A self-built house fire safety management system

Abstract

This invention belongs to the field of fire safety management technology, specifically relating to a fire safety management system for self-built houses. It includes: a data acquisition module, a cleaning module, a scoring system construction module, an initial scoring module, an indicator weight calculation module, a comprehensive score acquisition module, a calibration module, a standard threshold module, a hazard coefficient extraction module, a superposition coefficient calculation module, a risk level acquisition module, and a safety management result acquisition module. This invention achieves dynamic and accurate assessment of fire risks in self-built houses through comprehensive scoring and analysis of static inherent data and dynamic hazard data. It completes hazard classification, adaptive early warning, and closed-loop control, significantly improving the intelligence and refinement of fire safety management in self-built houses, reducing the incidence of fire accidents, and has good market application prospects.

Description

Technical Field

[0001] This invention belongs to the field of fire safety management technology, and specifically relates to a fire safety management system for self-built houses. Background Technology

[0002] Self-built houses generally suffer from low construction standards, arbitrary renovations, lack of fire-fighting facilities, and weak fire safety awareness among residents, making them a high-risk location for fire accidents in both urban and rural areas. Currently, existing fire safety management technologies for self-built houses have many deficiencies, severely restricting the accuracy and effectiveness of fire safety control.

[0003] In existing technologies, fire safety assessments of self-built houses mostly adopt static scoring models with fixed indicators and fixed weights. The scoring dimensions are singular, focusing only on basic explicit indicators such as fire-fighting equipment and unobstructed passageways. They do not take into account the differentiated dynamic factors such as the structural characteristics of the self-built house, the number of residents, the power load, the renovation period, and the business attributes. The scoring results are highly homogenized and cannot accurately distinguish the true fire risk level of different self-built houses. This easily leads to the problem of high-risk houses being missed and low-risk houses being misjudged.

[0004] Meanwhile, the weight calculation algorithms and risk warning algorithms used in traditional fire safety management have significant shortcomings: the conventional analytic hierarchy process (AHP) relies excessively on subjective human experience, resulting in highly subjective and biased weight assignments; the single entropy weight method relies solely on objective data calculations and cannot be combined with professional experience in the fire protection field, leading to poor adaptability. Furthermore, existing warning algorithms mostly use fixed thresholds to determine risk levels, failing to adapt to different scenarios such as self-built detached houses, multi-story buildings, and mixed-use commercial and residential buildings. This results in chaotic prioritization of hazard investigations, hindering hierarchical and precise control, leading to a heavy workload, low efficiency, and high difficulty in closed-loop management at the grassroots level.

[0005] In summary, the existing fire safety management system for self-built houses suffers from technical pain points such as a rigid scoring system, insufficient algorithm accuracy, and poor adaptability of control measures.

[0006] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0007] The purpose of this invention is to provide a fire safety management system for self-built houses to solve the problems of low accuracy of static scoring, imbalance of subjective and objective weights, and poor adaptability of risk warning in the existing fire safety management of self-built houses.

[0008] To achieve the above objectives, the present invention provides the following technical solution: A fire safety management system for self-built houses, comprising: The data acquisition module is used to collect static inherent data and dynamic potential hazard data, establish a temporary data cache pool, and realize the aggregation and summarization of all fire protection data of the same building; The cleaning module performs standardized cleaning processing on the collected raw data, sequentially completing outlier removal, duplicate data deduplication, missing data marking and completion, and format normalization. The scoring system construction module uses standardized cleaned data to build a scoring system for target housing scene classification indicators. The initial scoring module scores the standardized cleaning process data based on the target house scene classification index scoring system to obtain an initial comprehensive fire safety scoring matrix; The indicator weight calculation module calculates the indicator weights of the initial fire safety comprehensive scoring matrix to obtain the target indicator weights. The comprehensive score acquisition module optimizes the initial fire safety comprehensive score matrix according to the target indicator weights to obtain a dynamic fire safety comprehensive score; The calibration module performs a standardized calibration of the target's comprehensive score on a percentage scale, and outputs a standard score of 0-100 points. The standard threshold module uses a pre-trained machine learning threshold model to predict standard fire safety data and obtain basic standard thresholds. It then modifies the basic standard thresholds by combining four parameters: building type, number of floors, usage scenario, and population density, to generate adaptive risk assessment thresholds. The hazard coefficient extraction module extracts the hazard coefficient from the scoring system of the classification indicators of the housing scene to be processed according to the preset hazard coefficient extraction rules. The superposition coefficient calculation module calculates the superposition risk coefficient of hidden dangers based on the number of hidden dangers and the hazard coefficient collected; The risk level acquisition module is used to link three types of core data: adaptive risk judgment threshold, standard score, and risk coefficient of hidden danger superposition, to perform accurate four-level risk judgment. The safety management results acquisition module is used to receive the output building risk level labels and automatically match them with the corresponding control strategy database.

