Metro station city integration fire alarm prediction method and system based on artificial intelligence

Abstract

The application discloses a subway station-city integrated fire alarm prediction method and system based on artificial intelligence, which comprises the following steps: acquiring fire history data of a subway station and commercial buildings around the subway station, determining fire influencing factors and establishing a data set; constructing a fire risk prediction model according to the data set of the fire influencing factors; classifying fire scales and dividing fire alarm grades based on the fire risk prediction model and the fire history data; and determining the fire alarm grade based on the sudden fire information and the fire risk prediction model, wherein the fire risk prediction model is a risk prediction model based on building fire area, and the classification of the fire scale and the fire alarm grade is based on the fire area. The application can make the classification and prediction of the fire alarm grade more automatic, timely and accurate, so as to ensure the accurate and rapid scheduling of the subway station-city integrated fire extinguishing and rescue force, thereby improving the fire extinguishing and rescue efficiency of the fire rescue team.

Description

Technical Field

[0001] This invention relates to a fire risk prediction and disaster prevention system for large-scale building facilities, and more specifically, to an artificial intelligence-based method and system for predicting fire alarms in an integrated urban and subway system. Background Technology

[0002] In recent years, to promote the integrated development of urban and transportation functions and achieve high-quality and sustainable development, the process of subway station-city integration has been further promoted. Subway station-city integration refers to a development model that integrates subway space with urban development and construction, creating a completely new urban space through the fusion of station and urban spaces. The main functions of subway station-city integration are transportation transfer and commercial office space. It boasts diverse business formats and wide coverage, achieving efficient transportation conversion and high-quality space development, becoming a new urban vitality center and an inevitable requirement for my country's urban construction and economic and social development. However, due to the high density of people and complex surrounding environment in subway stations, they are also high-risk fire points. Once a fire gets out of control, it will cause serious casualties, huge property losses, and severe social impact.

[0003] my country classifies fire alarms into five levels, providing qualitative descriptions, but these are difficult to quantify. Even preliminary plans lack specificity. Existing fire alarm prediction methods, while quantifying fire severity, are still based on qualitative fire-influencing factors. They assign weights to the magnitude of fire impact and then use weighted averages or computer algorithms to classify fire alarm levels. This method fails to base its classification on physical quantities that characterize the actual fire situation, easily leading to inaccurate fire alarm level predictions. This results in significant arbitrariness and blindness in the deployment of resources during firefighting and rescue operations, frequently leading to unreasonable resource allocation and directly or indirectly affecting the efficiency of firefighting and rescue efforts.

[0004] In view of this, to address the above problems, it is necessary to design an AI-based method and system for predicting fire alarms in subway stations and urban areas, so as to make the classification and prediction of fire alarm levels more automatic, timely and accurate, and ensure the accurate and rapid dispatch of fire fighting and rescue forces in subway stations and urban areas, thereby improving the fire fighting and rescue efficiency of fire rescue teams. Summary of the Invention

[0005] The purpose of this invention is to provide an artificial intelligence-based method and system for predicting fire alarms in subway stations and urban areas, in order to solve the inaccuracies in fire alarm classification and fire alarm prediction in the existing technology, so as to ensure accurate and rapid dispatch of fire fighting and rescue forces in subway stations and urban areas, thereby improving the fire fighting and rescue efficiency of fire rescue teams.

[0006] To achieve the above objectives, this invention provides an artificial intelligence-based method for predicting fire alarms in an integrated urban subway station system, comprising the following steps:

[0007] Obtain historical fire data of subway stations and surrounding commercial buildings, identify fire influencing factors, and establish a dataset;

[0008] A fire risk prediction model is constructed based on a dataset of fire influencing factors.

[0009] Based on fire risk prediction models and historical fire data, fire scales are classified and fire alarm levels are determined.

[0010] Based on sudden fire information and a fire risk prediction model, the fire alarm level is determined. The fire risk prediction model is based on the fire-affected area of ​​the building. The classification of fire scale and fire alarm level is based on the fire-affected area. Where 0 ≤ fire-affected area ≤ 10m² 2 If the fire scale is 10m, then the fire scale is determined to be Level 1; if the fire scale is 10m 2 <Fire area ≤ 50m² 2 If the fire scale is 50m, then the fire scale is determined to be level two; if the fire scale is 50m 2 <Fire area ≤ 300m² 2 If the fire scale is 300m, then the fire scale is determined to be level three; if the fire scale is 300m 2 <Fire area ≤ 500m² 2 If the fire scale is 500m, then the fire scale is determined to be level four; if the fire scale is 500m 2 If the burned area is less than the fire-affected area, the fire scale is determined to be level five; among which,

[0011] For fires of scale one or two, the fire alarm level is determined as level one; for fires of scale three, the fire alarm level is determined as level two; for fires of scale four, the fire alarm level is determined as level three; and for fires of scale five, the fire alarm level is determined as level four.

[0012] By adopting the technical solution disclosed in this invention, the fire alarm level can be accurately and quantitatively predicted, so as to ensure the accurate and rapid dispatch of fire fighting and rescue forces in the integrated subway station city, thereby improving the fire fighting and rescue efficiency of the fire rescue team.

[0013] The aforementioned AI-based integrated fire alarm prediction method for subway stations considers factors such as combustible material type, fire load density, fire alarm time, first fire response time, and burned area.

[0014] The aforementioned AI-based integrated fire alarm prediction method for subway stations and cities further includes the following steps in constructing a fire risk prediction model based on a dataset of fire influencing factors: constructing corresponding probability density distribution models based on datasets of various fire influencing factors.

[0015] The aforementioned AI-based integrated fire alarm prediction method for subway stations and cities further includes the following steps in the process of constructing a fire risk prediction model based on a dataset of fire influencing factors: Based on the dataset of fire influencing factors, a fire area calculation model is derived by means of the metric value of total fire heat release rate and the formula for calculating instantaneous fire heat release rate; and a fire risk prediction model is constructed based on the fire area calculation model and the probability density distribution model.

[0016] The aforementioned AI-based integrated fire alarm prediction method for subway stations further includes the following steps for determining the fire alarm level based on sudden fire information and fire risk prediction models: if the sudden fire information includes data on fire influencing factors, the fire alarm level is determined using the mean method or the maximum value method based on the data on fire influencing factors and the fire risk prediction model; or, if the sudden fire information does not include data on fire influencing factors, the fire alarm level is determined using the mean method or the maximum value method based on the probability density distribution model and the fire risk prediction model.

[0017] The aforementioned AI-based integrated fire alarm prediction method for subway stations and cities further includes the following steps in constructing corresponding probability density distribution models based on datasets of various fire influencing factors: preprocessing the datasets of fire influencing factors using data cleaning methods; and constructing probability density distribution models based on fire influencing factors using statistical methods based on the preprocessed datasets.

[0018] The aforementioned AI-based integrated fire alarm prediction method for subway stations and cities further includes the following steps in constructing a probability density distribution model based on fire influencing factors using statistical methods: obtaining a probability density distribution map of fire influencing factors and their occurrence frequency and / or occurrence based on the preprocessed dataset; and obtaining a probability density distribution model of fire influencing factors based on the probability density distribution map using a data processing and curve fitting method.

