Modular prefabricated cabin substation intelligent comprehensive management system with multi-source data integration and collaborative management function

Abstract

The application relates to the technical field of intelligent management of transformer substations, in particular to a modular prefabricated cabin transformer substation intelligent comprehensive management system with multi-source data integration and collaborative management functions.The system comprises a multi-source data acquisition unit, an edge intelligent processing unit is used for locally preprocessing, feature extraction and abnormal early warning of collected transformer substation multi-source heterogeneous data, and through a lightweight AI algorithm and a mixed communication protocol carried by an edge computing terminal module, data noise reduction, abnormal identification and equipment health degree evaluation are realized; and a data integration and collaboration unit.Through an adaptive sampling strategy and a multi-modal feature fusion mechanism of the multi-source data acquisition unit, multidimensional data such as power parameters, dynamic environment, fire safety and security are integrated into a monitoring system, feature layer correlation analysis is realized by relying on Jousselme distance and Dempster synthesis rules, the evaluation limitation of traditional schemes with single equipment and single-dimensional data is broken through, and a more comprehensive state basis is provided for equipment health degree judgment.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent management of substations, and more specifically, to an intelligent integrated management system for modular prefabricated cabin substations with multi-source data integration and collaborative management functions. Background Art

[0002] With the development of smart grid technology, substation management systems are evolving from single-function monitoring to multi-source data fusion and collaborative control. Although existing substation management systems have achieved the collection and preliminary processing of data such as power parameters and environmental conditions, there are still technical bottlenecks in aspects such as real-time integration of multi-source heterogeneous data, cross-system linkage control, and equipment operation environment adaptability. For example, traditional systems often have fixed data sampling strategies and insufficient feature fusion accuracy, resulting in a lag in equipment anomaly warnings, and lack a modular environmental regulation mechanism, making it difficult to meet the reliable operation requirements under complex working conditions.

[0003] For example, Chinese Patent CN202411102635.2 discloses an intelligent management system and method for substation automated operation and maintenance, which includes obtaining the real-time operation parameters and oil fluid state parameters of a transformer through an operation and maintenance data collection module, obtaining a real-time operation and maintenance evaluation value through an operation and maintenance processing module, fitting a curve of the real-time operation and maintenance evaluation value changing with time through an operation and maintenance analysis module, obtaining the extreme value of the curve and using the time point corresponding to each extreme value as a marking time point, the time span between two adjacent marking time points is a detection period, obtaining the area enclosed by the curve of the real-time operation and maintenance evaluation value changing with time and the predicted curve of the operation and maintenance evaluation value changing with time within each detection period, obtaining whether the area is above or below the predicted curve of the operation and maintenance evaluation value changing with time, and analyzing to obtain the operation and maintenance indication coefficient of the transformer; an operation and maintenance execution module determines whether to perform predictive maintenance on the current transformer according to the operation and maintenance indication coefficient. Also, Chinese Patent CN202210976369.0 discloses an intelligent management system and method for substation safety monitoring, which includes a monitoring and collection module, a workstation host, an auxiliary module, and a switch. The monitoring and collection module is set on substation equipment and is used to collect information of the substation equipment. The auxiliary module is connected to the substation equipment and is used to perform intelligent auxiliary monitoring and management on the substation equipment. The substation equipment, the monitoring and collection module, and the auxiliary module are connected to the workstation host through the switch. The intelligent management system and method for substation safety monitoring provided by the present invention can achieve intelligent monitoring of substations, can achieve comprehensive control of the auxiliary module, can achieve monitoring of the switch, avoid waste of human resources, and improve safety.

[0004] While the aforementioned technical solutions each possess their own design advantages, they also suffer from the following technical shortcomings: Firstly, limitations in data dimensionality and fusion: Chinese patent CN202411102635.2 focuses on a single transformer device, collecting only operating parameters and oil status data, without covering environmental factors, fire signals, and cross-device correlation data, making it difficult to construct a multi-dimensional health profile of the equipment group; furthermore, it relies on historical data of a single device through curve fitting and area calculation for maintenance instructions, failing to integrate external data such as environment and operating conditions, resulting in a single evaluation dimension. Although Chinese patent CN202210976369.0 deploys multi-module monitoring, it only achieves data transmission through switches, failing to construct a semantic fusion model for cross-domain data such as power and environment (e.g., JSON-LD format conversion, evidence theory fusion), leaving the data still in the initial "collection-transmission" stage, without exploring the value of multi-source data correlation (e.g., the coupling relationship between equipment health and environmental temperature and humidity). Secondly, there is a lack of dynamic response and intelligence: CN202411102635.2 uses a fixed detection period (adjacent marked time points are the detection cycle), and does not dynamically adjust the sampling frequency and analysis cycle according to data fluctuation characteristics (such as sudden changes in electrical parameters and abnormal oil indicators), making it easy to miss abnormal data; moreover, the operation and maintenance instructions are only based on historical curve comparisons, lacking intelligent early warning under real-time operating conditions (such as an immediate response mechanism for sudden overload). CN202210976369.0 relies on the workstation host to centrally process data, and does not deploy lightweight AI algorithms (such as real-time feature extraction and anomaly recognition) on the edge side, resulting in high data transmission latency and failing to meet the millisecond-level response requirements of substations; moreover, the system does not support cross-system linkage (such as automatic triggering of fire alarms and fire extinguishing devices), and the linkage response relies on manual intervention, resulting in a low level of intelligence. Thirdly, there are shortcomings in environmental adaptability and modularity: CN202411102635.2 only focuses on equipment operation and maintenance assessment, without addressing the control design of the equipment operating environment (such as the impact of high temperature and high humidity on transformer oil performance). Under complex operating conditions, the equipment is prone to assessment deviations due to environmental interference, and there is no protective mechanism to ensure stable operation. CN202210976369.0's monitoring module is directly deployed on the substation equipment without adopting a modular cabin integration design. It lacks environmentally adaptable structures such as gradient insulation and intelligent moisture protection. Long-term exposure to harsh environments can accelerate aging and reduce system reliability. Therefore, we propose a modular prefabricated intelligent integrated management system for substations with multi-source data integration and collaborative management functions. Summary of the Invention

[0005] The purpose of this invention is to provide a modular prefabricated substation intelligent integrated management system with multi-source data integration and collaborative management functions, so as to solve the problems of data dimension and fusion limitations, lack of dynamic response and intelligence, and insufficient environmental adaptability and modularity mentioned in the background art.

[0006] To address the aforementioned technical problems, the present invention aims to provide a modular prefabricated substation intelligent integrated management system with multi-source data integration and collaborative management functions, comprising:

[0007] A multi-source data acquisition unit, which realizes real-time acquisition of multi-source heterogeneous data in the substation through a distributed sensor array and an adaptive sampling frequency adjustment module;

[0008] The substation's multi-source heterogeneous data includes electrical parameters such as voltage, current, and power; environmental parameters such as temperature, humidity, and water immersion; fire protection parameters such as the status of fire extinguishing devices and fire detection signals; and video security data such as equipment thermal imaging and personnel behavior.

[0009] The edge intelligent processing unit is used to perform local preprocessing, feature extraction and anomaly warning on the collected multi-source heterogeneous data of the substation, and realize data noise reduction, anomaly identification and equipment health assessment through the lightweight AI algorithm and hybrid communication protocol carried by the edge computing terminal module.

[0010] The data integration and collaboration unit is used to build a multi-source data fusion platform to realize the standardized fusion and linkage control of cross-domain data from power, environmental, fire protection and security systems, and to build a comprehensive management platform based on the HarmonyOS microkernel operating system to establish one-click linkage response for preset scenarios.

[0011] Modular cabin support unit, which provides a physical carrier for the integrated installation of system equipment, ensures the stability of the equipment operating environment, and integrates fire prevention, heat insulation and moisture-proof environmental control mechanisms through a steel structure frame that separates the high-pressure equipment cabin and the low-pressure equipment cabin;

[0012] The human-machine interaction control unit is used to realize remote system monitoring, intelligent decision-making and emergency command. Through the dual-end collaborative architecture of local workstation and mobile terminal, it supports three-dimensional visualization of equipment status, one-click linkage control and AI-assisted decision-making.

[0013] As a further improvement to this technical solution, the multi-source data acquisition unit includes a power parameter acquisition module, an environmental monitoring module, a fire protection data acquisition module, and a video security acquisition module, wherein:

[0014] The power parameter acquisition module is used to acquire the current, voltage, power factor and harmonic components of the substation bus and feeder in real time. It adopts high-precision current sensors, voltage sensors and signal conditioning circuits to support synchronous sampling and digital conversion of three-phase electrical parameters.

