Intelligent supervision system of IBMS building intelligent integration platform

Abstract

The invention relates to the technical field of intelligent supervision systems, and particularly discloses an intelligent supervision system of an IBMS building intelligent integration platform, and the system comprises a data fusion analysis module which is used for integrating data of a heterogeneous subsystem and storing real-time data and historical data through employing a distributed storage architecture; the intelligent analysis engine module is used for carrying out deep mining on the building operation data; the dynamic visualization module is used for generating a three-dimensional dynamic view of the operation state of the building; the self-adaptive regulation and control module is used for dynamically adjusting linkage logic among the heterogeneous subsystems according to building use requirements; the comprehensive evaluation module is used for carrying out quantitative evaluation on the overall performance of the building; and the edge computing node is used for locally processing high-frequency sensor data. The data of the heterogeneous subsystem is integrated through the data fusion analysis module, so that the problem of information isolated island is solved; and the intelligent analysis engine module is combined with time sequence analysis and machine learning technologies to realize accurate prediction of abnormal events.

Description

Technical Field

[0001] This invention relates to an intelligent monitoring system, specifically to an intelligent monitoring system for an IBMS building intelligence integration platform, belonging to the technical field of intelligent monitoring systems. Background Technology

[0002] In the field of building intelligence, the IBMS platform, as the core hub integrating subsystems such as building automation, security, fire protection, and energy management, undertakes the important responsibility of centralized monitoring and coordinated control. It achieves unified management of various devices and subsystems within a building through data acquisition, protocol conversion, and information fusion technologies. However, existing IBMS platforms have significant limitations in terms of supervision: large differences in data formats between heterogeneous subsystems lead to widespread information silos; traditional monitoring interfaces lack the capacity to process massive amounts of operational data and lack support for intelligent analysis and visualization; anomaly detection largely relies on preset thresholds, lacking predictive judgment capabilities based on historical data. Furthermore, the platform's monitoring of key indicators such as energy consumption and equipment failure often lags behind actual needs, making real-time diagnosis and optimization difficult. The linkage control logic between subsystems is relatively fixed, limiting its ability to adapt to dynamic building usage needs, thus hindering further improvements in overall energy efficiency and operational efficiency. To address these issues, there is an urgent need to develop an intelligent monitoring system for the IBMS platform with deep data perception, intelligent analysis, and adaptive control capabilities to meet the higher requirements of modern building intelligence development. Summary of the Invention

[0003] To address the problems in existing technologies, this invention provides an intelligent monitoring system for the IBMS building intelligence integration platform.

[0004] The objective of this invention can be achieved through the following technical solutions:

[0005] An intelligent monitoring system of IBMS building intelligence integration platform includes a data fusion and parsing module, which integrates data from heterogeneous subsystems, converts data of different formats into a unified standard format through a protocol adapter, and uses a distributed storage architecture to save real-time and historical data.

[0006] The intelligent analysis engine module is used to deeply mine building operation data, extract equipment operation patterns through time series analysis algorithms, and predict the probability of potential abnormal events by combining machine learning models.

[0007] The dynamic visualization module is used to generate a three-dimensional dynamic view of the building's operating status, and uses graphics rendering technology to present multi-source data intuitively in the form of heat maps and trend curves.

[0008] The adaptive control module is used to dynamically adjust the linkage logic between heterogeneous subsystems according to the building's usage requirements. It generates the optimal control strategy through the rule engine and sends it to the execution unit in real time.

[0009] The comprehensive assessment module is used to quantitatively evaluate the overall performance of a building, calculate energy efficiency scores based on a multi-dimensional indicator system, and generate optimization suggestion reports.

[0010] Edge computing nodes: used to process high-frequency sensor data locally, reduce cloud transmission latency, and ensure that tasks with high real-time requirements can be completed efficiently.

[0011] Optionally, the protocol adapter supports Modbus, BACnet, and OPC UA industrial communication protocols. The adapter uses an FPGA chip for hardware acceleration, and the time for a single data conversion does not exceed 10ms. The distributed storage architecture is built on the Hadoop ecosystem, supports petabyte-level data storage, and the data write latency does not exceed 50ms.

