vacuum pump operation data and ambient temperature and humidity data. The system generates preheating commands for the injection interface based on environmental parameters and SF6 gas characteristics. It then compares and verifies the initial control commands output by the environmental prediction and decision-making model with a preset environmental control rule base before outputting the final optimized control commands. If cloud communication is interrupted, it switches to local rule base mode to generate optimized control commands. Real-time simulation is performed using a digital twin model that is proportionally mapped to and synchronized with the physical installation site based on multimodal real-time data. The digital twin model includes the 3D geometry of the GIS equipment, personnel location trajectories, and the operating status of the control equipment. This model is used to simulate and predict environmental changes and control effects, and the optimized control commands are evaluated and corrected based on the simulation results. Finally, the system dynamically adjusts the installation environment based on the corrected optimized control commands.

[0007] Compared with existing technologies, the technical effects achieved by this solution are as follows: This invention, by acquiring multimodal real-time data including video image information, personnel positioning information, and environmental parameters, can comprehensively and in real-time perceive the environmental status and operational conditions of the installation site. This includes GIS multimodal real-time data, which also includes local micro-environmental monitoring data. Based on the acquisition of installation process stages using multimodal real-time data, it can accurately match the differentiated environmental requirements of different process stages, achieving synchronous adaptation between environmental control and the installation process. By judging and triggering events through video image information and personnel positioning information, and directly executing the bound preset environmental control actions upon triggering, it can provide immediate response to key risk scenarios, avoiding the risk of environmental runaway due to control lag. Inputting multimodal real-time data and installation process stages into an environmental prediction and decision-making model can intelligently generate control strategies adapted to the current working conditions. During the vacuuming and gas injection stages, feedforward temperature control commands are generated based on vacuum pump operation data and environmental temperature and humidity data, and based on environmental parameters and multimodal real-time data, including local micro-environmental monitoring data. SF6 multimodal real-time data, including local microenvironment monitoring data, generates preheating commands for the gas injection interface through gas characteristic simulation. This prevents temperature fluctuations and gas condensation in advance, ensuring process stability. The initial control commands output by the model are compared and verified with a preset environmental control rule base. In case of cloud communication interruption, the system switches to a local rule base mode, ensuring the safety and reliability of the control commands and improving the system's continuous operation capability in complex environments. Real-time simulation is performed using a digital twin model that is proportionally mapped and synchronized with the physical installation site based on multimodal real-time data. This allows for early prediction of environmental changes and control effects. The simulation results are used to evaluate and correct optimized control commands, achieving proactive and precise control. Finally, the control execution module dynamically adjusts the installation environment based on the corrected optimized control commands. This makes the environmental control throughout the installation process more precise, timely, and stable, significantly reducing the risk of equipment quality defects and operational failures caused by substandard environments, ensuring installation quality and long-term operational reliability.

[0008] In one embodiment of the present invention, the linkage triggering event includes at least one of the following: personnel entering a high-cleanliness area, opening of the dustproof booth, and the installation process entering a critical process stage; the installation process stage also includes: unpacking and cleaning and component placement stage and conductor connection and insertion stage.

[0009] Compared with existing technologies, the technical effects achieved by adopting this solution are as follows: By limiting the triggering events to at least one of the following: personnel entering a high-cleanliness area, opening a dustproof shed, or the installation process entering a critical stage, it can accurately cover high-risk scenarios that are most likely to cause environmental runaway and affect installation quality during the installation process, including multimodal real-time data, local micro-environment monitoring data, and GIS multimodal real-time data, including local micro-environment monitoring data. This enables rapid identification of critical links and sudden interferences. When the above events occur, preset environmental control actions are executed directly, skipping the conventional decision-making process and instantly activating a powerful protection strategy. This avoids a decrease in cleanliness, excessive humidity, or contaminant intrusion due to control lag, further improving the immediacy and reliability of environmental control.

[0010] By clearly defining the installation process stages, including unpacking and cleaning and component placement, conductor connection and insertion, the system can fully cover the core processes with significantly different environmental requirements throughout the entire installation process, including multimodal real-time data, local micro-environment monitoring data, and GIS multimodal real-time data, including local micro-environment monitoring data. This allows the system to implement differentiated and refined control for cleanliness, humidity, micro-positive pressure, and other requirements at different stages, ensuring a high degree of matching between environmental control strategies and actual installation procedures. This further enhances environmental adaptability and control accuracy, providing a stable and reliable environmental guarantee for the installation quality of each key process stage.

[0011] In one embodiment of the present invention, multimodal real-time data and installation process stages are input into an environmental prediction and decision model to generate optimized control instructions, including: determining whether the equipment is in the unpacking, cleaning, and component placement stage based on the installation process stage; if so, identifying unpacking or component movement based on video image information and determining the location of the pollution source in conjunction with personnel positioning information; wherein, the optimized control instructions include at least one of the following: an instruction to control the air supply system to form a dynamic air curtain at the pollution source location; and an instruction to control the mobile purification unit to approach the pollution source for purification.

[0012] Compared to existing technologies, this technical solution achieves the following technical effects: by integrating video motion recognition and precise personnel positioning, it enables real-time tracking and accurate location of dynamic pollution sources during operations such as unpacking and component movement. Based on this, the system can generate and execute targeted control commands centered on the pollution source, such as forming a dynamic air curtain for isolation at the pollution location or dispatching mobile purification units for close-range treatment. This changes the traditional extensive mode of overall environmental control, achieving a leap from "area control" to "source control," thereby significantly improving the immediacy and accuracy of pollution suppression, effectively reducing the risk of dust diffusion, and reducing unnecessary overall regional purification energy consumption.

