Sensor network based commercial complex fire hazard early warning system

Abstract

The present application relates to the technical field of fire monitoring and intelligent early warning, in particular to a commercial complex fire hazard early warning system based on a sensor network, comprising a data collection and screening module, a risk characteristic calculation module, a hazard level assessment module and a resource dynamic regulation module. The data collection and screening module collects multi-region fire parameters of the commercial complex and eliminates invalid data. The risk characteristic calculation module extracts dynamic characteristics of smoke, temperature and oxygen, and calculates the smoke diffusion rate, temperature gradient and oxygen concentration decline rate. The hazard level assessment module constructs a comprehensive measurement value and divides the hazard level. The resource dynamic regulation module calculates the fire resource starting weight according to the level, adjusts the equipment operating state, continuously collects new parameters and executes the process in a loop. The system can eliminate sensor data interference, accurately reflect the change trend in the early stage of a fire, accurately determine the hazard level and dynamically adapt the fire resources, form a closed-loop continuous early warning, and improve the accuracy of hazard identification and the rationality of disposal.

Description

Technical Field

[0001] This invention relates to the field of intelligent early warning technology for fire monitoring, and in particular to a fire hazard early warning system for commercial complexes based on sensor networks. Background Technology

[0002] Traditional fire alarm systems in commercial complexes collect fire parameters such as smoke, temperature, and oxygen by deploying sensors, and directly use instantaneous values ​​for threshold comparison to determine fire hazards. The system does not filter or process the raw sensor data; hazard assessment relies solely on single-point static parameters, failing to extract trend, rate, or gradient features based on parameter changes. Fire equipment linkage uses fixed threshold triggering methods, binding hazard levels to a single equipment action, and executing the same start-stop logic under different hazard situations.

[0003] Raw fire protection parameters contain invalid data due to environmental interference, and direct use of these parameters can easily lead to misjudgments and omissions. Static instantaneous values ​​cannot characterize the dynamic development patterns of a fire incubation stage, resulting in a lag in early hazard identification. Fixed linkage modes cannot match the stages of fire development, and the allocation of fire protection resources lacks flexibility. Single-detection judgments cannot adapt to dynamic changes in fire conditions, and early warning systems lack continuous iteration capabilities.

[0004] Invalid data needs to be removed from fire protection parameters and a preliminary screening needs to be completed. Based on the screened data, dynamic change characteristics of smoke, temperature, and oxygen need to be extracted and the corresponding rates and gradients calculated. A comprehensive metric needs to be constructed based on multiple dynamic characteristics to classify hazard levels. The activation weight of fire protection resources needs to be calculated according to the level, the operating status of equipment needs to be dynamically adjusted, and early warning and control can be completed through parameter updates in a cyclical manner. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a fire hazard early warning system for commercial complexes based on sensor networks.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a fire hazard early warning system for commercial complexes based on sensor networks, comprising: The data acquisition and filtering module obtains fire protection parameters from sensors deployed in different functional areas within the commercial complex, performs preliminary filtering of the fire protection parameters, and removes invalid data. The risk characteristic calculation module identifies features reflecting the trend of smoke accumulation and calculates the smoke diffusion rate from the fire parameters after preliminary screening; it identifies features reflecting the temperature change process and calculates the temperature gradient; and it identifies features reflecting the degree of oxygen consumption and calculates the oxygen concentration decrease rate. The hazard level assessment module constructs a comprehensive metric value characterizing the current fire hazard level of the space based on the smoke diffusion rate, the temperature gradient, and the oxygen concentration decrease rate. The comprehensive metric value is compared with multiple preset hazard level boundaries to determine the current fire hazard level. The resource dynamic control module calculates the activation weights to be allocated to different fire-fighting resources within the commercial complex based on the current level of fire hazard, and generates control signals based on the activation weights. The control signals are used to adjust the operating status of the corresponding fire-fighting resources within the commercial complex. Among them, after adjusting the operational status of fire-fighting resources, the system continuously acquires updated fire-fighting parameters and repeatedly executes the module's functions to achieve continuous early warning of fire hazards.

