Partition controller linkage method and device, equipment and storage medium

Abstract

The invention relates to the technical field of fire safety, and discloses a partition controller linkage method, device and equipment and a storage medium. The method comprises the following steps: acquiring original fire sensing data acquired by a plurality of fire-fighting sensors in each independent partition, and performing preprocessing of a plurality of sensing types; extracting space-time dimension features from the preprocessed data, generating fire behavior space-time feature vectors, and performing multi-source evidence weight fusion; performing fire abnormal mode identification and evaluation on the fused fire spatio-temporal feature vector to obtain a partition preliminary fire evaluation result, transmitting the partition preliminary fire evaluation result to a main controller, and performing global fire situation fusion based on a building physical structure and a partition connection relationship to obtain an overall fire situation evaluation result; and generating a multi-partition cooperative fire extinguishing execution scheme based on the overall fire situation assessment result, and controlling fire fighting equipment in each partition to perform linkage fire fighting. According to the invention, multi-source data fusion and cross-partition collaborative linkage of building fire are realized, and the fire detection accuracy and the fire-fighting response efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of fire safety, and particularly to a method, device, equipment and storage medium for linkage of zone controllers. Background Art

[0002] In the field of fire safety, the monitoring and prevention of building fires are key links in ensuring life and property safety, and the zone controller linkage system is the core component of fire prevention and control in large and complex buildings. Accurately monitoring the fire situation in the building and realizing multi-zone collaborative linkage are crucial for the effective operation of the fire protection system, which directly affects the timeliness of fire detection, the pertinence of fire extinguishing measures and the effectiveness of personnel evacuation. Therefore, an efficient method for linkage of zone controllers is of great significance for ensuring the overall response ability and protection effect of the building fire protection system.

[0003] Currently, most fire protection systems adopt simple linkage strategies based on threshold triggering, or attempt to use single-sensor data for fire situation judgment, and even start to apply artificial intelligence technology to fire situation assessment to improve the accuracy of fire recognition. However, these methods still face significant challenges in integrating multi-source sensing data, processing the spatio-temporal characteristics of fire situations and adapting to complex building environments. Traditional methods for linkage of zone controllers often neglect key factors such as the dynamic characteristics of fire evolution, the fusion mechanism of multi-sensor information and the physical relevance between zones, which have a decisive impact on fire situation judgment and linkage decision-making. That is, the existing methods for linkage of zone controllers lack in-depth fusion of multi-source sensing data and scientific assessment of the overall fire situation in the building, making it difficult to achieve accurate fire recognition and efficient multi-zone collaborative fire extinguishing, resulting in serious limitations in the response efficiency of the fire protection system in complex fire situations. Summary of the Invention

[0004] The main objective of the present invention is to solve the problem that the existing methods for linkage of zone controllers lack in-depth fusion of multi-source sensing data and scientific assessment of the overall fire situation in the building, making it difficult to achieve accurate fire recognition and efficient multi-zone collaborative fire extinguishing, resulting in serious limitations in the response efficiency of the fire protection system in complex fire situations.

[0005] The first aspect of the present invention provides a method for linking partition controllers, which is applied to a partition controller linkage system. The partition controller linkage system includes a plurality of partition controllers and a main controller. The method for linking partition controllers includes: obtaining the original fire sensing data collected by various fire sensors in the corresponding partitions received by the partition controllers of each independent partition in the building to be monitored, and using each partition controller to perform preprocessing of multiple sensing types on the original fire sensing data to obtain the preprocessed original fire sensing data. The area to be monitored includes a plurality of independent partitions, and each independent partition is configured with a partition controller and a variety of fire sensors are deployed; extracting the features of the time and space dimensions of the preprocessed original fire sensing data to generate a fire spatio-temporal feature vector for each independent partition, and performing multi-source evidence weight fusion within the partition on the fire spatio-temporal feature vectors of each independent partition to obtain a fused fire feature; performing fire anomaly pattern recognition and evaluation on the fire spatio-temporal feature vector to obtain a preliminary fire evaluation result for each partition, and transmitting the preliminary fire evaluation results of each partition controller to the main controller, and based on the building physical structure corresponding to the building to be monitored and the partition connection relationship corresponding to each independent partition, using the main controller to fuse the preliminary fire evaluation results of each partition to obtain an overall fire situation evaluation result; based on the overall fire situation evaluation result, using the main controller to generate a multi-partition collaborative fire extinguishing execution plan and transmit it to the corresponding partition controller, and based on the multi-partition collaborative fire extinguishing execution plan, each partition controller controls the fire fighting equipment in the corresponding independent partition to perform fire fighting linkage rescue to obtain a fire fighting linkage result.

[0006] Optionally, in the first implementation manner of the first aspect of the present invention, the original fire sensing data includes original temperature data, original smoke data, original harmful gas concentration data, and original flame image stream. The use of each of the partition controllers to perform preprocessing of multiple sensing types on the original fire sensing data to obtain preprocessed original fire sensing data includes: performing smoothing elimination of heat source fluctuations and extraction of fire heat characteristics on the original temperature data to obtain smoothed temperature data, and performing ventilation disturbance rejection and smoke purity improvement on the original smoke data to obtain stable smoke concentration data, and performing interference rejection of non-fire gases and extraction of fire characteristic gases on the original harmful gas concentration data to obtain corrected harmful gas concentration data, and performing non-fire light source filtering and flame image enhancement on the original flame image stream to obtain flame image enhancement data; based on the sensing types of the sensors corresponding to each of the partition controllers, performing acquisition time calibration and data time sequence matching on the smoothed temperature data, the stable smoke concentration data, the corrected harmful gas concentration data, and the flame image enhancement data to obtain time-aligned multi-source fire data, and performing fire index unification and quantification of fire characteristics on the time-aligned multi-source fire data to obtain preprocessed original fire sensing data.

[0007] Optionally, in the second implementation manner of the first aspect of the present invention, extracting the features of the time and space dimensions of the preprocessed original fire sensing data to generate the fire spatio-temporal feature vectors of each independent partition includes: dividing the preprocessed original fire sensing data into multi-scale time windows to obtain a multi-scale fire observation data set; calculating the temperature rise change rate and extracting the acceleration of the temperature data in the multi-scale fire observation data set to obtain the dynamic characteristics of fire temperature rise, and calculating the spatial distribution and identifying the propagation direction of the temperature data in the multi-scale fire observation data set to obtain the spatial diffusion characteristics of fire temperature, and calculating the concentration growth rate and extracting the distribution pattern of the smoke data in the multi-scale fire observation data set to obtain the dynamic characteristics of smoke diffusion, and calculating the gas ratio and extracting the change trend of the harmful gas sensor data in the multi-scale fire observation data set to obtain the gas characteristics of the combustion type, and extracting the characteristic parameters and identifying the combustion characteristics of the flame image in the multi-scale fire observation data set to obtain the visual characteristics of the fire source type; performing feature combination and unified format conversion of each independent partition on the dynamic characteristics of fire temperature rise, the spatial diffusion characteristics of fire temperature, the dynamic characteristics of smoke diffusion, the gas characteristics of the combustion type, and the visual characteristics of the fire source type to obtain the fire spatio-temporal feature vectors of each independent partition.

[0008] Optionally, in the third implementation manner of the first aspect of the present invention, the multi-source evidence weight fusion within the partition for the fire spatio-temporal feature vectors of each independent partition to obtain the fused fire feature includes: decomposing the fire spatio-temporal feature vectors based on the sensing types and feature categories of the sensors correspondingly connected to each partition controller to obtain a fire evidence source set composed of multiple independent fire evidence sources, and calculating the weight of each evidence source in the fire evidence source set based on a preset historical performance record library to obtain a weighted fire evidence set; performing fire probability quantization and unified fire fighting index conversion on each evidence in the weighted fire evidence set to obtain an initial fire probability distribution, and performing cross-comparison and conflict measurement on various fire index types in the initial fire probability distribution to obtain a fire evidence conflict matrix; performing threshold comparison and maximum conflict value positioning on the fire evidence conflict matrix to obtain a corrected fire probability distribution, and sorting each evidence source in the fire evidence source set based on the corresponding weight values in the weighted fire evidence set to obtain an evidence priority sequence, and performing evidence combination and feature integration on the fire index types in the corrected fire probability distribution based on the evidence priority sequence to obtain the fused fire feature.