[0012] Optionally, the time series analysis algorithm uses an ARIMA model combined with an LSTM neural network, with a model training cycle of no more than 2 hours and a prediction accuracy of no less than 95%; the machine learning model is built based on the random forest algorithm, and the feature selection process incorporates the SHAP value analysis method.

[0013] Optionally, the graphics rendering technology is based on DirectX 12, supports 4K resolution output, and maintains a stable frame rate of over 60 FPS; the color mapping range of the heatmap can be customized according to user needs, with a minimum resolution of 0.1℃ or 0.1kWh.

[0014] Optionally, the rule engine is implemented using the Drools framework, and the rule base contains no less than 500 predefined rules, supporting user-defined extensions; the control strategy generation process introduces a genetic algorithm optimization unit, and the time for a single strategy generation does not exceed 1 second.

[0015] Optionally, the multi-dimensional indicator system includes energy intensity, equipment health, environmental comfort, and safety risk level; the energy efficiency score calculation formula is: S=w1·E+w2·H+w3·C+w4·R, where S is the energy efficiency score, E is the normalized value of energy intensity, H is the normalized value of equipment health, C is the normalized value of environmental comfort, R is the normalized value of safety risk level, and the weight coefficients w1, w2, w3, and w4 satisfy w1+w2+w3+w4=1.

[0016] Optionally, the edge computing node uses an ARM Cortex-A72 processor with a main frequency of no less than 2.0 GHz and a memory capacity of no less than 8 GB; the node has a built-in lightweight database SQLite, which supports offline data caching and breakpoint resume functions.

[0017] Optionally, it also includes a data verification unit, which verifies the integrity of data transmission using a CRC check algorithm. The checksum length is 32 bits, and the bit error rate does not exceed 10. -6 The unit has a built-in time synchronization mechanism and uses the NTP protocol to ensure that the time error of each node does not exceed 1ms.

[0018] Optionally, anomaly prediction includes the following steps:

[0019] S1. Data preprocessing: Extract historical data for the target time period from the distributed storage architecture and remove noise and missing values;

[0020] S2. Feature Extraction: Extract time series features based on sliding window technology, with a window length of 1 hour and a step size of 10 minutes;

[0021] S3. Model Training: Input the extracted time series features into the ARIMA-LSTM hybrid model for training, with a training set to test set ratio of 7:3;

[0022] S4. Anomaly Detection: The trained hybrid model is used to predict real-time data. When the deviation between the predicted value and the actual value exceeds a set threshold, it is marked as a potential anomaly event.

[0023] S5. Risk Assessment: Calculate the risk level of abnormal events using a random forest model and output early warning information.

[0024] Optionally, the linkage logic optimization process includes the following steps:

[0025] S1. Status Acquisition: Obtain the real-time operating status of each subsystem through edge computing nodes;

[0026] S2. Rule Matching: Input the collected status data into the rule engine and match it with applicable predefined rules;

[0027] S3. Strategy Generation: Based on the matching results, the genetic algorithm optimization unit is invoked to generate an initial control strategy;

[0028] S4. Strategy Optimization: Iteratively optimize the initial strategy, with the objective function being to maximize the energy efficiency score;

[0029] S5. Policy Distribution: The optimized control policy is distributed to the execution unit via the MQTT protocol.

[0030] The beneficial effects of this invention are:

[0031] This invention solves the problem of information silos by integrating data from heterogeneous subsystems through a data fusion and analysis module; the intelligent analysis engine module combines time series analysis and machine learning techniques to achieve accurate prediction of abnormal events; the dynamic visualization module uses graphics rendering technology to intuitively present the building's operational status, improving the user experience; the adaptive control module generates optimal control strategies based on a rule engine and genetic algorithm, enhancing system flexibility; and the comprehensive evaluation module quantifies building performance through a multi-dimensional indicator system, providing a scientific basis for decision-making. The synergistic effect of these multiple technologies effectively improves the intelligence level and overall performance of the IBMS platform, meeting the higher requirements of modern building intelligence development. Attached Figure Description

[0032] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0033] Figure 1 This is a schematic diagram of the overall modular structure of the present invention.

