Multi-electrical cabinet collaborative fire protection and restart management system based on big data analysis

Abstract

The present invention discloses a multi-electrical cabinet collaborative fire prevention and restart management system based on big data analysis, which relates to the technical field of collaborative management of multiple electrical cabinets. It aims to solve the technical problems of fire warning lag, insufficient fault spread control, and lack of coordination in restart energy scheduling caused by abnormal environmental parameters, load fluctuations and fault conduction in the collaborative operation of multiple electrical cabinets in existing management systems. The system comprises: an information acquisition module for collecting environmental parameters, operating status, abnormal indicators and spatial correlation information in the electrical cabinet, specifically including temperature, humidity, smoke concentration, door opening frequency, load current value, leakage number, cabinet working status, position relationship of adjacent cabinets and real-time characteristics of power supply; an information processing module for real-time processing of data acquired by the information acquisition module, including data reception, verification, analysis and output. The present invention has the advantages of realizing early risk identification, collaborative fire prevention and efficient restart management in a multi-cabinet coupling scenario.

Description

Technical Field

[0001] The present invention relates to the technical field of collaborative management of multiple electrical cabinets, and more specifically, to a collaborative fire prevention and restart management system for multiple electrical cabinets based on big data analysis. Background Art

[0002] In the field of industrial automation and power systems, the coordinated operation of multiple electrical cabinets poses an increasingly significant risk of safety hazards and downtime caused by abnormal environmental parameters (such as excessive temperatures and smoke leaks), load fluctuations, or fault transmission. Existing technologies, which rely solely on monitoring and handling of individual cabinets, struggle to effectively address the coupling effects of thermal conduction and electrical interference between multiple cabinets. This results in delayed fire warnings, insufficient control of fault spread, and a lack of coordinated energy scheduling during the restart process, which can easily lead to secondary failures and waste resources. To address this, we propose a collaborative fire prevention and restart management system for multiple electrical cabinets based on big data analysis. Summary of the Invention

[0003] The purpose of the present invention is to provide a multi-electrical cabinet collaborative fire prevention and restart management system based on big data analysis, so as to solve the technical problems of fire warning lag, insufficient fault spread control and lack of coordination in restart energy scheduling caused by abnormal environmental parameters, load fluctuations and fault conduction in the collaborative operation of multiple electrical cabinets in the existing management system.

[0004] To solve the above technical problems, the present invention provides the following technical solutions: a multi-electrical cabinet collaborative fire prevention and restart management system based on big data analysis, a multi-electrical cabinet collaborative fire prevention and restart management system based on big data analysis, comprising:

[0005] The information collection module is used to collect environmental parameters, operating status, abnormal indicators and space-related information inside the electrical cabinet, including temperature, humidity, smoke concentration, door opening frequency, load current value, leakage number, cabinet working status, position relationship of adjacent cabinets and real-time power supply characteristics;

[0006] The information processing module is used to process the data obtained by the information acquisition module in real time, including data reception, verification, analysis and output;

[0007] The data storage module is used to store the collected raw data, historical data, static properties of electrical cabinets and multi-cabinet associations, as well as the analysis results and disposal strategies generated by the information analysis module;

[0008] The information analysis module analyzes the electrical cabinet status, abnormal risk, importance, fault impact and restart energy demand based on processed data and historical data, and generates a collaborative disposal strategy based on the coupling relationship between multiple cabinets.