[0009] Optionally, in the fourth implementation manner of the first aspect of the present invention, the fire spatio-temporal feature vectors are used for fire abnormal pattern recognition and evaluation to obtain a preliminary fire evaluation result for the partition, including: performing pattern feature matching on the fused fire feature based on a preset fire pattern feature library to obtain a fire type identifier and a matching confidence level, and calculating the temperature rise change rate and extracting the gas combustion stage characteristics of the fire temperature rise dynamic characteristics and combustion type gas characteristics in the fused fire feature to obtain a determination of the fire development stage and a prediction of fire evolution, and based on the fire temperature spatial diffusion characteristics and smoke diffusion dynamic characteristics in the fused fire feature, performing maximum temperature gradient positioning and concentration diffusion boundary calculation for each independent partition to obtain the fire source location and a fire influence range map, and using the fire source type visual characteristics and combustion type gas characteristics in the fused fire feature to perform fire source feature comparison and combustion substance judgment for each independent partition to obtain a fire source material and combustion feature recognition result; performing comprehensive hazard assessment and risk level division on the fire type identifier, the matching confidence level, the determination of the fire development stage, the prediction of fire evolution, the fire source location, the fire influence range map, the fire source material, and the combustion feature recognition result to obtain a fire risk level assessment; performing contextual assessment and data integration on the fire risk level assessment based on the partition scenario configuration data corresponding to each independent partition to generate a preliminary fire evaluation result for the partition.

[0010] Optionally, in the fifth implementation manner of the first aspect of the present invention, based on the building physical structure corresponding to the building to be monitored and the partition connection relationship corresponding to each independent partition, the master controller fuses the preliminary fire situation assessment results of each partition to obtain an overall fire situation assessment result, including: based on the building physical structure corresponding to the building to be monitored, performing positioning mapping of the building space coordinates on the preliminary fire situation assessment results of each independent partition to obtain an initial fire situation distribution map, and integrating the preset global combustion parameters for the fire source materials and combustion characteristic recognition results in the preliminary fire situation assessment results of the partition to obtain a variety of combustion propagation parameters, and based on the partition connection relationship corresponding to each independent partition, constructing a fire spread path network of the building to be monitored by using the initial fire situation distribution map; performing sequential calculation of the fire spread between independent partitions on the fire spread path network by using the combustion propagation parameters to obtain the fire spread prediction at multiple time points, and based on the fire spread prediction, performing isolation breakage risk assessment and multi-dimensional data integration on a variety of fire prevention and isolation facilities in the building physical structure to obtain an overall fire situation assessment result.

[0011] Optionally, in the sixth implementation manner of the first aspect of the present invention, based on the overall fire situation assessment result, the master controller generates a multi-partition collaborative fire extinguishing execution plan and transmits it to the corresponding partition controller, including: dividing the overall fire situation assessment result into fire risk levels to generate a partition response level allocation plan, and selecting a fire extinguishing medium and matching an inhibition strategy for the overall fire situation assessment result to generate a targeted fire extinguishing plan, and based on the fire protection resource configuration map preset in each independent partition, sorting the execution priorities and calculating the time nodes for the time series data in the overall fire situation assessment result to obtain a fire fighting equipment control time sequence table, and determining key points for the propagation path in the overall fire situation assessment result to generate a fire isolation control plan, and predicting the smoke flow direction and planning the clearance path for the overall fire situation assessment result to generate a smoke exhaust and ventilation control strategy, and planning a safe passage for the overall fire situation assessment result to generate a personnel evacuation guidance plan; based on the physical positions and control areas of each partition controller, screening the control instructions for the partition response level allocation plan, the targeted fire extinguishing plan, the fire fighting equipment control time sequence table, the fire isolation control plan, the smoke exhaust and ventilation control strategy and the personnel evacuation guidance plan to obtain a partition-level execution instruction set, and coordinating the execution time sequence and setting the linkage trigger threshold for the partition-level execution instruction set to generate a multi-partition collaborative fire extinguishing execution plan.

[0012] The second aspect of the present invention provides a partition controller linkage device, which is applied to a partition controller linkage system. The partition controller linkage system includes a plurality of partition controllers and a main controller. The partition controller linkage device includes: a preprocessing module, configured to obtain the original fire sensing data collected by various fire sensors in the corresponding partitions received by the partition controllers of each independent partition in the building to be monitored, and use each partition controller to perform preprocessing of multiple sensing types on the original fire sensing data to obtain the preprocessed original fire sensing data. The area to be monitored includes a plurality of independent partitions, and each independent partition is configured with a partition controller and a variety of fire sensors deployed; a feature fusion module, configured to extract the features in the time and space dimensions of the preprocessed original fire sensing data, generate the fire spatio-temporal feature vectors of each independent partition, and perform multi-source evidence weight fusion within each independent partition on the fire spatio-temporal feature vectors of each independent partition to obtain the fused fire features; a situation assessment module, configured to perform fire anomaly pattern recognition and assessment on the fire spatio-temporal feature vectors to obtain the preliminary fire assessment results of each partition, and transmit the preliminary fire assessment results of each partition controller to the main controller, and based on the building physical structure corresponding to the building to be monitored and the partition connection relationship corresponding to each independent partition, use the main controller to fuse the preliminary fire assessment results of each partition to obtain the overall fire situation assessment result; a linkage execution module, configured to generate a multi-partition collaborative fire extinguishing execution plan based on the overall fire situation assessment result by using the main controller and transmit it to the corresponding partition controller, and based on the multi-partition collaborative fire extinguishing execution plan, each partition controller controls the fire fighting equipment in the corresponding independent partition to perform fire fighting linkage rescue to obtain the fire fighting linkage result.

[0013] The third aspect of the present invention provides a partition controller linkage device, including: a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor calls the instructions in the memory so that the partition controller linkage device executes each step of the above-mentioned partition controller linkage method.

[0014] The fourth aspect of the present invention provides a computer-readable storage medium, in which instructions are stored. When it runs on a computer, it causes the computer to execute each step of the above-mentioned partition controller linkage method.

[0015] The above-mentioned partition controller linkage method, device, equipment and storage medium. In the embodiments of the present invention, by performing multi-source acquisition and type preprocessing on the fire sensor data of multiple independent partitions in the building to be monitored, the original fire sensing data after preprocessing is obtained. Then, spatio-temporal feature extraction and multi-source evidence weight fusion are performed on these data to obtain the fused fire feature. Next, fire anomaly pattern recognition and evaluation are carried out to generate the preliminary fire evaluation result of the partition, and the global fire situation is fused in combination with the building physical structure and partition connection relationship to form the overall fire situation evaluation result. Finally, a multi-partition collaborative fire extinguishing execution plan is generated based on the situation evaluation result, and the fire fighting equipment in each partition is controlled for linkage extinguishing. Through hierarchical data processing and feature analysis, accurate identification of building fire monitoring is realized. Especially in the aspects of multi-sensor data fusion and fire situation assessment, the dynamic characteristics and propagation laws of fire development are fully considered, effectively improving the accuracy of fire recognition; and a multi-level feature fusion and partition linkage strategy is adopted, which not only realizes accurate fire judgment within the partition, but also enhances the collaborative protection ability of the whole building; in addition, through global situation assessment and multi-partition collaborative execution, fire fighting resources are accurately organized and protection measures are optimized, so as to overall realize the efficient linkage prevention and control of building fires.

[0016] Other features and advantages of the present invention will be described in the following specification, and in part will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the specification, claims and drawings.

[0017] To make the above objectives, features and advantages of the present invention more obvious and understandable, the following specific preferred embodiments are given, and in conjunction with the accompanying drawings, the detailed description is as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 Schematic diagram of the first embodiment of the partition controller linkage method in the embodiments of the present invention; Figure 2 Schematic diagram of an embodiment of the partition controller linkage device in the embodiments of the present invention; Figure 3 Schematic diagram of an embodiment of the partition controller linkage equipment in the embodiments of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0020] As used in the embodiments of the present invention, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes other steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products, or devices.

[0021] For ease of understanding of this embodiment, the following describes the specific process of the embodiments of the present invention. Please refer to Figure 1 , the first embodiment of the partition controller linkage method in the embodiments of the present invention includes: 101. Obtain the original fire sensing data collected by various fire sensors in the corresponding partitions received by the partition controllers of each independent partition in the building to be monitored, and use each partition controller to perform preprocessing of various sensing types on the original fire sensing data to obtain the preprocessed original fire sensing data. The area to be monitored includes multiple independent partitions, and each independent partition is configured with a partition controller and a variety of fire sensors deployed; The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is a theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results.