[0034] Figure 2 This is a flowchart of the abnormal event prediction process of the intelligent analysis engine module of the present invention. Detailed Implementation

[0035] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] Please see Figure 1-2 As shown, an intelligent monitoring system of the IBMS building intelligence integration platform has the following modular structure: Figure 1 As shown, it includes a data fusion and parsing module, an intelligent analysis engine module, a dynamic visualization module, an adaptive control module, a comprehensive evaluation module, and edge computing nodes. These modules are interconnected through logical relationships and data flow directions to jointly achieve comprehensive monitoring of the building intelligence integration platform.

[0037] The data fusion and parsing module is the foundation of the entire system, with its core functions accomplished by a protocol adapter and a distributed storage architecture. The protocol adapter supports multiple industrial communication protocols such as Modbus, BACnet, and OPC UA, and utilizes FPGA chips for hardware acceleration, ensuring that a single data conversion takes no more than 10ms. The distributed storage architecture, built on the Hadoop ecosystem, supports petabyte-level data storage with a data write latency controlled within 50ms. The protocol adapter and the distributed storage architecture are connected via a high-speed data bus, converting data streams from heterogeneous subsystems into a unified standard format and storing them in real time. This design solves the information silo problem caused by data format differences in traditional systems, while ensuring high efficiency in data processing.

[0038] The intelligent analysis engine module is connected to the data fusion and parsing module via a data interface, used to extract historical and real-time data from a distributed storage architecture for in-depth analysis. The module's time series analysis algorithm employs an ARIMA model combined with an LSTM neural network. Feature extraction is based on a sliding window technique with a window length of 1 hour and a step size of 10 minutes. The model training cycle is no more than 2 hours, and the prediction accuracy reaches over 95%. Furthermore, the intelligent analysis engine module incorporates a random forest algorithm for anomaly event risk assessment, and uses SHAP value analysis during feature selection to ensure a balance between model interpretability and predictive performance. Figure 2 As shown, the abnormal event prediction process includes five steps: data preprocessing, feature extraction, model training, anomaly detection, and risk assessment. These steps are executed sequentially through internal logic, ultimately outputting early warning information.

[0039] The dynamic visualization module and the intelligent analysis engine module are connected via a data transmission channel to present analysis results graphically. This module utilizes DirectX 12 for graphics rendering, supporting 4K resolution output with a stable frame rate exceeding 60 FPS. The color mapping range of the heatmap can be customized according to user needs, with a minimum resolution of 0.1℃ or 0.1kWh. The dynamic visualization module displays the building's operational status through a 3D dynamic view, presenting multi-source data intuitively in the form of heatmaps, trend curves, and other formats. This design not only enhances the user experience but also facilitates maintenance personnel in quickly grasping the overall operational status of the building.

[0040] The adaptive control module and the dynamic visualization module are connected via a control signal interface to dynamically adjust the linkage logic between subsystems according to building usage requirements. The core components of this module are the rule engine and the genetic algorithm optimization unit. The rule engine is implemented using the Drools framework, with a rule base containing no fewer than 500 predefined rules and supporting user-defined extensions. The genetic algorithm optimization unit is responsible for generating initial control strategies and iteratively optimizing them, with the objective function being to maximize energy efficiency scores; a single strategy generation takes no more than one second. The adaptive control module distributes the optimized control strategy to the execution unit via the MQTT protocol, thereby achieving precise control of building equipment.