[0009] Preferably, the information collection module includes:

[0010] A distributed layout is adopted, with multiple monitoring points set at key heat points and heat dissipation channels in the electrical cabinet to support real-time temperature field reconstruction;

[0011] The humidity sensor uses a polymer film capacitive sensor, which is moisture-proof and anti-condensation, and can measure relative humidity and absolute humidity;

[0012] Smoke concentration sensor, which uses the principle of laser scattering to distinguish different types of smoke particles and provide early fire warning;

[0013] The door status monitoring module integrates a micro switch and an angle sensor, uses an event-triggered mechanism to record door opening time, door closing time, and door opening angle, and supports identification of abnormal door opening behavior;

[0014] The load current monitoring module uses a combination of Rogowski coils and Hall effect sensors, supports multi-range automatic switching, and has harmonic analysis and load characteristic identification functions;

[0015] The leakage monitoring module uses zero-sequence current transformer and digital signal processing technology to support leakage waveform feature extraction and fault location;

[0016] The condition monitoring module uses multi-sensor fusion technology to monitor power supply voltage fluctuations, equipment vibration frequency, and electromagnetic interference intensity parameters;

[0017] The spatial correlation monitoring unit uses positioning sensors and 3D modeling technology to obtain the location coordinates, adjacent spacing, and layout orientation of electrical cabinets, and construct the spatial topology relationship of multiple cabinets;

[0018] The power characteristics monitoring unit collects UPS output voltage / frequency, remaining capacity of the backup power supply, and grid fluctuation parameters in real time to monitor the dynamic characteristics of the power system.

[0019] Preferably, the information processing module includes:

[0020] Data receiving unit, with adaptive communication rate adjustment and protocol conversion functions;

[0021] The data verification unit adopts a three-level verification mechanism: the first level performs data format verification, the second level performs physical quantity range verification, and the third level performs data correlation verification. Suspicious data is marked and a retransmission mechanism is activated.

[0022] The data analysis unit performs trend analysis based on a sliding time window algorithm, uses an outlier detection algorithm to identify sudden abnormalities, and performs multi-parameter correlation analysis through a Bayesian network to achieve early fault prediction;

[0023] The data output unit uses message queue middleware to implement asynchronous communication and supports publish-subscribe mode;

[0024] A cross-cabinet parameter correlation analysis function has been added for multi-cabinet coupling scenarios. This function uses a cross-correlation function to calculate the time delay of temperature changes in adjacent cabinets and identify abnormal heat conduction paths.

[0025] Cluster-level abnormal pattern recognition algorithm triggers a regional environmental fault warning when ≥2 cabinets in the same area have similar parameter abnormalities.

[0026] Preferably, the data storage module includes:

[0027] The database uses a time series database to store high-frequency sampled real-time data, supports time series compression algorithms and multi-level indexing mechanisms, and uses a graph database to store the spatial topology and physical connection relationships of electrical cabinets;

[0028] Historical database, building a data warehouse architecture, including operational data storage, data marts, and data cubes, using machine learning models to perform knowledge mining on historical data to extract failure modes and treatment experience;

[0029] The graph database further constructs a directed graph model that includes the physical connections and business dependencies of electrical cabinets, and stores coupling parameters such as the thermal conductivity coefficient of adjacent cabinets and the weight of fault propagation paths.

[0030] Dynamic parameter database, real-time update of power system efficiency curve, equipment aging status and correction coefficient of ambient temperature on equipment startup characteristics.

[0031] Preferably, the information analysis module includes:

[0032] The electrical cabinet abnormality probability judgment unit builds a multi-source information fusion model based on DS evidence theory, performs weighted fusion on seven types of monitoring parameters, uses a hidden Markov model to predict abnormal states, and calculates the probability of abnormality through Monte Carlo simulation;

[0033] The electrical cabinet importance judgment unit establishes a hierarchical evaluation index system, including system level, equipment level and component level; uses network analysis method to handle the dependency relationship between indicators, and uses fuzzy hierarchical analysis method to determine the relative importance;

[0034] The electrical cabinet restart energy demand calculation unit establishes a dynamic energy model based on thermodynamic principles, considers the equipment startup characteristic curve, load inertia moment and power conversion efficiency, and uses the particle swarm optimization algorithm to solve the optimal startup sequence to reduce peak power demand.

[0035] Preferably, the information analysis module further includes:

[0036] The electrical cabinet fire extinguishing sequence judgment unit builds a three-dimensional fire spread simulation model, taking into account air flow, thermal radiation, and material combustion characteristics, and uses an ant colony algorithm to find the optimal fire extinguishing path. This balances fire extinguishing effectiveness and resource consumption through multi-objective optimization.