[0007] As a further aspect of the present invention, the preliminary screening of fire protection parameters to remove invalid data specifically includes: Receive raw data stream uploaded by the sensing device, the raw data stream containing timestamps and corresponding fire parameter measurement values; Within a continuous preset time window, the data stream of each sensor is sampled, and the dispersion of the measured values ​​within the time window is calculated. The degree of dispersion of the measured values ​​is compared with the dispersion threshold. If the degree of dispersion exceeds the dispersion threshold, it is determined that there is noise interference in the data within the time window. Measurements within the time window where noise interference is identified are marked as invalid and removed from the original data stream; measurements within the time window where the dispersion is below the threshold are retained. From the retained measurements, outliers that deviate from the historical normal data model by more than the tolerance range are identified, and these outliers are marked as data to be verified. From the retained measurements, data marked as pending verification are further removed to obtain the preliminary filtered fire protection parameters.

[0008] As a further aspect of the present invention, the step of identifying features reflecting smoke accumulation trends from the initially screened fire parameters and calculating the smoke diffusion rate specifically includes: From the initially screened fire parameters, the smoke concentration measurement value sequence was extracted and arranged in chronological order. In the time series, the difference in smoke concentration between adjacent time points is calculated to obtain the instantaneous change in smoke concentration; The instantaneous changes over multiple consecutive time periods are summed to calculate the cumulative increase in smoke concentration over a specified time period. The cumulative increase in smoke concentration is divided by the corresponding specified time length to obtain the average increase in smoke concentration per unit time, and the average increase is used as the initial smoke diffusion rate. Based on the topological location of the sensing device within the commercial complex space and the preliminary smoke diffusion rate, the rate at which smoke travels from the location of the sensing device with high concentration to a nearby location is calculated. The smoke diffusion rate of the entire monitoring area is calculated by combining the smoke transfer rates of multiple sensing devices and using a weighted average.

[0009] As a further aspect of the present invention, the step of identifying features reflecting the temperature change process and calculating the temperature gradient specifically includes: From the initial screening of fire parameters, temperature measurement values ​​from sensors at different locations were extracted and grouped. Within the same time period, select two adjacent locations within the commercial complex space and obtain the temperature measurements at these two locations. Calculate the temperature difference between two adjacent locations and divide the temperature difference by the actual physical distance between the two locations to obtain the local temperature change rate in the direction. Calculations were performed in multiple different directions within the commercial complex space to obtain the local temperature change rate in multiple different directions; Based on the distribution density of people and goods within the commercial complex, different weights are assigned to the local temperature change rate in different directions, and the weighted average spatial temperature gradient is calculated.

[0010] As a further aspect of the present invention, the step of identifying features reflecting the degree of oxygen consumption and calculating the rate of decrease in oxygen concentration includes: From the initially screened fire parameters, the oxygen concentration measurement value sequence of sensors at different locations was extracted; Based on the time reference, the oxygen concentration measurement sequence for each sensor is arranged in chronological order; Calculate the difference between the oxygen concentration measurement value of each sensor at the current time point and the preset historical normal baseline value to obtain the absolute change in oxygen concentration at the location of each sensor. Within a preset fixed time interval, the direction and magnitude of the absolute change in oxygen concentration at each sensor location are statistically analyzed, and sensors with the same direction of change and a change exceeding a set threshold are classified into the same change region. For each defined change region, the sensor with the largest absolute change in oxygen concentration within the change region is selected as a representative point, and the oxygen concentration measurement sequence of the representative point is obtained over the entire fixed time interval. By performing linear fitting on the oxygen concentration measurement sequence of representative points, the average rate of decrease of oxygen concentration over time in the changing region can be calculated. By combining the average rate of decrease across all affected areas, and using a weighted average of the area values, the oxygen concentration decrease rate for the entire monitoring area is obtained.

[0011] As a further aspect of the present invention, a comprehensive metric for characterizing the current fire hazard of the space is constructed based on the smoke diffusion rate, the temperature gradient, and the oxygen concentration decrease rate, specifically including: The calculated smoke diffusion rate is normalized and mapped to a numerical range of zero to one. The calculated temperature gradient is normalized and mapped to a numerical range of zero to one. The calculated rate of decrease in oxygen concentration was normalized and mapped to a numerical range of zero to one. Assign a weighting factor to the normalized smoke diffusion rate, a weighting factor to the normalized temperature gradient, and a weighting factor to the normalized oxygen concentration decrease rate. The normalized smoke diffusion rate is multiplied by its corresponding weighting coefficient, the normalized temperature gradient is multiplied by its corresponding weighting coefficient, and the normalized oxygen concentration decrease rate is multiplied by its corresponding weighting coefficient to obtain three weighted parameters. The three weighted parameters are linearly superimposed and summed to obtain a comprehensive metric value that characterizes the current fire hazard in the space.