[0041] The comprehensive assessment module and the adaptive control module are connected via a data feedback channel to quantitatively evaluate the overall performance of the building. This module calculates an energy efficiency score based on a multi-dimensional indicator system, including energy intensity, equipment health, environmental comfort, and safety risk level. The energy efficiency score calculation formula is: S = w1·E + w2·H + w3·C + w4·R, where S is the energy efficiency score, E is the normalized value of energy intensity, H is the normalized value of equipment health, C is the normalized value of environmental comfort, R is the normalized value of safety risk level, and the weighting coefficients w1, w2, w3, and w4 satisfy w1 + w2 + w3 + w4 = 1. The comprehensive assessment module generates an optimization suggestion report through five steps: data aggregation, indicator calculation, score calculation, problem identification, and suggestion generation, providing a scientific basis for building operation and maintenance decisions.

[0042] Edge computing nodes are distributed across key locations throughout the building, connected to the data fusion and parsing module, intelligent analysis engine module, and adaptive control module via a local area network. Each edge computing node utilizes an ARM Cortex-A72 processor with a clock speed of at least 2.0GHz and at least 8GB of memory. It incorporates the lightweight SQLite database, supporting offline data caching and resume functionality. These edge computing nodes process high-frequency sensor data, reducing cloud transmission latency and ensuring efficient completion of tasks with high real-time requirements. Furthermore, the edge computing nodes verify data transmission integrity using a CRC checksum algorithm with a 32-bit checksum length and an error rate controlled within 10%. -6 Within this range, it incorporates a built-in time synchronization mechanism and uses the NTP protocol to ensure that the time error of each node does not exceed 1ms.

[0043] The collaborative relationships between the above modules are as follows: The data fusion and analysis module first integrates data from heterogeneous subsystems and standardizes it through protocol adapters and a distributed storage architecture; the intelligent analysis engine module extracts data from the distributed storage architecture for in-depth mining and outputs analysis results; the dynamic visualization module presents the analysis results in graphical form; the adaptive control module generates the optimal control strategy based on the analysis results and optimizes and distributes the strategy through a rule engine and a genetic algorithm optimization unit; the comprehensive evaluation module quantitatively evaluates building performance based on a multi-dimensional indicator system and generates an optimization suggestion report; and the edge computing node is responsible for processing high-frequency sensor data and ensuring the real-time and integrity of data transmission. All modules interact and collaborate through data interfaces, control signal interfaces, and a local area network, thus forming a complete intelligent monitoring system.

[0044] In a practical application scenario, suppose a large commercial complex needs to monitor and optimize its energy consumption and equipment operation status in real time. First, the data fusion and parsing module converts data streams from heterogeneous subsystems such as air conditioning, lighting, and security systems into a unified format and stores them in a distributed storage architecture via a protocol adapter. Then, the intelligent analysis engine module extracts historical and real-time data from the distributed storage architecture, performs time series analysis using an ARIMA-LSTM hybrid model, and combines this with a random forest algorithm to predict the probability and risk level of potential abnormal events. The dynamic visualization module presents the analysis results in a 3D dynamic view, allowing maintenance personnel to intuitively understand the building's operational status through heatmaps and trend curves. When an abnormal energy consumption in a certain area of ​​the air conditioning system is detected, the adaptive control module matches applicable rules through a rule engine and calls a genetic algorithm optimization unit to generate the optimal control strategy. Finally, the strategy is distributed to the air conditioning system execution unit via the MQTT protocol to achieve energy-saving control. Simultaneously, the comprehensive evaluation module calculates energy efficiency scores based on a multi-dimensional indicator system and generates an optimization suggestion report, providing a reference for subsequent maintenance decisions. Edge computing nodes are responsible for processing high-frequency sensor data, ensuring that tasks with high real-time requirements can be completed efficiently, and guaranteeing the integrity and consistency of data transmission through CRC check algorithms and time synchronization mechanisms.