[0013] In one embodiment of the present invention, multimodal real-time data and installation process stages are input into an environmental prediction and decision model to generate optimized control instructions, including: multimodal real-time data also includes local microenvironment monitoring data; determining whether the equipment is in the conductor connection and plugging stage based on the installation process stage; if the determination is yes, then controlling the directional air supply system to form a clean airflow barrier based on personnel positioning information, and dynamic dehumidification compensation instructions generated based on local microenvironment monitoring data.

[0014] Compared with existing technologies, the technical effects achieved by this solution are as follows: Addressing the extreme sensitivity of cleanliness and humidity during the conductor connection stage, precise "person-area" linkage protection is achieved by integrating personnel positioning and local microenvironment monitoring data. On one hand, a directional clean airflow barrier is formed based on personnel location, actively isolating dust generated by the human body; on the other hand, dynamic compensation and adjustment based on real-time local humidity data ensures the extreme stability of the microenvironment at the operating interface, effectively guaranteeing the quality of the connection process from both spatial isolation and parameter accuracy perspectives.

[0015] In one embodiment of the present invention, environmental data, control command logs, key video clips, and identified process events during the installation process are associated with the unique identifier of the GIS equipment; a traceable chain of quality evidence is generated and stored in the full lifecycle digital archive of the GIS equipment.

[0016] Compared to existing technologies, the technical benefits of this solution are as follows: By structurally linking all environmental control data, operating instructions, video evidence, and process events throughout the entire process to the unique identifier of the GIS equipment, an immutable and fully traceable digital evidence chain for installation quality is automatically generated. This not only provides direct proof of process compliance and quality standards but also offers a valuable data foundation for subsequent equipment operation and maintenance, fault investigation, and full lifecycle management, realizing a deep extension of the value of environmental control data to digital asset management.

[0017] On the other hand, the present invention also provides an intelligent integrated environmental control system for GIS installation, used to implement the intelligent integrated environmental control method for GIS installation as described in any of the above examples, comprising: a data acquisition module for acquiring multimodal real-time data including video image information, personnel positioning information, and environmental parameters; a processing module for acquiring the installation process stage based on the multimodal real-time data, judging the linkage triggering event, calling the environmental prediction and decision model, combining the environmental control rule base to generate optimized control instructions, and driving the digital twin model to perform simulation to correct the optimized control instructions; and an execution module for dynamically adjusting the installation environment based on the corrected optimized control instructions, including the multimodal real-time data and local micro-environment monitoring data, and executing preset environmental control actions corresponding to the linkage triggering event; wherein, the environmental prediction and decision model is obtained by training based on a historical database, and the digital twin model is a three-dimensional virtual model that is proportionally mapped to the physical installation site and synchronized in real time.

[0018] Compared with existing technologies, the technical effects achieved by adopting this technical solution are as follows: it can achieve the technical effects corresponding to any of the above examples, which will not be elaborated here.

[0019] By adopting the technical solution of the present invention, the following technical effects can be achieved:

[0020] (1) It has achieved a leap from static threshold control to dynamic perception and adaptive regulation that is intelligently synchronized with the process. The system can automatically identify different installation stages and generate optimization instructions such as source suppression, precise protection and feedforward compensation for the core environmental risks of each stage, thereby significantly improving the timeliness, accuracy and effectiveness of environmental control;

[0021] (2) A decision-making and control system with high reliability, strong robustness, and safety redundancy was constructed. Through the dual decision-making mechanism of "model prediction + rule verification" and the dual-mode operation capability of "cloud collaboration + local offline", the safety and reliability of the control commands and the continuous and stable operation of the system were ensured. At the same time, simulation pre-evaluation was carried out using digital twin technology to further optimize the control strategy and improve the execution success rate;

[0022] (3) It has pioneered a new model of deep integration of environmental control and quality management. The system not only ensures the real-time installation environment, but also establishes a traceable full life cycle digital archive for GIS equipment by automatically generating a quality evidence chain containing multi-dimensional data with a unique ID of the associated equipment. This transforms process control data into a valuable digital twin foundation for assets, realizing the datafication and transparency of quality control. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating a specific method for intelligent integrated environmental control of GIS installation. Detailed Implementation

[0025] The following will refer to the appendix to this application. Figure 1 The technical solutions in this application are clearly and completely described herein. Obviously, the described embodiments are merely a part of the embodiments of this application, and not all of them. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0026] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0027] This invention provides an intelligent integrated environmental control method for GIS installation, comprising: acquiring multimodal real-time data including video image information, personnel positioning information, and environmental parameters; acquiring the installation process stage based on the multimodal real-time data; determining whether a linkage trigger event has occurred based on the video image information and personnel positioning information; if so, executing a preset environmental control action bound to the linkage trigger event; if not, inputting the multimodal real-time data and the installation process stage into an environmental prediction and decision model to generate optimized control instructions; generating optimized control instructions includes: the installation process stage includes a vacuuming and gas injection stage; when the equipment is in the vacuuming and gas injection stage, a feedforward temperature control instruction is generated based on the vacuum pump operation data and environmental temperature and humidity data, and... Based on environmental parameters and SF6 gas characteristics, a preheating command for the gas injection interface is generated through simulation. The initial control command output from the environmental prediction and decision-making model is compared and verified with a preset environmental control rule base before the final optimized control command is output. If cloud communication is interrupted, the system switches to local rule base mode to generate the optimized control command. Real-time simulation is performed based on a digital twin model that is proportionally mapped to and synchronized with the physical installation site using multimodal real-time data. The digital twin model includes the 3D geometry of GIS equipment, personnel location trajectories, and the operating status of the control equipment. This model is used to simulate and predict environmental changes and control effects, and the optimized control command is evaluated and corrected based on the simulation results. The control execution module dynamically adjusts the installation environment based on the corrected optimized control command.