[0015] The environmental monitoring module is installed in the cabin and key areas of the equipment. It realizes real-time acquisition of environmental parameters through a distributed sensor network. It includes a water immersion sensor for real-time monitoring of water leakage status and outputting switch signals, a temperature and humidity sensor for acquiring ambient temperature and relative humidity parameters, and a gas concentration sensor for monitoring gas leakage and concentration changes.

[0016] The fire data acquisition module integrates infrared smoke detectors, ultraviolet flame detectors, and fire extinguishing device status sensors to collect fire data.

[0017] The video security acquisition module uses dual-spectrum imaging technology to simultaneously acquire thermal imaging data and visible light video of the equipment, supporting monitoring of equipment temperature field distribution and analysis of personnel behavior.

[0018] As a further improvement to this technical solution, the adaptive sampling frequency adjustment module includes a data feature recognition submodule, a dynamic sampling decision submodule, and a data quality assurance submodule, wherein:

[0019] The data feature recognition submodule calculates the fluctuation characteristics of multi-source heterogeneous data in the substation in real time based on the sliding window algorithm, and identifies the steady-state, gradual change and sudden change states of the data.

[0020] The dynamic sampling decision submodule dynamically adjusts the sampling frequency according to the data type priority rules and the output of the data feature recognition submodule, and triggers a high-frequency sampling mode when the data changes abruptly, and switches to a low-frequency sampling mode when the data is in a steady state.

[0021] The data quality assurance submodule employs a data integrity verification algorithm to ensure data quality consistency at different sampling frequencies, and supports data breakpoint resume and abnormal data marking.

[0022] As a further improvement to this technical solution, the edge computing terminal module includes a data preprocessing submodule, a feature extraction submodule, and an early warning decision submodule, wherein:

[0023] The data preprocessing submodule uses wavelet transform algorithm to filter and denoise, identify outliers and normalize the format of multi-source heterogeneous data from the substation.

[0024] The feature extraction submodule is used to extract equipment operating status features from the multi-source heterogeneous data of the substation after preprocessing by the data preprocessing submodule, and fuse them to form an equipment health assessment feature vector.

[0025] The early warning decision submodule is used to generate graded early warning information for device status and dynamically adjust the data upload frequency according to the early warning level.

[0026] Furthermore, the early warning decision-making submodule is equipped with a three-level early warning mechanism:

[0027] Blue alert (abnormal status): Data upload frequency remains normal (10s level);

[0028] Yellow alert (fault risk): Data upload frequency increased to 1 second level;

[0029] Red Alert (Emergency Shutdown): Data upload frequency is increased to 10ms level, and the hardware-accelerated AI inference engine is triggered.

[0030] As a further improvement to this technical solution, the feature extraction of equipment operating status features and the fusion of health assessment feature vectors in the feature extraction submodule include the following steps:

[0031] S212.1 Feature Separation and Extraction:

[0032] Frequency domain features of the voltage and current waveform sequences are calculated using short-time Fourier transform to extract the fundamental amplitude. Phase difference of each harmonic Total harmonic distortion (THD); simultaneously calculate time-domain characteristics (effective voltage value) RMS value of current Peak factor ), forming an electric characteristic vector ;

[0033] Construct a 5-minute sliding window for the temperature and humidity series and calculate the average temperature. Average humidity Temperature standard deviation Humidity standard deviation Temperature change rate and humidity change rate Extract the percentage of time exceeding limits from gas concentration data to form an environmental feature vector. ;

[0034] The frequency of change of the switching signals of infrared smoke detectors and ultraviolet flame detectors per unit time is calculated, and the pressure deviation rate (the percentage difference between the current pressure and the rated pressure) is extracted from the pressure sensor data of the fire extinguishing device to form a fire protection feature vector. ;

[0035] The Gaussian pyramid algorithm was used to calculate the temperature characteristics of the ROI area of ​​the equipment surface thermal imaging and extract the mean temperature. Temperature gradient The proportion of hotspot areas (the percentage of pixels whose temperature exceeds the average temperature by 5°C); target features are calculated using the YOLOv5 target detection algorithm on visible light video to identify personnel behavior and equipment appearance defects, outputting a target category confidence vector C, thus forming a video feature vector. ;

[0036] S212.2, Multimodal Feature Fusion:

[0037] Electricity feature vector Environmental feature vectors Firefighting feature vector Video feature vectors Mapped to the corresponding evidence respectively , , , Furthermore, each piece of evidence contains a confidence distribution of the device status;

[0038] The difference between pieces of evidence is calculated using the Jousselme distance formula:

[0039] ;

[0040] in, The degree of difference between the two pieces of evidence; , The evidence body contains the confidence distribution of the equipment status; Representation matrix transpose; A distance matrix describing the differences in device states;

[0041] when When the value is greater than 0.6, weights are dynamically allocated based on sensor calibration accuracy and sampling frequency:

[0042] ;

[0043] in, For the first Data fusion weights; For the first indivual No. The calibration accuracy of each data point; , The first The, the The data collection frequency;

[0044] Based on the Dempster synthesis rules, the modified evidence is fused to generate fused evidence. Define the confidence level of the equipment's status in four dimensions: electrical performance, thermal stability, environmental adaptability, and security integrity.

[0045] S212.3, Generation and Dimensional Division of Health Feature Vectors:

[0046] Principal component analysis is performed on the fused high-dimensional feature vectors. Principal components whose cumulative contribution rate meets the preset threshold are retained to generate equipment health feature vectors. The preset threshold is 85%-98%. The weights of each dimension are allocated based on the analytic hierarchy process, and the sum of the weights of each dimension is 100%. The weight allocation is dynamically adjusted according to the type of substation equipment.

[0047] Furthermore, this embodiment allocates weights for each dimension based on the analytic hierarchy process (AHP), wherein:

[0048] Electrical performance accounts for 30%-50%.

[0049] Thermal stability accounts for 25%-35%.

[0050] The environmental adaptability dimension accounts for 15%-25%.

[0051] The security integrity dimension accounts for 5%-15%.

[0052] Furthermore, the weight allocation is dynamically adjusted based on the type of substation equipment, specifically including:

[0053] For GIS equipment, increase the weighting of the electrical performance dimension;

[0054] For outdoor equipment, increase the weighting of the environmental adaptability dimension;

[0055] For unattended substations, the weighting of the security integrity dimension is increased. The adjusted weights for each dimension satisfy the aforementioned interval constraints.

[0056] Furthermore, to adapt to the monitoring needs of different substation equipment types, this embodiment dynamically adjusts the weights of the four dimensions using the analytic hierarchy process (AHP), with the specific rules as follows:

[0057] Basic weighting range (applicable to conventional substations):

[0058] Electrical performance dimension: accounting for 30%–50% (focusing on the stability analysis of power parameters);

[0059] Thermal stability dimension: accounting for 25%–35% (related to equipment temperature field and aging risk);

[0060] Environmental adaptability dimension: accounting for 15%–25% (covering the effects of temperature, humidity, and gas concentration);

[0061] Security integrity dimension: accounting for 5%–15% (combined with intrusion and defect detection of video security);

[0062] The sum of the weights for each dimension is 100%.

[0063] Device type adaptation rules (dynamically corrected for specific scenarios):

[0064] GIS equipment: Because insulation performance has a more significant impact on operational safety, the weight of the electrical performance dimension is increased by 5%–10% (e.g., from 40% to 45%–50%).

[0065] Outdoor equipment is more susceptible to environmental factors (such as rain, snow, and dust), so the weight of the environmental adaptability dimension is increased by 5%–10% (e.g., from 20% to 25%–30%).

[0066] Unmanned substations: The security risks are higher, and the weight of the security integrity dimension is increased by 3%–8% (for example, from 10% to 13%–18%, while the proportion of other dimensions needs to be reduced at the same time to maintain the total of 100%).

[0067] Adjustment logic:

[0068] Weight adjustments must adhere to the principle of "prioritizing core dimensions": electrical performance should account for no less than 30%, and thermal stability should account for no less than 25%, to avoid weakening core features due to over-adaptation to special scenarios.