[0045] As can be seen from the above implementation methods, this invention solves the problem of information silos by integrating heterogeneous subsystem data through a data fusion and parsing module; the intelligent analysis engine module achieves accurate prediction of abnormal events by combining time series analysis and machine learning techniques; the dynamic visualization module enhances user experience by intuitively presenting the building's operating status through graphics rendering technology; the adaptive control module generates optimal control strategies based on a rule engine and genetic algorithm optimization unit, enhancing system flexibility; and the comprehensive evaluation module provides a scientific basis for decision-making by quantifying building performance through a multi-dimensional indicator system. The synergistic effect of these technologies effectively improves the intelligence level and comprehensive performance of the IBMS platform, meeting the higher requirements of modern building intelligence development.

[0046] To enable those skilled in the art to fully understand and implement this invention, the specific implementation principle of this invention is further explained below in conjunction with a specific application scenario.

[0047] In a large commercial complex, the building contains multiple heterogeneous subsystems, including air conditioning, lighting, and security systems. These subsystems have different operational data formats and require high real-time performance. To achieve comprehensive monitoring and optimized control of the building's overall operational status, the intelligent monitoring system provided by this invention operates according to the following steps:

[0048] First, the data fusion and parsing module receives raw data streams from various subsystems via a protocol adapter. The protocol adapter supports multiple industrial communication protocols such as Modbus, BACnet, and OPC UA, and employs an FPGA chip for hardware acceleration, ensuring that a single data conversion takes no more than 10ms. Subsequently, the protocol adapter transmits the converted, unified-format data to a distributed storage architecture via a high-speed data bus. This distributed storage architecture, built on the Hadoop ecosystem, supports petabyte-level data storage with a data write latency controlled within 50ms. This process solves the information silo problem caused by data format differences in traditional systems, while ensuring high-efficiency data processing.

[0049] Next, the intelligent analysis engine module extracts historical and real-time data from the distributed storage architecture for in-depth mining and analysis. The time series analysis algorithm in this module employs an ARIMA model combined with an LSTM neural network. The feature extraction process utilizes a sliding window technique with a window length of 1 hour and a step size of 10 minutes. The model training cycle does not exceed 2 hours, and the prediction accuracy can reach over 95%. Furthermore, the intelligent analysis engine module also introduces a random forest algorithm for risk assessment of anomalous events, and uses the SHAP value analysis method during feature selection to ensure a balance between model interpretability and predictive performance. Figure 2As shown, the abnormal event prediction process includes five steps: data preprocessing, feature extraction, model training, anomaly detection, and risk assessment. These steps are executed sequentially through internal logic, ultimately outputting early warning information. For example, when abnormal fluctuations occur in the energy consumption data of an air conditioning system, the intelligent analysis engine module can quickly identify potential problems and generate corresponding early warning information.

[0050] Subsequently, the dynamic visualization module receives the analysis results from the intelligent analysis engine module through the data transmission channel and presents them in a graphical form. This module utilizes DirectX 12 for graphics rendering, supporting 4K resolution output with a stable frame rate above 60 FPS. The color mapping range of the heatmap can be customized according to user needs, with a minimum resolution of 0.1℃ or 0.1kWh. The dynamic visualization module displays the building's operational status through a 3D dynamic view, presenting multi-source data intuitively in the form of heatmaps, trend curves, and other formats. Maintenance personnel can quickly grasp the overall operational status of the building through heatmaps and trend curves. For example, when the temperature distribution in a certain area is abnormal, maintenance personnel can intuitively locate the problem area using the heatmap.

[0051] When abnormal energy consumption of the air conditioning system in a certain area is detected, the adaptive control module matches applicable rules through the rule engine and calls the genetic algorithm optimization unit to generate the optimal control strategy. The rule engine is implemented using the Drools framework, with a rule base containing no fewer than 500 predefined rules and supporting user-defined extensions. The genetic algorithm optimization unit is responsible for generating the initial control strategy and iteratively optimizing the strategy, with the objective function being to maximize the energy efficiency score. A single strategy generation takes no more than 1 second. The adaptive control module distributes the optimized control strategy to the air conditioning system execution unit via the MQTT protocol, thereby achieving energy-saving control. For example, when the energy consumption of the air conditioning system in a certain area is too high, the adaptive control module will adjust the air conditioning operation mode according to the current environmental parameters to reduce energy consumption.

