Building intelligent energy hierarchical management system and method based on Internet of Things

Abstract

The invention relates to the technical field of intelligent buildings, in particular to a building intelligent energy hierarchical management system and method based on the Internet of Things, a building intelligent energy hierarchical architecture based on the Internet of Things comprises a sensing layer, an access layer, a network layer, an edge computing layer, a cloud platform layer and a service display layer, and the method comprises the following steps: step 1, in the sensing layer, the access layer, the network layer, the edge computing layer, the cloud platform layer and the service display layer; the method comprises the following steps: arranging a sensor for monitoring the operation state of key energy equipment of a building, collecting current, voltage, vibration, temperature and pressure operation data, adding a UTC timestamp and equipment metadata, and performing local preprocessing and encryption reporting; 2, in an access layer, protocol adaptation and semantic mapping are carried out on device data from a heterogeneous protocol; according to the invention, the problems of data island, processing delay, low control efficiency, inaccurate energy consumption prediction and the like in the prior art are effectively solved, hierarchical management, intelligent decision and energy-saving optimization of building energy are realized, and the stability, expandability and economical efficiency of an energy system are improved.

Description

Technical Field

[0001] This invention relates to an intelligent building energy hierarchical management system and method based on the Internet of Things (IoT), and particularly to an intelligent building energy hierarchical management system and method based on the Internet of Things (IoT), belonging to the field of intelligent building technology. Background Technology

[0002] Existing building energy management systems typically use a single control platform as the core to centrally monitor and dispatch energy facilities such as electricity, heating, air conditioning, and water systems within the building;

[0003] However, such systems often suffer from the following shortcomings: First, the system structure is usually relatively centralized, with limited interconnectivity and data interaction capabilities between sensing layer devices, making it difficult to achieve unified access and standardized data management for different types of energy devices. Second, the system's data processing relies heavily on a central server, resulting in high network transmission pressure and response latency, making it difficult to guarantee real-time performance in multi-building, multi-node environments. Furthermore, the lack of intelligent energy stratification analysis and prediction mechanisms prevents the implementation of hierarchical calculation, dynamic scheduling, and strategy optimization of energy consumption data, leading to low energy utilization, serious energy waste, and poor system scalability and security. Therefore, there is an urgent need to improve the IoT-based intelligent building energy stratification management system and methods to address the aforementioned problems. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent building energy tiered management system and method based on the Internet of Things (IoT) to address the following issues: First, the system structure is usually relatively centralized, with limited interconnectivity and data interaction capabilities between sensing layer devices, making it difficult to achieve unified access and standardized data management for different types of energy devices. Second, the system's data processing relies heavily on a central server, resulting in high network transmission pressure and response latency, making it difficult to guarantee real-time performance in multi-building, multi-node environments. Furthermore, the lack of an intelligent energy tiered analysis and prediction mechanism prevents the implementation of tiered calculation, dynamic scheduling, and strategy optimization of energy consumption data, leading to low energy utilization, serious energy waste, and poor system scalability and security.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] The IoT-based building intelligent energy hierarchical management system and method, the IoT-based building intelligent energy hierarchical architecture includes a sensing layer, access layer, network layer, edge computing layer, cloud platform layer and service presentation layer, and the method includes the following steps:

[0007] Step 1: At the sensing layer, sensors are deployed to monitor the operating status of key energy equipment in the building, collecting operating data such as current, voltage, vibration, temperature, and pressure. After adding UTC timestamps and equipment metadata, the data is preprocessed and encrypted locally before being reported.

[0008] Step 2: At the access layer, perform protocol adaptation and semantic mapping on device data from heterogeneous protocols to achieve time synchronization, edge caching, retransmission after disconnection, and event priority reporting;

[0009] Step S3: At the edge computing layer, perform sliding window feature extraction and online anomaly detection on the real-time sensor stream. If the anomaly score exceeds the confidence threshold, execute closed-loop protection action and report the anomaly event.

[0010] Step S4: At the cloud platform layer, the degradation model and the remaining useful life prediction (RUL) model are trained based on historical operating data and maintenance records, and the models are distributed to edge nodes in a standardized format for real-time inference.