[0028] Specifically, multimodal real-time data refers to real-time data acquired from various types of sensors and acquisition devices at the installation site, including multimodal real-time data and local microenvironment monitoring data (GIS multimodal real-time data and local microenvironment monitoring data), which comprehensively reflects the current environmental status, personnel activities, and equipment operating conditions. It includes video image information, personnel positioning information, temperature, humidity, cleanliness, micro-positive pressure, vacuum pump operating parameters, etc., and can provide a complete, accurate, and real-time data foundation for process identification, event judgment, and intelligent control.

[0029] Furthermore, the installation process stage refers to the current installation progress of the equipment based on the multimodal real-time data automatically identified by multimodal real-time data, which also includes local micro-environment monitoring data and GIS multimodal real-time data, including local micro-environment monitoring data. Specifically, it includes the unpacking and cleaning and component placement stage, the conductor connection and insertion stage, and the vacuuming and gas injection stage. Different stages have different environmental requirements for cleanliness, humidity, and temperature and humidity stability, so differentiated and precise environmental control is required for each stage.

[0030] Furthermore, the linkage-triggered event refers to a critical event that may rapidly affect the installation environment, detected based on video image recognition and personnel positioning. When such an event is detected, the system directly executes the preset strong control action without waiting for the model to make a decision, thereby achieving instantaneous protection and preventing the environment from getting out of control.

[0031] Furthermore, the environmental prediction and decision-making model is an intelligent decision-making model trained based on historical installation environmental data, process data, and control results. It is used to combine current multimodal data and process stages to automatically output appropriate initial control commands, thereby achieving predictive and adaptive control.

[0032] Furthermore, during the vacuuming and gas injection stages, feedforward temperature control commands are generated based on vacuum pump operating data and ambient temperature and humidity data. This can suppress temperature drift caused by vacuum pump heating or environmental fluctuations in advance, ensuring a stable vacuum process. Based on environmental parameters and multimodal real-time data, including local microenvironment monitoring data, and SF6 multimodal real-time data including local microenvironment monitoring data, gas characteristic simulation generates preheating commands for the gas injection interface. This can prevent gas condensation due to temperature differences during injection, ensuring gas injection quality and equipment insulation reliability.

[0033] Furthermore, the preset environmental control rule base contains safety thresholds and standard control logic preset for the field. The initial control commands output by the model are compared and verified with the rule base to eliminate unreasonable or out-of-range commands. When cloud communication is interrupted, the system automatically switches to local rule base mode to run independently, ensuring uninterrupted and unfailed control, and improving field adaptability and reliability.

[0034] Furthermore, the digital twin model is a scaled virtual model created using 3D modeling tools based on the actual dimensions of the installation site, equipment placement, air outlet layout, sensor coordinates, and dustproof canopy structure, including multimodal real-time data, local micro-environment monitoring data, GIS multimodal real-time data, and local micro-environment monitoring data. The model contains real, callable multimodal real-time data, including local micro-environment monitoring data, GIS multimodal real-time data, and local micro-environment monitoring data, the 3D geometric structure of the equipment, purification units, air supply ducts, personnel positioning coordinates, monitoring points, and operating status parameters of control equipment. It maintains spatial structure consistency with the physical site, achieving real-time mapping of all elements between the virtual scene and the physical site.

[0035] Before the system is put into operation, the digital twin model completes the on-site coordinate calibration and data interface binding. The video image recognition results, personnel positioning coordinates, temperature and humidity, cleanliness, micro-positive pressure, vacuum pump operation status, dehumidifier status, and air supply parameters are written into the model in real time at a fixed frequency, so that the virtual scene and the physical scene are kept updated synchronously.

[0036] The operation and management of digital twin models includes data access management, simulation task scheduling, instruction pre-play management, result output management, and anomaly locking mechanism.

[0037] In one embodiment, the digital twin model refreshes the model state at a frequency of once per second / time, including local microenvironment monitoring data, based on multimodal real-time data and local microenvironment monitoring data. A complete environmental simulation is performed once per second / time to ensure that the model does not lag, drift, or lose data.

[0038] During the simulation, the digital twin model calculates and outputs three actual results in real time based on the current environmental parameters, the control commands to be executed, personnel locations, and process stages: the temperature and humidity change trends and stabilization time after the control commands are executed; the cleanliness distribution and the range of pollution source diffusion; and whether there are short circuits, dead zones, or disturbances in the airflow organization.

[0039] Based on the simulation results of temperature, humidity, cleanliness, and airflow, the system quantitatively evaluates the optimized control commands. If the simulation shows that the parameters after adjustment do not meet the standards, the system automatically corrects the air supply volume, dehumidification volume, preheating power, or air curtain position. If the simulation shows local environmental degradation, the system automatically supplements and compensates with control commands to ensure that the commands finally issued to the execution module are authentic, effective, and meet the standards on the first attempt. Furthermore, the optimized control commands are the final control commands after model generation, rule base verification, and digital twin simulation correction. They are used to precisely guide the execution module to complete actions such as temperature, humidity, cleanliness, airflow, and preheating to ensure that the environment meets the current process requirements.

[0040] Furthermore, the execution module includes equipment such as air conditioners, dehumidifiers, purification units, directional air supply devices, dynamic air curtains, preheating devices, and mobile purification units. It is used to receive optimization and control commands and complete actual environmental adjustment actions, realizing dynamic, accurate, and stable control of the installation environment, including multimodal real-time data and local microenvironment monitoring data.