[0009] Preferably, the static inherent data consists of the building's lifetime fixed or altered data plus fire protection infrastructure parameters; the building's lifetime fixed or altered data includes: location of the self-built house, land area, number of floors, structural material, construction year, renovation records, usage attributes, and maximum occupancy; the fire protection infrastructure parameters include: fire extinguishers, emergency lights, evacuation signs, and fire water sources; the dynamic hazard data consists of real-time hazard parameters, including: peak power load, electrical circuit status, gas sealing status, degree of passage blockage, quantity of flammable and explosive materials stored, illegal use of fire and electricity, and illegal overnight stays / business operations.

[0010] As a preferred option, the scoring system construction module constructs a target housing scene classification index scoring system using standardized and cleaned data, specifically including the following steps: The data from the standardized cleaning process is classified according to the attributes of the housing scene, housing scene classification labels are generated and entered into the database; According to the preset house scene classification index optimization rules, irrelevant and redundant indicators are removed from different scenes, and scene-specific high-risk indicators are added to generate an initial house scene classification index scoring system. The initial housing scene classification index scoring system is scored and deducted according to the preset housing scene classification scoring rules to obtain the housing scene classification index scoring system to be processed. The scoring system for classifying housing scenarios was updated based on seasonal changes, new fire regulations, and fire accident cases to obtain the target housing scenario classification scoring system.

[0011] Preferably, the indicator weight calculation module calculates the indicator weights of the initial fire safety comprehensive scoring matrix to obtain the target indicator weights, as shown in the following formula: ; in, Let be the target indicator weight corresponding to the initial comprehensive fire safety scoring index in column j, and the normalized sum of the weights of all indicators is 1. The scene adaptive fusion coefficient has three fixed adaptive values: 0.3, 0.5, and 0.7. The subjective weight corresponding to the j-th initial comprehensive fire safety scoring index; represents the objective weight corresponding to the initial comprehensive fire safety scoring index in column j.

[0012] Preferably, the comprehensive scoring module optimizes the initial fire safety comprehensive scoring matrix according to the target indicator weights to obtain a dynamic fire safety comprehensive score, as shown in the following formula: ; ; in, A comprehensive score for the target; The initial overall fire safety score is the one that has not been revised. Let J represent the target indicator weights corresponding to the initial comprehensive fire safety scoring index in column j. The initial comprehensive fire safety score corresponds to the initial comprehensive fire safety score index in column j. This represents the total number of initial comprehensive fire safety scoring indicators; This is the hazard superposition coefficient. When there are two or more moderate to severe hazards, the value range is [0.8, 0.95], and it is fixed at 1.0 when there are no superposition hazards. The seasonal correction factor is set to [0.9, 0.98] for the summer and winter fire seasons when fires are most frequent, and fixed at 1.0 for the spring and autumn seasons when the risk is relatively stable.

[0013] Preferably, the standard threshold module modifies the basic standard threshold to generate an adaptive risk assessment threshold, as shown in the following formula: ; in, The adaptive risk assessment threshold; The basic standard threshold; This is a correction factor for building type; This is a correction factor for the number of building floors; Adjustment coefficients for housing usage scenarios; This is a correction factor for population density.

[0014] Preferably, the superposition coefficient calculation module calculates the superposition risk coefficient of hidden dangers based on the number of collected hidden dangers and the hazard coefficient, as shown in the following formula: ; in, The risk coefficient is used to assess potential hazards, with a value range of 0-1. The smaller the value, the higher the risk. This represents the total number of existing effective hidden dangers in self-built houses. Let be the k-th hazard coefficient, where minor hazards are assigned a value of 0.05, general hazards are assigned a value of 0.15, relatively serious hazards are assigned a value of 0.3, and severe hazards are assigned a value of 0.5.

[0015] Preferably, the risk level acquisition module uses three core data points—adaptive risk judgment threshold, standard score, and risk coefficient of hazard superposition—to perform a four-level accurate risk judgment, as detailed below: Standard score from 0 to Score, or a score higher than Divide but Houses with an elevation of ≤0.8 are classified as having a major risk, triggering a red emergency alert. Standard score to And 0.8 < Houses with a value ≤0.9 are considered to have a higher risk and trigger an orange alert. Standard score to And 0.9 < Houses with an elevation of ≤0.98 are classified as having general risk and trigger a yellow alert. Standard score Up to 100 points and Houses with a risk level greater than 0.98 are classified as low-risk and trigger a green normal status.