[0011] Step S5: At the cloud platform layer, construct a maintenance scheduling optimization model based on the remaining useful life prediction (RUL), equipment importance, and maintenance resource constraints, and solve for the optimal maintenance plan;

[0012] Step S6: Generate and dispatch maintenance work orders through the business presentation layer, and collect execution feedback and post-verification operation data.

[0013] Preferably, the following steps are also included:

[0014] Step S7: Evaluate model performance online, detect data drift, and perform model retraining, rollback, or manual review processes;

[0015] Step S8: The method achieves security management between layers through TLS encrypted communication and role-based access control.

[0016] Preferably, in step 3, the anomaly detection model uses a sliding window mechanism to extract features from the time-series data stream, and its window function is defined as:

[0017]

[0018] In the formula, w is the window length. Let be the eigenvector at time t, and the function be... Calculate the statistical characteristics of mean, variance, spectral energy, and kurtosis;

[0019] The formula for calculating the anomaly score is:

[0020]

[0021] when Anomalies are detected in time. This is the confidence threshold.

[0022] Preferably, in step 4, the remaining lifetime prediction model employs an improved temporal deep hybrid network, combining a long short-term memory network and a temporal convolutional network for joint modeling. Its core recursive formula is:

[0023]

[0024]

[0025] In the formula, For activation function, For trainable parameters, the function Output the remaining lifetime value for the linear mapping layer.

[0026] Preferably, in step 5, the maintenance scheduling optimization problem is defined as:

[0027]

[0028] In the formula, For maintenance costs, Losses due to unplanned downtime For energy loss, The weighting coefficients are used for the constraints, which include the availability of maintenance personnel, equipment priority, and time window limits.

[0029] Preferably, the model performance monitoring employs a drift detection algorithm, which calculates distribution differences based on Kullback-Leibler divergence:

[0030]

[0031] when When this happens, it triggers a model retraining or rollback operation.

[0032] Preferably, the anomaly detection and inference process of edge nodes is deployed using a lightweight model distillation method, with the teacher model trained in the cloud and the student model inferred on the edge side, thereby reducing communication bandwidth usage and latency.

[0033] Preferably, MQTT or AMQP message middleware is used at the access layer to realize asynchronous transmission of data streams, and device identifiers, timestamps and semantic tags are defined through a unified namespace.

[0034] Preferably, security measures include device-side certificate authentication, trusted platform module verification, operation auditing, and role-based access control to ensure data integrity and access security.

[0035] Preferably, a complete energy consumption optimization closed loop is constructed based on a multi-layer collaborative distributed architecture, from perception to decision execution;

[0036] The system as a whole includes: a sensing layer, which is used to collect the operating status data of energy equipment in the building. The sensing layer includes current sensors, vibration sensors, temperature sensors, pressure sensors and oil temperature sensors. The sensors are used to collect operating data at a preset sampling frequency and add timestamps and equipment identification information. After local preprocessing and encryption, the data is output to the access layer.

[0037] The access layer is used to perform protocol adaptation and semantic mapping on device data from different manufacturers or with different communication protocols, and to achieve time synchronization, edge caching, retransmission after disconnection and event priority reporting. It also transmits the standardized time-series data to the edge computing layer or cloud platform layer via message middleware.

[0038] The network layer is used to establish secure communication links between the perception layer, access layer, edge computing layer and cloud platform layer, and to achieve secure data transmission through TLS encryption protocol and device-side certificate authentication.

[0039] The edge computing layer is used to receive the standardized time-series data and run a lightweight online anomaly detection model. It performs feature extraction and anomaly score calculation on the device operation characteristics within the sliding time window. When the anomaly score exceeds the preset confidence threshold, it executes local protection actions and reports the anomaly event.

[0040] The cloud platform layer is used to store building operation data and maintenance records, and to train degradation models and remaining life prediction models based on historical data. It performs inference and life prediction on the received real-time feature data to generate failure probability and equipment health status indicators.

[0041] The optimization scheduling module, located at the cloud platform layer, is used to build a maintenance scheduling optimization model based on the RUL prediction results, equipment importance indicators, and maintenance resource constraints, and generate maintenance plans with the goal of minimizing maintenance costs, unplanned downtime losses, and energy consumption losses.

[0042] The business presentation layer is used to automatically generate maintenance plans and dispatch them as maintenance work orders to the operation and maintenance system, and to receive maintenance execution feedback and post-operation data for online model evaluation and retraining.