[0037] The electrical cabinet restart determination unit establishes a fault propagation model to analyze cascading effects, uses a decision tree algorithm to assess the necessity of restart, and formulates differentiated restart strategies based on business recovery time objectives and data loss tolerance.

[0038] The cascading fault prediction subunit uses a Bayesian network to build a fault propagation model based on the multi-cabinet coupling parameters stored in the graph database, and calculates in real time the probability of risk transmission from a cabinet's abnormal parameters to adjacent cabinets.

[0039] The coupled risk assessment module simultaneously analyzes the thermal radiation impact and electrical interference risk on adjacent cabinets when a single cabinet triggers an abnormal warning, and generates a multi-cabinet linkage risk index.

[0040] Preferably, the electrical cabinet fire extinguishing sequence judgment unit combines the physical connection relationship stored in the graph database to preferentially cut off key nodes that may cause fault spread, forming a layered fire extinguishing strategy. The key node judgment formula is:

[0041] ;

[0042] in, For nodes The fault propagation sensitivity of For adjacent cabinets, is the connection weight, It is the importance level of adjacent cabinets.

[0043] Preferably, the electrical cabinet restart energy demand calculation unit includes:

[0044] The dynamic load modeling module dynamically corrects the device startup impedance through the thermal resistance network model based on the real-time collected load current waveform and ambient temperature data. The startup impedance temperature correction formula is:

[0045] ;

[0046] in, is the starting impedance of the device at room temperature, is the temperature sensitivity coefficient, is the real-time temperature, is the reference temperature;

[0047] The redundant resource scheduling subunit scans the idle capacity of adjacent electrical cabinets and generates a cross-cabinet load sharing plan to reduce the peak power demand of the target cabinet. The energy sharing formula considering redundant equipment is:

[0048] ;

[0049] in, is the total restart energy after correction, is the initial energy requirement of the target cabinet, It is the idle capacity of redundant cabinets. is the rated capacity of the target cabinet, and the load sharing ratio is calculated as follows:

[0050] ;

[0051] in, For the The load borne by the redundant cabinets, is the rated capacity, is the current load, is a collection of available redundant cabinets. is the excess power requirement of the target cabinet.

[0052] Preferably, the specific strategy of the redundant resource scheduling subunit is:

[0053] When the restart energy demand of the target cabinet exceeds 80% of its rated capacity, the cross-cabinet scheduling mechanism is triggered to dynamically adjust the cross-cabinet load distribution ratio.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] 1. The present invention uses a distributed sensor network and multi-parameter fusion analysis, combined with a three-dimensional fire spread simulation model and an ant colony algorithm, to achieve real-time identification of abnormal heat conduction paths in multiple cabinets and optimal fire extinguishing path planning, prioritize cutting off key nodes of fault propagation, significantly improve the timeliness and accuracy of fire prevention and control, and avoid regional chain accidents caused by single cabinet failures. It solves the problems of existing management systems in the coordinated operation of multiple electrical cabinets, such as delayed fire warning, insufficient fault spread control, and lack of coordination in restart energy scheduling caused by abnormal environmental parameters, load fluctuations, and fault conduction.

[0056] 2. The present invention is also based on a thermodynamic dynamic energy model and a particle swarm optimization algorithm to dynamically correct the equipment startup impedance and construct a cross-cabinet load sharing strategy. When the restart energy demand exceeds 80% of the rated capacity, cross-cabinet scheduling is triggered, and the idle capacity of adjacent cabinets is used to reduce the peak power demand, thereby solving the power supply overload problem caused by concentrated energy consumption in traditional restart strategies.