[0041] Specifically, the multimodal real-time data provided in this embodiment also includes local microenvironment monitoring data. The installation process of the intelligent integrated environmental control method is as follows: First, multimodal real-time data containing video image information, personnel positioning information, and environmental parameters is acquired through on-site deployed cameras, positioning base stations, temperature and humidity sensors, particulate matter sensors, micro-positive pressure sensors, and equipment acquisition units.

[0042] Subsequently, based on real-time multimodal data, intelligent analysis is performed to automatically identify the current installation process stage. Simultaneously, the system uses video images and personnel location information to determine in real-time whether a trigger event has occurred. If triggered, preset controlled environment actions are executed directly; otherwise, an intelligent decision-making process is initiated.

[0043] The system inputs multimodal real-time data and installation process stages into the environmental prediction and decision-making model to generate initial control commands. During the vacuuming and gas injection stages, the system automatically generates feedforward temperature control commands and gas injection interface preheating commands to ensure the stability of key processes.

[0044] Subsequently, the initial commands are verified against the preset environment control rule base to ensure that the commands are safe and compliant; if cloud communication is interrupted, the system will automatically switch to local rule base control.

[0045] Next, the system drives the digital twin model to simulate and preview the verified instructions, predict changes in temperature, humidity, cleanliness, and airflow organization, and evaluates and corrects the instructions based on the simulation results to obtain the final optimized control instructions.

[0046] Finally, the revised optimized control instructions are sent to the execution module to dynamically and precisely adjust the installation environment, ensuring that the environment throughout the entire installation process meets the stringent requirements of each process stage, thereby improving installation quality and equipment operational reliability.

[0047] In one embodiment of the present invention, the installation process includes: unpacking and cleaning and component placement stage and conductor connection and insertion stage; the linkage triggering event includes at least one of the following: personnel entering a high-cleanliness area, opening of the dustproof booth, and the installation process entering a critical process stage.

[0048] Specifically, video image information refers to real-time video streams or image frames acquired by visual sensors deployed within the GIS installation area. This information provides intuitive visual data for identifying the state of objects, personnel behavior, and changes in equipment components on-site. For example, image recognition technology can be used to detect the opening and closing status of dustproof shed doors or to identify the installation progress of critical equipment components.

[0049] Furthermore, personnel location information refers to data obtained through real-time location tracking of workers within the GIS installation area using specific technical means. Its purpose is to accurately determine the range and specific location of personnel's activities in order to determine whether they have entered areas with special cleanliness requirements.

[0050] Preferably, ultra-wideband (UWB) positioning technology is used, which achieves high-precision positioning at the centimeter or sub-meter level by setting up UWB base stations in the area and having personnel wear UWB tags.

[0051] Based on video image information and personnel location information, the system determines whether a preset trigger event has occurred. This determination process aims to monitor the on-site situation in real time, identify critical events that may have an immediate impact on environmental cleanliness or installation quality, and enable the system to respond quickly.

[0052] In one embodiment, an event recognition model is established, which is a trained machine learning model. For the event of "personnel entering a high-cleanliness area," the system continuously compares the personnel's location information with the preset electronic fence boundary of the high-cleanliness area. Once the personnel's location enters the area, the event is triggered. For the event of "dustproof shed opening," the system analyzes video images to identify the opening and closing status of the dustproof shed door. When an opening action is detected, the event is triggered. For the event of "installation process entering a critical stage," the installation status of critical equipment components can be identified through video images. For example, it can identify that the casing of a GIS device has been opened, or it can combine personnel location information to determine whether personnel are performing critical operations at a specific workstation.

[0053] Furthermore, if a linked trigger event occurs, the optimized control instructions generated by the environmental prediction and decision-making model are ignored, and instead, the preset environmental control actions bound to that event are executed. This mechanism ensures that in emergency or critical scenarios, the system can immediately execute predefined, targeted control strategies, avoiding the risks caused by potential response lags or inappropriate decisions in the environmental prediction and decision-making model. The system can maintain a mapping table internally. When a linked trigger event is detected, the system queries this table to find the corresponding preset environmental control action and directly sends the action instruction to the execution module, while pausing or overriding the currently generated optimized control instructions from the environmental prediction and decision-making model. For example, when the event "personnel enter a high-cleanliness requirement area" is triggered, the system immediately executes the preset action "activate the area-oriented purification air curtain" without waiting for the model to calculate new control instructions.

[0054] The triggering events include at least one of the following: personnel entering a high-cleanliness area, opening of the dustproof enclosure, or the installation process entering a critical technological stage. Personnel entering a high-cleanliness area refers to workers entering the core installation area of ​​the GIS equipment. This area has extremely high cleanliness requirements, and personnel entry may introduce dust or contamination risks. Detection can be achieved through a combination of personnel location information and electronic fences. Opening of the dustproof enclosure refers to the opening of the door of the dustproof enclosure used to isolate the GIS installation area from the external environment, which may allow external contamination to enter. Detection can be achieved through video image analysis or door magnetic sensors. The installation process entering a critical technological stage refers to stages in the GIS installation process that are particularly sensitive to environmental conditions, such as the unpacking and cleaning and component placement stage, the conductor connection and insertion stage, and the vacuuming and gas injection stage. This can be determined by identifying the installation status of key equipment components through video image recognition or by obtaining process flow information through integration with the installation management system.

[0055] Through the above technical solution, this application introduces a linkage-triggered event judgment mechanism based on specific real-time data. When a preset event occurs, it directly executes preset environmental control actions, effectively solving the problem of control lag caused by untimely response of environmental prediction and decision-making models in sudden or critical scenarios. Multimodal real-time data, especially video image information and personnel positioning information, provides richer real-time monitoring capabilities. Video image information can intuitively capture changes in the status of on-site equipment, and personnel positioning information can accurately track personnel locations, thus laying a solid data foundation for event judgment.