[0022] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0023] In this embodiment, the original fire sensing data includes original temperature data, original smoke data, original harmful gas concentration data, and original flame image stream. The original temperature data is subjected to smoothing elimination of heat source fluctuations and extraction of fire heat characteristics to obtain smoothed temperature data. The original smoke data is subjected to elimination of ventilation disturbances and improvement of smoke purity to obtain stable smoke concentration data. The original harmful gas concentration data is subjected to elimination of interference from non-fire gases and extraction of fire characteristic gases to obtain corrected harmful gas concentration data. The original flame image stream is subjected to filtering of non-fire light sources and enhancement of flame images to obtain enhanced flame image data. Based on the sensing types of the sensors connected to each partition controller, the smoothed temperature data, stable smoke concentration data, corrected harmful gas concentration data, and enhanced flame image data are subjected to acquisition time calibration and data timing matching to obtain multi-source fire data with time alignment. Then, the multi-source fire data with time alignment is subjected to unification of fire fighting indicators and quantification of fire characteristics to obtain the preprocessed original fire sensing data.

[0024] In practical applications, first, the entire building to be monitored is pre-divided into multiple functionally independent zones, such as office areas, computer rooms, warehouses, etc. Each independent zone is equipped with a dedicated zone controller responsible for managing all the sensing devices and actuating devices in that area. And various types of fire sensors are deployed within each zone, including temperature sensors, smoke sensors, harmful gas sensors, and flame image acquisition devices. These sensors are distributed at key positions in the zone, such as ceilings, corners, around equipment, and near ventilation openings, forming a three-dimensional monitoring network. The zone controller establishes connections with various sensors through field buses or wireless communication methods to obtain real-time raw fire sensing data (the raw fire sensing data includes raw temperature data, raw smoke data, raw harmful gas concentration data, and raw flame image streams). Among them, for the raw temperature data collected by the temperature sensors, which contains fluctuations and noises generated by normal heat sources indoors (such as air conditioners, electrical equipment), an adaptive median filtering algorithm is used to smooth the temperature data, thereby automatically adjusting the filtering window size according to the temperature change rate, effectively suppressing the fluctuations of slow-changing normal heat sources, while retaining the characteristics of sudden fire temperature rise. And the temperature gradient analysis method is applied to extract the fast-rising mode with fire characteristics by calculating the first and second derivatives of temperature in the time dimension, and finally obtain smooth temperature data that highlights fire characteristics (i.e., smoothed temperature data); and for the raw data collected by the smoke sensors, since air conditioning air supply, fresh air systems, and human activities may all cause short-term fluctuations in smoke concentration, affecting the accuracy of judgment, it is necessary to handle the influence brought by the air flow disturbance of the ventilation system. Through a filtering method based on frequency domain analysis, the periodic disturbances generated by the ventilation system are identified and removed through Fourier transform. At the same time, combined with time series correlation analysis, the random fluctuations caused by ventilation are distinguished from the continuous smoke growth mode generated by real fires to improve the purity of the smoke concentration data and obtain a more stable and reliable smoke concentration index; and for the data collected by the harmful gas sensors, since it is necessary to distinguish the characteristic gases generated by fires from the interfering gases existing in the environment, and vehicle exhaust, industrial emissions, kitchen fumes, etc. in the daily environment may all contain gases such as carbon monoxide and carbon dioxide, which are similar in gas composition to those generated by fires. By adopting the multi-gas ratio analysis method, by calculating the ratio of carbon monoxide / carbon dioxide and its change trend, combined with the analysis of gas concentration growth rate, the gases from fire combustion and non-fire sources are effectively distinguished, and the data is dynamically corrected through a pre-established gas baseline adaptive correction mechanism (this mechanism can regularly learn the normal gas concentration level in the environment), and finally obtain harmful gas concentration data that accurately reflects fire characteristics;For the collected flame image data, since the original flame images are often interfered by non-fire light sources such as sunlight changes, indoor lighting, and reflective objects, through a flame recognition algorithm based on spectral and color features, analyze the unique flicker frequency (3 - 10 Hz) and color distribution pattern of the flame, effectively distinguish real flames from other light sources, and at the same time apply image enhancement technology. Through adaptive histogram equalization and edge sharpening processing, enhance the visibility of flame features, improve the sensitivity and accuracy of fire source recognition, and generate enhanced flame image data. Furthermore, by pre-setting a unified clock source synchronization mechanism, provide a standard time reference for all sensors, and combine data resampling technology to convert sensor data with different sampling frequencies to a unified time scale. Thus, through interpolation and time window matching, achieve precise temporal alignment of different types of sensing data, ensure the consistency of multi-source data in the time dimension, and perform standardization processing on the temporally aligned multi-source fire data. Convert sensing data with different physical quantities and dimensions into a unified fire protection index system, and through feature quantization technology, map the sensing data to a standard scale of 0 - 100 to obtain the preprocessed original fire sensing data.;

[0025] 102. Extract the features of the time and space dimensions of the preprocessed original fire sensing data, generate the fire spatio-temporal feature vectors for each independent partition, and perform multi-source evidence weight fusion within each partition on the fire spatio-temporal feature vectors of each independent partition to obtain the fused fire features; In this embodiment, the pre - processed original fire - sensing data is divided into multi - scale time windows to obtain a multi - scale fire - observation data set; the temperature rise change rate is calculated and the acceleration is extracted from the temperature data in the multi - scale fire - observation data set to obtain the dynamic characteristics of fire temperature rise, and the spatial distribution is calculated and the propagation direction is identified for the temperature data in the multi - scale fire - observation data set to obtain the spatial diffusion characteristics of fire temperature, and the concentration growth rate is calculated and the distribution pattern is extracted for the smoke data in the multi - scale fire - observation data set to obtain the dynamic characteristics of smoke diffusion, and the gas ratio is calculated and the change trend is extracted for the harmful gas sensor data in the multi - scale fire - observation data set to obtain the gas characteristics of combustion types, and the characteristic parameters are extracted and the combustion characteristics are identified for the flame images in the multi - scale fire - observation data set to obtain the visual characteristics of fire source types; the dynamic characteristics of fire temperature rise, the spatial diffusion characteristics of fire temperature, the dynamic characteristics of smoke diffusion, the gas characteristics of combustion types, and the visual characteristics of fire source types are subjected to characteristic combination and unified format conversion for each independent partition to obtain the spatio - temporal fire - situation feature vectors for each independent partition; based on the sensing types and characteristic categories of the sensors connected to the corresponding partition controllers, the spatio - temporal fire - situation feature vectors are decomposed to obtain a fire - situation evidence source set composed of multiple independent fire - situation evidence sources, and based on a preset historical performance record library, the weight of each evidence source in the fire - situation evidence source set is calculated to obtain a weighted fire - situation evidence set; each evidence in the weighted fire - situation evidence set is subjected to fire - situation probability quantization and unified fire - fighting index conversion to obtain an initial fire - situation probability distribution, and cross - comparison and conflict measurement are performed on various fire - situation indexes in the initial fire - situation probability distribution to obtain a fire - situation evidence conflict matrix; threshold comparison and maximum conflict value positioning are performed on the fire - situation evidence conflict matrix to obtain a corrected fire - situation probability distribution, and based on the corresponding weight values in the weighted fire - situation evidence set, each evidence source in the fire - situation evidence source set is sorted to obtain an evidence priority sequence, and based on the evidence priority sequence, the fire - situation indexes in the corrected fire - situation probability distribution are subjected to evidence combination and feature integration to obtain the fused fire - situation features. The above - mentioned historical performance record library refers to a library that stores the detection performance data of different types of sensors in various fire scenarios, including indexes such as true positive rate, false positive rate, and detection delay.