[0057] 3. The present invention also constructs a directed graph of multi-cabinet spatial topology and business dependencies through a graph database, combines Bayesian networks with Monte Carlo simulation, calculates the fault conduction probability in real time, and combines it with the multi-cabinet linkage risk index to achieve advanced warning and differentiated handling of cascading failures, breaking through the limitations of the traditional independent monitoring mode in perceiving multi-cabinet coupling risks and reducing the risk of systemic downtime. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 Schematic diagram of the system framework of the present invention. DETAILED DESCRIPTION

[0059] To facilitate those skilled in the art to understand the technical solution of the present invention, the technical solution of the present invention is further described with reference to the accompanying drawings.

[0060] Example 1, as Figure 1 As shown, the present invention provides a multi-electrical cabinet collaborative fire prevention and restart management system based on big data analysis, including:

[0061] The information collection module is used to collect environmental parameters, operating status, abnormal indicators and space-related information inside the electrical cabinet, including temperature, humidity, smoke concentration, door opening frequency, load current value, leakage number, cabinet working status, position relationship of adjacent cabinets and real-time power supply characteristics;

[0062] The information processing module is used to process the data obtained by the information acquisition module in real time, including data reception, verification, analysis and output;

[0063] The data storage module is used to store the collected raw data, historical data, static properties of electrical cabinets and multi-cabinet associations, as well as the analysis results and disposal strategies generated by the information analysis module;

[0064] The information analysis module analyzes the electrical cabinet status, abnormal risk, importance, fault impact and restart energy demand based on processed data and historical data, and generates a collaborative disposal strategy based on the coupling relationship between multiple cabinets.

[0065] In an embodiment of the present invention, the information collection module includes:

[0066] A distributed layout is adopted, with multiple monitoring points set at key heat points and heat dissipation channels in the electrical cabinet to support real-time temperature field reconstruction;

[0067] The humidity sensor uses a polymer film capacitive sensor, which is moisture-proof and anti-condensation, and can measure relative humidity and absolute humidity;

[0068] Smoke concentration sensor, which uses the principle of laser scattering to distinguish different types of smoke particles and provide early fire warning;

[0069] The door status monitoring module integrates a micro switch and an angle sensor. It uses an event trigger mechanism to record the door opening time, door closing time, and door opening angle. It supports the recognition of abnormal door opening behavior. The door opening frequency calculation formula is:

[0070] ;

[0071] in, The number of door openings during the monitoring period. is the duration of the monitoring cycle;

[0072] The load current monitoring module uses a combination of Rogowski coils and Hall effect sensors, supports multi-range automatic switching, and has harmonic analysis and load characteristic identification functions. The load characteristic identification formula is:

[0073] ;

[0074] in, is the harmonic distortion rate, is the effective value of harmonic current, is the effective value of fundamental current;

[0075] The leakage monitoring module uses zero-sequence current transformer and digital signal processing technology to achieve microampere-level leakage detection and support leakage waveform feature extraction and fault location;

[0076] The status monitoring module uses multi-sensor fusion technology to monitor power supply voltage fluctuations, equipment vibration frequency, and electromagnetic interference intensity parameters to achieve comprehensive perception of the operating status of the electrical cabinet;

[0077] The spatial correlation monitoring unit uses positioning sensors and 3D modeling technology to obtain the location coordinates, adjacent spacing, and layout orientation of electrical cabinets, and construct the spatial topology relationship of multiple cabinets;

[0078] The power characteristics monitoring unit collects UPS output voltage / frequency, remaining capacity of the backup power supply, and grid fluctuation parameters in real time to monitor the dynamic characteristics of the power system.