[0056] Based on this real-time data, it can determine whether a preset trigger event has occurred. By utilizing the visual recognition capabilities of video image information and the spatial positioning capabilities of personnel location information, high-risk events can be detected in real time, ensuring the timeliness and accuracy of event identification and avoiding missing critical opportunities due to model delays.

[0057] If a linkage trigger event occurs, the optimization and control instructions generated by the environmental prediction and decision-making model are ignored, and instead, the preset environmental control actions bound to the event are executed. This mechanism directly bypasses the model decision-making when a specific event occurs and adopts a predefined targeted control strategy (such as strengthening purification in a high-cleanliness area or adjusting environmental parameters at a critical stage), eliminating the model's adaptability problem in sudden scenarios and improving the immediacy and reliability of the response.

[0058] The linked trigger events cover the core risk points in the GIS installation process. The preset environmental control actions can address these scenarios in a targeted manner, ensuring that environmental control is closely matched with process requirements. This significantly reduces the risk of equipment quality defects and operational failures caused by substandard environmental conditions, thus guaranteeing the installation quality of GIS equipment.

[0059] Optionally, the visual sensor can be a visible light camera, an infrared camera, or a fisheye camera to capture images or videos.

[0060] Further multimodal real-time data, especially video image information and personnel positioning information, provides richer real-time monitoring capabilities. Video image information can intuitively capture changes in the status of on-site equipment, and personnel positioning information can accurately track personnel locations, thus laying a solid data foundation for event judgment. Based on this real-time data, it is possible to determine whether a preset linkage trigger event has occurred. By utilizing the visual recognition capabilities of video image information and the spatial positioning capabilities of personnel positioning information, high-risk events can be detected in real time, ensuring the timeliness and accuracy of event identification and avoiding missing critical opportunities due to model delays.

[0061] If a triggering event occurs, the optimized control instructions generated by the environmental prediction and decision-making model are ignored. Instead, pre-defined environmental control actions bound to the event are executed. This mechanism bypasses model decision-making directly when a specific event occurs, employing predefined targeted control strategies. This eliminates the model's adaptability issues in sudden scenarios and improves the immediacy and reliability of the response. The triggering events cover the core risk points in the GIS installation process. The pre-defined environmental control actions can specifically address these scenarios, ensuring a close match between environmental control and process requirements. This significantly reduces the risk of equipment quality defects and operational failures caused by substandard environmental conditions, thus guaranteeing the installation quality of GIS equipment.

[0062] Based on this, the current installation process stage of the equipment is determined according to video image information. This step aims to automatically identify the specific progress of the GIS equipment installation process through the analysis of real-time video image information and categorize it into predefined installation process stages. To achieve this goal, a deep learning-based image recognition model can be used, such as training a convolutional neural network (CNN) or other deep learning models, enabling them to recognize features such as GIS equipment components, tools, personnel postures, and environmental changes in video frames, and determine the current installation process stage based on these features. For example, the stage can be determined by recognizing visual events such as unpacking actions, component hoisting, bolt connections, and vacuum pump operation status.

[0063] Meanwhile, the installation process includes: unpacking and cleaning and component placement, conductor connection and insertion, and vacuuming and gas injection. These stages clearly define several key phases in the GIS installation process with significantly different environmental control requirements, providing refined input for subsequent environmental prediction and decision-making. Specifically, the unpacking and cleaning and component placement phase visually represents the opening of the GIS equipment enclosure, the removal of internal components, cleaning operations (such as wiping and vacuuming), and the initial placement or hoisting of components at the installation location; the conductor connection and insertion phase visually represents the precise alignment of internal conductors within the GIS, bolt tightening, and the installation of connectors; the vacuuming and gas injection phase visually represents the connection and operation of the vacuum pump equipment, changes in pressure gauge readings, and the connection and operation of the SF6 gas filling equipment. By breaking down the complex installation process into these clearly defined stages, the environmental control system can perform customized and refined adjustments based on the specific environmental requirements of each stage.

[0064] In one embodiment of the present invention, multimodal real-time data and installation process stages are input into an environmental prediction and decision model to generate optimized control instructions, including: determining whether the equipment is in the unpacking, cleaning, and component placement stage based on the installation process stage; if so, identifying unpacking or component movement based on video image information and determining the location of the pollution source in conjunction with personnel positioning information; wherein the generated optimized control instructions include at least one of the following: an instruction to control the air supply system to form a dynamic air curtain at the pollution source location; and an instruction to control the mobile purification unit to approach the pollution source for purification.

[0065] Specifically, the system determines whether the equipment is in the unpacking, cleaning, and component placement stage based on the installation process phase. The installation process phase describes the current installation progress of the GIS equipment, and the "unpacking, cleaning, and component placement stage" is a specific stage in the GIS installation process that requires high cleanliness and is prone to contamination. Determining whether the equipment is currently in this stage aims to trigger corresponding precise environmental control strategies. This determination is achieved by combining video image analysis (e.g., recognizing unpacking actions and component movement trajectories) or sensor data (e.g., component weight changes and installation tool usage) to automatically determine the current process phase.

[0066] If the system is determined to be in the unpacking, cleaning, and component placement stage, then the unpacking or component movement is identified based on video image information. Video image information refers to real-time video stream data collected by cameras deployed within the installation area. Identifying unpacking or component movement involves using image processing and machine learning algorithms to detect the opening of the GIS equipment enclosure or the handling and installation of internal components from the video stream.