[0026] In practical applications, first, multi-scale time window partitioning is implemented on the preprocessed original fire sensing data. That is, by simultaneously constructing three observation windows with different time scales: a short time window (about 10 seconds), a medium time window (about 1 minute), and a long time window (about 10 minutes). This multi-scale partitioning can take into account both the real-time and trend characteristics of fire changes. The short time window is used to capture the instantaneous characteristics of sudden fires, the medium time window is used to monitor the development trend of the fire, and the long time window is used to observe the overall evolution process of the fire, so as to generate a multi-scale fire observation data set containing information with different time spans. Then, based on the multi-scale fire observation data set, for the temperature data in the multi-scale fire observation data set, by calculating the first derivative (temperature rise change rate) and the second derivative (temperature rise acceleration) of the temperature, where the temperature rise change rate reflects the speed of temperature rise and will show an obvious positive value in typical fire scenarios; the temperature rise acceleration reveals the mode of temperature change, such as accelerating rise, stable rise, or decelerating rise, etc. These two indicators together constitute the dynamic characteristics of fire temperature rise. And in the spatial dimension, using the data of multiple temperature sensors distributed at different positions, calculate the spatial gradient vector of the temperature and the direction of heat flow diffusion, and construct the temperature distribution field within the entire partition through a three-dimensional interpolation algorithm, and then identify the main direction and rate of heat propagation, and complete the extraction of the spatial diffusion characteristics of fire temperature to determine the fire source location and the heat propagation trend. And for the smoke data in the multi-scale fire observation data set, by calculating the smoke concentration growth rate within different time windows based on the multi-scale data set, and comparing it with a preset threshold, while analyzing the fluctuation frequency and distribution pattern of the smoke concentration. And because the diffusion of smoke in space has obvious fluid characteristics, by establishing a spatial distribution model of smoke concentration, identify the main paths and diffusion speeds of smoke flow, and extract the dynamic characteristics of smoke diffusion through a spatio-temporal combined analysis method (to provide a basis for determining the smoke source location and predicting the diffusion path). And for the data of harmful gas sensors in the multi-scale fire observation data set, since different types of fires will produce specific gas composition ratios (for example, electrical fires usually produce a higher carbon monoxide / carbon dioxide ratio, while the combustion of organic substances often shows a lower ratio), by calculating the concentration ratios of key gases such as carbon monoxide and carbon dioxide in real time, and tracking their change trends within different time windows, and integrating this information to form the gas characteristics of the combustion type.For the flame image data in the multi-scale fire observation dataset, since flames generated by different combustible materials have different visual characteristics (e.g., the color, shape, and flickering pattern of a wood-burning flame are significantly different from those of an electrical fire or a chemical combustion), by extracting parameters such as color distribution, flickering frequency, flame shape, and boundary characteristics from the flame image sequence in the multi-scale dataset, and comparing and analyzing these visual parameters with the preset fire source type templates (referring to the pre-established standard visual feature sets of different combustible materials and fire types, which include the flame characteristic parameters of various typical fire scenarios (such as electrical fires, wood material combustion, liquid combustion, chemical substance combustion, etc.)) to identify the possible types of combustible materials and generate the visual characteristics of the fire source type; furthermore, for each independent partition, organize the dynamic characteristics of fire temperature rise, the spatial diffusion characteristics of fire temperature, the dynamic characteristics of smoke diffusion, the gas characteristics of the combustion type, and the visual characteristics of the fire source type according to the predefined data structure (e.g., the rate of temperature change may be assigned to the first few element positions of the feature vector, while the smoke diffusion characteristics are placed in subsequent specific positions), and use matrix splicing and vector normalization processing to convert the characteristics of different physical quantities and different dimensions into a unified format of numerical representation, where each feature dimension is accompanied by a confidence index reflecting the reliability of the data, and finally form a high-dimensional fire spatio-temporal feature vector, which comprehensively characterizes the temporal evolution and spatial distribution characteristics of the fire situation within the partition.

[0027] Secondly, structurally decompose the fire spatio-temporal feature vector according to the types of sensors actually connected to each partition controller and the feature categories, split the mixed high-dimensional vector into multiple independent evidence sources, and consider the physical properties and measurement principles of the sensors during the decomposition process. For example, extract the temperature-related features as a temperature evidence source separately, extract the smoke-related features as a smoke evidence source separately, and similarly, the gas features and flame visual features are also extracted as independent evidence sources respectively, forming a fire evidence source set containing multiple independent members, and each member represents the basis for fire judgment provided by a type of sensor; furthermore, based on the preset historical performance record library, use the improved AHP-entropy weight method combined model to calculate the weight of each evidence source in the fire evidence source set, and its calculation formula is: ; where, is the weight value of the i-th evidence source; is the expert evaluation weight obtained based on the analytic hierarchy process, representing the subjective evaluation result of the expert; is the objective weight calculated based on information entropy; is the accuracy rate of this evidence source in the historical data; is the average detection delay of the evidence source; α is the balance factor between subjective weight and objective weight (0 ≤ α ≤ 1); β is the detection delay penalty index (for example, in an office area with well - arranged smoke and temperature sensors, the smoke evidence source may obtain a relatively high weight (such as 0.4), while in the kitchen area, due to more smoke interference sources, its weight may decrease (such as 0.25), and the weight of the temperature evidence source will increase accordingly). Thus, through this dynamic weight assignment mechanism, a weighted fire evidence set containing weight information is generated; then, the characteristic data of each evidence source in the weighted fire evidence set is converted into a fire probability representation, that is, based on a preset probability mapping function, the characteristic values of different physical quantities (i.e., the characteristic data of each evidence source) are uniformly mapped to the fire probability values in the [0, 1] interval. At the same time, different types of indicators are converted into standardized fire evaluation indicators, such as fire confidence, fire type probability distribution, and fire development stage probability distribution, etc. These probability values and standardized indicators together constitute the initial fire probability distribution, and by cross - comparing the judgment results of different evidence sources, the degree of conflict between judgments is calculated. For example, when the temperature sensor indicates a high fire probability while the smoke sensor indicates a low fire probability, this conflict is identified and its degree is quantified. By constructing an N×N evidence source comparison matrix, the conflict values between each pair of evidence sources are recorded to form a complete fire evidence conflict matrix; then, the conflict values in the fire evidence conflict matrix are compared with a threshold to identify serious conflicts exceeding the allowable threshold, and the maximum conflict set is located through graph - theoretic algorithms to find the evidence sources that may have faults or abnormalities. Then, after determining the source of the conflict, targeted correction strategies are taken, such as reducing the weight of the evidence source with serious conflict, excluding the obviously abnormal evidence source, or adjusting the relative influence between conflicting evidences. Through these correction operations, a more consistent and reliable corrected fire probability distribution is generated. At the same time, based on the weight values in the weighted fire evidence set, all evidence sources are ranked according to their importance to generate an evidence priority sequence arranged from high to low (where this sequence determines the order and influence degree of evidence fusion); then, according to the evidence priority sequence, the ordered weighted average (OWA) algorithm is adopted: ; where, is the weight at the i - th position ( ); is the i - th evidence value sorted by priority, n is the total number of evidence sources, and the position weight is calculated using the method guided by linguistic quantifiers, and its formula is: ; Among them, Q is the fuzzy quantifier function corresponding to the linguistic quantifier (for example, for the linguistic quantifier "most", Q(r) = r^(1 / 2); for "at least half", Q(r) = r^2), so as to fuse according to the reliability order of the evidence, and perform final feature integration on the fusion result to generate a fused fire feature containing comprehensive information such as the certainty degree of the fire, the determination of the fire type, the estimation of the development stage, and the description of the spatial distribution, so as to reflect all aspects of the fire situation in the partition (for example: assume that a certain office partition is equipped with temperature, smoke, carbon monoxide gas, and flame image sensors at the same time, extract the spatio-temporal fire feature vector from the preprocessed data, and decompose it into four independent evidence sources; furthermore, based on the historical performance records, assign a weight of 0.4 to the smoke evidence source, a weight of 0.3 to the temperature evidence source, a weight of 0.2 to the gas evidence source, and a weight of 0.1 to the flame image evidence source; then map each evidence source to the fire probability, and obtain a fire probability of 0.85 indicated by the smoke, a probability of 0.25 indicated by the temperature, a probability of 0.4 indicated by the gas, and a probability of 0.1 indicated by the flame; then through conflict analysis, it is found that there is a high conflict between the smoke and other evidence (K = 0.72), and after PCR5 conflict processing, the smoke probability is corrected to 0.65. Finally, perform OWA fusion according to the evidence priority sequence (smoke > temperature > gas > flame) to obtain the fused fire feature of the partition, the comprehensive fire probability is 0.48, and the spatio-temporal feature information of each evidence source is retained).