[0069] As a further improvement to this technical solution, the data integration and collaboration unit includes a cross-domain data standardization module, a fusion middleware service bus module, and a linkage control execution module, wherein:

[0070] The cross-domain data standardization module is used to construct a data mapping dictionary for power, environment, fire protection, and security, converting multi-source heterogeneous data from substations into a unified JSON-LD semantic format; and based on the metadata registration mechanism, assigning a unique identifier to each data point to achieve cross-system data traceability.

[0071] The integrated middleware service bus module is based on the distributed soft bus technology of the HarmonyOS microkernel operating system, establishes a low-latency data channel, supports concurrent access of tens of thousands of devices, and implements a data caching and partitioning strategy to store different types of data differently.

[0072] Furthermore, the data caching partitioning strategy specifically includes:

[0073] Real-time alarm data is stored in an in-memory database;

[0074] Historical monitoring data is stored in a time-series database.

[0075] The linkage control execution module converts monitored events into multi-system collaborative control instructions according to preset event-action mapping rules, resolves conflicts of multi-system linkage requests triggered simultaneously through a priority decision mechanism, and triggers dynamic adjustment when the execution result deviates from a preset threshold based on closed-loop feedback of the execution status.

[0076] As a further improvement to this technical solution, the data integration and collaboration unit also includes a security isolation module. The security isolation module is used to convert the communication protocols of the multi-source data acquisition units (such as Modbus, IEC61850, MQTT) into a unified internal communication protocol. The internal communication protocol is based on the JSON-LD semantic format and supports cross-system data semantic interoperability.

[0077] The security isolation module implements fine-grained access control policies based on a role-based access control model. These access control policies include at least data type level, operation level, and time window permission control.

[0078] The security isolation module uses the national cryptographic SM4 algorithm to encrypt data transmitted across systems and supports automatic key updates.

[0079] As a further improvement to this technical solution, the modular cabin support unit includes a structural support module and an environmental control module, wherein:

[0080] The structural support module adopts a steel structure frame that separates the high-pressure equipment compartment and the low-pressure equipment compartment, and a fireproof isolation wall is set between the high-pressure equipment compartment and the low-pressure equipment compartment.

[0081] The environmental control module is integrated within the steel structure frame and is used to achieve fire prevention, heat insulation, and moisture-proof functions.

[0082] As a further improvement to this technical solution, the environmental control module includes a gradient heat insulation submodule, an intelligent moisture-proof submodule, an adaptive ventilation submodule, and a redundant fireproof submodule, wherein:

[0083] The gradient insulation submodule forms a gradient insulation barrier with increasing thermal resistance by alternately layering aerogel felt and vacuum insulation panels on the inner wall of the steel structure frame, along the side walls and top of the high-pressure chamber and the low-pressure chamber.

[0084] The intelligent moisture-proof sub-module integrates a graphene electrothermal film and a biomimetic microchannel drainage system, which extends along the slope of the cabin and converges at the corner drain outlet.

[0085] The adaptive ventilation submodule sets up independent ventilation loops in the high-pressure equipment compartment and the low-pressure equipment compartment, and configures shape memory alloy driven variable cross-section regulating valves in the high-pressure equipment compartment and the low-pressure equipment compartment.

[0086] The redundant fireproof submodule has a fireproof valve driven by a magnetic latching relay installed in the middle of the partition wall between the high-pressure chamber and the low-pressure chamber. The mechanical locking mechanism of the fireproof valve is arranged facing the low-pressure chamber side. When the power is off, the valve plate is closed by attraction through a permanent magnet.

[0087] As a further improvement to this technical solution, the human-computer interaction control unit includes a dual-end collaboration module, a three-dimensional visualization module, an intelligent control module, and an emergency command module, wherein:

[0088] The dual-end collaboration module is used to build a dual-end data interaction architecture between the local workstation and the mobile terminal. Through the distributed data processing capabilities of the local workstation and the multi-network adaptive switching communication link of the mobile terminal, the dual-end real-time synchronization of device status data is achieved.

[0089] The three-dimensional visualization module is used to construct a three-dimensional scene of the system, and to map the current and voltage data of the power parameter acquisition module through building information modeling technology, and to realize graded early warning of abnormal equipment status using a red-yellow-blue color coding mechanism;

[0090] The intelligent control module is used to perform device linkage control. It includes a strategy library of at least 50 preset linkage scenarios. Through a three-stage verification mechanism of operation permission verification, device status pre-detection and execution consequence simulation, it realizes one-click control of device anomaly, multi-component linkage and alarm push.

[0091] The emergency command module is used to dispatch emergency resources and trigger warnings. It stores no fewer than 20 emergency plans and can quickly dispatch emergency resources (such as fire-fighting equipment and maintenance personnel) based on the spatial location of the equipment. When receiving an emergency command, it simultaneously triggers an audible and visual alarm on the equipment and a red flashing warning on the mobile terminal operation interface.

[0092] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0093] 1. This invention incorporates multi-dimensional data such as power parameters, environmental dynamics, and fire safety into the monitoring system through an adaptive sampling strategy and multi-modal feature fusion mechanism of multi-source data acquisition units. It relies on Jousselme distance and Dempster synthesis rules to achieve feature layer correlation analysis, breaking through the limitations of traditional solutions that rely on single equipment and single-dimensional data for evaluation, and providing a more comprehensive status basis for equipment health assessment.

[0094] 2. This invention utilizes the lightweight AI deployment and hierarchical early warning design of the edge intelligent processing unit to complete data noise reduction, feature extraction and real-time anomaly identification at the edge. This reduces the data transmission load in the cloud while compressing the device anomaly response latency to the millisecond level, avoiding the congestion risk of centralized computing and solving the problems of slow dynamic response and weak edge computing power in traditional solutions.

[0095] 3. This invention is based on the cross-domain data integration and collaboration unit of HarmonyOS microkernel. Through distributed soft bus and JSON-LD semantic interoperability technology, it breaks down the data barriers of power, environmental and fire protection systems, realizes one-click response of "equipment anomaly - multi-system linkage", replaces the manual decentralized operation mode, and improves the systematicness and timeliness of substation emergency response.

[0096] 4. This invention constructs an "environmental buffer layer" for equipment operation through the modular intelligent cabin's gradient heat insulation, intelligent moisture-proof and adaptive ventilation design. It can regulate parameters such as temperature, humidity and gas concentration inside the cabin in real time, reduce the corrosive interference of extreme environments on the equipment, solve the problem of insufficient environmental adaptability of traditional solutions, and ensure the long-term stable operation of the system. Attached Figure Description

[0097] Figure 1 This is a system framework diagram of the present invention;

[0098] The meanings of the labels in the diagram are as follows:

[0099] 100. Multi-source data acquisition unit; 110. Power parameter acquisition module; 120. Environmental monitoring module; 130. Fire protection data acquisition module; 140. Video security acquisition module; 150. Adaptive sampling frequency adjustment module; 151. Data feature recognition submodule; 152. Dynamic sampling decision submodule; 153. Data quality assurance submodule;

[0100] 200. Edge intelligent processing unit; 210. Edge computing terminal module; 211. Data preprocessing submodule; 212. Feature extraction submodule; 213. Early warning decision submodule;

[0101] 300. Data integration and collaboration unit; 310. Cross-domain data standardization module; 320. Fusion middleware service bus module; 330. Linkage control and execution module; 340. Security isolation module;

[0102] 400. Modular cabin support unit; 410. Structural support module; 420. Environmental control module; 421. Gradient thermal insulation submodule; 422. Intelligent moisture-proof submodule; 423. Adaptive ventilation submodule; 424. Redundant fireproof submodule;

[0103] 500. Human-computer interaction control unit; 510. Dual-end collaboration module; 520. 3D visualization module; 530. Intelligent control module; 540. Emergency command module. Detailed Implementation

[0104] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0105] like Figure 1 As shown, this embodiment provides a modular prefabricated substation intelligent integrated management system with multi-source data integration and collaborative management functions, including:

[0106] The multi-source data acquisition unit 100, through a distributed sensor array and an adaptive sampling frequency adjustment module 150, realizes real-time acquisition of multi-source heterogeneous data in the substation.