[0019] The aforementioned AI-based integrated fire alarm prediction method for subway stations also includes a step of automatically correcting the fire alarm level based on the on-site fire situation. The fire situation includes one or more of the following: the number of people trapped or injured, the activation status of the automatic sprinkler system, and / or the type of fire.

[0020] The aforementioned AI-based fire alarm prediction method for integrated subway stations also includes a step to automatically raise the fire alarm level when special circumstances occur in the integrated subway station area. These special circumstances include: the integrated subway station area is located during major holidays, important political events, or politically sensitive or important areas; the integrated subway station area experiences severe weather conditions; the number of fire alarm calls in the integrated subway station area continues to increase, indicating a clear trend towards disaster; a fire occurs in underground commercial buildings, high-rise buildings, or integrated complexes above subway stations in the integrated subway station area, or there is a risk of combustion and explosion within the buildings; and / or, a fire occurs in a subway car in the integrated subway station area, rendering the subway train unable to move and causing it to remain in the tunnel section.

[0021] The aforementioned AI-based integrated fire alarm prediction method for subway stations also includes the step of automatically dispatching corresponding fire fighting and rescue forces based on the fire alarm level.

[0022] To better achieve the objectives of this invention, this invention also provides an artificial intelligence-based integrated fire alarm prediction system for subway stations, comprising:

[0023] The data acquisition module is used to acquire historical fire data of subway stations and surrounding commercial buildings, identify fire influencing factors, and establish a dataset.

[0024] The model building module is used to build fire risk prediction models based on a dataset of fire influencing factors.

[0025] The classification module is used to classify the scale of fires and divide fire alarm levels based on fire risk prediction models and historical fire data.

[0026] The determination module is used to determine the fire alarm level based on sudden fire information and a fire risk prediction model. The fire risk prediction model is a risk prediction model based on the burned area of ​​the building. The classification of fire scale and fire alarm level is based on the burned area.

[0027] If 0 ≤ burned area ≤ 10m² 2 If the fire scale is 10m, then the fire scale is determined to be Level 1; if the fire scale is 10m 2 <Fire area ≤ 50m² 2 If the fire scale is 50m, then the fire scale is determined to be level two; if the fire scale is 50m 2 <Fire area ≤ 300m² 2 If the fire scale is 300m, then the fire scale is determined to be level three; if the fire scale is 300m 2 <Fire area ≤ 500m² 2 If the fire scale is 500m, then the fire scale is determined to be level four; if the fire scale is 500m 2 If the burned area is less than the fire-affected area, the fire scale is determined to be level five; among which,

[0028] For fires of scale one or two, the fire alarm level is determined as level one; for fires of scale three, the fire alarm level is determined as level two; for fires of scale four, the fire alarm level is determined as level three; and for fires of scale five, the fire alarm level is determined as level four.

[0029] The aforementioned AI-based integrated fire alarm prediction system for subway stations and cities includes fire-related factors in its data acquisition module, such as the type of combustible material, fire load density, fire alarm time, first fire dispatch time, and burned area.

[0030] The aforementioned AI-based integrated fire alarm prediction system for subway stations and cities further includes the following steps in the model building module: constructing a fire risk prediction model based on a dataset of fire influencing factors, and constructing corresponding probability density distribution models based on datasets of various fire influencing factors.

[0031] The aforementioned AI-based integrated fire alarm prediction system for subway stations and cities further includes the following steps in its model building module: First, based on the dataset of fire influencing factors, a fire area calculation model is derived using the measurement of total fire heat release rate and the formula for calculating instantaneous fire heat release rate. Second, a fire risk prediction model is constructed based on the fire area calculation model and the probability density distribution model.

[0032] The aforementioned AI-based integrated fire alarm prediction system for subway stations and cities further includes the following steps in the determination module: if the sudden fire information includes data on fire influencing factors, the fire alarm level is determined using the mean method or the maximum value method based on the data on fire influencing factors and the fire risk prediction model; or, if the sudden fire information does not include data on fire influencing factors, the fire alarm level is determined using the mean method or the maximum value method based on the probability density distribution model and the fire risk prediction model.

[0033] The aforementioned AI-based integrated fire alarm prediction system for subway stations and cities further includes the following steps in the model building module: constructing corresponding probability density distribution models based on datasets of various fire influencing factors; preprocessing the datasets of fire influencing factors using data cleaning methods; and constructing probability density distribution models based on fire influencing factors using statistical methods based on the preprocessed datasets.

[0034] The aforementioned AI-based integrated fire alarm prediction system for subway stations and cities further includes the following steps in its model building module: constructing a probability density distribution model based on fire influencing factors using statistical methods; obtaining a probability density distribution map of fire influencing factors and their occurrence frequency and / or occurrence based on the preprocessed dataset; and obtaining a probability density distribution model of fire influencing factors based on the probability density distribution map using a data processing and curve fitting method.

[0035] The aforementioned AI-based integrated fire alarm prediction system for subway stations and cities is characterized in that the determination module is further used to automatically correct the fire alarm level based on the on-site fire situation, wherein the fire situation includes one or more of the following: the number of people trapped or injured, the activation status of the automatic sprinkler system, and / or the type of fire.

[0036] The aforementioned AI-based integrated subway station fire alarm prediction system also includes a fire alarm escalation module, which automatically raises the fire alarm level when special circumstances occur in the integrated subway station area. These special circumstances include: the integrated subway station area is located during major holidays, important political events, or politically sensitive or important areas; the integrated subway station area experiences severe weather conditions; the number of fire alarm calls in the integrated subway station area continues to increase, indicating a clear trend towards disaster; a fire occurs in underground commercial buildings, high-rise buildings, or integrated complexes above subway stations in the integrated subway station area, or there is a risk of combustion and explosion within the buildings; and / or, a fire occurs in a subway car in the integrated subway station area, rendering the subway train unable to move and causing it to remain in the tunnel section.

[0037] The aforementioned AI-based integrated fire alarm prediction system for subway stations and cities also includes a dispatch module, which automatically dispatches corresponding fire fighting and rescue forces based on the fire alarm level.

[0038] To better achieve the objectives of this invention, this invention also provides a computer-readable storage medium having a computer program stored thereon, the computer program being configured to execute the above-described method at runtime.

[0039] To better achieve the objectives of this invention, the present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the above-described method.

[0040] To better achieve the objectives of this invention, this invention also provides a fire alarm prediction intelligent robot, which is equipped with or connected via a network to the aforementioned artificial intelligence-based integrated subway station and city fire alarm prediction system.

[0041] Of course, the AI-based integrated fire alarm prediction system for subway stations and cities, storage medium, electronic device, and intelligent robot provided by this invention have the same beneficial technical effects as the methods described above.

[0042] To provide a better understanding of the above and other aspects of the present invention, specific embodiments are described below in conjunction with the accompanying drawings, but these are not intended to limit the scope of protection of the present invention. Attached Figure Description

[0043] Figure 1 This is a flowchart illustrating the steps of an AI-based urban-subway fire alarm prediction method according to an embodiment of the present invention.