[0043] The security management module runs through all layers of the system and is used to perform identity authentication, access control and audit trail. The security management module includes a trusted platform module, a role-based access control and operation log recording unit.

[0044] The layers are collectively configured through standardized data interfaces to form a closed-loop control system for energy data acquisition, anomaly detection, fault prediction, optimized scheduling, and intelligent operation and maintenance.

[0045] The present invention has at least the following beneficial effects:

[0046] This invention constructs a multi-layered distributed architecture comprising a perception layer, an access layer, a network layer, an edge computing layer, a cloud platform layer, and a service presentation layer. At the perception layer, the system achieves unified data acquisition and protocol adaptation across multiple energy terminals; at the edge computing layer, it enables real-time response and traffic distribution of data through local analysis and load prediction; at the cloud platform layer, it performs energy consumption modeling, optimized scheduling, and policy learning; and at the service presentation layer, it achieves cross-building and cross-system energy status visualization and policy-linked control. This effectively solves problems such as data silos, processing delays, low control efficiency, and inaccurate energy consumption prediction in existing technologies, realizing hierarchical management, intelligent decision-making, and energy-saving optimization of building energy, and improving the stability, scalability, and economy of the energy system. Attached Figure Description

[0047] Figure 1 This is a flowchart of the IoT-based intelligent building energy hierarchical management method of the present invention;

[0048] Figure 2 This is a system block diagram of the IoT-based intelligent building energy hierarchical management system of the present invention. Detailed Implementation

[0049] The following will describe in detail the implementation of this application with reference to the accompanying drawings and embodiments, so that the implementation process of how this application uses technical means to solve technical problems and achieve technical effects can be fully understood and implemented accordingly.

[0050] The IoT-based intelligent building energy hierarchical management system and method are based on a multi-layer collaborative distributed architecture to build a complete energy consumption optimization closed loop from perception to decision execution.

[0051] The system as a whole includes: a sensing layer, which is used to collect the operating status data of energy equipment in the building. The sensing layer includes current sensors, vibration sensors, temperature sensors, pressure sensors and oil temperature sensors, etc. The sensors are used to collect operating data at a preset sampling frequency and add timestamps and equipment identification information. After local preprocessing and encryption, the data is output to the access layer.

[0052] The access layer is used to perform protocol adaptation and semantic mapping on device data from different manufacturers or with different communication protocols, and to achieve time synchronization, edge caching, retransmission after disconnection and event priority reporting. It also transmits the standardized time-series data to the edge computing layer or cloud platform layer via message middleware.

[0053] The network layer is used to establish secure communication links between the perception layer, access layer, edge computing layer and cloud platform layer, and to achieve secure data transmission through TLS encryption protocol and device-side certificate authentication.

[0054] The edge computing layer is used to receive the standardized time-series data and run a lightweight online anomaly detection model. It performs feature extraction and anomaly score calculation on the device operation characteristics within the sliding time window. When the anomaly score exceeds the preset confidence threshold, it executes local protection actions and reports the anomaly event.

[0055] The cloud platform layer is used to store building operation data and maintenance records, and to train degradation models and remaining useful life (RUL) prediction models based on historical data. It performs inference and life prediction on the received real-time feature data to generate failure probability and equipment health status indicators.

[0056] The optimization scheduling module, located at the cloud platform layer, is used to build a maintenance scheduling optimization model based on the RUL prediction results, equipment importance indicators, and maintenance resource constraints, and generate maintenance plans with the goal of minimizing maintenance costs, unplanned downtime losses, and energy consumption losses.

[0057] The business presentation layer is used to automatically generate maintenance plans and dispatch them as maintenance work orders to the operation and maintenance system, and to receive maintenance execution feedback and post-operation data for online model evaluation and retraining.

[0058] The security management module runs through all layers of the system and is used to perform identity authentication, access control and audit trail. The security management module includes a Trusted Platform Module (TPM), a Role-Based Access Control (RBAC) module and an operation log recording unit.

[0059] The layers are coordinated through standardized data interfaces to form a closed-loop control system for energy data acquisition, anomaly detection, fault prediction, optimized scheduling, and intelligent operation and maintenance, thereby realizing real-time perception, predictive maintenance, and energy efficiency optimization of building energy operation status.