[0028] 103. Perform fire anomaly pattern recognition and evaluation on the spatio-temporal fire feature vector to obtain the preliminary fire evaluation result of the partition, and transmit the preliminary fire evaluation results of each partition controller to the main controller. Based on the building physical structure corresponding to the building to be monitored and the partition connection relationship corresponding to each independent partition, use the main controller to fuse the preliminary fire evaluation results of each partition to obtain the overall fire situation evaluation result; In this embodiment, based on a preset fire mode feature library, pattern feature matching is performed on the fused fire characteristics to obtain a fire type identifier and a matching confidence level. In addition, the rate of temperature rise is calculated for the fire temperature rise dynamic characteristics and the gas characteristics of the combustion type in the fused fire characteristics, and the characteristics of the gas combustion stage are extracted to obtain the determination of the fire development stage and the prediction of fire evolution. Based on the fire temperature spatial diffusion characteristics and the smoke diffusion dynamic characteristics in the fused fire characteristics, the maximum value of the temperature gradient is located for each independent zone, and the concentration diffusion boundary is calculated to obtain the fire source location and the fire influence range map. Moreover, using the visual characteristics of the fire source type and the gas characteristics of the combustion type in the fused fire characteristics, the fire source characteristics are compared for each independent zone, and the combustible material is judged to obtain the identification results of the fire source material and the combustion characteristics. A comprehensive hazard assessment and risk level division are performed on the fire type identifier, the matching confidence level, the determination of the fire development stage, the prediction of fire evolution, the fire source location, the fire influence range map, the fire source material, and the identification results of the combustion characteristics to obtain the fire risk level assessment. Based on the partition scenario configuration data corresponding to each independent zone, a contextual assessment and data integration are performed on the fire risk level assessment to generate a preliminary fire situation assessment result for the partition. Based on the building physical structure corresponding to the building to be monitored, the preliminary fire situation assessment results for each independent zone are mapped to the building space coordinates to obtain the initial fire situation distribution map. In addition, the identification results of the fire source material and the combustion characteristics in the preliminary fire situation assessment results for the partition are integrated with preset global combustion parameters to obtain various combustion propagation parameters. Based on the partition connection relationship corresponding to each independent zone, the fire spread path network of the building to be monitored is constructed using the initial fire situation distribution map. The sequential calculation of the fire spread between each independent zone is performed on the fire spread path network using the combustion propagation parameters to obtain the prediction of the fire spread at multiple time points. Based on the prediction of the fire spread, a risk assessment of the isolation breakage and multi-dimensional data integration are performed on various fire prevention isolation facilities in the building physical structure to obtain the overall fire situation assessment result. The above global combustion parameters refer to a standardized parameter set that describes the physical and chemical characteristics exhibited by different materials and substances during the combustion process (such as the heat release rate and the flame propagation speed, etc.).

[0029] In practical applications, first, based on a pre-established fire pattern feature library (which contains feature templates for various typical fire types, such as electrical fires, organic fires, liquid fires, and smoldering fires, etc. Each fire type has its unique multi-dimensional feature fingerprints, including temperature rise patterns, gas release characteristics, smoke characteristics, and flame visual characteristics, etc.), a multi-level pattern matching algorithm is adopted to calculate the similarity between the fused fire characteristics and each template in the library, and a comprehensive score is carried out through various measurement methods such as cosine similarity and Mahalanobis distance. Finally, the template with the highest similarity is selected as the fire type identifier, and at the same time, a matching confidence value is generated to indicate the reliability of the recognition result; and analyze the temporal pattern of the temperature rise change rate of the fire temperature rise dynamic characteristics and combustion type gas characteristics in the fused fire characteristics. Based on the classical four-stage model of fire development (initial stage, development stage, full stage, and decline stage), perform stage matching on the current temperature rise characteristics, and at the same time analyze indicators such as the carbon monoxide / carbon dioxide ratio and oxygen consumption rate in the gas characteristics. These indicators show obvious differences in different combustion stages. For example, in the initial stage, incomplete combustion often produces a relatively high carbon monoxide concentration, while in the full stage, oxygen consumption intensifies. Thus, by comprehensively analyzing the temporal variation laws of the temperature rise dynamics and gas characteristics, the determination of the fire development stage and the prediction of fire evolution (such as the risk assessment of possible flashover or deflagration and the prediction of the development rate) are obtained; and based on the fire temperature spatial diffusion characteristics and smoke diffusion dynamic characteristics in the fused fire characteristics, use the temperature gradient distribution for reverse tracking, calculate the position of the maximum gradient value in the temperature field, which is usually closest to the real fire source. At the same time, use the spatial distribution pattern of smoke concentration, combined with the airflow dynamics model in the building, to analyze the source and path of smoke diffusion, thereby cross-verify the fire source position obtained by the temperature method, calculate the influence range of the fire, and draw the boundary contour of the fire influence by analyzing the critical value distribution of temperature and smoke concentration, forming an intuitive fire influence range map (this map not only shows the currently affected area but also indicates the possible spreading direction through gradient analysis); and use the fire source type visual characteristics and combustion type gas characteristics in the fused fire characteristics, by comparing the current gas characteristics with the preset material combustion template library, combined with the judgment result of the visual characteristics, finally determine the possible fire source material type and specific combustion characteristics (where the fire source material type and specific combustion characteristics are crucial for selecting an appropriate fire extinguishing strategy); and then use the fire type identifier, matching confidence, fire development stage determination, fire evolution prediction, fire source position, fire influence range map, fire source material, and combustion characteristic recognition results as input parameters, and calculate the comprehensive hazard degree through a multi-factor risk assessment model. The model processing formula is: ; where R is the overall risk index, is the frequency or probability of the i-th risk factor, is its severity, is its rate of change (reflecting the development trend), is its scope of influence, is the weight coefficient of the i-th risk factor, which is used to adjust the influence degree of this factor on the overall risk. β is the weight coefficient of the rate of change, and γ is the weight coefficient of the scope of influence, is the total number of risk factors. Thus, factors such as fire type, development stage, and scope of influence are comprehensively considered to calculate a unified risk score, and based on this, it is divided into five risk levels: weak, slight, medium, severe, and catastrophic (for example: in the fire situation assessment of a computer room, due to involving electrical fire (high-risk type), being in the initial development stage (medium time urgency), and the scope of influence including key equipment (high-value target), the calculated risk index is 78.6, and it is determined to be in the "severe" risk level); furthermore, based on the partition scenario configuration data corresponding to each independent partition, the fire risk assessment result is combined with the specific partition scenario for contextual assessment, that is, information such as the function type, importance level, personnel density, evacuation difficulty, and special hazard sources of the current partition is extracted from the partition configuration database, and the risk level is contextually adjusted in combination with these scenario characteristics (for example: the risk assessment of a fire of the same level in a crowded area will be higher than that in an equipment room, and the risk assessment of an area containing flammable and explosive substances will be higher than that of an ordinary area), so as to generate a more practically meaningful preliminary fire situation assessment result for the partition, which includes a comprehensive description of the fire situation status, risk level assessment, and response suggestions, etc. Finally, each partition controller transmits the generated preliminary fire situation assessment result for the partition to the main controller through a secure communication channel.

[0030] Secondly, based on the building's physical structure corresponding to the building to be monitored (this building physical structure includes detailed information such as wall positions, door and window distributions, floor heights, ventilation ducts, and fire protection facilities), the main controller performs spatial coordinate mapping on the preliminary fire situation assessment results of each independent zone, accurately locating the scattered fire situation information in the actual physical space of the building (that is, by establishing an association between the zone logical numbers and the building coordinate system, projecting fire situation information such as the fire source location, influence range, and risk level onto the building physical structure), forming an intuitive initial fire situation distribution map. This distribution map is presented in the form of a heat map, with different colors representing different risk levels and fire intensities, making the overall fire situation clearly visible; and the main controller combines the identified fire source materials and combustion characteristics in the preliminary fire situation assessment results of the zones with a preset global combustion parameter library to obtain the combustion characteristic parameters of various substances under different conditions (these parameters include key indicators such as heat release rate (heat released per unit area), flame propagation speed, radiant heat flux, smoke generation rate, and toxic gas release coefficient), and by integrating these combustion characteristic parameters, generates a comprehensive set of combustion propagation parameters to reflect the speed and manner of fire spread; and based on the zone connection relationships corresponding to each independent zone, constructs a fire spread path network for the building, where this network is represented by a graph theory model, with each zone as a node and the physical connections between zones (such as doors, windows, ventilation ducts, cable shafts, ceiling spaces, etc.) as edges. The attributes of each edge include information such as connection type, material characteristics, fire protection rating, and size. At the same time, connections are also made to the "weak points" between zones, such as non-fire doors, open channels, shared air ducts, etc. These positions are often the key paths for fire spread across zones. In this way, a complete fire spread path network is formed; then, the obtained combustion propagation parameters are used to perform a time-series calculation on the fire spread path network, and an improved Monte Carlo method is used to simulate the diffusion process of the fire between different zones, taking into account the influence of various uncertain factors, including differences in material ignition points, changes in air flow, and fluctuations in combustion states, so as to generate fire spread predictions at multiple time points (such as after 5 minutes, 15 minutes, 30 minutes), describing the possible development paths, spread speeds, and influence ranges of the fire (where these prediction results are intuitively presented in the form of dynamic heat maps and diffusion front lines); then, based on the fire spread predictions, calculate the heat load and exposure time that each fire protection isolation facility in the building physical structure, such as fire doors, firewalls, fire shutters, and fire compartments, may withstand during the development of the fire, evaluate the risk probability of their failure or damage, and pay special attention to those isolation facilities on the key paths of fire spread. The integrity of these facilities directly affects the scope and speed of fire spread, thus integrating and correlating multi-dimensional data such as the initial fire situation distribution map, fire spread path network, multi-time point fire spread predictions, and key fire protection facility risk assessments to generate a comprehensive and detailed overall fire situation assessment result.