[0079] In an embodiment of the present invention, the information processing module includes:

[0080] The data receiving unit supports industrial bus protocols such as Modbus, Profibus, and CAN bus, as well as wireless communication protocols such as 4G / 5G, Wi-Fi, and LoRa, and has adaptive communication rate adjustment and protocol conversion functions;

[0081] The data verification unit adopts a three-level verification mechanism: the first level performs data format verification, the second level performs physical quantity range verification, and the third level performs data correlation verification. Suspicious data is marked and a retransmission mechanism is activated. The verification formula is:

[0082] ;

[0083] in, For parameters and The Pearson correlation coefficient, and is the mean;

[0084] The data analysis unit performs trend analysis based on a sliding time window algorithm, uses an outlier detection algorithm to identify sudden abnormalities, and performs multi-parameter correlation analysis through a Bayesian network to achieve early fault prediction. The multi-parameter correlation probability formula is:

[0085] ;

[0086] in, 、 、 To monitor parameter abnormal events;

[0087] The data output unit uses message queue middleware to implement asynchronous communication, supports the publish-subscribe model, and can perform data format conversion and information summary extraction according to the needs of the receiver;

[0088] The cross-cabinet parameter correlation analysis function added for multi-cabinet coupling scenarios calculates the time delay of temperature changes in adjacent cabinets through the cross-correlation function and identifies abnormal heat conduction paths. The time delay calculation formula is:

[0089] ;

[0090] in, is the temperature sequence of adjacent cabinets and The cross-correlation function, For time delay;

[0091] Cluster-level abnormal pattern recognition algorithm: When ≥2 cabinets in the same area have similar parameter abnormalities, it triggers a regional environmental fault warning instead of a single cabinet alarm.

[0092] In an embodiment of the present invention, the data storage module includes:

[0093] The database uses a time series database to store high-frequency sampled real-time data, supports time series compression algorithms and multi-level indexing mechanisms, and uses a graph database to store the spatial topology and physical connection relationships of electrical cabinets;

[0094] Historical database, building a data warehouse architecture, including operational data storage, data marts, and data cubes, supporting OLAP analysis, using machine learning models to perform knowledge mining on historical data, extracting failure modes and handling experience;

[0095] The graph database further constructs a directed graph model that includes the physical connections and business dependencies of electrical cabinets, and stores coupling parameters such as the thermal conductivity coefficient of adjacent cabinets and the weight of fault propagation paths.

[0096] Dynamic parameter database, real-time update of power system efficiency curve, equipment aging status and correction coefficient of ambient temperature on equipment startup characteristics.

[0097] In an embodiment of the present invention, the information analysis module includes:

[0098] The electrical cabinet abnormality probability judgment unit builds a multi-source information fusion model based on DS evidence theory, performs weighted fusion on seven types of monitoring parameters, uses a hidden Markov model to predict abnormal states, and calculates the probability of abnormality through Monte Carlo simulation. The evidence theory fusion formula is as follows:

[0099] set up To identify the framework, for on The basic probability distribution functions, the focal elements are , the basic probability distribution function after fusion for:

[0100] ;

[0101] in, is a normalization constant used to eliminate evidence conflicts, and its calculation formula is:

[0102] ;

[0103] in, For the fusion proposition The basic probability distribution of It is single sensor evidence;

[0104] The electrical cabinet importance judgment unit establishes a hierarchical evaluation index system, including the system layer, equipment layer, and component layer. The network analysis method is used to process the dependency relationship between indicators, and the relative importance is determined by the fuzzy hierarchical analysis method. The hierarchical analysis weight formula is:

[0105] ;

[0106] in, For the The weight of the indicator, For indicators and indicators The relative importance of the judgment matrix elements, is the total number of indicators;

[0107] The electrical cabinet restart energy demand calculation unit establishes a dynamic energy model based on thermodynamic principles, considers the equipment startup characteristic curve, load inertia moment, and power conversion efficiency, and uses a particle swarm optimization algorithm to solve the optimal startup sequence to reduce peak power demand;

[0108] The electrical cabinet fire extinguishing sequence judgment unit builds a three-dimensional fire spread simulation model, taking into account air flow, thermal radiation, and material combustion characteristics, and uses an ant colony algorithm to find the optimal fire extinguishing path. This balances fire extinguishing effectiveness and resource consumption through multi-objective optimization.