[0107] In this embodiment, the multi-source data acquisition unit 100 includes a power parameter acquisition module 110, an environmental monitoring module 120, a fire protection data acquisition module 130, and a video security acquisition module 140, wherein:

[0108] The power parameter acquisition module 110 is used to acquire the current, voltage, power factor, and harmonic components of the substation bus and feeders in real time. It adopts high-precision current sensors, voltage sensors, and signal conditioning circuits to support synchronous sampling and digital conversion of three-phase electrical parameters. The sensors adopt a three-phase synchronous sampling architecture, with a matching filter circuit to suppress high-frequency interference, and generate digital signals after analog-to-digital conversion. At the same time, the power parameter acquisition module 110 supports real-time caching and standardized encapsulation of electrical parameters, and transmits data to the edge intelligent processing unit 200 through an industrial Ethernet interface to ensure the synchronization and reliability of the three-phase electrical parameters.

[0109] The environmental monitoring module 120 is installed in key areas of the cabin and equipment. It uses a distributed sensor network to collect environmental parameters in real time. This includes a water immersion sensor for real-time monitoring of water leakage and outputting switch signals, a temperature and humidity sensor for collecting ambient temperature and relative humidity parameters, and a gas concentration sensor for monitoring gas leaks and concentration changes. The water immersion sensor is installed in the cable trench and at the bottom of the equipment cabin to monitor water leakage and output switch signals in real time. The temperature and humidity sensors are arranged at preset intervals inside the equipment cabin to collect ambient temperature and humidity parameters. The gas concentration sensor is located near equipment areas where leaks may occur to monitor changes in gas concentration. All sensors form a star topology through a wireless communication network, directly aggregating environmental data to the signal processing circuit of the environmental monitoring module 120, enabling real-time perception of environmental conditions and early warning of anomalies.

[0110] The fire data acquisition module 130 integrates infrared smoke detectors, ultraviolet flame detectors, and fire extinguishing device status sensors to collect fire data. It adopts a combination of ceiling-mounted and directional deployment to cover key areas such as the main transformer area and switchgear area. When the smoke detector and flame detector are triggered simultaneously, the fire data acquisition module 130 immediately generates a fire alarm signal and simultaneously collects the operating status data of the fire extinguishing device.

[0111] The video security acquisition module 140 adopts dual-spectrum imaging technology to simultaneously acquire thermal imaging data and visible light video of the equipment, supporting monitoring of equipment temperature field distribution and analysis of personnel behavior.

[0112] In this embodiment, the adaptive sampling frequency adjustment module 150 includes a data feature recognition submodule 151, a dynamic sampling decision submodule 152, and a data quality assurance submodule 153, wherein:

[0113] The data feature recognition submodule 151 calculates the fluctuation characteristics of multi-source heterogeneous data in substations in real time based on the sliding window algorithm, and identifies the steady-state, gradual change and sudden change states of the data.

[0114] As a further explanation of this embodiment, the data feature recognition submodule 151 in this embodiment is based on the principle of the sliding window algorithm to perform fluctuation feature analysis on the real-time data stream. The sliding window algorithm segments the data by setting a time window, and calculates the fluctuation amplitude and rate of change of the data within the window in real time to identify the state of the data—when the data fluctuation amplitude is small and the rate of change is low, it is determined to be a steady state; when the fluctuation amplitude or rate of change gradually increases, it is determined to be a gradual change state; and when the data shows significant jumps or abnormal fluctuations, it is determined to be a sudden change state. This process provides key state basis for dynamically adjusting the sampling frequency.

[0115] The dynamic sampling decision submodule 152 dynamically adjusts the sampling frequency according to the data type priority rules and the output of the data feature recognition submodule 151, and triggers a high-frequency sampling mode when the data changes abruptly, and switches to a low-frequency sampling mode when the data is in a steady state.

[0116] As a further explanation of this embodiment, the dynamic sampling decision submodule 152 establishes data type priority rules and implements a differentiated sampling strategy based on the status identifier output by the data feature identification submodule 151. For critical data such as power fault parameters, a high priority is set, and a high-frequency sampling mode is immediately triggered when a sudden change in state is detected to ensure complete capture of abnormal data. For non-urgent data such as environmental data during steady-state operation, a low-frequency sampling mode is switched to reduce transmission load. This mechanism of dynamically adjusting the sampling frequency based on data importance and status effectively balances data acquisition accuracy and system resource consumption.

[0117] The data quality assurance submodule 153 adopts a data integrity verification algorithm to ensure the consistency of data quality under different sampling frequencies, and supports data breakpoint resume and abnormal data marking.

[0118] As a further explanation of this embodiment, the data quality assurance submodule 153 in this embodiment adopts a data integrity verification algorithm to perform consistency verification on data at different sampling frequencies. This algorithm ensures the accuracy of data during transmission by embedding verification information in the data frame, and also supports a data interruption resumption mechanism—when the network is interrupted, the untransmitted data is temporarily buffered and automatically retransmitted after the network is restored, and abnormal data is marked, thereby ensuring the reliability and traceability of data throughout the entire sampling period.

[0119] The edge intelligent processing unit 200 is used to perform local preprocessing, feature extraction and anomaly warning on the multi-source heterogeneous data collected from the substation, and realize data noise reduction, anomaly identification and equipment health assessment through the lightweight AI algorithm and hybrid communication protocol carried by the edge computing terminal module 210.

[0120] In this embodiment, the edge computing terminal module 210 includes a data preprocessing submodule 211, a feature extraction submodule 212, and an early warning decision submodule 213, wherein:

[0121] Data preprocessing submodule 211 uses wavelet transform algorithm to perform filtering, noise reduction, outlier identification and format normalization on multi-source heterogeneous data from substations;

[0122] As a further explanation of this embodiment, the data preprocessing submodule 211 in this embodiment uses a wavelet transform algorithm to perform filtering, noise reduction, outlier identification, and format normalization on the multi-source heterogeneous data from the substation. Wavelet transform, as a time-domain-frequency domain analysis method, decomposes a signal into high-frequency detail components and low-frequency approximation components using wavelet functions of different scales, thereby achieving multi-resolution analysis of non-stationary signals. In this embodiment, the algorithm improves data quality for noisy data such as substation current waveforms, voltage waveforms, and environmental parameters through the following process:

[0123] First, by utilizing the multi-scale decomposition characteristics of wavelet transform, the original data is decomposed into components of different frequency bands, effectively separating high-frequency interference components such as power frequency interference and switching operation noise from low-frequency effective signals characterizing the equipment's operating status. Second, by setting thresholds to filter high-frequency coefficients, abnormal fluctuations corresponding to noise are suppressed, while retaining effective components reflecting the true characteristics of the data. Finally, the denoised data is normalized to map the outputs of different types of sensors (such as power parameter sensors and temperature and humidity sensors) to a standard value range, eliminating the influence of dimensional differences on subsequent feature extraction.

[0124] By applying this algorithm to the data preprocessing submodule, the reliability and consistency of multi-source data in substations can be improved, providing high-quality input data for the equipment status feature analysis of the feature extraction submodule 212. It is especially suitable for processing sudden change signals during short-circuit faults and gradual change data during long-term operation, ensuring the accuracy of subsequent health assessment and anomaly warning.

[0125] The feature extraction submodule 212 is used to extract equipment operating status features from the substation multi-source heterogeneous data preprocessed by the data preprocessing submodule 211, and fuse them to form an equipment health assessment feature vector;

[0126] The early warning decision submodule 213 is used to generate graded early warning information for equipment status and dynamically adjust the data upload frequency according to the early warning level.

[0127] Furthermore, the early warning decision-making submodule 213 is equipped with a three-level early warning mechanism:

[0128] Blue alert (abnormal status): Data upload frequency remains normal (10s level);

[0129] Yellow alert (fault risk): Data upload frequency increased to 1 second level;

[0130] Red Alert (Emergency Shutdown): Data upload frequency is increased to 10ms level, and the hardware-accelerated AI inference engine is triggered.