[0031] 104. Based on the evaluation result of the overall fire situation, the main controller generates a multi-zone collaborative fire extinguishing execution plan and transmits it to the corresponding zone controllers. Based on the multi-zone collaborative fire extinguishing execution plan, each zone controller controls the fire-fighting equipment in the corresponding independent zone to carry out fire-fighting linkage rescue, and obtains the fire-fighting linkage result.

[0032] In this embodiment, the evaluation result of the overall fire situation is used to divide the fire risk level, generate a zoning response level allocation plan, and select the fire extinguishing medium and match the suppression strategy for the evaluation result of the overall fire situation to generate a targeted fire extinguishing plan. Based on the pre-set fire-fighting resource configuration map in each independent zone, the time series data in the evaluation result of the overall fire situation is sorted by execution priority and time nodes are calculated to obtain the fire-fighting equipment control time sequence table. The key points of the propagation path in the evaluation result of the overall fire situation are determined to generate a fire isolation control plan. The smoke flow direction is predicted and the evacuation path is planned for the evaluation result of the overall fire situation to generate a smoke exhaust and ventilation control strategy. The safe passage is planned for the evaluation result of the overall fire situation to generate a personnel evacuation guidance plan. Based on the physical locations and control areas of each zone controller, the control instructions of the zoning response level allocation plan, the targeted fire extinguishing plan, the fire-fighting equipment control time sequence table, the fire isolation control plan, the smoke exhaust and ventilation control strategy, and the personnel evacuation guidance plan are screened to obtain a zone-level execution instruction set, and the execution time sequence of the zone-level execution instruction set is coordinated and the linkage trigger threshold is set to generate a multi-zone collaborative fire extinguishing execution plan.

[0033] In practical applications, the main controller conducts a global-level fire risk level classification for the overall fire situation assessment results. Based on factors such as the severity of the fire, the spread speed, and the threat level, it usually includes seven levels: observation level, warning level, preparation level, alert level, vigilance level, action level, and emergency level. Considering the fire situation status of each zone and its positional relationship in the overall fire, for example, the directly burning area is usually classified as the emergency level, the adjacent areas along the fire spread path may be classified as the vigilance level or the action level, while the relatively safe distal areas may be classified as the observation level or the warning level. Each zone is divided into different response levels to generate a comprehensive zone response level allocation plan, clarifying the roles and task priorities of each zone in the coordinated action. And according to the fire type and fire source material information in the overall fire situation assessment results, targeted fire extinguishing strategies are designed (for example, electrical fires are suitable for using carbon dioxide or dry powder fire extinguishers and not suitable for water-based fire extinguishers; oil fires are suitable for using foam or dry powder fire extinguishers; metal fires require special fire extinguishers for metal fires). Thus, the most suitable fire extinguishing plan for the current fire characteristics is selected from the fire extinguishing strategy library, and the required fire extinguishing dosage and release method are determined according to the fire scale and development trend, finally generating a targeted fire extinguishing plan including the selection of fire extinguishing medium, release location, dosage estimation, and operation method. Also, the fire protection resource configuration information of each zone is extracted from the building fire protection system database, including the type, quantity, location, and control method of fire extinguishing equipment, etc. Then, these resource information is matched with the time series data in the overall fire situation assessment results to calculate the optimal startup time and operation sequence of each fire protection equipment, and then the startup of direct fire extinguishing and fire control equipment is prioritized, such as sprinkler systems, gas fire extinguishing devices, etc. Subsequently, the activation of protective equipment is arranged, such as fireproof rolling shutters, smoke prevention and exhaust facilities, etc., to generate a detailed fire protection equipment control time sequence table, clearly specifying the startup time, operation duration, and associated conditions of each equipment; and for the propagation path in the overall fire situation assessment results, key fire breakthrough points and propagation bottlenecks are identified, such as the locations of fire doors, pipe penetration points, ventilation valves, and cable wells, etc. Then, corresponding isolation control strategies are formulated for these key points, such as closing the fire doors in specific areas, lowering the fireproof rolling shutters, cutting off the ventilation system, or starting local gas fire extinguishing devices, etc., to form a complete fire isolation control plan to establish an effective fireproof isolation belt around the fire source area and block the spread of the fire to a larger area. And based on the building structure data and fire development prediction in the overall fire situation assessment results, the flow path and diffusion pattern of smoke in the building are simulated. Thus, for the predicted smoke flow direction, the optimal smoke exhaust path is planned, and by controlling the combined operation of smoke exhaust fans, positive pressure air supply systems, and ventilation valves, the smoke is guided to be discharged from the building along the predetermined path. At the same time, positive pressure areas are established in key areas (such as evacuation passages, refuge floors) to prevent the intrusion of smoke and ensure the safety of personnel. These strategies together constitute the smoke exhaust and ventilation control strategy;Based on the evaluation results of the overall fire situation, identify the areas affected by the fire and smoke, determine the unsafe areas and safe areas, and then calculate the optimal evacuation routes for the personnel in each area based on the evacuation passage layout of the building, avoiding dangerous areas and optimizing the evacuation efficiency. At the same time, considering the personnel density, differences in mobility, and the needs of special groups, formulate a hierarchical and batch-by-batch evacuation plan, and plan the control strategy of the evacuation indication system, including emergency broadcast content, evacuation indicator display, and elevator operation mode, etc., to form a comprehensive personnel evacuation guidance plan; furthermore, based on the above six global plans, the master controller decomposes and screens the formulated plans according to the physical locations and control scopes of each zone controller, extracts the control instructions directly related to each zone, generates a zone-level execution instruction set (where the instruction set for each zone contains specific operations, execution times, expected effects, etc. that need to be executed in that zone), and coordinates the timing and optimizes the triggering conditions of the execution instruction sets for each zone to ensure the synchronization and coherence of multi-zone operations, especially for fine-tuning cross-zone collaborative operations (such as linked smoke exhaust, pressure balance, and evacuation guidance), setting appropriate linked trigger thresholds and interlock conditions, and finally forming a complete multi-zone collaborative fire extinguishing execution plan, and then transmitting the corresponding parts to each zone controller through a secure communication channel. After receiving the execution plan, each zone controller controls the fire-fighting equipment in its jurisdiction to perform corresponding operations, including starting fire extinguishing equipment, controlling fire isolation facilities, adjusting the smoke exhaust system, and activating evacuation indicators, etc. Through this collaborative and linked method, a comprehensive and efficient control of the fire is achieved, and finally an ideal fire-fighting linkage result is obtained.;

[0034] In the embodiment of the present invention, through multi-source collection and type preprocessing of the fire-fighting sensor data of multiple independent zones in the building to be monitored, the original fire situation sensing data after preprocessing is obtained, and then the spatio-temporal features of these data are extracted and the multi-source evidence weights are fused to obtain the fused fire situation features; then the fire abnormal mode is identified and evaluated to generate the preliminary fire situation evaluation results for each zone, and the global fire situation is fused in combination with the building physical structure and the zone connection relationship to form the overall fire situation evaluation results; finally, based on the situation evaluation results, a multi-zone collaborative fire extinguishing execution plan is generated to control the fire-fighting equipment in each zone to carry out joint fire fighting. Through hierarchical data processing and feature analysis, accurate identification of the building fire situation is realized. Especially in the aspect of multi-sensor data fusion and fire situation assessment, the dynamic characteristics and propagation laws of fire development are fully considered, effectively improving the accuracy of fire situation identification; and a multi-level feature fusion and zone linkage strategy is adopted, which not only realizes accurate fire situation judgment within the zone, but also enhances the collaborative protection ability of the overall building; in addition, through global situation assessment and multi-zone collaborative execution, fire-fighting resources are accurately organized and protection measures are optimized, so as to achieve efficient linkage prevention and control of building fires as a whole.