[0109] The electrical cabinet restart determination unit establishes a fault propagation model to analyze cascading effects, uses a decision tree algorithm to assess the necessity of a restart, and considers business recovery time objectives and data loss tolerance to develop differentiated restart strategies. The specific method is to retrieve the business dependency graph from the graph database, analyze whether the target cabinet failure has caused more than 30% of associated equipment to shut down, query the real-time status of the backup equipment in adjacent cabinets through the dynamic parameter database, and prioritize partial equipment replacement over a full restart strategy if more than three redundant devices are available.

[0110] The cascading fault prediction subunit uses a Bayesian network to build a fault propagation model based on the multi-cabinet coupling parameters stored in the graph database. It calculates the risk transmission probability of a cabinet's abnormal parameters to adjacent cabinets in real time. The fault transmission probability calculation formula is:

[0111] ;

[0112] in, Source cabinet Target cabinet when abnormal The probability of cascading failures, Target cabinet Fault causes source cabinet The conditional probability of anomaly, Target cabinet Prior probability of failure;

[0113] The coupled risk assessment module analyzes the thermal radiation impact and electrical interference risk on adjacent cabinets when a single cabinet triggers an abnormality warning, and generates a multi-cabinet linkage risk index. The heat conduction rate formula is:

[0114] ;

[0115] in, is the heat conduction rate, is the heat transfer coefficient between adjacent cabinets, is the contact area, is the temperature difference, The distance between cabinets.

[0116] In an embodiment of the present invention, the electrical cabinet fire extinguishing sequence determination unit combines the physical connection relationships stored in the graph database to prioritize cutting off key nodes that may cause fault spread, forming a layered fire extinguishing strategy. The key node determination formula is:

[0117] ;

[0118] in, For nodes The fault propagation sensitivity of For adjacent cabinets, is the connection weight, It is the importance level of adjacent cabinets.

[0119] In an embodiment of the present invention, the electrical cabinet restart energy demand calculation unit includes:

[0120] The dynamic load modeling module dynamically corrects the device startup impedance through the thermal resistance network model based on the real-time collected load current waveform and ambient temperature data. The startup impedance temperature correction formula is:

[0121] ;

[0122] in, is the starting impedance of the device at room temperature, is the temperature sensitivity coefficient, is the real-time temperature, is the reference temperature;

[0123] The redundant resource scheduling subunit scans the idle capacity of adjacent electrical cabinets and generates a cross-cabinet load sharing plan to reduce the peak power demand of the target cabinet. The energy sharing formula considering redundant equipment is:

[0124] ;

[0125] in, is the total restart energy after correction, is the initial energy requirement of the target cabinet, It is the idle capacity of redundant cabinets. is the rated capacity of the target cabinet, and the load sharing ratio is calculated as follows:

[0126] ;

[0127] in, For the The load borne by the redundant cabinets, is the rated capacity, is the current load, is a collection of available redundant cabinets. is the excess power requirement of the target cabinet.

[0128] In an embodiment of the present invention, the specific strategy of the redundant resource scheduling subunit is:

[0129] When the target cabinet restart energy demand exceeds 80% of its rated capacity, the cross-cabinet scheduling mechanism is triggered;

[0130] Cabinets that are physically adjacent and have a current load rate lower than 50% are preferentially selected as redundant carriers. The load rate calculation formula is:

[0131] ;

[0132] in, For cabinet Load factor;

[0133] Dynamically adjust the load distribution ratio across cabinets to ensure that the real-time load of all cabinets does not exceed 90% of the rated capacity during the restart process.

[0134] The embodiments disclosed in the present invention are preferred embodiments, but are not limited to them. Ordinary technicians in this field can easily understand the spirit of the present invention based on the above embodiments and make different extensions and changes. As long as they do not deviate from the spirit of the present invention, they are all within the scope of protection of the present invention.