[0131] In this embodiment, the feature extraction of device operating status and the fusion of health assessment feature vectors in the feature extraction submodule 212 include the following steps:

[0132] S212.1 Feature Separation and Extraction:

[0133] Short-time Fourier transform was used to calculate the frequency domain characteristics of the voltage and current waveform sequences, and the fundamental amplitude was extracted. Phase difference of each harmonic Total harmonic distortion (THD); simultaneously calculate time-domain characteristics (effective voltage value) RMS value of current Peak factor ), forming an electric characteristic vector ;

[0134] Construct a 5-minute sliding window for the temperature and humidity series and calculate the mean temperature. Average humidity Temperature standard deviation Humidity standard deviation Temperature change rate and humidity change rate Extract the percentage of time exceeding limits from gas concentration data to form an environmental feature vector. ;

[0135] The frequency of change of the switching signals of infrared smoke detectors and ultraviolet flame detectors per unit time is calculated, and the pressure deviation rate (the percentage difference between the current pressure and the rated pressure) is extracted from the pressure sensor data of the fire extinguishing device to form a fire protection feature vector. ;

[0136] The Gaussian pyramid algorithm was used to calculate the temperature characteristics of the ROI area of ​​the equipment surface thermal imaging and extract the average temperature. Temperature gradient The proportion of hotspot areas (the percentage of pixels whose temperature exceeds the average temperature by 5°C); target features are calculated using the YOLOv5 target detection algorithm on visible light video to identify personnel behavior and equipment appearance defects, outputting a target category confidence vector C, thus forming a video feature vector. ;

[0137] S212.2, Multimodal Feature Fusion:

[0138] Electricity feature vector Environmental feature vectors Firefighting feature vector Video feature vectors Mapped to the corresponding evidence respectively , , , Furthermore, each piece of evidence contains a confidence distribution of the device status;

[0139] The difference between pieces of evidence is calculated using the Jousselme distance formula:

[0140] ;

[0141] in, The degree of difference between the two pieces of evidence; , The evidence body contains the confidence distribution of the equipment status; Representation matrix Transpose of; A distance matrix describing the differences in device states;

[0142] when When the value is greater than 0.6, weights are dynamically allocated based on sensor calibration accuracy and sampling frequency:

[0143] ;

[0144] in, For the first Data fusion weights; For the first indivual No. The calibration accuracy of each data point; , The first The, the The data collection frequency;

[0145] Based on the Dempster synthesis rules, the modified evidence is fused to generate fused evidence. Define the confidence level of the equipment's status in four dimensions: electrical performance, thermal stability, environmental adaptability, and security integrity.

[0146] S212.3, Generation and Dimensional Division of Health Feature Vectors:

[0147] Principal component analysis is performed on the fused high-dimensional feature vectors. Principal components whose cumulative contribution rate meets the preset threshold are retained to generate equipment health feature vectors. The preset threshold is 85%-98%. The weights of each dimension are allocated based on the analytic hierarchy process, and the sum of the weights of each dimension is 100%. The weight allocation is dynamically adjusted according to the type of substation equipment.

[0148] In this embodiment, the weight allocation is dynamically adjusted according to the type of substation equipment, specifically including:

[0149] For GIS equipment, increase the weighting of the electrical performance dimension;

[0150] For outdoor equipment, increase the weighting of the environmental adaptability dimension;

[0151] For unattended substations, the weighting of the security integrity dimension is increased. The adjusted weights for each dimension satisfy the aforementioned interval constraints.

[0152] In this embodiment, to adapt to the monitoring needs of different substation equipment types, the weights of the four dimensions are dynamically adjusted using the analytic hierarchy process (AHP). The specific rules are as follows:

[0153] Basic weighting range (applicable to conventional substations):

[0154] Electrical performance dimension: accounting for 30%–50% (focusing on the stability analysis of power parameters);

[0155] Thermal stability dimension: accounting for 25%–35% (related to equipment temperature field and aging risk);

[0156] Environmental adaptability dimension: accounting for 15%–25% (covering the effects of temperature, humidity, and gas concentration);

[0157] Security integrity dimension: accounting for 5%–15% (combined with intrusion and defect detection of video security);

[0158] The sum of the weights for each dimension is 100%.

[0159] Device type adaptation rules (dynamically corrected for specific scenarios):

[0160] GIS equipment: Because insulation performance has a more significant impact on operational safety, the weight of the electrical performance dimension is increased by 5%–10% (e.g., from 40% to 45%–50%).

[0161] Outdoor equipment is more susceptible to environmental factors (such as rain, snow, and dust), so the weight of the environmental adaptability dimension is increased by 5%–10% (e.g., from 20% to 25%–30%).

[0162] Unmanned substations: The security risks are higher, and the weight of the security integrity dimension is increased by 3%–8% (for example, from 10% to 13%–18%, while the proportion of other dimensions needs to be reduced at the same time to maintain the total of 100%).

[0163] Adjustment logic:

[0164] Weight adjustments must adhere to the principle of "prioritizing core dimensions": electrical performance should account for no less than 30%, and thermal stability should account for no less than 25%, to avoid weakening core features due to over-adaptation to special scenarios.

[0165] It should be added that the evidence in this embodiment The method for quantifying the confidence distribution of multi-source data is divided into offline training and real-time computation stages, in which:

[0166] Offline training phase: Collect historical data of normal device operation (time span ≥ 30 days, sampling frequency ≥ 100Hz), and extract time-domain features (mean, variance, kurtosis, covariance) from the current sensor. Wei, denoted as ), extract frequency domain features (dominant frequency, bandwidth, m = dimensions in total, denoted as) from the temperature sensor. The initial evidence set is generated using the K-means clustering algorithm (cluster number k=3, corresponding to "normal, abnormal, and faulty" states) and stored as prior knowledge on the edge device.

[0167] Real-time computing phase:

[0168] The feature vectors are dynamically updated using a sliding time window: the window length is set to 10 minutes (with a step size of 5 seconds to balance computational accuracy and computational power consumption), and the current and temperature feature vectors of the current window are combined into a real-time evidence body. This is used for subsequent distance calculations.

[0169] The data integration and collaboration unit 300 is used to build a multi-source data fusion platform to realize the standardized fusion and linkage control of cross-domain data from power, environmental, fire protection and security systems. It also builds a comprehensive management platform based on the HarmonyOS microkernel operating system to establish one-click linkage response for preset scenarios.

[0170] In this embodiment, the data integration and collaboration unit 300 includes a cross-domain data standardization module 310, a converged middleware service bus module 320, and a linkage control execution module 330, wherein:

[0171] The cross-domain data standardization module 310 is used to construct a data mapping dictionary for power, environment, fire protection and security, converting multi-source heterogeneous data from substations into a unified JSON-LD semantic format; and based on the metadata registration mechanism, assigning a unique identifier to each data point to achieve cross-system data traceability;

[0172] The integrated middleware service bus module 320 is based on the distributed soft bus technology of the HarmonyOS microkernel operating system to establish a low-latency data channel, supporting concurrent access of tens of thousands of devices; and implements a data caching and partitioning strategy to store different types of data differently.

[0173] As a further explanation of this embodiment, the integrated middleware service bus module 320 in this embodiment is based on the distributed soft bus technology of the HarmonyOS microkernel operating system to establish a low-latency data channel and support concurrent access of tens of thousands of devices. This service bus adopts a publish-subscribe model to manage data transmission, classifying data such as power parameters and fire alarms into different topics. It automatically identifies access devices and establishes communication links through HarmonyOS's SDP service discovery protocol. The data caching partitioning strategy implements differentiated storage based on data timeliness: real-time data (within 10 minutes) is stored in an in-memory database, recent data (10 minutes to 24 hours) is archived to a distributed database, and historical data (>24 hours) is migrated to object storage. The cache partitioning period is dynamically adjusted through scheduled tasks to optimize storage efficiency and access speed.

[0174] The linkage control execution module 330 converts monitored events into multi-system collaborative control instructions according to preset event-action mapping rules, resolves conflicts of multi-system linkage requests triggered simultaneously through a priority decision mechanism, and triggers dynamic adjustment when the execution result deviates from the preset threshold based on the closed-loop feedback of the execution status.

[0175] As a further explanation of this embodiment, the linkage control execution module 330 in this embodiment implements one-click linkage response for preset scenarios based on a state machine model. For example, taking a fire alarm scenario, when a fire alarm signal is received from the fire alarm data, the module executes the following process: first, it triggers the delayed start of the gas extinguishing device (delayed for 30 seconds to ensure personnel evacuation), and simultaneously cuts off non-fire-fighting power and turns on emergency lighting. The linkage logic adopts a priority control mechanism, where fire linkage (highest priority) can interrupt other low-priority linkage tasks to ensure the timeliness of emergency response. The linkage control execution module 330 controls external devices through various methods such as dry contact output, RS485 interface, and network protocols (such as HTTP, Modbus), supporting a hybrid linkage mode of hard-wiring and soft communication, improving system compatibility and reliability.

[0176] In this embodiment, the data integration and collaboration unit 300 further includes a security isolation module 340. The security isolation module 340 is used to convert the communication protocols (such as Modbus, IEC61850, MQTT) of the multi-source data acquisition unit 100 into a unified internal communication protocol. The internal communication protocol is based on the JSON-LD semantic format and supports cross-system data semantic interoperability.