[0060] Specific process: Step 1: Data acquisition and preprocessing at the sensing layer: Deploy various types of sensors on the main energy-consuming equipment in the building (such as chillers, fans, water pumps, elevators, lighting and air conditioning units), including but not limited to: current sensors and voltage sensors for monitoring electrical power; acceleration sensors and vibration sensors for detecting the mechanical state of rotating equipment; temperature and pressure sensors for sensing the operating parameters of the thermal system; oil temperature and bearing temperature sensors for detecting the aging or lubrication status of the equipment;

[0061] Each sensor uses a preset sampling frequency Data is collected periodically, and the collected data is appended with a UTC timestamp and device metadata (ID, location, model, operating condition, etc.). After local noise filtering and normalization, the data is encrypted and reported.

[0062] To ensure data security and integrity, verification is performed before transmission using the TLS communication encryption protocol and in conjunction with the device-side certificate authentication mechanism.

[0063] Step 2: Access Layer Data Adaptation and Semantic Mapping: Data from devices from different manufacturers may use various industrial protocols (such as Modbus, BACnet, OPC UA, CAN, 485, etc.).

[0064] At this layer, the gateway node implements: Protocol adaptation: parsing heterogeneous protocols into a unified format;

[0065] Semantic mapping: Based on the equipment model, the original point table is mapped to semantic labels (such as "cooler outlet temperature" and "compressor vibration amplitude").

[0066] Time synchronization: Synchronize data timestamps using NTP or GPS;

[0067] Retransmission and caching: Temporarily caching data when the network is interrupted;

[0068] Event priority reporting: Assign higher transmission priority to alarms and critical events;

[0069] The standardized data is uploaded to the edge layer or cloud platform via message middleware (MQTT / AMQP / Kafka);

[0070] Step 3: Anomaly Detection and Protection Response at the Edge Computing Layer: Edge nodes receive data streams from the access layer in real time, and perform feature extraction for each device using a sliding time window, defined as:

[0071]

[0072] In the formula, w is the window length. Let be the eigenvector at time t, and the function be... Calculate the statistical characteristics of mean, variance, spectral energy, and kurtosis;

[0073] The formula for calculating online anomaly scores is as follows:

[0074]

[0075] when When the threshold is reached, the system identifies an anomaly and triggers:

[0076] Closed-loop protection actions (such as shutdown, load reduction, and alarm) are performed on the edge side.

[0077] Abnormal events, abnormal scores, and feature vectors are uploaded to the cloud platform for subsequent analysis.

[0078] This mechanism can respond to sudden failures in milliseconds, avoiding energy waste and cascading damage;

[0079] Step 4: Degradation and Remaining Life (RUL) Prediction at the Cloud Platform Layer: The cloud platform integrates historical operational data, maintenance work orders, and expert-annotated samples from various buildings to train a degradation and remaining life prediction model. This model adopts an improved temporal deep hybrid structure (LSTM + TCN), combining long-term dependencies and local temporal convolutional features.

[0080] Its core recursive formula is:

[0081]

[0082]

[0083] In the formula, For activation function, For trainable parameters, the function Output the remaining lifetime value for the linear mapping layer;

[0084] After the model is trained in the cloud, it is delivered to the edge nodes as a lightweight model file (ONNX / TFLite format) for real-time inference.

[0085] Step 5: Maintenance Scheduling Optimization and Energy Efficiency Decision-Making: Based on the remaining life prediction results (RULtRUL_tRULt) of each device, failure probability, equipment importance indicators, and maintenance resource constraints, the cloud platform constructs a maintenance scheduling optimization model:

[0086]

[0087] In the formula, For maintenance costs, Losses due to unplanned downtime For energy loss, The weighting coefficients are used; the constraints include the availability of maintenance personnel, equipment priority, and time window limits. The system uses a genetic algorithm or reinforcement learning method to solve for the optimal maintenance plan, thereby balancing energy efficiency and maintenance costs.

[0088] Step 6: Maintenance Plan Generation and Work Order Closure: After the platform generates a maintenance plan, the system automatically generates a corresponding work order through the business presentation layer and synchronizes it to the building operation and maintenance system. Operation and maintenance personnel receive the work order, confirm its execution, and send back the following information: actual maintenance results, replacement component information, maintenance time and energy consumption records, and post-operational data. This feedback data is automatically archived by the cloud platform for subsequent model retraining.