[0016] Preferably, the safety management result acquisition module receives and outputs the building risk level label, and automatically matches it with the corresponding control strategy database, as follows: Red-level major risks are matched with emergency shutdown and control rules; Orange-level areas with higher risks are subject to rectification within a specified period. Yellow indicates general risk matching and routine inspection rules; Green and low-risk practices align with standard regulatory rules.

[0017] Compared with the prior art, the present invention has the following beneficial effects: (1) The self-built house fire safety management system of the present invention constructs a target house scene classification index scoring system through standardized cleaning and processing data, breaks through the limitations of traditional static scoring, adapts to different types of self-built houses with differentiated scenes, and combines dynamic gradient deduction of hidden dangers to completely solve the problems of homogenization of scoring and misjudgment and omission of risk, and improves the matching degree between scoring results and real fire risk.

[0018] (2) The self-built house fire safety management system of the present invention calculates the target index weight by performing index weight calculation on the initial fire safety comprehensive scoring matrix, balances the professional experience of human and the objective laws of data, adapts to different scenarios through adaptive fusion coefficient, avoids the defects of single algorithm, reduces the index weight error, and effectively improves the reliability of the evaluation system.

[0019] (3) The self-built house fire safety management system of the present invention performs adaptive risk judgment threshold analysis on standard fire safety data through a pre-trained machine learning threshold model, realizes the scenario-based judgment of risk level, can identify multiple hidden risks superimposed, and has a clear priority for rectification of hidden dangers. It solves the problem of one-size-fits-all control caused by the use of fixed thresholds to determine risk level in traditional early warning algorithms, greatly improves the efficiency of grassroots fire management, and improves the rectification closure rate.

[0020] (4) The self-built house fire safety management system of the present invention performs four-level risk accurate judgment by linking three core data: adaptive risk judgment threshold, standard score, and hidden danger superposition risk coefficient. It automatically matches the corresponding control strategy database, covering all urban and rural self-built house scenarios. It does not require manual customization and can adapt to dynamic scenarios such as seasonal changes, house renovation, and change of use. It can stably adapt to the refined management needs of self-built house fire safety in the long term and effectively reduce the probability of fire accidents. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the system modules according to an embodiment of the present invention. Detailed Implementation

[0022] The technical solution of this invention patent will be clearly and completely described below. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this invention.

[0023] See attached document Figure 1 This application provides a fire safety management system for self-built houses, specifically including: a data acquisition module, a cleaning module, a scoring system construction module, an initial scoring module, an indicator weight calculation module, a comprehensive score acquisition module, a calibration module, a standard threshold module, a hazard coefficient extraction module, a superposition coefficient calculation module, a risk level acquisition module, and a safety management result acquisition module.

[0024] The data acquisition module collects two types of data: static inherent data (lifetime fixed or change update data of the building + fire protection infrastructure parameters) and dynamic hazard data (real-time hazard parameters). It achieves accurate collection by category and integrates data entered on-site by manual grid workers (i.e., fire protection infrastructure parameters), real-time data uploaded by power / gas / smoke sensors (i.e., real-time hazard parameters), and official ledger data from housing and construction and emergency departments (i.e., lifetime fixed or change update data of the building). It establishes a temporary data cache pool, breaks down the barriers between multiple data sources, and realizes the collection and summarization of all fire protection data of the same building, avoiding the problems of data dispersion and fragmentation.

[0025] The permanent or updated data for a building includes: location of the self-built house, land area, number of floors, structural material, construction year, renovation records, usage attributes, and maximum occupancy; fire infrastructure parameters include: fire extinguishers, emergency lights, evacuation signs, and fire water sources; real-time hazard parameters include: peak power load, electrical wiring status, gas sealing status, degree of passage blockage, amount of flammable and explosive materials stored, illegal use of fire and electricity, and illegal overnight stays and business operations.

[0026] The cleaning module performs standardized cleaning processing on the collected raw data, sequentially completing outlier removal, duplicate data deduplication, missing data marking and completion, and format normalization. Specifically, it removes erroneous monitoring data exceeding reasonable ranges, deletes redundant data that has been repeatedly entered, performs secondary verification and supplementation of data with missing key fields, and standardizes all data storage formats and unit standards. A unique identification code is assigned to each self-built house, and an independent electronic fire safety file and dedicated database are established based on the cleaned standardized data, binding core identifiers such as house location, ownership, and usage attributes to achieve one file per house and data traceability. The system dynamically updates the database data regularly, providing accurate, complete, and effective raw data support for subsequent indicator modeling, algorithm calculation, risk scoring, and early warning management.