[0177] The security isolation module 340 implements fine-grained access control policies based on a role-based access control model. The access control policies include at least data type level, operation level, and time window permission control.

[0178] The security isolation module 340 uses the national cryptographic SM4 algorithm to encrypt data transmitted across systems and supports automatic key updates.

[0179] As a further explanation of this embodiment, the security isolation module 340 in this embodiment converts the Modbus, IEC61850, MQTT, and other protocol data from the multi-source data acquisition unit 100 into an internal communication protocol based on the JSON-LD semantic format. The conversion process is as follows:

[0180] First, the protocol type of the input data (such as the function code 0x03 of the Modbus frame) is identified by the protocol parser, and the data fields (such as the current value corresponding to register address 0x0001) are extracted.

[0181] Then, based on the predefined protocol-semantic mapping table, the fields are mapped to JSON-LD entities;

[0182] Finally, a security header (including source address, timestamp, and encryption identifier) ​​is added to the data to form a unified internal protocol frame;

[0183] In addition, the security isolation module 340 supports dynamically loading protocol parsing plugins. By adding conversion rules for new protocols through the configuration tool, semantic interoperability of heterogeneous systems can be ensured.

[0184] As a further explanation of this embodiment, the security isolation module 340 in this embodiment adopts a role-based access control model, implementing three-dimensional permission control based on data type level, operation level, and time window. The access control policy is implemented in the following way:

[0185] Role definition: Preset roles include administrator, maintenance personnel, and auditor. Each role corresponds to a set of permissions (e.g., administrators can read and write all power parameters, while maintenance personnel can only read environmental data).

[0186] Permission rules: Permissions are defined using a 5-tuple of {user, role, data type, operation, time window}.

[0187] Access decision: When a data request arrives, the module parses the user identifier, data type, and operation type in the request, matches it with a preset policy, allows or denies access, and records an audit log.

[0188] It should be added that the security isolation module 340 in this embodiment uses the national cryptographic SM4 algorithm to encrypt data transmitted across systems. The specific implementation method is as follows:

[0189] Encryption parameters: A 128-bit key (generated by a hardware random number generator) is used, employing CBC (Cryptographic Block Chaining) mode, and the initialization vector IV is generated by combining the system timestamp and the device ID;

[0190] Key Management: Supports an automatic key update mechanism with a default update cycle of 24 hours, which can be configured to 1-7 days through the management interface; during updates, a key negotiation protocol (based on the ECDH algorithm) is used to securely distribute new keys, ensuring the security of the key transmission process;

[0191] Encryption process: The internal protocol frame in JSON-LD format is first Base64 encoded, then encrypted using the SM4 algorithm, and finally a MAC checksum (based on the HMAC-SM3 algorithm) is added to prevent data tampering.

[0192] Modular cabin support unit 400 is used to provide a physical carrier for the integrated installation of system equipment, ensure the stability of the equipment operating environment, and integrate fire prevention, heat insulation and moisture-proof environmental control mechanisms through the steel structure frame that separates the high-pressure equipment cabin and the low-pressure equipment cabin.

[0193] In this embodiment, the modular cabin support unit 400 includes a structural support module 410 and an environmental control module 420, wherein:

[0194] The structural support module 410 adopts a steel structure frame that separates the high-voltage equipment compartment and the low-voltage equipment compartment, and a fireproof isolation wall is set between the high-voltage equipment compartment and the low-voltage equipment compartment;

[0195] As a further illustration of this embodiment, the structural support module 410 adopts a high-strength steel frame structure, divided into independent high-pressure equipment compartments and low-pressure equipment compartments. The compartment frame adopts a compartmentalized design, with a fireproof isolation wall installed between the high-pressure and low-pressure compartments to ensure the safety of the separate installation of high and low-pressure equipment. The frame structure is customized according to the equipment installation requirements, meeting the equipment load-bearing and installation space requirements, while the fireproof isolation wall enhances the fire resistance of the compartment.

[0196] The environmental control module 420 is integrated into the steel structure frame to achieve fire prevention, heat insulation and moisture protection functions.

[0197] In this embodiment, the environmental control module 420 includes a gradient thermal insulation submodule 421, an intelligent moisture-proof submodule 422, an adaptive ventilation submodule 423, and a redundant fireproof submodule 424, wherein:

[0198] The gradient thermal insulation submodule 421 forms a gradient thermal insulation barrier with increasing thermal resistance by alternately layering aerogel felt and vacuum insulation panels on the inner wall of the steel structure frame, along the side walls and top of the high-pressure chamber and the low-pressure chamber.

[0199] As a further explanation of this embodiment, the gradient insulation submodule 421 in this embodiment is formed by alternately laying insulation materials on the inner wall of the steel structure frame, the side walls of the cabin, and the top. It adopts a layered structure of aerogel felt and vacuum insulation panels, forming a gradient insulation structure with increasing thermal resistance from the outside to the inside of the cabin. This effectively blocks external environmental heat radiation, reduces temperature conduction between the inside and outside of the cabin, and provides a stable temperature environment for equipment operation.

[0200] The intelligent moisture-proof sub-module 422 integrates a graphene electrothermal film and a biomimetic microchannel drainage system. The biomimetic microchannel drainage system extends along the slope of the cabin and converges at the corner drain outlet.

[0201] Understandably, this embodiment integrates a graphene electrothermal film and a biomimetic microchannel drainage system. The graphene electrothermal film is laid on the inner wall of the chamber and is automatically activated when the humidity exceeds the standard, through linkage control with a humidity sensor, to prevent the formation of condensation inside the chamber; the biomimetic microchannels extend along the slope of the chamber and converge at the drain outlet, forming an efficient drainage path to ensure that moisture and condensation inside the chamber are discharged in a timely manner.

[0202] The adaptive ventilation submodule 423 sets up independent ventilation loops in the high-pressure equipment compartment and the low-pressure equipment compartment, and configures shape memory alloy driven variable cross-section regulating valves in the high-pressure equipment compartment and the low-pressure equipment compartment.

[0203] It is understood that this embodiment sets up independent ventilation loops for the high-pressure chamber and the low-pressure chamber, and is equipped with variable cross-section regulating valves driven by shape memory alloys. The ventilation loops automatically adjust the ventilation volume according to changes in the chamber temperature: when the chamber temperature rises, the shape memory alloy drives the regulating valve to increase the ventilation cross-section and improve ventilation and heat dissipation efficiency; when the temperature drops, the cross-section is automatically reduced to maintain a stable chamber environment.

[0204] The redundant fireproof submodule 424 is equipped with a fireproof valve driven by a magnetic latching relay in the middle of the partition wall between the high-pressure compartment and the low-pressure compartment. The mechanical locking mechanism of the fireproof valve is arranged facing the low-pressure compartment side. When the power is cut off, the valve plate is closed by attraction through a permanent magnet.

[0205] It is understandable that a fire damper driven by a magnetic latching relay is installed in the middle of the partition wall between the high-pressure compartment and the low-pressure compartment. The fire damper adopts a mechanical locking mechanism and a permanent magnet adsorption design. When energized, it remains open. When the power is cut off or a fire alarm signal is received, the permanent magnet drives the valve plate to close, and the mechanical locking ensures that the fire damper is sealed, preventing the spread of flames and smoke between compartments and improving the fire safety of the compartment.

[0206] The Human-Machine Interaction Control Unit 500 is used to realize remote system monitoring, intelligent decision-making and emergency command. Through the dual-end collaborative architecture of local workstation and mobile terminal, it supports three-dimensional visualization of equipment status, one-click linkage control and AI-assisted decision-making.

[0207] In this embodiment, the human-computer interaction control unit 500 includes a dual-end collaboration module 510, a three-dimensional visualization module 520, an intelligent control module 530, and an emergency command module 540, wherein:

[0208] The dual-end collaboration module 510 is used to build a dual-end data interaction architecture between the local workstation and the mobile terminal. Through the distributed data processing capabilities of the local workstation and the multi-network adaptive switching communication link of the mobile terminal, the dual-end real-time synchronization of device status data is achieved.

[0209] As a further explanation of this embodiment, a message queue mechanism (such as the MQTT protocol) can be used to manage data transmission. The local workstation acts as a data hub, preprocessing and compressing the collected multi-source data, and pushing it to the mobile terminal through incremental synchronization to ensure data consistency and real-time performance between the two ends. The mobile terminal supports an offline caching mechanism, temporarily storing operation instructions when the network is interrupted and automatically resending them after the network is restored, ensuring continuous interaction.