[0044] Figure 2 A flowchart illustrating the steps of constructing a probability density distribution model according to an embodiment of the present invention.

[0045] Figure 3 This is a flowchart illustrating the steps of a statistical method for constructing a probability density distribution model according to an embodiment of the present invention.

[0046] Figure 4 This is a cumulative frequency curve of the fire-affected area of ​​a commercial building according to an embodiment of the present invention.

[0047] Figure 5 A flowchart illustrating the steps for determining a fire alarm level according to an embodiment of the present invention.

[0048] Figure 6 This is a structural block diagram of an AI-based integrated fire alarm prediction system for subway stations and cities, according to an embodiment of the present invention.

[0049] In the attached figures, the following labels are used:

[0050] S1~S4: Steps of the AI-based Integrated Subway City Fire Alarm Prediction Method

[0051] S21, S211~S212: Steps for constructing a probability density distribution model

[0052] S2121~S2122: Steps in the statistical methods for constructing probability density distribution models

[0053] L1 - Accumulated frequency of automatic sprinkler system when effective

[0054] L2 - Accumulated frequency when automatic sprinkler system is not configured

[0055] 1-AI-based Integrated Fire Alarm Prediction System for Subways and Cities

[0056] 11-Data Acquisition Module

[0057] 12-Model Building Module

[0058] 13-Grading Module

[0059] 14-Determine Module

[0060] 15-Fire Alarm Upgrade Module

[0061] 16-Scheduling Module Detailed Implementation

[0062] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments to further understand the purpose, solution and beneficial technical effects of the present invention, but this is not intended to limit the scope of protection of the appended claims. It should be noted that in this specification, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0063] The core of this invention is to provide an artificial intelligence-based method and system for predicting fire alarms in an integrated urban and subway system. This method classifies fire alarm levels based on the burned area and accurately predicts the fire alarm level by combining information on sudden fires.

[0064] Please see Figure 1 , Figure 1 This is a flowchart illustrating the steps of an AI-based integrated fire alarm prediction method for subway stations according to an embodiment of the present invention. Specifically, it includes the following steps:

[0065] S1: Obtain historical fire data for subway stations and surrounding commercial buildings, identify fire influencing factors, and establish a dataset. The historical fire data includes fire data and case studies of subway stations and surrounding commercial buildings both domestically and internationally.

[0066] As a preferred implementation method, fire influencing factors include the type of combustible material, fire load density, fire alarm time, first fire dispatch time, and burned area. Here, fire load density is the fire load per unit building area; fire alarm time is the time from ignition to alarm (excluding smoldering time); first fire dispatch time is the time from receiving the dispatch order to arriving at the fire scene; and burned area is the area affected by the high temperature of the fire.

[0067] Within integrated subway station buildings, the subway concourse and platform public areas contain almost no flammable materials, resulting in a relatively low fire risk. Subway station ancillary rooms are mostly fire-separated, with few flammable materials inside, and all rooms are equipped with fire-fighting facilities, further reducing the fire risk. Subway cars themselves are constructed of non-combustible or flame-retardant materials; the flammable materials inside the cars are mainly the floor, concealed electrical equipment and cables, and passengers' luggage. During peak hours, the cars are densely populated, posing a significant fire risk. The buildings above the integrated station-city complex are primarily commercial, and commercial buildings are characterized by abundant flammable materials, high passenger flow, and difficulties in evacuation and rescue after a fire, posing a substantial fire risk. Therefore, subway cars and commercial buildings represent the main fire risks associated with integrated station-city development.

[0068] Specifically, based on historical fire data, big data analysis is used to identify fire-influencing factors. In one specific embodiment, the types of combustible materials in the historical fire data are identified as fire-influencing factors. By obtaining survey data on combustible materials in subway stations, analysis reveals that the main combustible materials in the public areas of subway station halls and platforms are displays, monitoring instruments, vending machines, signage, and passenger belongings. These combustible materials are relatively few and discontinuously distributed, resulting in a low fire hazard. Regarding subway cars, early subway cars typically used soft fabric sofa seats and flammable decorations, while modern subway cars and their interiors typically use non-combustible or halogen-free, low-smoke flame-retardant materials. Seats are made of flame-retardant or non-combustible materials such as stainless steel and flame-retardant plastics (high-end locations such as airport lines often use sofa-style soft seats). Electrical wires and cables are made of flame-retardant and fire-resistant materials. The main combustible materials in the cars are floors, billboards, decorative materials, concealed electrical equipment and electrical wires and cables, and passenger luggage.

[0069] Furthermore, by obtaining survey data on combustible materials in commercial buildings surrounding subway stations, analysis revealed that the main combustible materials in commercial buildings include clothing, shoes, bags, furniture, and bedding. The materials used are typically cotton, linen, silk, wool, and chemical fibers. Product packaging is usually paper or a combination of plastic and paper / non-woven fabric; product displays are typically stacked, hung, or folded. Therefore, the combustion characteristics and fire spread rates of combustible materials within commercial buildings vary significantly. For example, hanging fires spread more easily into three-dimensional fires than stacked fires, and items in plastic packaging may melt when burning. Thus, data analysis can identify the types of combustible materials in the integrated subway station and commercial building complex, and based on the combustion characteristics of these types of combustible materials, the fire growth types corresponding to different functional areas within the integrated subway station and commercial building complex can be derived. The fire growth types are shown in Table 1.

[0070] Table 1. Types of Fire Increases in Subway Station-City Integration

[0071]

[0072] After obtaining the fire growth type based on the combustible material types in the integrated subway station and city, the fire growth coefficient α corresponding to each fire growth type is further obtained by combining combustible material combustion data from historical fire data. Specifically, when the fire growth type is a slow fire, the fire growth coefficient α is 0.00278 kW / s. 2 When the fire growth type is medium-speed fire, the fire growth coefficient α is 0.011 kW / s. 2 When the fire growth type is rapid fire, the fire growth coefficient α is 0.044 kW / s. 2 When the fire growth type is ultra-fast fire, the fire growth coefficient α is 0.178 kW / s. 2 .

[0073] In addition, the above methods can be used to determine fire load density, fire alarm time, first fire dispatch time, and burned area in historical fire data as fire influencing factors, and further establish a dataset of other fire influencing factors for subway stations and surrounding commercial buildings. Among them, the burned area data also includes data on the effective and ineffective times of the automatic sprinkler system of commercial buildings.

[0074] S2: Construct a fire risk prediction model based on the dataset of fire influencing factors. The fire risk prediction model is a risk prediction model based on the fire area of ​​the building.

[0075] As a preferred embodiment, the step of constructing a fire risk prediction model based on a dataset of fire influencing factors further includes:

[0076] S21: Construct corresponding probability density distribution models based on datasets of various fire-influencing factors.