[0067] The location of pollution sources is determined by combining personnel location information. The location of a pollution source refers to the specific spatial area where pollutants are generated or may be generated during the GIS installation process. Determining the location of pollution sources by combining personnel location information involves associating identified actions such as unpacking or moving components with the real-time location information of the personnel performing these actions, thereby accurately pinpointing the specific spatial point or area where pollution occurs. This approach aims to provide precise spatial coordinates of the pollution source, providing a basis for subsequent precise environmental control and avoiding blind or large-scale control measures.

[0068] The method to achieve this determination is to obtain the location information of personnel near the area where the action occurs during the time period when the video image recognizes the action of opening the box or moving the component, and use the personnel location as the center point or range of the pollution source.

[0069] The generated optimization and control instructions include at least one of the following: instructions to control the air supply system to form a dynamic air curtain at the pollution source location; and instructions to control the mobile purification unit to approach the pollution source for purification. A dynamic air curtain refers to a high-speed airflow barrier formed by the air supply system in a specific area to prevent external pollutants from entering or internal pollutants from spreading. The instructions for the dynamic air curtain are to control the air supply system to adjust the airflow direction, airflow speed, and air supply area. Its purpose is to quickly isolate the pollution source, prevent pollutants from spreading throughout the installation area, and maintain the cleanliness of other areas.

[0070] In one embodiment, the installation area is divided into multiple air supply zones, each equipped with an independent air valve and fan. When the location of the pollution source is determined, the system sends a command to the air supply system of the corresponding zone to adjust the air valve opening and fan speed, forming a local high wind speed area around the pollution source to achieve an air curtain effect.

[0071] In another embodiment, an air outlet with adjustable guide vanes or nozzles is used. An electric actuator adjusts the airflow direction and angle in real time according to commands, ensuring precise airflow coverage of the pollution source area and forming a dynamic air curtain. The mobile purification unit is a mobile device with air filtration and purification functions, capable of autonomously moving to a designated location. The command to control the mobile purification unit to approach the pollution source for purification guides the unit's movement and activates its purification function. Its purpose is to perform localized, efficient, on-site purification of the pollution source, rapidly reducing the particulate matter concentration in the polluted area.

[0072] In another embodiment, purification modules such as high-efficiency air filters (HEPA) and fans are integrated into the chassis of an AGV (Automated Guided Vehicle). Upon receiving an instruction, the AGV autonomously moves to the vicinity of the pollution source based on the pollution source location information, using methods such as SLAM (Simultaneous Localization and Mapping) or magnetic navigation, and then activates the purification function. Alternatively, operators can remotely control the vehicle or use a system-provided path plan to guide it to the vicinity of the pollution source and activate its built-in purification system for localized purification.

[0073] In one embodiment of the present invention, multimodal real-time data and installation process stages are input into an environmental prediction and decision model to generate optimized control instructions, including: multimodal real-time data also includes local microenvironment monitoring data; determining whether the equipment is in the conductor connection and plugging stage based on the installation process stage; if the determination is yes, then controlling the directional air supply system to form a clean airflow barrier based on personnel positioning information, and dynamic dehumidification compensation instructions generated based on local microenvironment monitoring data.

[0074] Specifically, based on UWB (Ultra-Wideband) technology, high-precision indoor positioning is achieved by deploying positioning base stations in the installation area and having personnel wear positioning tags; local micro-environment monitoring data refers to environmental parameter data collected at specific local locations within the GIS equipment installation area, and by deploying a network of micro-environment sensors, parameters such as temperature, humidity, and dust particle concentration in specific areas are monitored in real time.

[0075] Furthermore, dynamic dehumidification compensation commands are generated based on local microenvironment monitoring data. These commands dynamically adjust the operating parameters of the dehumidification equipment according to real-time humidity monitoring data of the local area, precisely controlling the humidity of the local microenvironment to prevent condensation or exceeding the standard.

[0076] In one embodiment, the dehumidification compensation command calculates the required dehumidification amount in real time by analyzing the deviation between local microenvironment monitoring data and preset humidity thresholds or target values, and sends it to the dehumidifier or air conditioning system to adjust its operating mode.

[0077] In another embodiment, by combining an environmental prediction model, not only can the current humidity change be responded to, but the humidity trend in the future can also be predicted, and dehumidification compensation instructions can be generated in advance to achieve forward-looking regulation and avoid humidity fluctuations.

[0078] In one embodiment of the present invention, multimodal real-time data and installation process stages are input into an environmental prediction and decision model to generate optimized control instructions, including: multimodal real-time data including installation equipment status information and environmental parameters; determining whether the equipment is in the vacuuming and gas injection stage based on the installation process stage; if the determination is yes, then a feedforward temperature control instruction is generated based on vacuum pump operation data and environmental temperature and humidity data, and a gas injection interface preheating instruction is generated based on environmental parameters and SF6 gas characteristic simulation.

[0079] Specifically, installation equipment status information refers to the operational status data of key internal or external components of GIS equipment during installation, such as the start / stop status, operating power, vacuum pressure curve, valve opening / closing status, and heater operating status. This information can be directly acquired through sensors integrated into the equipment, PLCs (Programmable Logic Controllers), or SCADA (Supervisory and Data Acquisition) systems, or through data exchange via interfaces provided by the equipment manufacturer. Environmental parameters refer to various physical environmental indicators at the installation site, such as ambient temperature, humidity, atmospheric pressure, cleanliness of the local microenvironment (e.g., particulate matter concentration), and airflow velocity. These parameters can be monitored and collected in real time through various environmental sensors deployed in the installation area (e.g., temperature and humidity sensors, particulate matter sensors, wind speed sensors, etc.).