[0210] The 3D visualization module 520 is used to construct the 3D scene of the system. It maps the current and voltage data of the power parameter acquisition module 110 through building information modeling technology and realizes graded early warning of abnormal equipment status using a red-yellow-blue color coding mechanism.

[0211] As a further explanation of this embodiment, this embodiment uses a red-yellow-blue color coding mechanism to achieve graded early warning of abnormal equipment states: normal states are displayed in blue, abnormal states (such as voltage fluctuations) are highlighted in yellow, and emergency faults (such as short circuits) are indicated by flashing red. The 3D visualization module 520 also supports interactive operation of the 3D scene. Users can view equipment details by dragging and zooming with the mouse, and click on the 3D model to retrieve the corresponding equipment's operating parameters and historical data in real time, forming an intuitive association display of "spatial location - operating status".

[0212] The intelligent control module 530 is used to perform device linkage control. It includes a strategy library of at least 50 preset linkage scenarios. Through a three-stage verification mechanism of operation permission verification, device status pre-detection and execution consequence simulation, it realizes one-click control of device anomaly, multi-component linkage and alarm push.

[0213] As a further explanation of this embodiment, when performing linkage control, this embodiment ensures operational safety through a three-stage verification mechanism, specifically including:

[0214] First, perform operation permission verification (role-based access control model).

[0215] Secondly, check whether the current state of the target device meets the linkage conditions;

[0216] Finally, the consequences of coordinated execution are simulated using a digital twin model.

[0217] For example, taking a transformer overload scenario as an example, the intelligent control module 530 automatically triggers the linkage process of "adjusting the cooling fan - adjusting the load distribution - sending an alarm", and simultaneously displays the linkage steps and expected effects on the operation interface, realizing closed-loop control of "anomaly identification - strategy matching - linkage execution - result feedback".

[0218] The emergency command module 540 is used to dispatch emergency resources and trigger alarms. It stores no less than 20 emergency plans and can quickly dispatch emergency resources (such as fire-fighting equipment, maintenance personnel, etc.) based on the spatial location of the equipment. When receiving emergency instructions, it simultaneously triggers the audible and visual alarm on the equipment and the red flashing warning on the mobile terminal operation interface.

[0219] As a further explanation of this embodiment, when an emergency command is received, the emergency command module 540 simultaneously triggers the device-side audible and visual alarm and the red flashing warning on the mobile terminal operation interface, displaying the emergency steps and precautions in a graphic and textual manner.

[0220] For example, when a fire is detected in the main transformer, the emergency command module 540 automatically retrieves the corresponding fire extinguishing emergency plan, highlights the location of fire-fighting equipment and evacuation routes in the 3D scene, pushes step-by-step instructions such as "start gas extinguishing - cut off power - report to dispatch" to the mobile terminals of operation and maintenance personnel, and tracks the execution status of emergency measures in real time, thereby improving the efficiency of emergency response in substations.

[0221] It should be added that the specific interaction relationships of the multiple units in this embodiment are as follows:

[0222] The multi-source data acquisition unit 100 collects multi-source data such as power parameters and environmental conditions through a distributed sensor array, and transmits it to the edge intelligent processing unit 200 via industrial Ethernet. The edge intelligent processing unit 200 performs preprocessing on the data, such as wavelet transform denoising and feature extraction, to generate equipment health feature vectors and hierarchical early warning information. Then, through the distributed soft bus technology of the HarmonyOS microkernel, the processed data is transmitted to the data integration and collaboration unit 300.

[0223] The data integration and collaboration unit 300 performs JSON-LD semantic standardization conversion on the data output by the edge intelligent processing unit 200, and establishes a cross-domain data channel through the fusion middleware service bus to uniformly store and manage data from power, fire protection, and other systems. The integrated standardized data is pushed to the human-machine interaction control unit 500. The 3D visualization module maps the data to the equipment status in the 3D scene of the substation, such as displaying the abnormality level using red-yellow-blue color coding. The intelligent control module 530 triggers preset linkage strategies based on the data, and the emergency command module 540 dispatches emergency resources according to the data.

[0224] After receiving user operation commands, the human-machine interaction control unit 500 parses and forwards them to the edge intelligent processing unit 200 via the data integration and coordination unit 300, or directly controls the execution device. For example, if a user triggers a "fire linkage" command on their local workstation, after semantic conversion by the data integration and coordination unit 300, the edge intelligent processing unit 200 executes a high-frequency sampling strategy. Simultaneously, the data integration and coordination unit 300 sends environmental control commands, such as activating the fire damper closing mechanism, to the modular cabin support unit 400. The environmental control module 420 of the modular cabin support unit 400 collects data such as cabin temperature and humidity, and fire prevention status in real time, and uploads it to the system via the multi-source data acquisition unit 100, forming a closed-loop control link of "data acquisition-analysis-decision-execution".

[0225] Those skilled in the art will understand that the process of implementing all or part of the steps of the above embodiments can be carried out by hardware or by a program instructing the relevant hardware.

[0226] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A modular prefabricated substation intelligent integrated management system with multi-source data integration and collaborative management functions, characterized in that: include: A multi-source data acquisition unit (100) is provided, which realizes real-time acquisition of multi-source heterogeneous data of the substation through a distributed sensor array and an adaptive sampling frequency adjustment module (150). Edge intelligent processing unit (200) is used to perform local preprocessing, feature extraction and anomaly warning on the collected multi-source heterogeneous data of the substation, and realize data noise reduction, anomaly identification and equipment health assessment through the lightweight AI algorithm and hybrid communication protocol carried by the edge computing terminal module (210); The edge computing terminal module (210) includes a feature extraction submodule (212); The feature extraction submodule (212) is used to extract equipment operating status features from the substation multi-source heterogeneous data preprocessed by the data preprocessing submodule (211), and fuse them to form an equipment health assessment feature vector; The feature extraction submodule (212) for device operating status feature extraction and health assessment feature vector fusion includes the following steps: S212.1 Feature Separation and Extraction: Short-time Fourier transform was used to calculate the frequency domain characteristics of the voltage and current waveform sequences, and the fundamental amplitude was extracted. Phase difference of each harmonic Total harmonic distortion (THD); simultaneously, time-domain characteristics are calculated to form an electric power characteristic vector. ; Construct a 5-minute sliding window for the temperature and humidity series and calculate the mean temperature. Average humidity Temperature standard deviation Humidity standard deviation Temperature change rate and humidity change rate Extract the percentage of time exceeding limits from gas concentration data to form an environmental feature vector. ; The frequency of change of the switching signals of infrared smoke detectors and ultraviolet flame detectors per unit time is calculated, and the pressure deviation rate is extracted from the pressure sensor data of the fire extinguishing device to form a fire protection feature vector. ; The Gaussian pyramid algorithm was used to calculate the temperature characteristics of the ROI area of ​​the equipment surface thermal imaging and extract the average temperature. Temperature gradient The percentage of hotspot areas; target features are calculated using the YOLOv5 target detection algorithm on visible light video to identify personnel behavior and equipment appearance defects, and the target category confidence vector C is output to form the video feature vector. ; S212.2, Multimodal Feature Fusion: Electricity feature vector Environmental feature vectors Firefighting feature vector Video feature vectors Mapped to the corresponding evidence respectively , , , Furthermore, each piece of evidence contains a confidence distribution of the device status; The difference between pieces of evidence is calculated using the Jousselme distance formula: ； in, The degree of difference between the two pieces of evidence; , The evidence body contains the confidence distribution of the equipment status; Represents the transpose of a matrix; A distance matrix describing the differences in device states; The degree of difference between the evidence corresponding to the power feature vector and the video feature vector When the value is greater than 0.6, weights are dynamically allocated based on sensor calibration accuracy and sampling frequency: ； in, For the first Data fusion weights; For the first indivual No. The calibration accuracy of each data point; , The first The, the The data collection frequency; Based on the Dempster synthesis rules, the modified evidence is fused to generate fused evidence. The confidence level of the equipment's status in four dimensions—electrical performance, thermal stability, environmental adaptability, and security integrity—is clearly defined. S212.3, Generation and Dimensional Division of Health Feature Vectors: Principal component analysis is performed on the fused high-dimensional feature vectors, and principal components whose cumulative contribution rate meets the preset threshold are retained to generate equipment health feature vectors; weights for each dimension are assigned based on the analytic hierarchy process, and the sum of the weights for each dimension is 100%; and the weight allocation is dynamically adjusted according to the type of substation equipment. The data integration and collaboration unit (300) is used to build a multi-source data fusion platform to realize the standardized fusion and linkage control of cross-domain data of power, environmental, fire protection and security systems, and to build a comprehensive management platform based on the HarmonyOS microkernel operating system to establish a one-click linkage response for preset scenarios; the data integration and collaboration unit (300) includes a security isolation module (340), which is used to convert the communication protocol of the multi-source data acquisition unit (100) into a unified internal communication protocol, and the internal communication protocol is based on the JSON-LD semantic format to support cross-system data semantic interoperability; The security isolation module (340) implements fine-grained access control policies based on the role-based access control model. The access control policies include at least data type level, operation level, and time window permission control. The security isolation module (340) uses the national cryptographic SM4 algorithm to encrypt data transmitted across systems and supports automatic key updates; Modular cabin support unit (400), the modular cabin support unit (400) is used to provide a physical carrier for the integrated installation of system equipment, and integrates fire prevention, heat insulation and moisture-proof environmental control mechanism through the steel structure frame that separates the high-pressure equipment cabin and the low-pressure equipment cabin; The human-machine interaction control unit (500) is used to realize remote monitoring, intelligent decision-making and emergency command of the system. Through the dual-end collaborative architecture of local workstation and mobile terminal, it supports three-dimensional visualization of equipment status, one-click linkage control and AI-assisted decision-making.