[0077] Please see Figure 2 and Figure 3 , Figure 2 This is a flowchart illustrating the steps of constructing a probability density distribution model according to an embodiment of the present invention. Figure 3This is a flowchart illustrating the steps of a statistical method for constructing a probability density distribution model according to an embodiment of the present invention. In one specific embodiment, the combustible material type, fire load density, fire alarm time, first fire dispatch time, and burned area in historical fire data are determined as fire influencing factors, and a corresponding probability distribution model is constructed based on the historical data of fire load density, first fire dispatch time, and burned area. The combustible material type is obtained through combustible material survey data of subway stations and surrounding commercial buildings, and no probability distribution model needs to be constructed there. The fire alarm time is the time from ignition to alarm (excluding smoldering time). According to comprehensive fire alarm time statistics, the average ignition time is 4 minutes, and the average alarm time is 2 minutes and 30 seconds, i.e., the average fire alarm time is 6 minutes and 30 seconds, and no probability distribution model needs to be constructed there.

[0078] Furthermore, the steps of constructing corresponding probability density distribution models based on datasets of various fire influencing factors further include:

[0079] S211: Preprocess the dataset of fire influencing factors using data cleaning methods;

[0080] S212: Based on the preprocessed dataset, construct a probability density distribution model based on fire influencing factors using statistical methods. The specific steps of the statistical methods are as follows:

[0081] S2121: Based on the preprocessed dataset, obtain the probability density distribution map of fire influencing factors and their occurrence frequency and / or occurrence frequency;

[0082] S2122: Based on the probability density distribution map, a probability density distribution model of fire influencing factors is obtained through data processing and curve fitting.

[0083] In one specific embodiment, when constructing the probability distribution models for fire load density, first fire response time, and burned area, inconsistencies arise because historical fire data includes data from multiple years and originates from various departments. Therefore, preprocessing of the dataset is necessary to remove incomplete, inaccurate, or irrelevant data to ensure the accuracy of the data used to construct the fire risk prediction model. This data cleaning process can be achieved through big data analytics and will not be elaborated upon here.

[0084] After obtaining a preprocessed dataset of various fire-influencing factors, statistical methods are used to obtain the frequency and / or occurrence of each fire-influencing factor in historical data, thereby obtaining a probability density distribution map, which is a histogram.

[0085] Based on historical data of fire influencing factors, the probability density distribution model of each fire influencing factor is obtained by data processing and curve fitting method. The probability density distribution model is shown below.

[0086] Probability density distribution model of load density in subway car fires: Where x is the fire load density of the subway car (MJ / m³) 2 f(x) is the probability of fire load density.

[0087] Probability density distribution model of fire load density in commercial buildings: Where x is the fire load density of the commercial building (MJ / m³) 2 f(x) is the probability of fire load density.

[0088] Firefighters' first response time probability density distribution model:

[0089]

[0090] Where t is the first fire dispatch time (s), and f(t) is the probability of the first fire dispatch time. The first fire dispatch time is related to the distance from the fire station to the fire scene, but is independent of whether the fire occurs in a subway car or a commercial building. Therefore, there is no need to distinguish between the probability density distribution models of the first fire dispatch time for subway cars and commercial buildings.

[0091] In integrated subway station projects, subway cars typically do not have automatic sprinkler systems, while commercial buildings usually do. The probability density model of the fire area in commercial buildings when the automatic sprinkler system is effective and ineffective is as follows:

[0092] Probability density distribution model of fire-affected area in commercial buildings when automatic sprinkler systems are effective:

[0093]

[0094] Where x is the fire-affected area (m²) of the commercial building when the automatic sprinkler system is effective. 2 f(x) is the probability of the burned area.

[0095] Probability density distribution model of burned area in commercial buildings when automatic sprinkler systems are ineffective:

[0096]

[0097] Where x is the fire-affected area (m²) of the commercial building when the automatic sprinkler system is ineffective. 2 f(x) is the probability of the burned area.

[0098] In this embodiment of the invention, by obtaining a probability density distribution model, the probability and / or frequency of various fire-influencing factors occurring in historical fires can be obtained. Even when the information or data in the information on sudden fires is incomplete, data on various fire-influencing factors can still be obtained based on the probability density distribution model, thereby accurately predicting the fire risk level.

[0099] As a preferred embodiment, the step of constructing a fire risk prediction model based on a dataset of fire influencing factors further includes the following steps:

[0100] S22: Based on a dataset of fire influencing factors, a fire area calculation model is derived by using the metric of total fire heat release rate and the calculation formula for instantaneous fire heat release rate.

[0101] S23: Construct a fire risk prediction model based on the fire area calculation model and the probability density distribution model.

[0102] In one specific embodiment, the total heat release rate of the fire is measured based on the total heat release rate of the fire, and the instantaneous heat release rate of the fire is calculated using the formula Q = αt. 2 Based on the assumption that the fire spreads outwards after a unit of combustible material burns, and using the fire-affecting factors data obtained by the method described above, a calculation model for the fire-affected area of ​​a building can be derived:

[0103]

[0104] Where Q is the instantaneous heat release rate of the fire, α is the fire growth coefficient, and A is the burned area (m²) when the fire brigade arrives. 2 ), t is the fire growth time, t1 is the fire alarm time, which is 390s as mentioned above, t2 is the first fire dispatch time, δ is the combustion efficiency factor, which is a comprehensive data factor representing the degree of incomplete combustion of combustibles. Based on combustion experiment data and fire history data, the combustion efficiency factor is obtained as 0.3-0.9. When the amount of combustibles burned reaches 80% of the total amount of combustibles, the fire enters the decay stage, and the fire combustion efficiency factor usually does not exceed 0.5. Therefore, the combustion efficiency factor in this formula is 0.5, and q is the fire load density.

[0105] Please see Figure 4 , Figure 4 This is a cumulative frequency curve of the fire-affected area of ​​a commercial building according to an embodiment of the present invention. In one specific embodiment, after obtaining the probability density distribution map and distribution model of the commercial building when the automatic sprinkler system is effective and ineffective according to step S2, a cumulative frequency curve of the fire-affected area of ​​the commercial building can be further obtained. Figure 4 As shown in L1, when the automatic sprinkler system is effective, approximately 90% of fires can be contained within 5 meters. 2Below, 95% of fires can be controlled within 10m. 2 Below, approximately 99% are controlled within 100m. 2 The following; such as Figure 4 As shown in L2, when the automatic sprinkler system is ineffective, approximately 90% of fires are contained within 10 square meters, and 95% of fires are contained within 50 square meters. 2 Below, 99% of fires can be controlled within 500m. 2 Therefore, based on the fire area calculation model and the fire area probability density distribution model, the fire risk prediction model for the integrated subway station and city based on the fire area is constructed as follows:

[0106]

[0107] In this embodiment of the invention, the fire risk prediction model is based on the burned area, making the fire risk prediction based on a physical quantity that represents the characteristics of a fire at a real-world location. By calculating the burned area, the fire alarm level is predicted, making fire risk prediction more targeted. Furthermore, the fire risk prediction model is based on historical fire data and a probability density distribution model, resulting in high accuracy.

[0108] S3: Based on the fire risk prediction model and historical fire data, the fire scale is classified and the fire alarm level is divided. The classification of fire scale and fire alarm level is based on the burned area.