[0052] Meanwhile, the comprehensive assessment module quantifies the overall performance of the building based on a multi-dimensional indicator system. This module generates an optimization suggestion report through five steps: data aggregation, indicator calculation, score calculation, problem identification, and suggestion generation. The indicator system includes energy intensity, equipment health, environmental comfort, and safety risk level. The energy efficiency score calculation formula is: S = w1·E + w2·H + w3·C + w4·R, where S is the energy efficiency score, E is the normalized value of energy intensity, H is the normalized value of equipment health, C is the normalized value of environmental comfort, R is the normalized value of safety risk level, and the weighting coefficients w1, w2, w3, and w4 satisfy w1 + w2 + w3 + w4 = 1. The comprehensive assessment module generates an optimization suggestion report through the above steps, providing a scientific basis for subsequent operation and maintenance decisions. For example, when the energy efficiency score of a certain area is low, the comprehensive assessment module will analyze the specific reasons and propose targeted optimization suggestions.

[0053] Edge computing nodes are distributed across key locations within the building, connected to the data fusion and parsing module, intelligent analysis engine module, and adaptive control module via a local area network. Each edge computing node utilizes an ARM Cortex-A72 processor with a clock speed of at least 2.0GHz and at least 8GB of memory. It incorporates the lightweight SQLite database, supporting offline data caching and breakpoint resume functionality. These nodes process high-frequency sensor data, reducing cloud transmission latency and ensuring efficient completion of tasks with high real-time requirements. Furthermore, the edge computing nodes verify data transmission integrity using a CRC checksum algorithm with a 32-bit checksum length and an error rate controlled within 10⁻⁶ × 10⁻⁶. They also incorporate a time synchronization mechanism, employing the NTP protocol to ensure that the time error between nodes does not exceed 1ms. For example, if data transmission from a sensor is interrupted, the edge computing node can resume transmission through the breakpoint resume function, ensuring data integrity.

[0054] As can be seen from the above steps, this invention solves the problem of information silos by integrating heterogeneous subsystem data through a data fusion and analysis module; the intelligent analysis engine module achieves accurate prediction of abnormal events by combining time series analysis and machine learning techniques; the dynamic visualization module enhances user experience by intuitively presenting the building's operating status through graphics rendering technology; the adaptive control module enhances system flexibility by generating optimal control strategies based on a rule engine and genetic algorithm optimization unit; and the comprehensive evaluation module provides a scientific basis for decision-making by quantifying building performance through a multi-dimensional indicator system. The synergistic effect of these technologies effectively improves the intelligence level and overall performance of the IBMS platform, meeting the higher requirements of modern building intelligence development.

[0055] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. An intelligent monitoring system for an IBMS building intelligence integration platform, characterized in that, include: The data fusion and parsing module is used to integrate data from heterogeneous subsystems. It converts data of different formats into a unified standard format through a protocol adapter and uses a distributed storage architecture to save real-time and historical data. The intelligent analysis engine module is used to deeply mine building operation data, extract equipment operation patterns through time series analysis algorithms, and predict the probability of potential abnormal events by combining machine learning models. The dynamic visualization module is used to generate a three-dimensional dynamic view of the building's operating status, and uses graphics rendering technology to present multi-source data intuitively in the form of heat maps and trend curves. The adaptive control module is used to dynamically adjust the linkage logic between heterogeneous subsystems according to the building's usage requirements. It generates the optimal control strategy through the rule engine and sends it to the execution unit in real time. The comprehensive assessment module is used to quantitatively evaluate the overall performance of a building, calculate energy efficiency scores based on a multi-dimensional indicator system, and generate optimization suggestion reports. Edge computing nodes are used to process high-frequency sensor data locally, reducing transmission latency to the cloud.

2. The intelligent monitoring system of the IBMS building intelligence integration platform according to claim 1, characterized in that, The protocol adapter supports Modbus, BACnet, and OPC UA industrial communication protocols. The adapter uses an FPGA chip for hardware acceleration, and the time for a single data conversion is no more than 10ms. The distributed storage architecture is built on the Hadoop ecosystem, supports petabyte-level data storage, and the data write latency is no more than 50ms.