[0027] In this embodiment, the scoring system construction module constructs a target housing scene classification index scoring system using standardized and cleaned data, specifically as follows: The standardized cleaning data is classified into housing scene attributes. The standardized cleaning data is retrieved, and through the attribute tag matching algorithm, all self-built houses are automatically and accurately divided into three categories: pure residential self-built houses, mixed commercial and residential self-built houses, and commercial self-built houses. Housing scene classification tags are generated and entered into the database. Combined with the housing scene classification tags, the indicators are adaptively filtered and adapted.

[0028] Based on the preset housing scenario classification index optimization rules, irrelevant and redundant indicators are removed and scenario-specific high-risk indicators are added for different scenarios to generate an initial housing scenario classification index scoring system. For example, for commercial self-built houses, special indicators such as annual fire inspection, employee drills, and passage compliance are added; for purely residential self-built houses, the commercial-specific indicators are simplified while the core home safety indicators are retained to form a scenario-differentiated index subset, avoiding the problems of indicator redundancy and poor adaptability.

[0029] The initial housing scenario classification index scoring system is scored according to the preset housing scenario classification scoring rules to obtain the housing scenario classification index scoring system to be processed. For all selected sub-indicators, a three-dimensional quantitative scoring mechanism is established, quantifying and assigning values ​​based on three dimensions: the severity of the hazard, the probability of the hazard occurring, and the difficulty of hazard rectification. All hazards are divided into four levels: minor, moderate, relatively serious, and severe, with corresponding gradient deduction values. A rule for cumulative deduction for repeated occurrences of the same type of hazard is also set, replacing the traditional fixed deduction model. This achieves precise matching between the degree of hazard risk and the deduction value, completing the quantification and solidification of the scoring rule data.

[0030] The scoring system for classifying housing scenarios is updated based on seasonal changes, new fire regulations, and fire accident cases to obtain the target housing scenario classification scoring system. External data such as seasonal changes, new fire regulations, and fire accident cases are captured in real time, and the indicator library and deduction rule parameters are dynamically fine-tuned. In summer, scoring indicators are updated to accommodate high-temperature electricity use and the risk of flammable materials being exposed to sunlight; in winter, scoring rules are optimized to accommodate illegal heating and the risk of blocked escape routes; and compliance thresholds for indicators are iterated synchronously when policies and standards are updated, achieving a normalized dynamic update of the scoring indicator system. This ensures that the indicator system aligns with real-time fire safety management needs, ultimately forming a complete, accurate, and adaptive hierarchical dynamic scoring indicator system.

[0031] The initial scoring module scores the standardized cleaning process data based on the target house scene classification index scoring system, and obtains the initial fire safety comprehensive scoring matrix.

[0032] The indicator weight calculation module calculates the indicator weights from the initial fire safety comprehensive scoring matrix to obtain the target indicator weights, specifically: The eigenvalues ​​and eigenvectors are solved using the square root method. First, the product of each row's elements is calculated and its nth root is taken. After normalization, the precise subjective weights of each sub-indicator are obtained. To ensure that subjective weights align with the professional control logic of fire protection, the normalized calculation formula for subjective weights is as follows: ; In the formula: Let be the subjective weight corresponding to the i-th initial comprehensive fire safety scoring index, and the sum of the subjective weights of all indicators is 1; Assign a value to the importance of the i-th initial fire safety comprehensive score index relative to the j-th initial fire safety comprehensive score index in the initial fire safety comprehensive score matrix; This represents the total number of initial comprehensive fire safety scoring indicators; The maximum eigenvalue of the comprehensive fire safety scoring matrix is ​​calculated using the following formula:

[0033] in, , in, The largest eigenvalue of the fire safety comprehensive scoring matrix. Let i be the eigenvalue corresponding to the i-th initial comprehensive fire safety score index. denoted as the subjective weight corresponding to the j-th initial comprehensive fire safety scoring index.

[0034] The CR consistency ratio data is obtained by calculating using the consistency verification formula, as follows: , in, , in, The consistency ratio; As a consistency indicator; It is a random consistency indicator, and its value is a fixed standard value in the industry.