[0109] When an automatic sprinkler system is effective, 95% of the fire area can be controlled within 10m². 2 The following analysis shows that when automatic sprinkler systems are ineffective, they can more accurately reflect the trend and scale of a fire. Therefore, based on the integrated station-city fire risk prediction model based on burned area, and combined with the 90%, 95%, 98%, and 99% quantiles of burned area statistics when automatic sprinkler systems in commercial buildings are ineffective, the fire scale level is determined. The fire growth time range corresponding to different burned area levels is obtained based on the burned area calculation formula and fire load density calculation. The corresponding fire heat release rate range is then calculated based on the instantaneous fire heat release rate formula.

[0110] In one specific embodiment, based on the aforementioned historical fire data and probability density distribution model, the fire scale of the subway-city integration is divided into five levels according to the fire risk prediction model, and the corresponding heat release rate is obtained. The fire scale and its corresponding heat release rate are shown in Table 2.

[0111] Table 2 Fire Scale Classification

[0112]

[0113] Based on historical fire data and survey data, one water cannon can extinguish fires at a depth of 15-20 meters. 2 For a fire, a fire tanker truck carries two water hoses, and a firefighting team carries at least four water hoses. A firefighting team can extinguish a fire of 60 square meters. Therefore, based on the firefighting team and the fire scale classification based on the burned area, combined with the above-mentioned historical fire data and probability density distribution model, the fire alarm level of the subway station-city integration is further divided into four levels according to the fire risk prediction model, and the burned area corresponding to the first to fourth fire alarm levels is determined. The fire alarm level classification is shown in Table 3.

[0114] Table 3 Fire Alarm Levels

[0115] <![CDATA[Burned area (m 2 )]]> [0,50] (50,300] (300,500] (500,+∞)

[0116] In this embodiment of the invention, the fire-affected area is the range involved by the high temperature effect of the fire. The fire scale and fire alarm level are divided by the fire-affected area. The fire scale and fire alarm level are based on a physical quantity that actually represents the characteristics of the fire. The level is divided in a quantitative way, making the level division more reasonable.

[0117] S4: Determine the fire alarm level based on sudden fire information and fire risk prediction models. Sudden fire information includes data of the same type as historical fire data, but differs in that it is real data obtainable when an actual fire occurs.

[0118] As a preferred embodiment, the step of determining the fire alarm level based on sudden fire information and a fire risk prediction model further includes:

[0119] If the information regarding a sudden fire includes data on fire-influencing factors, the fire alarm level is determined using either the mean method or the maximum value method, based on the data on these factors and the fire risk prediction model; or...

[0120] If the information on a sudden fire does not include data on factors influencing the fire, the fire alarm level is determined using the mean method or the maximum value method based on the probability density distribution model and the fire risk prediction model.

[0121] Please see Figure 5 , Figure 5 This is a flowchart illustrating the steps for determining a fire alarm level according to an embodiment of the present invention. In one specific embodiment, a fire risk prediction model is constructed using the above method, and fire alarm levels are classified, so that when a fire occurs, the fire alarm level can be predicted based on sudden fire information.

[0122] Specifically, when predicting fire alarm levels, the first step is to determine whether the sudden fire information includes data on fire influencing factors used to calculate the burned area. If the sudden fire information includes data on the type of combustibles in the subway station complex, the first fire response time, fire load density, and combustion factors, the burned area can be obtained using existing data and a fire risk prediction model. Based on the type of combustibles in the subway station complex, the corresponding fire growth coefficient α can be obtained; based on the first fire response time data, the average first fire response time t2 for that subway station complex can be obtained; based on the fire load density of the subway cars and commercial buildings, the fire load density q used to calculate the burned area is obtained by calculating the mean or maximum value. Finally, combining the combustion efficiency factor δ of the comprehensive data factors of the subway station complex, the actual fire alarm time t1, and the effectiveness of the automatic sprinkler system in the commercial buildings, the burned area is obtained through the fire risk prediction model, and the fire alarm level is determined.

[0123] Please refer to the following: Figure 5 In another specific embodiment, unlike the embodiments described above, when a fire occurs, data on all fire-influencing factors at the fire location cannot be obtained based on the sudden fire information. Therefore, the burned area cannot be directly determined based on existing data to ascertain the fire alarm level. Thus, it is necessary to obtain data on these factors using a probability density distribution model of various fire-influencing factors obtained through the aforementioned method, in order to calculate the burned area. Specifically, based on a preset quantile value of the probability density distribution model, the first fire response time t2 of the subway station area and the fire load density data of the subway cars and commercial buildings can be obtained. Then, by calculating the mean or maximum value, the fire load density q of the subway station area used to calculate the burned area is obtained. Furthermore, based on the sudden fire information, the type of combustible material at the fire location can be obtained, along with the corresponding fire growth coefficient α of the subway station area. Finally, combining the combustion efficiency factor δ of the comprehensive data factors of the subway station area, the actual fire alarm time t1, and the effectiveness of the automatic sprinkler system in the commercial buildings, the burned area is calculated using a fire risk prediction model to determine the fire alarm level. Furthermore, the preset quantile value of the probability density distribution model can be 95% or other values. The higher the quantile value, the greater the reliability of the obtained data. The preset quantile value is obtained through big data analysis and can be adjusted as needed, but this invention is not limited to this. It should be noted that the combustible material type, fire alarm time data, and the effectiveness of the automatic sprinkler system in the commercial building required for calculating the fire area can be obtained from the information on the sudden fire, without needing to be obtained through the probability density distribution model.

[0124] In this embodiment of the invention, a fire risk prediction model based on burned area and fire alarm level can be used to calculate the burned area and determine the fire alarm level based on existing sudden fire information data at the fire site. If sudden fire information data at the fire site is unavailable, the burned area can also be calculated based on historical fire data using a probability density distribution model, resulting in high flexibility and accuracy in fire alarm prediction.

[0125] As a preferred embodiment, the AI-based integrated subway station fire alarm prediction method further includes step S5: automatically correcting the fire alarm level based on the on-site fire situation, wherein the fire situation includes one or more of the following: the number of trapped or injured persons, the activation status of the automatic sprinkler system, and / or the fire type. In a specific embodiment, the determination of the fire alarm level after considering the on-site fire situation is shown in Table 4:

[0126] Table 4 Fire Alarm Levels Based on On-Site Fire Circumstances

[0127]

[0128] Specifically, if the burned area at the time of the fire brigade's arrival is obtained using the above method, it is 20m². 2 The fire alarm level was initially set at Level 1. Based on the on-site fire situation, it was determined that there were two people trapped and the fire type was a cinema / theater with high foot traffic; therefore, the fire alarm level was automatically adjusted to Level 2. If the burned area upon the fire department's arrival was determined to be 60m², the fire alarm level was automatically adjusted accordingly. 2 If the fire alarm level is determined to be Level 2, and then based on the on-site fire situation, it is found that there are no trapped or injured personnel, the automatic sprinkler system is effectively activated, and the fire type is combustible materials inside the shop, then the fire alarm level is automatically corrected to Level 1. This invention is not limited to this.

[0129] In this embodiment of the invention, after obtaining the burned area and confirming the fire alarm level through the fire risk prediction model, the fire alarm level can be further automatically corrected by combining the on-site fire situation at the fire location and the division in Table 4. Therefore, the prediction of the corrected fire alarm level is more comprehensive and accurate.