Claims

1. A multi-electrical cabinet collaborative fire prevention and restart management system based on big data analysis, characterized in that: include: The information collection module is used to collect environmental parameters, operating status, abnormal indicators and space-related information inside the electrical cabinet, including temperature, humidity, smoke concentration, door opening frequency, load current value, leakage number, cabinet working status, position relationship of adjacent cabinets and real-time power supply characteristics; The information processing module is used to process the data obtained by the information acquisition module in real time, including data reception, verification, analysis and output; The data storage module is used to store the collected raw data, historical data, static properties of electrical cabinets and multi-cabinet associations, as well as the analysis results and disposal strategies generated by the information analysis module; The information analysis module analyzes the electrical cabinet status, abnormal risk, importance, fault impact, and restart energy requirements based on processed data and historical data, and generates a collaborative disposal strategy based on the coupling relationship between multiple cabinets; The information analysis module includes: The electrical cabinet abnormality probability judgment unit builds a multi-source information fusion model based on DS evidence theory, performs weighted fusion on seven types of monitoring parameters, uses a hidden Markov model to predict abnormal states, and calculates the probability of abnormality through Monte Carlo simulation; The electrical cabinet importance judgment unit establishes a hierarchical evaluation index system, including system level, equipment level and component level; uses network analysis method to handle the dependency relationship between indicators, and uses fuzzy hierarchical analysis method to determine the relative importance; The electrical cabinet restart energy demand calculation unit establishes a dynamic energy model based on thermodynamic principles, considers the equipment startup characteristic curve, load inertia moment, and power conversion efficiency, and uses a particle swarm optimization algorithm to solve the optimal startup sequence to reduce peak power demand; The electrical cabinet restart energy demand calculation unit includes: The dynamic load modeling module dynamically corrects the device startup impedance through the thermal resistance network model based on the real-time collected load current waveform and ambient temperature data. The startup impedance temperature correction formula is: ； in, is the starting impedance of the device at room temperature, is the temperature sensitivity coefficient, is the real-time temperature, is the reference temperature; The redundant resource scheduling subunit scans the idle capacity of adjacent electrical cabinets and generates a cross-cabinet load sharing plan to reduce the peak power demand of the target cabinet. The energy sharing formula considering redundant equipment is: ； in, is the total restart energy after correction, is the initial energy requirement of the target cabinet, It is the idle capacity of redundant cabinets. is the rated capacity of the target cabinet, and the load sharing ratio is calculated as follows: ； in, For the The load borne by the redundant cabinets, is the rated capacity, is the current load, is a collection of available redundant cabinets. is the excess power requirement of the target cabinet.

2. The multi-electrical cabinet collaborative fire prevention and restart management system based on big data analysis according to claim 1 is characterized in that: The information collection module includes: A distributed layout is adopted, with multiple monitoring points set at key heat points and heat dissipation channels in the electrical cabinet to support real-time temperature field reconstruction; The humidity sensor uses a polymer film capacitive sensor, which is moisture-proof and anti-condensation, and can measure relative humidity and absolute humidity; Smoke concentration sensor, which uses the principle of laser scattering to distinguish different types of smoke particles and provide early fire warning; The door status monitoring module integrates a micro switch and an angle sensor, uses an event-triggered mechanism to record door opening time, door closing time, and door opening angle, and supports identification of abnormal door opening behavior; The load current monitoring module uses a combination of Rogowski coils and Hall effect sensors, supports multi-range automatic switching, and has harmonic analysis and load characteristic identification functions; The leakage monitoring module uses zero-sequence current transformer and digital signal processing technology to support leakage waveform feature extraction and fault location; The condition monitoring module uses multi-sensor fusion technology to monitor power supply voltage fluctuations, equipment vibration frequency, and electromagnetic interference intensity parameters; The spatial correlation monitoring unit uses positioning sensors and 3D modeling technology to obtain the location coordinates, adjacent spacing, and layout orientation of electrical cabinets, and construct the spatial topology relationship of multiple cabinets; The power characteristics monitoring unit collects UPS output voltage / frequency, remaining capacity of the backup power supply, and grid fluctuation parameters in real time to monitor the dynamic characteristics of the power system.