[0035] The consistency of the fire safety comprehensive scoring matrix is ​​checked. If CR > 0.1, it is determined that the initial fire safety comprehensive scoring matrix has subjective bias, and the process returns to the indicator weight calculation module to reconstruct the target house scene classification indicator scoring system. If CR ≤ 0.1, the data of the initial fire safety comprehensive scoring matrix is ​​determined to be valid, and the process proceeds to the next calculation process to completely eliminate subjective bias data.

[0036] According to the preset positive and negative rules, the initial fire safety comprehensive score in the initial fire safety comprehensive score matrix is ​​divided into positive index score and negative index score; Positive and negative indicator scores were preprocessed using positive and negative normalization to eliminate computational interference caused by differences in the units and numerical ranges of different indicators, uniformly mapping all data to the 0-1 numerical range. Different normalization formulas were used for positive indicators (higher values ​​indicate better fire safety) and negative indicators (higher values ​​indicate higher fire risk).

[0037]

[0038] In the formula: The normalized index score is the corresponding to the initial fire safety comprehensive score index in the i-th row and j-th column of the initial fire safety comprehensive score matrix. The score is the score of the indicator corresponding to the initial fire safety comprehensive score indicator in the i-th row and j-th column of the initial fire safety comprehensive score matrix. This represents the maximum value of the initial fire safety comprehensive score index in the j-th column of the initial fire safety comprehensive score matrix. This represents the minimum value of the initial fire safety comprehensive score index in the j-th column of the initial fire safety comprehensive score matrix.

[0039] After normalization, calculate the weight p of the initial comprehensive fire safety score index in the i-th row and j-th column:

[0040] In the formula: The value of the initial comprehensive fire safety score index in the i-th row and j-th column is the proportion of the initial comprehensive fire safety score matrix. The normalized index score is the corresponding to the initial fire safety comprehensive score index in the i-th row and j-th column of the initial fire safety comprehensive score matrix. This represents the total number of rows in the initial fire safety comprehensive scoring matrix.

[0041] The information entropy value E is calculated based on the weighting of indicators. The formula for calculating the entropy value is as follows:

[0042] In the formula: is the information entropy value corresponding to the initial comprehensive fire safety scoring index in column j. The smaller the entropy value, the stronger the risk differentiation ability of the index. This represents the total number of rows in the initial fire safety comprehensive scoring matrix; The value of the initial comprehensive fire safety score index in the i-th row and j-th column is the proportion of the initial comprehensive fire safety score matrix. This is the operator for the natural logarithm.

[0043] Finally, the objective weights corresponding to the initial comprehensive fire safety scoring index in column j are calculated by combining the entropy values. The formula is as follows:

[0044] In the formula: Let be the objective weight corresponding to the initial comprehensive fire safety scoring index in column j, and the sum of the objective weights of all indicators is 1; Let be the information entropy value corresponding to the initial comprehensive fire safety scoring index in column j; This represents the total number of initial comprehensive fire safety scoring indicators. The smaller the information entropy value, the higher the dispersion of the indicator data, the stronger the risk differentiation ability, and the higher the corresponding objective weight ratio, ensuring that the weight assignment is consistent with the actual characteristics of the hidden danger data.

[0045] Subjective and objective weights are calculated to obtain the target indicator weights, and adaptive fusion coefficients are configured for different self-built house scenarios. , The value range is fixed at 0.3-0.7 to achieve a dynamic weighting of subjective and objective factors: For purely residential self-built houses, the following settings are made: =0.3, weakening the weight of subjective human experience and amplifying the weight of objective data to avoid excessive human judgment bias; for self-built houses with mixed commercial and residential uses, set =0.5, achieving a balanced ratio of subjective and objective weights; for commercial self-built houses, set =0.7, strengthening the weight of professional subjective judgment and prioritizing the management of potential hazards in high-risk business scenarios. The formula for calculating the target indicator weight is as follows:

[0046] In the formula: Let be the target indicator weight corresponding to the initial comprehensive fire safety scoring index in column j, and the normalized sum of the weights of all indicators is 1. The scene adaptive fusion coefficient has three fixed adaptive values: 0.3, 0.5, and 0.7. The subjective weight corresponding to the j-th initial comprehensive fire safety scoring index; Let be the objective weight corresponding to the initial comprehensive fire safety scoring index in column j. The final weights of all indicators satisfy the normalization condition. This completes the accurate weighting of all indicators, providing core algorithmic support for subsequent comprehensive scoring.