[0130] As a preferred embodiment, the AI-based subway-city integrated fire alarm prediction method further includes step S6: when a special situation occurs in the subway-city integrated area, the fire alarm level is automatically raised, wherein the special situation includes:

[0131] The subway station-city integration area is located during major holidays, important political events, or in politically sensitive or important areas.

[0132] The integrated subway station area encountered severe weather conditions.

[0133] The number of fire alarm calls in the integrated subway station area continues to increase, and the trend of disaster is obvious.

[0134] A fire occurs in the underground commercial area of ​​a subway station integrated development; a fire occurs in a Class I high-rise building or a mixed-use development above a subway station; or there is a risk of fire or explosion within the building; and / or,

[0135] A fire broke out in a subway car within the integrated subway station area, rendering the subway train immobile and stranded in the tunnel. Among these, Class I high-rise buildings are residential buildings with a height exceeding 54 meters, including residential buildings with commercial service outlets.

[0136] Specifically, the fire alarm level determined and automatically corrected by the above method is Level II. If a large-scale celebration is being held in the metro station-city integration area, it is determined that the metro station-city integration area is in a major holiday, an important political event, or a politically sensitive or important area. Therefore, the fire alarm level is automatically raised by one level to Level III. If the metro station-city integration area encounters strong winds and extreme weather, and the number of fire alarm calls in the area continues to increase, it is determined that the metro station-city integration area is experiencing severe weather conditions and that the number of fire alarm calls in the area continues to increase, indicating a clear trend of disaster. Therefore, the fire alarm level is automatically raised by two levels to Level IV. However, this invention is not limited to these limitations.

[0137] When a fire occurs, its development and changes are usually influenced by other factors in the surrounding area. Therefore, in this embodiment of the invention, by incorporating special circumstances affecting the fire alarm level into the fire alarm level prediction, and after determining the fire alarm level based on the burned area obtained from the fire influencing factors, the fire alarm level can be automatically increased based on preset special circumstances. This allows for a more realistic and accurate determination of the fire alarm level, avoiding inaccuracies in the subsequent dispatch of rescue forces.

[0138] As a preferred embodiment, the AI-based integrated subway station and city fire alarm prediction method further includes step S7: automatically dispatching corresponding fire fighting and rescue forces according to the fire alarm level.

[0139] This invention provides an AI-based integrated fire alarm prediction method for subway stations and surrounding commercial buildings. By analyzing historical fire data of subway stations and surrounding commercial buildings, it determines fire influencing factors based on the data and the degree of fire impact. Then, based on these factors, it constructs an integrated fire risk prediction model for subway stations and predicts fire levels. The fire risk prediction model and fire alarm levels are based on physical quantities that actually characterize fire features, specifically the fire-affected area of ​​buildings. In contrast, while existing technologies also use quantitative methods to classify fire alarm levels, they are still based on qualitative fire influencing factors. They assign weights to the magnitude of fire impact and then use weighted averages or computer algorithms to classify fire alarm levels. Therefore, this invention can more automatically, timely, and accurately classify and predict fire alarm levels, reducing misjudgments and ensuring accurate and rapid dispatch of fire-fighting and rescue forces in integrated subway stations and cities, thereby improving the efficiency of fire-fighting and rescue teams.

[0140] Please see Figure 6 , Figure 6 A structural block diagram of an AI-based integrated subway station fire alarm prediction system 1 according to an embodiment of the present invention is shown, comprising:

[0141] Data acquisition module 11 is used to acquire historical fire data of subway stations and surrounding commercial buildings, determine fire influencing factors, and establish a dataset. The historical fire data includes fire data and case studies of subway stations and surrounding commercial buildings both domestically and internationally.

[0142] In a preferred embodiment, the fire influencing factors in the data acquisition module 11 include the type of combustible material, fire load density, fire alarm time, first fire dispatch time, and burned area. The specific implementation and beneficial technical effects of the data acquisition module 11 are as described in step S1 above, and will not be repeated here.

[0143] Model building module 12 is used to build a fire risk prediction model based on a dataset of fire influencing factors.

[0144] In a preferred embodiment, the step of constructing a fire risk prediction model based on a dataset of fire influencing factors in the model construction module 12 further includes: constructing corresponding probability density distribution models based on datasets of various fire influencing factors.

[0145] In a preferred embodiment, the step of constructing corresponding probability density distribution models based on datasets of various fire influencing factors in model building module 12 further includes:

[0146] Data cleaning methods were used to preprocess the dataset of fire influencing factors;

[0147] Based on the preprocessed dataset, a probability density distribution model based on the fire influencing factors is constructed using statistical methods.

[0148] In a preferred embodiment, the step of constructing a probability density distribution model based on fire influencing factors using statistical methods in the model building module 12 further includes:

[0149] Based on the preprocessed dataset, a probability density distribution map of fire influencing factors and their occurrence frequency and / or occurrence frequency is obtained.

[0150] Based on the probability density distribution map, a probability density distribution model of fire influencing factors is obtained through data processing and curve fitting.

[0151] In a preferred embodiment, the step of constructing a fire risk prediction model based on a dataset of fire influencing factors in the model building module 12 further includes the following steps:

[0152] Based on a dataset of fire influencing factors, a model for calculating the burned area is derived by using the metric of total fire heat release rate and the formula for calculating instantaneous fire heat release rate.

[0153] A fire risk prediction model is constructed based on the fire area calculation model and the probability density distribution model.

[0154] The specific implementation method and beneficial technical effects of the model building module 12 are as described in step S2 above, and will not be repeated here.

[0155] The grading module 13 is used to classify the scale of fires and assign fire alarm levels based on the fire risk prediction model and historical fire data. The specific implementation and beneficial technical effects of the grading module 13 are as described in step S3 above, and will not be repeated here.

[0156] Module 14 is used to determine the fire alarm level based on sudden fire information and a fire risk prediction model. The fire risk prediction model is a risk prediction model based on the burned area of ​​the building. The classification of fire scale and fire alarm level is based on the burned area.

[0157] If 0 ≤ burned area ≤ 10m² 2 If the fire scale is 10m, then the fire scale is determined to be Level 1; if the fire scale is 10m 2 <Fire area ≤ 50m² 2 If the fire scale is 50m, then the fire scale is determined to be level two; if the fire scale is 50m 2 <Fire area ≤ 300m² 2 If the fire scale is 300m, then the fire scale is determined to be level three; if the fire scale is 300m 2 <Fire area ≤ 500m² 2 If the fire scale is 500m, then the fire scale is determined to be level four; if the fire scale is 500m 2If the burned area is less than the fire-affected area, the fire scale is determined to be level five; among which,

[0158] For fires of scale one or two, the fire alarm level is determined as level one; for fires of scale three, the fire alarm level is determined as level two; for fires of scale four, the fire alarm level is determined as level three; and for fires of scale five, the fire alarm level is determined as level four.

[0159] In a preferred embodiment, the step of determining the fire alarm level based on the sudden fire information and the fire risk prediction model in the determining module 14 further includes:

[0160] If the information regarding a sudden fire includes data on fire-influencing factors, the fire alarm level is determined using either the mean method or the maximum value method, based on the data on these factors and the fire risk prediction model; or...