3. The multi-electrical cabinet collaborative fire prevention and restart management system based on big data analysis according to claim 2 is characterized in that: The information processing module includes: Data receiving unit, with adaptive communication rate adjustment and protocol conversion functions; The data verification unit adopts a three-level verification mechanism: the first level performs data format verification, the second level performs physical quantity range verification, and the third level performs data correlation verification. Suspicious data is marked and a retransmission mechanism is activated. The data analysis unit performs trend analysis based on a sliding time window algorithm, uses an outlier detection algorithm to identify sudden abnormalities, and performs multi-parameter correlation analysis through a Bayesian network to achieve early fault prediction; The data output unit uses message queue middleware to implement asynchronous communication and supports publish-subscribe mode; A cross-cabinet parameter correlation analysis function has been added for multi-cabinet coupling scenarios. This function uses a cross-correlation function to calculate the time delay of temperature changes in adjacent cabinets and identify abnormal heat conduction paths. Cluster-level abnormal pattern recognition algorithm triggers a regional environmental fault warning when ≥2 cabinets in the same area have similar parameter abnormalities.

4. The multi-electrical cabinet collaborative fire prevention and restart management system based on big data analysis according to claim 3 is characterized in that: The data storage module includes: The database uses a time series database to store high-frequency sampled real-time data, supports time series compression algorithms and multi-level indexing mechanisms, and uses a graph database to store the spatial topology and physical connection relationships of electrical cabinets; Historical database, building a data warehouse architecture, including operational data storage, data marts, and data cubes, using machine learning models to perform knowledge mining on historical data to extract failure modes and treatment experience; The graph database further constructs a directed graph model that includes the physical connections and business dependencies of electrical cabinets, and stores coupling parameters such as the thermal conductivity coefficient of adjacent cabinets and the weight of fault propagation paths. Dynamic parameter database, real-time update of power system efficiency curve, equipment aging status and correction coefficient of ambient temperature on equipment startup characteristics.

5. The multi-electrical cabinet collaborative fire prevention and restart management system based on big data analysis according to claim 4 is characterized in that: The information analysis module also includes: The electrical cabinet fire extinguishing sequence judgment unit builds a three-dimensional fire spread simulation model, taking into account air flow, thermal radiation, and material combustion characteristics, and uses an ant colony algorithm to find the optimal fire extinguishing path. This balances fire extinguishing effectiveness and resource consumption through multi-objective optimization. The electrical cabinet restart determination unit establishes a fault propagation model to analyze cascading effects, uses a decision tree algorithm to assess the necessity of restart, and formulates differentiated restart strategies based on business recovery time objectives and data loss tolerance. The cascading fault prediction subunit uses a Bayesian network to build a fault propagation model based on the multi-cabinet coupling parameters stored in the graph database, and calculates in real time the probability of risk transmission from a cabinet's abnormal parameters to adjacent cabinets. The coupled risk assessment module simultaneously analyzes the thermal radiation impact and electrical interference risk on adjacent cabinets when a single cabinet triggers an abnormal warning, and generates a multi-cabinet linkage risk index.

6. The multi-electrical cabinet coordinated fire prevention and restart management system based on big data analysis according to claim 5 is characterized in that: The electrical cabinet fire extinguishing sequence judgment unit combines the physical connection relationships stored in the graph database to prioritize cutting off key nodes that may cause fault spread, forming a layered fire extinguishing strategy. The key node judgment formula is: ； in, For nodes The fault propagation sensitivity of For adjacent cabinets, is the connection weight, It is the importance level of adjacent cabinets.

7. The multi-electrical cabinet coordinated fire prevention and restart management system based on big data analysis according to claim 6 is characterized in that: The specific strategy of the redundant resource scheduling subunit is: When the restart energy demand of the target cabinet exceeds 80% of its rated capacity, the cross-cabinet scheduling mechanism is triggered to dynamically adjust the cross-cabinet load distribution ratio.

Images