[0047] The comprehensive scoring module optimizes the initial fire safety comprehensive scoring matrix based on the target indicator weights to obtain a dynamic fire safety comprehensive score, specifically: The initial fire safety comprehensive score in the initial fire safety comprehensive score matrix is ​​weighted and multiplied with the corresponding target indicator weights. All initial fire safety comprehensive scores are summed, and the uncorrected initial comprehensive score is calculated using the initial score formula. Complete the basic score aggregation calculation as follows:

[0048] In the formula: The initial overall fire safety score is the one that has not been revised. Let J represent the target indicator weights corresponding to the initial comprehensive fire safety scoring index in column j. The initial comprehensive fire safety score corresponds to the initial comprehensive fire safety score index in column j. This represents the total number of initial comprehensive fire safety scoring indicators.

[0049] According to the preset matching correction coefficient rules, correction coefficients are extracted for the scoring system of the housing scene classification indicators to be processed and the scoring system of the target housing scene classification indicators to obtain the hidden danger superposition coefficient. and seasonal correction factor It can automatically identify the current overlapping state of hidden dangers in a building and the season, and adaptively match and correct the parameters: when two or more moderate to severe hidden dangers are detected, it automatically matches... ∈[0.8,0.95], when there is no hidden danger accumulation =1.0; Automatically matches when identifying high-risk summer and winter fire seasons. ∈[0.9,0.98], Spring and Autumn are stable seasons =1.0, complete the precise assignment of the correction parameter.

[0050] The uncorrected initial composite score Hidden danger superposition coefficient Seasonal correction factor Substituting into the correction formula, a second correction calculation is performed to avoid scoring deviations caused by multiple overlapping risks and seasonal high-incidence risks, resulting in an accurate comprehensive target score, as shown in the following formula:

[0051] In the formula: A comprehensive score for the target; The initial overall fire safety score is the one that has not been revised. This is the hazard superposition coefficient. When there are two or more moderate to severe hazards, the value range is [0.8, 0.95], and it is fixed at 1.0 when there are no superposition hazards. The seasonal correction coefficient is set between [0.9, 0.98] for the summer and winter fire seasons, while a fixed value of 1.0 is used for the spring and autumn seasons when the risk is relatively stable. This dual-coefficient correction effectively mitigates scoring biases caused by the accumulation of hidden risks from multiple hazard events and the high incidence of seasonal fires.

[0052] The calibration module performs standardized calibration of the target comprehensive score on a 100-point scale, uniformly outputting a standard score of 0-100 points, with higher scores indicating better fire safety status. It stores each scoring result in real time, generates a score change curve daily, dynamically tracks the rise and fall of the building's fire safety status, accurately captures the dynamic processes of new hazards, hazard rectification, and status restoration, and provides intuitive and accurate quantitative scoring basis for subsequent risk warnings.

[0053] The standard threshold module retrieves a pre-trained machine learning threshold model and uses it to predict standard fire safety data, thereby obtaining the basic standard threshold. It automatically matches four types of correction coefficients based on four parameters: building type, number of floors, usage scenario, and population density. , , , The adaptive threshold calculation is performed using a threshold calculation formula to generate an adaptive risk judgment threshold T, replacing the traditional unified fixed threshold.

[0054]

[0055] In the formula: The adaptive risk assessment threshold; The basic standard threshold; This is a correction factor for building type; This is a correction factor for the number of building floors; Adjustment coefficients for housing usage scenarios; The coefficients are used to correct for population density. Each coefficient can be obtained through iterative fitting of machine learning on massive sample data.

[0056] The hazard coefficient extraction module extracts the hazard coefficient from the scoring system of the classification index of the housing scene to be processed according to the preset hazard coefficient extraction rules. The superposition coefficient calculation module calculates the superposition risk coefficient of hazards by substituting the collected number of hazards and their hazard coefficients into the superposition coefficient calculation formula. This quantifies the hidden high-risk factors resulting from the superposition of multiple hidden dangers, avoiding the problem of under-judgment due to low scores for individual hidden dangers and high superposition risks, as shown in the following formula:

[0057] In the formula: The risk coefficient is used to assess potential hazards, with a value range of 0-1. The smaller the value, the higher the risk. This represents the total number of existing effective hidden dangers in self-built houses. Let be the k-th hazard coefficient, where minor hazards are assigned a value of 0.05, general hazards 0.15, relatively serious hazards 0.3, and severe hazards 0.5. The more hazards there are, the higher the hazard level. The smaller the value, the higher the overall risk, enabling accurate identification of multiple hidden risks and eliminating missed detections.