[0161] If the information on a sudden fire does not include data on factors influencing the fire, the fire alarm level is determined using the mean method or the maximum value method based on the probability density distribution model and the fire risk prediction model.

[0162] In a preferred embodiment, the determining module 14 is further configured to automatically correct the fire alarm level based on the on-site fire situation, wherein the fire situation includes one or more of the following: the number of people trapped or injured, the activation status of the automatic sprinkler system, and / or the type of fire.

[0163] The specific implementation method and beneficial technical effects of module 14 are as described in steps S4 and S5 above, and will not be repeated here.

[0164] In a preferred embodiment, the AI-based subway station-city integrated fire alarm prediction system 1 provided by the present invention further includes a fire alarm escalation module 15, used to automatically escalate the fire alarm level when special circumstances occur in the subway station-city integrated area, wherein the special circumstances include:

[0165] The subway station-city integration area is located during major holidays, important political events, or in politically sensitive or important areas.

[0166] The integrated subway station area encountered severe weather conditions.

[0167] The number of fire alarm calls in the integrated subway station area continues to increase, and the trend of disaster is obvious.

[0168] A fire occurs in the underground commercial area of ​​a subway station integrated development; a fire occurs in a Class I high-rise building or a mixed-use development above a subway station; or there is a risk of fire or explosion within the building; and / or,

[0169] A fire broke out in a subway car in the integrated subway station area, rendering the subway train unable to move and causing it to stop in the tunnel section.

[0170] The specific implementation method and beneficial technical effects of the fire alarm upgrade module 15 are as described in step S6 above, and will not be repeated here.

[0171] As a preferred embodiment, the AI-based integrated subway station and city fire alarm prediction system 1 provided by the present invention further includes a dispatch module 16, which is used to automatically dispatch corresponding fire fighting and rescue forces according to the fire alarm level.

[0172] The specific implementation method and beneficial technical effects of the scheduling module 16 are as described in step S7 above, and will not be repeated here.

[0173] Since the embodiments of the system part correspond to the embodiments of the method part, please refer to the description of the embodiments of the method part for the embodiments of the system part, which will not be repeated here. It should be noted that the system proposed above can also be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of the modules described above is merely a logical functional division. In actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed.

[0174] An embodiment of the present invention also provides a computer-readable storage medium having a computer program stored thereon, the computer program being configured to execute the AI-based urban-subway fire alarm prediction method as described above when running, which will not be repeated here.

[0175] An embodiment of the present invention also provides an electronic device including a memory and a processor. The memory stores a computer program, and the processor is configured to run the computer program to execute the AI-based urban-subway fire alarm prediction method as described above, which will not be repeated here.

[0176] An embodiment of the present invention also provides a fire alarm prediction intelligent robot, which is equipped with or connected via a network to the above-described artificial intelligence-based integrated fire alarm prediction system for subway stations and cities.

[0177] In this embodiment of the invention, by setting up an AI-based integrated fire alarm prediction system for subway stations and cities on the local end of the intelligent robot, the fire alarm level can be predicted directly on the local end in combination with the fire situation. It can operate normally even when there is no network or the network is unstable, thus ensuring the stability of the prediction system.

[0178] Furthermore, an AI-based integrated fire alarm prediction system for subway stations and cities can be set up in the cloud via a network connection, allowing it to be used without local setup. The intelligent robot only needs to perform the functions of connecting to the network or the host computer. Connecting to the cloud system via the network improves the convenience and economy of using the intelligent robot. The operation and maintenance of the cloud system are not limited by time or location; only the cloud server needs maintenance, which can respond to and report problems in a timely manner, thus improving the efficiency of operation and maintenance.

[0179] The above provides a detailed description of the AI-based integrated fire alarm prediction method and system for subway stations and urban areas provided by this invention. The various embodiments in the specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems, readable storage media, electronic devices, and robots disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple, and relevant details can be found in the method section.

[0180] Of course, the present invention may have other various embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding changes and modifications according to the present invention, but these corresponding changes and modifications should all fall within the protection scope of the appended claims.

Claims

1. An AI-based method for predicting fire alarms in an integrated urban subway station system, comprising the following steps: Obtain historical fire data of subway stations and surrounding commercial buildings, identify fire influencing factors, and establish a dataset; A fire risk prediction model is constructed based on the dataset of fire influencing factors. Based on the fire risk prediction model and historical fire data, the fire scale is classified and the fire alarm level is determined. Based on the information on sudden fires and the fire risk prediction model, the fire alarm level is determined; characterized in that, The step of constructing a fire risk prediction model based on the dataset of fire influencing factors further includes: Based on datasets of various fire-influencing factors, corresponding probability density distribution models are constructed respectively; based on the burned area calculation model, the fire risk prediction model is constructed; and, The step of determining the fire alarm level based on the sudden fire information and the fire risk prediction model further includes: If the sudden fire information includes data on the fire influencing factors, the burned area is calculated based on the data on the fire influencing factors and the fire risk prediction model to determine the fire alarm level; If the sudden fire information does not include data on the fire influencing factors, the data on the fire influencing factors are obtained according to the probability density distribution model, and the burned area is calculated according to the fire risk prediction model to determine the fire alarm level; wherein... The fire risk prediction model is based on the fire-affected area of ​​a building, and the classification of fire scale and fire alarm level is based on the fire-affected area. If 0 ≤ burned area ≤ 10m² 2 If so, the fire scale is determined to be Level 1; If 10m 2 <Fire area ≤ 50m² 2 If so, the fire scale is determined to be level two; If 50m 2 <Fire area ≤ 300m² 2 If so, the fire scale is determined to be level three; If 300m 2 <Fire area ≤ 500m² 2 If so, the fire scale is determined to be level four; If 500m 2 If the burned area is less than the fire-affected area, the fire scale is determined to be level five; among which, If the fire scale is level one or level two, the fire alarm level is determined to be level one; If the fire scale is level three, then the fire alarm level is determined to be level two; If the fire scale is level four, then the fire alarm level is determined to be level three; For a fire of magnitude five, the fire alarm level is determined to be level four; wherein the fire influencing factors include the type of combustible material, fire load density, fire alarm time, first fire response time, and the burned area; and the fire risk prediction model is... A represents the burned area (m²) when the fire brigade arrives. 2 ), Fire growth coefficient (kW / s) for the type of combustible material 2 ), t1 is the fire alarm time (s), t2 is the first fire dispatch time (s), δ is the combustion efficiency factor, and q is the fire load density (MJ / m³). 2 ).

2. The AI-based integrated fire alarm prediction method for subway stations as described in claim 1, characterized in that, The step of constructing a fire risk prediction model based on the dataset of fire influencing factors further includes the following steps: Based on the dataset of fire influencing factors, and according to the metric of total fire heat release rate and the calculation formula of instantaneous fire heat release rate, the fire area calculation model is derived.

3. The AI-based integrated fire alarm prediction method for subway stations and urban areas according to claim 1, characterized in that, The step of constructing corresponding probability density distribution models based on datasets of various fire-influencing factors further includes: The dataset of fire influencing factors was preprocessed using data cleaning methods; Based on the preprocessed dataset, a probability density distribution model based on the fire influencing factors is constructed using statistical methods.