2. The modular prefabricated substation intelligent integrated management system with multi-source data integration and collaborative management functions according to claim 1, characterized in that, The multi-source data acquisition unit (100) includes a power parameter acquisition module (110), an environmental monitoring module (120), a fire protection data acquisition module (130), and a video security acquisition module (140), wherein: The power parameter acquisition module (110) is used to acquire the current, voltage, power factor and harmonic components of the substation bus and feeder in real time. It adopts high-precision current sensor, voltage sensor and signal conditioning circuit to support synchronous sampling and digital conversion of three-phase electrical parameters. The environmental monitoring module (120) realizes real-time acquisition of environmental parameters through a distributed sensor network, including a water immersion sensor, a temperature and humidity sensor and a gas concentration sensor. The fire data acquisition module (130) integrates an infrared smoke detector, an ultraviolet flame detector, and a fire extinguishing device status sensor for collecting fire data. The video security acquisition module (140) adopts dual-spectrum imaging technology to simultaneously acquire thermal imaging data and visible light video of the equipment, supporting equipment temperature field distribution monitoring and personnel behavior analysis.

3. The modular prefabricated substation intelligent integrated management system with multi-source data integration and collaborative management functions according to claim 1, characterized in that, The adaptive sampling frequency adjustment module (150) includes a data feature recognition submodule (151), a dynamic sampling decision submodule (152), and a data quality assurance submodule (153), wherein: The data feature recognition submodule (151) calculates the fluctuation characteristics of multi-source heterogeneous data of the substation in real time based on the sliding window algorithm, and identifies the steady state, gradual change and sudden change states of the data. The dynamic sampling decision submodule (152) dynamically adjusts the sampling frequency according to the data type priority rules and the output of the data feature recognition submodule (151), and triggers a high-frequency sampling mode when the data changes abruptly, and switches to a low-frequency sampling mode when the data is in a steady state. The data quality assurance submodule (153) adopts a data integrity verification algorithm to ensure the consistency of data quality under different sampling frequencies, and supports data breakpoint resume and abnormal data marking.

4. The modular prefabricated substation intelligent integrated management system with multi-source data integration and collaborative management functions according to claim 1, characterized in that, The edge computing terminal module (210) further includes a data preprocessing submodule (211) and an early warning decision submodule (213), wherein: The data preprocessing submodule (211) uses wavelet transform algorithm to filter and denoise, identify outliers and normalize the format of multi-source heterogeneous data from the substation. The early warning decision submodule (213) is used to generate hierarchical early warning information of equipment status and dynamically adjust the data upload frequency according to the early warning level.

5. The modular prefabricated substation intelligent integrated management system with multi-source data integration and collaborative management functions according to claim 1, characterized in that, The data integration and collaboration unit (300) further includes a cross-domain data standardization module (310), a converged middleware service bus module (320), and a linkage control execution module (330), wherein: The cross-domain data standardization module (310) is used to construct a data mapping dictionary for power-environment-fire protection-security, converting multi-source heterogeneous data from substations into a unified JSON-LD semantic format; and based on the metadata registration mechanism, assigning a unique identifier to each data point to achieve cross-system data traceability. The integrated middleware service bus module (320) is based on the distributed soft bus technology of the HarmonyOS microkernel operating system to establish a low-latency data channel, support the concurrent access of tens of thousands of devices, and implement a data caching and partitioning strategy to store different types of data in a differentiated manner. The linkage control execution module (330) converts the monitored events into multi-system collaborative control instructions according to the preset event-action mapping rules, resolves conflicts of multi-system linkage requests triggered simultaneously through a priority decision mechanism, and triggers dynamic adjustment when the execution result deviates from the preset threshold based on the closed-loop feedback of the execution status.

6. The modular prefabricated substation intelligent integrated management system with multi-source data integration and collaborative management functions according to claim 1, characterized in that: The modular cabin support unit (400) includes a structural support module (410) and an environmental control module (420), wherein: The structural support module (410) adopts a steel structure frame that separates the high-pressure equipment compartment and the low-pressure equipment compartment, and a fireproof isolation wall is set between the high-pressure equipment compartment and the low-pressure equipment compartment; The environmental control module (420) is integrated within the steel structure frame and is used to achieve fire prevention, heat insulation and moisture-proof functions.

7. The modular prefabricated substation intelligent integrated management system with multi-source data integration and collaborative management functions according to claim 6, characterized in that, The environmental control module (420) includes a gradient heat insulation submodule (421), an intelligent moisture-proof submodule (422), an adaptive ventilation submodule (423), and a redundant fireproof submodule (424), wherein: The gradient insulation submodule (421) forms a gradient insulation barrier with increasing thermal resistance by alternately layering aerogel felt and vacuum insulation board on the inner wall of the steel frame, along the side walls and top of the high-pressure chamber and the low-pressure chamber. The intelligent moisture-proof sub-module (422) integrates a graphene electrothermal film and a biomimetic microchannel drainage system, wherein the biomimetic microchannel drainage system extends along the slope of the cabin and converges to the corner drain outlet. The adaptive ventilation submodule (423) sets up independent ventilation loops in the high-pressure equipment compartment and the low-pressure equipment compartment, and configures shape memory alloy driven variable cross-section regulating valves in the high-pressure equipment compartment and the low-pressure equipment compartment; The redundant fireproof submodule (424) is equipped with a fireproof valve driven by a magnetic latching relay in the middle of the partition wall between the high-pressure chamber and the low-pressure chamber. The mechanical locking mechanism of the fireproof valve is arranged facing the low-pressure chamber side. When the power is off, the valve plate is closed by the attraction of a permanent magnet.

8. The modular prefabricated substation intelligent integrated management system with multi-source data integration and collaborative management functions according to claim 1, characterized in that, The human-computer interaction control unit (500) includes a dual-end collaboration module (510), a three-dimensional visualization module (520), an intelligent control module (530), and an emergency command module (540), wherein: The dual-end collaboration module (510) is used to construct a dual-end data interaction architecture between the local workstation and the mobile terminal. Through the distributed data processing capability of the local workstation and the multi-network adaptive switching communication link of the mobile terminal, the dual-end real-time synchronization of device status data is realized. The three-dimensional visualization module (520) is used to construct a three-dimensional scene of the system and to map the current and voltage data of the power parameter acquisition module (110) through building information modeling technology, and to realize the graded early warning of abnormal equipment status using a red-yellow-blue color coding mechanism; The intelligent control module (530) is used to perform device linkage control, including a strategy library of at least 50 preset linkage scenarios. Through a three-stage verification mechanism of operation permission verification, device status pre-detection and execution consequence simulation, it realizes one-click control of device abnormality - multi-component linkage - alarm push. The emergency command module (540) is used to dispatch emergency resources and trigger warnings, store no less than 20 emergency plans, and quickly dispatch emergency resources based on the spatial location of the equipment; when receiving emergency instructions, it simultaneously triggers the sound and light alarm on the equipment and the red flashing warning on the mobile terminal operation interface.

Images