[0058] The risk level acquisition module is linked to the adaptive risk judgment threshold T, standard score, and risk coefficient of hazard superposition. Three types of core data are used to accurately determine four levels of risk: Standard score from 0 to... Score, or a score higher than Divide but Houses with a score ≤0.8 (high-risk hazard superposition) are classified as major risks, triggering a red emergency alert; standard score And 0.8 < Houses with a score ≤0.9 are considered to have a higher risk and trigger an orange alert; standard score to And 0.9 < Houses with a score ≤0.98 are classified as having general risk and trigger a yellow alert; standard score Up to 100 points and Houses with a risk level greater than 0.98 are classified as low-risk and trigger a green normal status.

[0059] The safety management results acquisition module receives the building risk level labels (i.e., four-level risk precise judgment) output by the risk level acquisition module, and automatically matches them with the corresponding control strategy database. Red major risks are matched with emergency shutdown control rules, orange relatively large risks are matched with time-limited rectification rules, yellow general risks are matched with routine inspection rules, and green low risks are matched with regular supervision rules. This achieves adaptive and precise matching between risk levels and control plans. The results are pushed to grid members and building responsible persons through the system terminal, forming an online traceable rectification task ledger, thereby obtaining the safety management results.

[0060] The self-built house fire safety management system of the present invention achieves dynamic and accurate assessment of fire risk in self-built houses by comprehensively scoring and analyzing static inherent data and dynamic hidden danger data. It completes the classification and grading of hidden dangers, adaptive early warning and closed-loop management, which greatly improves the intelligence and refinement of fire safety management in self-built houses and reduces the incidence of fire accidents.

[0061] The foregoing description of specific exemplary embodiments of the invention is for illustrative and explanatory purposes. These descriptions are not intended to limit the invention to the precise forms disclosed, and it will be apparent that many changes and variations can be made in accordance with the foregoing teachings. The exemplary embodiments were chosen and described in order to explain the specific principles of the invention and its practical application, thereby enabling those skilled in the art to implement and utilize various different exemplary embodiments of the invention, as well as various different choices and variations. The scope of the invention is intended to be defined by the claims and their equivalents.

Claims

1. A fire safety management system for self-built houses, characterized in that, include: The data acquisition module is used to collect static inherent data and dynamic potential hazard data, establish a temporary data cache pool, and realize the aggregation and summarization of all fire protection data of the same building; The cleaning module performs standardized cleaning processing on the collected raw data, sequentially completing outlier removal, duplicate data deduplication, missing data marking and completion, and format normalization. The scoring system construction module uses standardized cleaned data to build a target housing scene classification index scoring system. The initial scoring module scores the standardized cleaning process data based on the target house scene classification index scoring system to obtain an initial comprehensive fire safety scoring matrix; The indicator weight calculation module calculates the indicator weights of the initial fire safety comprehensive scoring matrix to obtain the target indicator weights. The comprehensive score acquisition module optimizes the initial fire safety comprehensive score matrix according to the target indicator weights to obtain a dynamic fire safety comprehensive score; The calibration module performs a standardized calibration of the target's comprehensive score on a percentage scale, and outputs a standard score of 0-100 points. The standard threshold module uses a pre-trained machine learning threshold model to predict standard fire safety data and obtain basic standard thresholds. By combining four parameters—building type, number of floors, usage scenario, and population density—the basic standard threshold is modified to generate an adaptive risk assessment threshold. The hazard coefficient extraction module extracts the hazard coefficient from the scoring system of the classification index of the house scene to be processed according to the preset hazard coefficient extraction rules; The superposition coefficient calculation module calculates the superposition risk coefficient of hidden dangers based on the number of collected hidden dangers and the hazard coefficient; The risk level acquisition module is used to link three types of core data: adaptive risk judgment threshold, standard score, and risk coefficient of hidden danger superposition, to perform accurate four-level risk judgment. The safety management results acquisition module is used to receive the output building risk level labels and automatically match them with the corresponding control strategy database.

2. The self-built house fire safety management system according to claim 1, characterized in that, The static inherent data consists of the building's lifetime fixed or change-up data plus fire protection infrastructure parameters. The building's lifetime fixed or change-up data includes: location of the self-built house, land area, number of floors, structural material, construction year, renovation records, usage attributes, and maximum occupancy. The fire protection infrastructure parameters include: fire extinguishers, emergency lights, evacuation signs, and fire water sources. The dynamic hazard data consists of real-time hazard parameters, including: peak power load, electrical circuit status, gas sealing status, degree of passage blockage, amount of flammable and explosive materials stored, illegal use of fire and electricity, and illegal overnight stays and businesses.