4. The AI-based integrated fire alarm prediction method for subway stations as described in claim 3, characterized in that, The step of constructing a probability density distribution model based on the fire influencing factors using statistical methods further includes: Based on the preprocessed dataset, a probability density distribution map of the fire influencing factors and their occurrence frequency is obtained. Based on the probability density distribution map, the probability density distribution model of the fire influencing factors is obtained by data processing and curve fitting methods.

5. The AI-based integrated fire alarm prediction method for subway stations and urban areas according to claim 1, characterized in that, It also includes a step of automatically correcting the fire alarm level based on the on-site fire situation, wherein the fire situation includes one or more of the following: the number of people trapped or injured, the activation status of the automatic sprinkler system, and / or the type of fire.

6. The AI-based integrated fire alarm prediction method for subway stations according to claim 1 or 5, characterized in that, It also includes a step of automatically raising the fire alarm level when special circumstances occur in the integrated subway station area, wherein the special circumstances include: The subway station-city integration area is located during major holidays, important political events, or politically sensitive or important areas. The integrated subway station area encountered severe weather conditions. The number of fire alarm calls in the area where the subway station is integrated with the city continues to increase, and the trend of disaster is obvious. A fire occurred in the underground commercial area of ​​the subway station-city integrated area; a fire occurred in a Class I high-rise building or a complex built above a subway station; or there was a risk of fire and explosion within the building; and / or, A fire broke out in a subway car in the integrated subway station area, rendering the subway train unable to move and causing it to remain in the tunnel section.

7. The AI-based integrated fire alarm prediction method for subway stations according to claim 1 or 5, characterized in that, It also includes the step of automatically dispatching the corresponding fire fighting and rescue forces based on the fire alarm level.

8. An artificial intelligence-based integrated fire alarm prediction system for subway stations, comprising: The data acquisition module is used to acquire historical fire data of subway stations and surrounding commercial buildings, determine fire influencing factors, and establish a dataset. The model building module is used to build a fire risk prediction model based on the dataset of fire influencing factors. The classification module is used to classify the scale of a fire and classify the fire alarm level based on the fire risk prediction model and historical fire data. The determination module is used to determine the fire alarm level based on sudden fire information and the fire risk prediction model; characterized in that, in the model construction module, the step of constructing the fire risk prediction model based on the dataset of fire influencing factors further includes: Based on datasets of various fire-influencing factors, corresponding probability density distribution models are constructed respectively; based on the burned area calculation model, the fire risk prediction model is constructed; and, In the determining module, the step of determining the fire alarm level based on the sudden fire information and the fire risk prediction model further includes: If the sudden fire information includes data on the fire influencing factors, the burned area is calculated based on the data on the fire influencing factors and the fire risk prediction model to determine the fire alarm level; If the sudden fire information does not include data on the fire influencing factors, the data on the fire influencing factors are obtained according to the probability density distribution model, and the burned area is calculated according to the fire risk prediction model to determine the fire alarm level; wherein... The fire risk prediction model is based on the fire-affected area of ​​a building, and the classification of fire scale and fire alarm level is based on the fire-affected area. If 0 ≤ burned area ≤ 10m² 2 If so, the fire scale is determined to be Level 1; If 10m 2 <Fire area ≤ 50m² 2 If so, the fire scale is determined to be level two; If 50m 2 <Fire area ≤ 300m² 2 If so, the fire scale is determined to be level three; If 300m 2 <Fire area ≤ 500m² 2 If so, the fire scale is determined to be level four; If 500m 2 If the burned area is less than the fire-affected area, the fire scale is determined to be level five; among which, If the fire scale is level one or level two, the fire alarm level is determined to be level one; If the fire scale is level three, then the fire alarm level is determined to be level two; If the fire scale is level four, then the fire alarm level is determined to be level three; For a fire of level five in scale, the fire alarm level is determined to be level four; wherein, in the data acquisition module, the fire influencing factors include combustible material type, fire load density, fire alarm time, first fire dispatch time, and the burned area; and, in the model construction module, the fire risk prediction model is... A represents the burned area (m²) when the fire brigade arrives. 2 ), Fire growth coefficient (kW / s) for the type of combustible material 2 ), t1 is the fire alarm time (s), t2 is the first fire dispatch time (s), δ is the combustion efficiency factor, and q is the fire load density (MJ / m³). 2 ).

9. A subway station integrated fire alarm prediction system based on artificial intelligence according to claim 8, characterized in that, In the model building module, the step of building a fire risk prediction model based on the dataset of fire influencing factors further includes the following steps: Based on the dataset of fire influencing factors, and according to the metric of total fire heat release rate and the calculation formula of instantaneous fire heat release rate, the fire area calculation model is derived.

10. The AI-based integrated fire alarm prediction system for subway stations and cities according to claim 8, characterized in that, In the model building module, the step of constructing corresponding probability density distribution models based on datasets of various fire influencing factors further includes: The dataset of fire influencing factors was preprocessed using data cleaning methods; Based on the preprocessed dataset, a probability density distribution model based on the fire influencing factors is constructed using statistical methods.

11. The AI-based integrated fire alarm prediction system for subway stations according to claim 10, characterized in that, In the model building module, the step of constructing a probability density distribution model based on the fire influencing factors using statistical methods further includes: Based on the preprocessed dataset, a probability density distribution map of the fire influencing factors and their occurrence frequency is obtained. Based on the probability density distribution map, the probability density distribution model of the fire influencing factors is obtained by data processing and curve fitting methods.

12. The AI-based integrated fire alarm prediction system for subway stations and urban areas according to claim 8, characterized in that, The determining module is also used to further automatically correct the fire alarm level based on the on-site fire situation, wherein the fire situation includes one or more of the following: the number of people trapped or injured, the activation status of the automatic sprinkler system, and / or the type of fire.

13. The AI-based integrated fire alarm prediction system for subway stations and cities according to claim 8 or 12, characterized in that, It also includes a fire alarm escalation module, used to automatically raise the fire alarm level when special circumstances occur in the integrated subway station area, wherein the special circumstances include: The subway station-city integration area is located during major holidays, important political events, or politically sensitive or important areas. The integrated subway station area encountered severe weather conditions. The number of fire alarm calls in the area where the subway station is integrated with the city continues to increase, and the trend of disaster is obvious. A fire occurred in the underground commercial area of ​​the subway station-city integrated area; a fire occurred in a Class I high-rise building or a complex built above a subway station; or there was a risk of fire and explosion within the building; and / or, A fire broke out in a subway car in the integrated subway station area, rendering the subway train unable to move and causing it to remain in the tunnel section.

14. The AI-based integrated fire alarm prediction system for subway stations and cities according to claim 8 or 12, characterized in that, It also includes a dispatch module, which is used to automatically dispatch the corresponding fire fighting and rescue forces according to the fire alarm level.

15. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program is configured to execute the method of any one of claims 1-7 when it is run.

16. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the method of any one of claims 1-7.

17. A fire alarm prediction intelligent robot, equipped with or connected via a network to an AI-based integrated subway and city fire alarm prediction system as described in any one of claims 8-14.

Images