[0012] As a further aspect of the present invention, the comprehensive measurement value is compared with multiple preset hazard level boundaries to determine the current level of fire hazard, specifically including: A set of incremental hazard level boundary values ​​is pre-defined, with each hazard level boundary value corresponding to a fire hazard level; Obtain the calculated comprehensive metric value, starting from the lowest hazard level boundary value, and compare the comprehensive metric value with the preset hazard level boundary value in sequence; When the comprehensive metric value is greater than or equal to a certain hazard level boundary value, but less than the next higher hazard level boundary value, the current level of fire hazard is determined to be the level corresponding to the hazard level boundary value. If the comprehensive metric value is less than the lowest hazard level boundary value, then the current fire hazard level is determined to be a safe level; The current level of fire hazard is output as input for subsequent resource allocation decisions.

[0013] As a further aspect of the present invention, based on the current level of fire hazard, the activation weights to be allocated to different fire response resources within the commercial complex are calculated, specifically including: The fire-fighting resources include smoke exhaust fans, fire pumps, and emergency lighting and evacuation guidance systems; Establish a fire response resource list, which includes all available fire-fighting equipment and their basic attributes within the commercial complex; For different fire hazard levels, corresponding resource allocation strategies are preset, and the resource allocation strategies define the expected activation intensity of various fire-fighting equipment under each level; Based on the current level of fire hazard, query the corresponding resource allocation strategy and obtain the expected activation intensity requirements for each type of fire-fighting equipment under the current level; Calculate the response efficiency coefficient for each type of fire-fighting equipment to reach the designated risk area based on its current availability and physical location; The expected activation intensity requirement is combined with the corresponding response efficiency coefficient to calculate the theoretical activation weight of each type of fire-fighting equipment in the current situation. The theoretical activation weights of all fire-fighting equipment of the same type within the commercial complex are summarized and standardized to obtain the final activation weights required for each type of fire-fighting resource.

[0014] As a further aspect of the present invention, a control signal is generated based on the activation weight. This control signal is used to adjust the operational status of corresponding fire-fighting resources within the commercial complex, specifically including: The calculated final activation weight of each type of fire-fighting resource is converted into an instruction format that can be recognized by the corresponding fire-fighting equipment control unit. For exhaust fans, control signals containing the number of fans to be started and the speed level are generated based on the start weights assigned to them. For fire pumps, a control signal containing the number of pumps to be started and the set value of the outlet pressure is generated according to the start weight assigned to them. For emergency lighting and evacuation guidance systems, control signals containing lighting brightness levels and guidance paths are generated based on the activation weights assigned to them. The generated control signals are distributed to the corresponding fire equipment control units through the fire protection network within the commercial complex. The fire equipment control unit receives and parses control signals, drives the actuators of the fire equipment, and adjusts the operating status of the fire equipment to the target state specified by the control signal.

[0015] As a further aspect of the present invention, after adjusting the operating status of fire-fighting resources, continuously acquiring updated fire-fighting parameters and repeatedly executing the module's function to achieve continuous early warning of fire hazards specifically includes: At a set time interval after the fire-fighting equipment adjusts its operating status according to the control signal, new fire-fighting parameters are re-obtained from the sensing devices deployed in different functional areas inside the commercial complex. The complete process of performing preliminary screening, feature identification, comprehensive metric construction, hazard level determination, activation weight calculation and control signal generation for new fire protection parameters; The activation weights of fire response resources generated by the new process are compared and analyzed with the activation weights used in the previous round of adjustments to calculate the direction and magnitude of the weight changes. If the weight changes indicate an increasing trend in fire hazard, the newly generated activation weights and control signals will be immediately used to strengthen the adjustment of the operational status of fire response resources. If the weight changes indicate that the fire hazard is decreasing, the newly generated activation weights will be compared with the current operating status to fine-tune the operating status of fire response resources in a gradual manner. The process of acquiring, analyzing, making decisions, and adjusting is continuously repeated until all fire parameters acquired from the sensing devices fall back to within the preset safety threshold range.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: Preliminary screening of fire protection parameters collected by sensors in different areas of the commercial complex is performed to eliminate invalid data and abnormal fluctuations, reducing the impact of external environment and sensor errors on the analysis results. Based on valid fire protection parameters, the dynamic change characteristics of smoke, temperature, and oxygen are extracted, and the quantitative calculations of smoke diffusion rate, temperature gradient, and oxygen concentration decrease rate are completed. Through precise calculation of parameter change rates and spatial gradients, the dynamic evolution of the fire protection environment can be accurately reflected, reducing the judgment error caused by single instantaneous detection values, mitigating the lag of traditional threshold detection, and improving the accuracy and sensitivity of hazard identification.