[0080] Furthermore, vacuum pump operating data includes pumping time, pumping rate, current vacuum level, and motor temperature. This data reflects the real-time progress of the vacuuming process and the equipment load. Ambient temperature and humidity data refers to the real-time temperature and humidity values ​​within the installation area. Feedforward temperature control is a predictive control strategy designed to detect or predict system disturbances in advance and take appropriate control measures before the disturbances affect the system. For example, based on the vacuum pump's expected operating time and power, as well as ambient temperature and humidity data, a preset physical model or empirical curve can be used to predict the temperature change trend of the installation area over a future period, and advance adjustment commands can be issued to the air conditioning system or heater to maintain the target temperature range.

[0081] Simultaneously, a preheating command for the injection interface is generated based on environmental parameters and SF6 (sulfur hexafluoride) gas characteristic simulation. SF6 gas characteristic simulation refers to using computer models to simulate the physicochemical properties of SF6 gas under different temperature and pressure conditions, particularly its thermodynamic behavior and flow characteristics. For example, a state equation model and heat and mass transfer model for SF6 gas can be established, or computational fluid dynamics (CFD) methods can be used to simulate the flow and temperature distribution of SF6 gas at the injection pipeline and interface. The injection interface preheating command is a control command that preheats the injection pipeline, valve, or equipment interface before SF6 gas is injected into the GIS equipment. Its purpose is to ensure that the SF6 gas is maintained at a suitable temperature during injection, avoiding liquefaction, uneven density, or adverse effects on the equipment's insulation performance due to excessively low temperatures. This command can be generated based on real-time environmental parameters (such as ambient temperature and humidity) and SF6 gas characteristic simulation results, calculating the preheating temperature and time required to reach the target injection temperature, and controlling heating devices (such as electric heating belts or hot air guns) accordingly for precise preheating. For example, a lookup table or regression model can be built based on simulation results, with the current ambient temperature and target injection temperature as input, and the required preheating power and duration as output.

[0082] In one embodiment of the present invention, generating optimized control instructions includes: comparing and logically verifying the initial control instructions output by the environmental prediction and decision model with a preset environmental control rule base; outputting the final optimized control instructions based on the verification results; wherein, when communication with the cloud is interrupted, switching to a local control mode based on the environmental control rule base to generate optimized control instructions.

[0083] Specifically, the initial control commands output by the environmental prediction and decision-making model are compared and logically verified with a pre-set environmental control rule base to ensure the validity and safety of the initial control commands generated by the model. The environmental prediction and decision-making model may generate commands based on complex algorithms, but these commands may not fully conform to actual operating procedures, safety thresholds, or specific process requirements. By comparing and logically verifying with the pre-set environmental control rule base, unreasonable, unsafe, or inconsistent commands can be filtered out.

[0084] Meanwhile, when communication with the cloud is interrupted, the system switches to a local control mode based on an environmental control rule base to generate optimized control instructions. This mechanism aims to improve the robustness and continuity of the system, ensuring that the environmental control system can still operate normally in the event of unreliable or interrupted network connectivity, thus avoiding control failure due to communication failure. The local control mode provides an emergency and backup solution.

[0085] In one embodiment of the present invention, a digital twin model of the GIS installation area is simulated in real time based on multimodal real-time data; the simulation results are used to evaluate and correct the optimization and control instructions.

[0086] A digital twin model is a virtual digital model that is synchronized with a physical entity in real time, simulating the environmental conditions and equipment operation of the GIS installation area. Specifically, this digital twin model can be achieved by constructing a three-dimensional virtual environment that includes a geometric model of the GIS equipment, environmental sensor data (such as temperature, humidity, and cleanliness), airflow models, and personnel activity models. This model can receive multimodal real-time data from the actual installation area, such as video image information, personnel location information, local microenvironment monitoring data, installation equipment status information, and environmental parameters, and update its internal state in real time based on this data, thereby accurately reflecting internal dynamic changes.

[0087] In one embodiment of the present invention, environmental data, control command logs, key video clips, and identified process events during the installation process are associated with the unique identifier of the GIS equipment; a traceable chain of quality evidence is generated and stored in the full lifecycle digital archive of the GIS equipment.

[0088] Specifically, environmental data, control command logs, key video clips, and identified process events during the installation process are linked to the unique identifier of the GIS equipment. This aims to ensure that all data related to the installation process of a specific GIS equipment can be accurately traced back to that equipment.

[0089] As attached Figure 1 As shown above, the following is a detailed process for a GIS installation intelligent integrated environmental control method:

[0090] S1. Acquire multimodal real-time data: By deploying various sensors and data acquisition devices at the installation site of the multimodal real-time data GIS (including local micro-environment monitoring data), multi-dimensional information including video images, personnel positioning, environmental parameters, and equipment operating status is collected in real time, and the data is synchronously pushed to the processing unit and digital twin model.

[0091] S2. Identify the installation process stage: Based on video image information, the system automatically determines the specific installation process stage of the equipment through intelligent analysis, and simultaneously synchronizes the process stage information to the digital twin model for scene matching.

[0092] S3. Generate optimized control commands: Input multimodal real-time data and the identified installation process stages into a pre-trained environmental prediction and decision-making model. The model performs comprehensive analysis and prediction, and outputs initial control commands. Verify the initial commands against a preset environmental control rule base to ensure the commands are safe and compliant. Based on a digital twin model driven by multimodal real-time data and mapped proportionally to the physical site in real time, simulate the changes in temperature, humidity, cleanliness, airflow field, and equipment operating status after control. Evaluate, optimize, and correct the commands based on the simulation results to obtain the final executable optimized control commands.