3. The intelligent monitoring system of the IBMS building intelligence integration platform according to claim 1, characterized in that, The time series analysis algorithm uses an ARIMA model combined with an LSTM neural network. The model training cycle is no more than 2 hours, and the prediction accuracy is no less than 95%. The machine learning model is built based on the random forest algorithm, and the SHAP value analysis method is introduced in the feature selection process.

4. The intelligent monitoring system of the IBMS building intelligence integration platform according to claim 1, characterized in that, The graphics rendering technology is based on DirectX 12, supports 4K resolution output, and maintains a stable frame rate of over 60 FPS; the color mapping range of the heatmap can be customized according to user needs, with a minimum resolution of 0.1℃ or 0.1kWh.

5. The intelligent monitoring system of the IBMS building intelligence integration platform according to claim 1, characterized in that, The rule engine is implemented using the Drools framework, and the rule base contains no fewer than 500 predefined rules, supporting user-defined extensions; the control strategy generation process incorporates a genetic algorithm optimization unit, and the time for a single strategy generation does not exceed 1 second.

6. The intelligent monitoring system of the IBMS building intelligence integration platform according to claim 1, characterized in that, The multi-dimensional indicator system includes energy intensity, equipment health, environmental comfort, and safety risk level. The energy efficiency score calculation formula is: S = w1·E + w2·H + w3·C + w4·R, where S is the energy efficiency score, E is the normalized value of energy intensity, H is the normalized value of equipment health, C is the normalized value of environmental comfort, R is the normalized value of safety risk level, and the weight coefficients w1, w2, w3, and w4 satisfy w1 + w2 + w3 + w4 = 1.

7. The intelligent monitoring system of the IBMS building intelligence integration platform according to claim 1, characterized in that, The edge computing node uses an ARM Cortex-A72 processor with a main frequency of no less than 2.0GHz and a memory capacity of no less than 8GB; the node has a built-in lightweight database SQLite, which supports offline data caching and breakpoint resume functions.

8. The intelligent monitoring system of the IBMS building intelligence integration platform according to claim 1, characterized in that, It also includes a data verification unit, which verifies the integrity of data transmission using a CRC check algorithm. The checksum length is 32 bits, and the error rate does not exceed 10. -6 The unit has a built-in time synchronization mechanism and uses the NTP protocol to ensure that the time error of each node does not exceed 1ms.

9. The intelligent monitoring system of the IBMS building intelligence integration platform according to claim 1, characterized in that, Abnormal event prediction includes the following steps: S1. Data preprocessing: Extract historical data for the target time period from the distributed storage architecture and remove noise and missing values; S2. Feature Extraction: Extract time series features based on sliding window technology, with a window length of 1 hour and a step size of 10 minutes; S3. Model Training: Input the extracted time series features into the ARIMA-LSTM hybrid model for training, with a training set to test set ratio of 7:3; S4. Anomaly Detection: The trained hybrid model is used to predict real-time data. When the deviation between the predicted value and the actual value exceeds a set threshold, it is marked as a potential anomaly event. S5. Risk Assessment: Calculate the risk level of abnormal events using a random forest model and output early warning information.

10. The intelligent monitoring system of the IBMS building intelligence integration platform according to claim 1, characterized in that, The linkage logic optimization process includes the following steps: S1. Status Acquisition: Obtain the real-time operating status of each subsystem through edge computing nodes; S2. Rule Matching: Input the collected status data into the rule engine and match it with applicable predefined rules; S3. Strategy Generation: Based on the matching results, the genetic algorithm optimization unit is invoked to generate an initial control strategy; S4. Strategy Optimization: Iteratively optimize the initial strategy, with the objective function being to maximize the energy efficiency score; S5. Policy Distribution: The optimized control policy is distributed to the execution unit via the MQTT protocol.

Images