[0035] The above describes the zoning controller linkage method in the embodiments of the present invention. Next, the zoning controller linkage device in the embodiments of the present invention will be described. Please refer to Figure 2 An embodiment of the zoning controller linkage device in the embodiments of the present invention includes: A preprocessing module 201, configured to obtain the original fire sensing data collected by a variety of fire sensors in the corresponding partition from the partition controllers of each independent partition in the building to be monitored, and use each of the partition controllers to perform preprocessing of multiple sensing types on the original fire sensing data to obtain preprocessed original fire sensing data. The area to be monitored includes multiple independent partitions, and each independent partition is configured with a partition controller and a variety of fire sensors are deployed; A feature fusion module 202, configured to extract the time and space dimension features of the preprocessed original fire sensing data, generate the fire spatio-temporal feature vectors of each independent partition, and perform multi-source evidence weight fusion within the partition on the fire spatio-temporal feature vectors of each independent partition to obtain fused fire features; A situation assessment module 203, configured to perform fire abnormal pattern recognition and assessment on the fire spatio-temporal feature vectors to obtain a preliminary fire assessment result of the partition, and transmit the preliminary fire assessment results of each partition controller to the main controller, and based on the building physical structure corresponding to the building to be monitored and the partition connection relationship corresponding to each independent partition, use the main controller to fuse the preliminary fire assessment results of each partition to obtain an overall fire situation assessment result; A linkage execution module 204, configured to generate a multi-partition collaborative fire extinguishing execution plan based on the overall fire situation assessment result by using the main controller and transmit it to the corresponding partition controller, and based on the multi-partition collaborative fire extinguishing execution plan, each partition controller controls the fire fighting equipment in the corresponding independent partition to perform fire fighting linkage rescue to obtain a fire fighting linkage result.

[0036] In the embodiments of the present invention, by performing multi-source acquisition and type preprocessing on the fire sensor data of multiple independent zones in the building to be monitored, the original fire sensing data after preprocessing is obtained. Then, spatio-temporal feature extraction and multi-source evidence weight fusion are performed on these data to obtain the fused fire characteristics. Next, fire anomaly pattern recognition and evaluation are carried out to generate the preliminary fire situation evaluation results for each zone, and the global fire situation is fused in combination with the building physical structure and the connection relationship between zones to form the overall fire situation evaluation result. Finally, a multi-zone collaborative fire extinguishing execution plan is generated based on the situation evaluation result, and the fire-fighting equipment in each zone is controlled for linkage extinguishing. Through hierarchical data processing and feature analysis, accurate identification of the building fire situation is achieved. Especially in the aspects of multi-sensor data fusion and fire situation assessment, the dynamic characteristics and propagation laws of fire development are fully considered, effectively improving the accuracy of fire situation identification. And by adopting a multi-level feature fusion and zone linkage strategy, accurate fire situation judgment within the zone is realized, and the collaborative protection ability of the overall building is enhanced. In addition, through global situation assessment and multi-zone collaborative execution, fire-fighting resources are accurately organized and protection measures are optimized, thus realizing the efficient linkage prevention and control of building fires as a whole.

[0037] Above Figure 2 The linkage device of the zone controller in the embodiments of the present invention is described in detail from the perspective of modular functional entities. Next, the zone controller linkage device in the embodiments of the present invention is described in detail from the perspective of hardware processing.

[0038] Figure 3 FIG. is a schematic structural diagram of a zone controller linkage device provided by an embodiment of the present invention. The zone controller linkage device 300 may vary greatly due to configuration or performance differences, and may include one or more processors (central processing units, CPUs) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 for storing application programs 333 or data 332 (for example, one or more mass storage devices). Among them, the memory 320 and the storage media 330 may be transient storage or persistent storage. The program stored in the storage media 330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the zone controller linkage device 300. Further, the processor 310 may be configured to communicate with the storage media 330 and execute a series of instruction operations in the storage media 330 on the zone controller linkage device 300.

[0039] The partition controller linkage device 300 may further include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input / output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, and so on. Those skilled in the art can understand that Figure 3 The shown structure of the partition controller linkage device does not constitute a limitation on the partition controller linkage device, and it may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0040] The present invention also provides a partition controller linkage device. The computer device includes a memory and a processor. When computer-readable instructions stored in the memory are executed by the processor, the processor executes each step of the partition controller linkage method in the above-mentioned various embodiments.

[0041] The present invention also provides a computer-readable storage medium. The computer-readable storage medium may be a non-volatile computer-readable storage medium, or may also be a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions run on a computer, the computer is enabled to execute each step of the partition controller linkage method.

[0042] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated here.

[0043] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions to enable a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0044] This application can be used in numerous general-purpose or special-purpose computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This application can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0045] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for linkage of partition controllers, which is applied to a partition controller linkage system, and is characterized in that, The partition controller linkage system includes multiple partition controllers and a main controller. The partition controller linkage method includes: Obtaining the original fire sensing data collected by various fire sensors in the corresponding partitions received by the partition controllers of each independent partition in the building to be monitored, and using each of the partition controllers to perform preprocessing of multiple sensing types on the original fire sensing data to obtain the preprocessed original fire sensing data. The area to be monitored includes multiple independent partitions, and each independent partition is configured with a partition controller and is deployed with various fire sensors; Extracting the features of the time and space dimensions of the preprocessed original fire sensing data to generate the fire spatio-temporal feature vectors of each independent partition, and performing multi-source evidence weight fusion within each independent partition on the fire spatio-temporal feature vectors of each independent partition to obtain the fused fire features; Performing fire anomaly pattern recognition and evaluation on the fire spatio-temporal feature vectors to obtain the preliminary fire evaluation results of each partition, and transmitting the preliminary fire evaluation results of each partition controller to the main controller. Based on the building physical structure corresponding to the building to be monitored and the partition connection relationship corresponding to each independent partition, using the main controller to perform global fire situation fusion on the preliminary fire evaluation results of each partition to obtain the overall fire situation evaluation result; Based on the overall fire situation evaluation result, using the main controller to generate a multi-partition collaborative fire extinguishing execution plan and transmit it to the corresponding partition controller. Based on the multi-partition collaborative fire extinguishing execution plan, each partition controller controls the fire fighting equipment in the corresponding independent partition to perform fire fighting linkage rescue to obtain the fire fighting linkage result.

2. The partition controller linkage method according to claim 1, wherein The original fire sensing data includes original temperature data, original smoke data, original harmful gas concentration data, and original flame image stream. Using each of the partition controllers to perform preprocessing of multiple sensing types on the original fire sensing data to obtain the preprocessed original fire sensing data includes: Performing smoothing elimination of heat source fluctuations and extraction of fire heat characteristics on the original temperature data to obtain smoothed temperature data, performing ventilation disturbance removal and smoke purity improvement on the original smoke data to obtain stable smoke concentration data, performing interference removal of non-fire gases and extraction of fire characteristic gases on the original harmful gas concentration data to obtain corrected harmful gas concentration data, and performing filtering of non-fire light sources and flame image enhancement on the original flame image stream to obtain flame image enhancement data; Based on the sensing types of the sensors corresponding to each partition controller, performing acquisition time calibration and data time series matching on the smoothed temperature data, the stable smoke concentration data, the corrected harmful gas concentration data, and the flame image enhancement data to obtain time-aligned multi-source fire data, and performing fire fighting index unification and quantification of fire characteristics on the time-aligned multi-source fire data to obtain the preprocessed original fire sensing data.