[0089] Step 7: Online Model Evaluation and Adaptive Updates

[0090] The platform monitors the model inference results online. If performance degradation or data distribution drift is detected, it uses the Kullback-Leibler divergence as a criterion.

[0091]

[0092] when When this happens, the system will automatically trigger: model retraining, model version rollback,

[0093] Alternatively, a manual review process may be used, and all version information, training data range, and audit logs are recorded in the cloud.

[0094] Step 8: Security Protection and Access Control: To prevent data leakage and unauthorized access, the system adopts the following measures across the entire chain: device-side certificate authentication, transport layer encryption, trusted platform modules for gateways and edge nodes, role-based access control, operation logs, and security auditing mechanisms.

[0095] The above methods enable the integration of energy consumption visualization and equipment health monitoring, improve fault prediction accuracy by more than 30%, reduce sudden downtime, enable intelligent scheduling of operation and maintenance resources, reduce maintenance and energy costs, and provide real-time response at the edge, thereby improving system stability and security. At the same time, the model is continuously optimized to achieve a data-driven energy management closed loop.

[0096] If certain terms are used in the specification and claims to refer to specific components, those skilled in the art will understand that hardware manufacturers may use different names to refer to the same component. This specification and claims do not distinguish components based on differences in name, but rather on differences in function. The term "comprising" as used throughout the specification and claims is an open-ended term and should be interpreted as "comprising but not limited to." "Approximately" means that within an acceptable margin of error, those skilled in the art can solve the technical problem and substantially achieve the technical effect within a certain margin of error.

[0097] It should be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a product or system comprising a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a product or system. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the product or system that includes that element.

[0098] The foregoing description illustrates and describes several preferred embodiments of the present invention. However, as previously stated, it should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the inventive concept described herein through the foregoing teachings or techniques or knowledge in related fields. Any modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

Claims

1. A building intelligent energy hierarchical management system and method based on the Internet of Things (IoT), wherein the building intelligent energy hierarchical management method based on the Internet of Things is characterized by, The IoT-based building intelligent energy hierarchical architecture includes a sensing layer, access layer, network layer, edge computing layer, cloud platform layer, and service presentation layer. The method includes the following steps: Step 1: At the sensing layer, sensors are deployed to monitor the operating status of key energy equipment in the building, collecting operating data such as current, voltage, vibration, temperature, and pressure. After adding UTC timestamps and equipment metadata, the data is preprocessed and encrypted locally before being reported. Step 2: At the access layer, perform protocol adaptation and semantic mapping on device data from heterogeneous protocols to achieve time synchronization, edge caching, retransmission after disconnection, and event priority reporting; Step S3: At the edge computing layer, perform sliding window feature extraction and online anomaly detection on the real-time sensor stream. If the anomaly score exceeds the confidence threshold, execute closed-loop protection action and report the anomaly event. Step S4: At the cloud platform layer, the degradation model and the remaining useful life prediction (RUL) model are trained based on historical operating data and maintenance records, and the models are distributed to edge nodes in a standardized format for real-time inference. Step S5: At the cloud platform layer, construct a maintenance scheduling optimization model based on the remaining useful life prediction (RUL), equipment importance, and maintenance resource constraints, and solve for the optimal maintenance plan; Step S6: Generate and dispatch maintenance work orders through the business presentation layer, and collect execution feedback and post-verification operation data.

2. The building intelligent energy hierarchical management system and method based on the Internet of Things as described in claim 1, characterized in that: It also includes the following steps: Step S7: Evaluate model performance online, detect data drift, and perform model retraining, rollback, or manual review processes; Step S8: The method achieves security management between layers through TLS encrypted communication and role-based access control.

3. The building intelligent energy hierarchical management system and method based on the Internet of Things as described in claim 1, characterized in that: In step 3, the anomaly detection model uses a sliding window mechanism to extract features from the time-series data stream. The window function is defined as follows: ； In the formula, w is the window length. Let be the eigenvector at time t, and the function be... Calculate the statistical characteristics of mean, variance, spectral energy, and kurtosis; The formula for calculating the anomaly score is: ； when Anomalies are detected in time. This is the confidence threshold.