3. The self-built house fire safety management system according to claim 1, characterized in that, The scoring system construction module builds a target housing scene classification index scoring system using standardized and cleaned data, specifically including the following steps: The data from the standardized cleaning process is classified according to the attributes of the housing scene, housing scene classification labels are generated and entered into the database; According to the preset house scene classification index optimization rules, irrelevant and redundant indicators are removed from different scenes, and scene-specific high-risk indicators are added to generate an initial house scene classification index scoring system. The initial housing scene classification index scoring system is scored and deducted according to the preset housing scene classification scoring rules to obtain the housing scene classification index scoring system to be processed. The scoring system for classifying housing scenarios was updated based on seasonal changes, new fire regulations, and fire accident cases to obtain the target housing scenario classification scoring system.

4. The self-built house fire safety management system according to claim 1, characterized in that, The indicator weight calculation module calculates the indicator weights of the initial fire safety comprehensive scoring matrix to obtain the target indicator weights, as shown in the following formula: ； in, Let be the target indicator weight corresponding to the initial comprehensive fire safety scoring index in column j, and the normalized sum of the weights of all indicators is 1. The scene adaptive fusion coefficient has three fixed adaptive values: 0.3, 0.5, and 0.

7. The subjective weight corresponding to the j-th initial comprehensive fire safety scoring index; represents the objective weight corresponding to the initial comprehensive fire safety scoring index in column j.

5. The self-built house fire safety management system according to claim 4, characterized in that, The comprehensive scoring module optimizes the initial fire safety comprehensive scoring matrix based on the target indicator weights to obtain a dynamic fire safety comprehensive score, as shown in the following formula: ； ； in, A comprehensive score for the target; The initial overall fire safety score is the one that has not been revised. Let J represent the target indicator weights corresponding to the initial comprehensive fire safety scoring index in column j. The initial comprehensive fire safety score corresponds to the initial comprehensive fire safety score index in column j. This represents the total number of initial comprehensive fire safety scoring indicators; This is the hazard superposition coefficient. When there are two or more moderate to severe hazards, the value range is [0.8, 0.95], and it is fixed at 1.0 when there are no superposition hazards. The seasonal correction factor is set to [0.9, 0.98] for the summer and winter fire seasons when fires are most frequent, and fixed at 1.0 for the spring and autumn seasons when the risk is relatively stable.

6. The self-built house fire safety management system according to claim 1, characterized in that, The standard threshold module modifies the basic standard threshold to generate an adaptive risk assessment threshold, as shown in the following formula: ； in, The adaptive risk assessment threshold; The basic standard threshold; This is a correction factor for building type; This is a correction factor for the number of building floors; Adjustment coefficients for housing usage scenarios; This is a correction factor for population density.

7. The fire safety management system for self-built houses according to claim 1, characterized in that, The superposition coefficient calculation module calculates the superposition risk coefficient of hidden dangers based on the number of collected hidden dangers and the hazard coefficient, as shown in the following formula: ； in, The risk coefficient is used to assess potential hazards, with a value range of 0-1. The smaller the value, the higher the risk. This represents the total number of existing effective hidden dangers in self-built houses. Let be the k-th hazard coefficient, where minor hazards are assigned a value of 0.05, general hazards are assigned a value of 0.15, relatively serious hazards are assigned a value of 0.3, and severe hazards are assigned a value of 0.

5.

8. The self-built house fire safety management system according to claim 1, characterized in that, The risk level acquisition module integrates three core data points—adaptive risk assessment threshold, standard score, and risk coefficient of hazard superposition—to perform a precise four-level risk assessment, as detailed below: Standard score from 0 to Score, or a score higher than Divide but Houses with an elevation of ≤0.8 are classified as having a major risk, triggering a red emergency alert. Standard score to And 0.8 < Houses with a value ≤0.9 are considered to have a higher risk and trigger an orange alert. Standard score to And 0.9 < Houses with an elevation of ≤0.98 are classified as having general risk and trigger a yellow alert. Standard score Up to 100 points and Houses with a risk level greater than 0.98 are classified as low-risk and trigger a green normal status.

9. The self-built house fire safety management system according to claim 8, characterized in that, The safety management result acquisition module receives and outputs the building risk level label, and automatically matches it with the corresponding control strategy database, as follows: Red-level major risks are matched with emergency shutdown and control rules; Orange-level areas with higher risks are subject to rectification within a specified period. Yellow indicates general risk matching and routine inspection rules; Green and low-risk practices align with standard regulatory rules.

Images