3. The zoning controller linkage method according to claim 2, wherein The extracting the features of the time and space dimensions of the preprocessed original fire sensing data to generate the fire spatio-temporal feature vectors of each independent partition includes: Divide the preprocessed original fire sensing data into multi-scale time windows to obtain a multi-scale fire observation data set; Calculate the temperature rise change rate and extract the acceleration of the temperature data in the multi-scale fire observation data set to obtain the dynamic characteristics of fire temperature rise, and calculate the spatial distribution and identify the propagation direction of the temperature data in the multi-scale fire observation data set to obtain the spatial diffusion characteristics of fire temperature, and calculate the concentration growth rate and extract the distribution pattern of the smoke data in the multi-scale fire observation data set to obtain the dynamic characteristics of smoke diffusion, and calculate the gas ratio and extract the change trend of the harmful gas sensor data in the multi-scale fire observation data set to obtain the gas characteristics of combustion types, and extract the characteristic parameters and identify the combustion characteristics of the flame images in the multi-scale fire observation data set to obtain the visual characteristics of fire source types; Perform feature combination and unified format conversion for each independent partition on the dynamic characteristics of fire temperature rise, the spatial diffusion characteristics of fire temperature, the dynamic characteristics of smoke diffusion, the gas characteristics of combustion types, and the visual characteristics of fire source types to obtain the fire spatio-temporal feature vectors for each independent partition.

4. The method for interlocking partition controllers according to claim 1, wherein, The multi-source evidence weight fusion within each independent partition of the fire spatio-temporal feature vectors to obtain the fused fire characteristics includes: Based on the sensing types and characteristic categories of the sensors connected to each partition controller, decompose the fire spatio-temporal feature vectors to obtain a fire evidence source set composed of multiple independent fire evidence sources, and calculate the weight of each evidence source in the fire evidence source set based on a preset historical performance record library to obtain a weighted fire evidence set; Quantify the fire probability and perform unified fire protection index conversion for each evidence in the weighted fire evidence set to obtain an initial fire probability distribution, and perform cross-comparison and conflict measurement on various fire index in the initial fire probability distribution to obtain a fire evidence conflict matrix; Perform threshold comparison and maximum conflict value positioning on the fire evidence conflict matrix to obtain a corrected fire probability distribution, and sort each evidence source in the fire evidence source set based on the corresponding weight values in the weighted fire evidence set to obtain an evidence priority sequence, and perform evidence combination and feature integration on the fire index in the corrected fire probability distribution based on the evidence priority sequence to obtain the fused fire characteristics.

5. The zoning controller linkage method according to claim 3, wherein The fire spatio-temporal feature vectors are used for fire abnormal pattern recognition and evaluation to obtain a preliminary fire evaluation result for each partition, including: Based on a preset fire mode feature library, perform pattern feature matching on the fused fire situation features to obtain a fire type identifier and a matching confidence level, and calculate the temperature rise change rate and extract the gas combustion stage features of the fire temperature rise dynamic feature and the combustion type gas feature in the fused fire situation features to obtain a determination of the fire development stage and a prediction of fire evolution. Also, based on the fire temperature spatial diffusion feature and the smoke diffusion dynamic feature in the fused fire situation features, perform positioning of the maximum temperature gradient and calculation of the concentration diffusion boundary for each independent partition to obtain the fire source location and a fire influence range map. Additionally, use the fire source type visual feature and the combustion type gas feature in the fused fire situation features to perform a comparison of fire source features and a determination of the combustible material for each independent partition to obtain an identification result of the fire source material and combustion characteristics; Perform a comprehensive hazard assessment and risk level classification on the fire type identifier, the matching confidence level, the determination of the fire development stage, the prediction of fire evolution, the fire source location, the fire influence range map, the fire source material, and the identification result of the combustion characteristics to obtain a fire risk level assessment; Based on the partition scenario configuration data corresponding to each independent partition, perform a contextual assessment and data integration on the fire risk level assessment to generate a preliminary fire situation assessment result for the partition.

6. The zoning controller linkage method according to claim 1, wherein Based on the building physical structure corresponding to the building to be monitored and the partition connection relationships corresponding to each independent partition, use the main controller to fuse the preliminary fire situation assessment results of each partition to obtain an overall fire situation assessment result, including: Based on the building physical structure corresponding to the building to be monitored, perform a positioning mapping of the building space coordinates for the preliminary fire situation assessment results of each independent partition to obtain an initial fire situation distribution map, and integrate the preset global combustion parameters for the identification results of the fire source material and combustion characteristics in the preliminary fire situation assessment results of the partition to obtain multiple combustion propagation parameters. Also, based on the partition connection relationships corresponding to each independent partition, use the initial fire situation distribution map to construct a fire spread path network for the building to be monitored; Use the combustion propagation parameters to perform a time-series calculation of the fire spread between independent partitions for the fire spread path network to obtain a fire spread prediction at multiple time points. And based on the fire spread prediction, perform an isolation breakage risk assessment and multi-dimensional data integration on various fire prevention isolation facilities in the building physical structure to obtain an overall fire situation assessment result.

7. The method for linkage of the zoning controller according to claim 6, wherein Based on the overall fire situation assessment result, use the main controller to generate a multi-partition collaborative fire extinguishing execution plan and transmit it to the corresponding partition controller, including: Conduct a fire risk level division for the evaluation results of the overall fire situation, generate a zoning response level allocation plan, and select a fire extinguishing medium and match a suppression strategy for the evaluation results of the overall fire situation to generate a targeted fire extinguishing plan. Based on the pre-set fire protection resource allocation maps in each independent zone, perform a priority ranking and time node calculation on the time series data in the evaluation results of the overall fire situation to obtain a fire fighting equipment control time sequence table. Determine key points for the propagation path in the evaluation results of the overall fire situation to generate a fire isolation control plan. Predict the smoke flow direction and plan a clearing path for the evaluation results of the overall fire situation to generate a smoke exhaust and ventilation control strategy. Plan a safe passage for the evaluation results of the overall fire situation to generate a personnel evacuation guidance plan; Based on the physical locations and control areas of each zone controller, screen control instructions for the zoning response level allocation plan, the targeted fire extinguishing plan, the fire fighting equipment control time sequence table, the fire isolation control plan, the smoke exhaust and ventilation control strategy, and the personnel evacuation guidance plan to obtain a zone-level execution instruction set, and perform execution time sequence coordination and linkage trigger threshold setting on the zone-level execution instruction set to generate a multi-zone collaborative fire extinguishing execution plan.

8. A partition controller linkage device is applied to a partition controller linkage system, and is characterized in that The zone controller linkage system includes multiple zone controllers and a main controller, and the zone controller linkage device includes: A preprocessing module, configured to obtain the original fire situation sensing data collected by various fire sensors in the corresponding zone received by the zone controllers of each independent zone in the building to be monitored, and use each zone controller to perform preprocessing of various sensing types on the original fire situation sensing data to obtain preprocessed original fire situation sensing data. The area to be monitored includes multiple independent zones, and each independent zone is configured with a zone controller and is deployed with various fire sensors; A feature fusion module, configured to extract the time and space dimension features of the preprocessed original fire situation sensing data to generate a fire situation spatio-temporal feature vector for each independent zone, and perform multi-source evidence weight fusion within each independent zone on the fire situation spatio-temporal feature vectors of each independent zone to obtain a fused fire situation feature; A situation evaluation module, configured to perform fire anomaly pattern recognition and evaluation on the fire situation spatio-temporal feature vector to obtain a preliminary fire situation evaluation result for each zone, and transmit the preliminary fire situation evaluation results of each zone controller to the main controller, and based on the building physical structure corresponding to the building to be monitored and the zone connection relationship corresponding to each independent zone, use the main controller to perform global fire situation integration on the preliminary fire situation evaluation results of each zone to obtain an overall fire situation evaluation result; A linkage execution module, configured to generate a multi-zone collaborative fire extinguishing execution plan based on the overall fire situation evaluation result by using the main controller and transmit it to the corresponding zone controller, and based on the multi-zone collaborative fire extinguishing execution plan, each zone controller controls the fire fighting equipment in the corresponding independent zone to perform fire fighting linkage rescue to obtain a fire fighting linkage result.

9. A partition controller linkage device, characterized in that, The partition controller linkage device includes: a memory and at least one processor, and instructions are stored in the memory; The at least one processor invokes the instructions in the memory to cause the partition controller linkage device to execute each step of the partition controller linkage method according to any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that, When the instructions are executed by the processor, each step of the partition controller linkage method according to any one of claims 1-7 is implemented.

Images