[0093] S4. Perform dynamic environment adjustment: Based on the optimized control instructions corrected by digital twin simulation, drive the corresponding execution modules to perform precise and real-time dynamic adjustments to the installation environment, and send the execution results back to the digital twin model in real time to form a closed loop.

[0094] On the other hand, the present invention also provides an intelligent integrated environmental control system for GIS installation, used to implement the intelligent integrated environmental control method for GIS installation as described in any of the above examples, comprising: a data acquisition module for acquiring multimodal real-time data including video image information, personnel positioning information, and environmental parameters; a processing module for acquiring the installation process stage based on the multimodal real-time data, judging the linkage triggering event, calling the environmental prediction and decision model, combining the environmental control rule base to generate optimized control instructions, and driving the digital twin model to perform simulation to correct the optimized control instructions; and an execution module for dynamically adjusting the installation environment based on the corrected optimized control instructions, including the multimodal real-time data and local micro-environment monitoring data, and executing preset environmental control actions corresponding to the linkage triggering event; wherein, the environmental prediction and decision model is obtained by training based on a historical database, and the digital twin model is a three-dimensional virtual model that is proportionally mapped to the physical installation site and synchronized in real time.

[0095] Specifically, a historical database refers to a collection that stores a large amount of data related to past GIS installation projects. This data includes historical environmental parameters, control records, process stage information, abnormal events during installation, and final installation quality assessment results, providing a foundation for the training and optimization of environmental prediction and decision-making models.

[0096] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can make various modifications and alterations without departing from the spirit and scope of the invention; therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.

Claims

1. A method for intelligent integrated environmental control of GIS installation, characterized in that, include: Acquire multimodal real-time data including video image information, personnel positioning information, and environmental parameters; The installation process stage is obtained based on the aforementioned multimodal real-time data; Based on the video image information and the personnel location information, determine whether a linkage trigger event has occurred; If so, then execute the preset environment control action bound to the aforementioned linkage trigger event; If not, the multimodal real-time data will be combined with the installation process stage input environment prediction and decision model to generate optimized control instructions; The generated optimization control instructions include: The installation process includes a vacuuming and gas injection stage. When the equipment is in the vacuuming and gas injection stage, a feedforward temperature control command is generated based on the vacuum pump operation data and the ambient temperature and humidity data, and a gas injection interface preheating command is generated based on the simulation of the environmental parameters and SF6 gas characteristics. The initial control command output by the environmental prediction and decision-making model is compared and verified with the preset environmental control rule base, and then the final optimized control command is output. If the cloud communication is interrupted, the local rule base mode is switched to generate the optimized control command. Real-time simulation is performed based on a digital twin model that is proportionally mapped to and synchronized with the physical installation site using the multimodal real-time data. The digital twin model includes the three-dimensional geometric structure of the GIS equipment, the personnel location trajectory, and the operating status of the control equipment. It is used to simulate and predict environmental changes and control effects, and to evaluate and correct the optimized control commands based on the simulation results. The installation environment is dynamically adjusted by the control execution module according to the revised optimized control instructions. The linkage triggering events include at least one of the following: personnel entering a high-cleanliness area, dustproof booth opening, and the installation process entering a critical process stage. The installation process also includes: unpacking and cleaning and component placement stage and conductor connection and insertion stage; The step of inputting the multimodal real-time data and the installation process stage into the environmental prediction and decision-making model to generate optimized control instructions also includes: Based on the aforementioned installation process stages, determine whether the equipment is in the unpacking, cleaning, and component placement stage; If the determination is yes, then the unpacking or component movement is identified based on the video image information, and the location of the pollution source is determined in combination with the personnel positioning information; The optimized control command includes at least one of the following: a command to control the air supply system to form a dynamic air curtain at the pollution source location; Command the mobile purification unit to move closer to the pollution source to perform purification.

2. The intelligent integrated environmental control method for GIS installation according to claim 1, characterized in that, The step of inputting the multimodal real-time data and the installation process stage into the environmental prediction and decision-making model to generate optimized control instructions includes: The multimodal real-time data also includes local microenvironment monitoring data; Based on the aforementioned installation process stages, determine whether the equipment is in the conductor connection and insertion stage; If the determination is yes, then the system controls the directional air supply system to form a clean airflow barrier based on the personnel positioning information, and generates a dynamic dehumidification compensation instruction based on the local microenvironment monitoring data.

3. The intelligent integrated environmental control method for GIS installation according to claim 1, characterized in that, Also includes: Associate environmental data, control command logs, key video clips, and identified process events during the installation process with the unique identifier of the GIS equipment; Generate a traceable chain of quality evidence and store it in the full lifecycle digital archive of the GIS device.

4. A GIS installation intelligent integrated environmental control system, used to implement the GIS installation intelligent integrated environmental control method according to any one of claims 1-3, characterized in that, include: The data acquisition module is used to acquire multimodal real-time data including the video image information, the personnel positioning information, and the environmental parameters; The processing module is used to acquire the installation process stage based on the multimodal real-time data, determine the linkage triggering event, call the environmental prediction and decision model, combine the environmental control rule base to generate optimization control instructions, and drive the digital twin model to perform simulation to correct the optimization control instructions. The execution module is used to dynamically adjust the installation environment of the multimodal real-time data, including local microenvironment monitoring data, according to the modified optimization and control instructions, and to execute the preset environmental control action corresponding to the linkage trigger event. The environmental prediction and decision-making model is trained based on a historical database, and the digital twin model is a three-dimensional virtual model that is proportionally mapped to the physical installation site and synchronized in real time.

Images