4. The building intelligent energy hierarchical management system and method based on the Internet of Things as described in claim 1, characterized in that: In step 4, the remaining lifetime prediction model employs an improved temporal deep hybrid network, combining a long short-term memory network and a temporal convolutional network for joint modeling. Its core recursive formula is: ； ； In the formula, For activation function, For trainable parameters, the function Output the remaining lifetime value for the linear mapping layer.

5. The building intelligent energy hierarchical management system and method based on the Internet of Things as described in claim 1, characterized in that: In step 5, the maintenance scheduling optimization problem is defined as: ； In the formula, For maintenance costs, Losses due to unplanned downtime For energy loss, The weighting coefficients are used for the constraints, which include the availability of maintenance personnel, equipment priority, and time window limits.

6. The building intelligent energy hierarchical management system and method based on the Internet of Things as described in claim 1, characterized in that: The model performance monitoring employs a drift detection algorithm, which calculates distribution differences based on Kullback-Leibler divergence. ； when When this happens, it triggers a model retraining or rollback operation.

7. The building intelligent energy hierarchical management system and method based on the Internet of Things as described in claim 1, characterized in that: The anomaly detection and inference process at edge nodes is deployed using a lightweight model distillation method. The teacher model is trained in the cloud, while the student model performs inference at the edge, thereby reducing communication bandwidth usage and latency.

8. The building intelligent energy hierarchical management system and method based on the Internet of Things according to claim 1, characterized in that: The access layer uses MQTT or AMQP message middleware to achieve asynchronous transmission of data streams, and defines device identifiers, timestamps and semantic tags through a unified namespace.

9. The building intelligent energy hierarchical management system and method based on the Internet of Things as described in claim 1, characterized in that: Security measures include device-side certificate authentication, trusted platform module verification, operation auditing, and role-based access control to ensure data integrity and access security.

10. The building intelligent energy hierarchical management system and method based on the Internet of Things as described in any one of claims 1-9, characterized in that: Based on a multi-layered collaborative distributed architecture, a complete energy consumption optimization closed loop is constructed, from perception to decision execution; The system as a whole includes: a sensing layer, which is used to collect the operating status data of energy equipment in the building. The sensing layer includes current sensors, vibration sensors, temperature sensors, pressure sensors and oil temperature sensors. The sensors are used to collect operating data at a preset sampling frequency and add timestamps and equipment identification information. After local preprocessing and encryption, the data is output to the access layer. The access layer is used to perform protocol adaptation and semantic mapping on device data from different manufacturers or with different communication protocols, and to achieve time synchronization, edge caching, retransmission after disconnection and event priority reporting. It also transmits the standardized time-series data to the edge computing layer or cloud platform layer via message middleware. The network layer is used to establish secure communication links between the perception layer, access layer, edge computing layer and cloud platform layer, and to achieve secure data transmission through TLS encryption protocol and device-side certificate authentication. The edge computing layer is used to receive the standardized time-series data and run a lightweight online anomaly detection model. It performs feature extraction and anomaly score calculation on the device operation characteristics within the sliding time window. When the anomaly score exceeds the preset confidence threshold, it executes local protection actions and reports the anomaly event. The cloud platform layer is used to store building operation data and maintenance records, and to train degradation models and remaining life prediction models based on historical data. It performs inference and life prediction on the received real-time feature data to generate failure probability and equipment health status indicators. The optimization scheduling module, located at the cloud platform layer, is used to build a maintenance scheduling optimization model based on the RUL prediction results, equipment importance indicators, and maintenance resource constraints, and generate maintenance plans with the goal of minimizing maintenance costs, unplanned downtime losses, and energy consumption losses. The business presentation layer is used to automatically generate maintenance plans and dispatch them as maintenance work orders to the operation and maintenance system, and to receive maintenance execution feedback and post-operation data for online model evaluation and retraining. The security management module runs through all layers of the system and is used to perform identity authentication, access control and audit trail. The security management module includes a trusted platform module, a role-based access control and operation log recording unit. The layers are collectively configured through standardized data interfaces to form a closed-loop control system for energy data acquisition, anomaly detection, fault prediction, optimized scheduling, and intelligent operation and maintenance.

Images