[0017] A comprehensive fire hazard metric is constructed based on smoke diffusion rate, temperature gradient, and oxygen concentration decrease rate. Hazard levels are then classified by comparing these metrics with preset multi-level boundaries. Activation weights for various fire-fighting resources are calculated according to the hazard level, and corresponding control signals are generated based on these weights to achieve precise adjustment of fire equipment operation. Real-time fire parameters are continuously collected after equipment status adjustments, and the data filtering, feature calculation, level assessment, and resource control processes are repeated. Differentiated scheduling of fire resources is achieved through weight allocation to match the response needs of different hazard levels. Closed-loop iterative calculations can synchronously update assessment results and control commands according to changes in the environmental situation, maintaining the continuity of the early warning process, enhancing the system's adaptability to the fire development process, and improving the synergy between early warning and response. Attached Figure Description

[0018] Figure 1 This is a timing diagram of the fire hazard early warning system for commercial complexes based on sensor networks described in this invention. Figure 2 A flowchart for preliminary screening and elimination of invalid data for fire protection parameters; Figure 3 A flowchart for identifying temperature change characteristics and calculating temperature gradients. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0021] See Figure 1 This invention provides a fire hazard early warning system for commercial complexes based on sensor networks. The specific system includes: The data acquisition and filtering module obtains fire parameters from sensors deployed in different functional areas within the commercial complex, and performs preliminary filtering to eliminate invalid data. The risk characteristic calculation module identifies features reflecting smoke accumulation trends and calculates the smoke diffusion rate from the pre-filtered fire parameters, identifies features reflecting temperature changes and calculates the temperature gradient, and identifies features reflecting oxygen consumption and calculates the oxygen concentration decrease rate. The hazard level assessment module constructs a comprehensive metric value characterizing the current fire hazard of the space based on the smoke diffusion rate, temperature gradient, and oxygen concentration decrease rate, and compares this comprehensive metric value with multiple preset hazard level boundaries to determine the current fire hazard level. The resource dynamic control module calculates the activation weights required for different fire response resources within the commercial complex based on the current fire hazard level, generates control signals based on the activation weights to adjust the operating status of the corresponding fire response resources within the commercial complex, and continuously acquires updated fire parameters after adjusting the operating status of the fire response resources, repeating the functions of the above modules to achieve continuous early warning of fire hazards.

[0022] In one embodiment of the present invention, see [reference] Figure 2 The data acquisition and filtering module receives the raw data stream uploaded by the sensing devices. The raw data stream contains timestamps and corresponding fire parameter measurement values. Within a continuous preset time window, the module samples the data stream of each sensing device and calculates the dispersion of the measurement values ​​within the time window. The dispersion of the measurement values ​​is compared with a dispersion threshold. If the dispersion exceeds the dispersion threshold, it is determined that the data within the time window contains noise interference. The measurement values ​​within the time window that are determined to contain noise interference are marked as invalid and removed from the raw data stream, retaining the measurement values ​​within the time window whose dispersion is below the threshold. From the retained measurement values, outliers that deviate from the historical normal data model beyond the tolerance range are identified and marked as data to be verified. The data marked as data to be verified is further removed from the retained measurement values ​​to obtain the fire parameters after preliminary filtering.

[0023] From the initially screened fire safety parameters, a sequence of smoke concentration measurements was extracted and arranged chronologically. The difference in smoke concentration between adjacent time points was calculated to obtain the instantaneous change in smoke concentration. The instantaneous changes over multiple consecutive time periods were summed to calculate the cumulative increase in smoke concentration over a specified time period. The cumulative increase in smoke concentration was divided by the corresponding specified time period to obtain the average increase in smoke concentration per unit time, which was used as the initial smoke diffusion rate. Based on the topological location of the sensors within the commercial complex, the rate at which smoke travels from a high-concentration sensor location to a neighboring location was calculated using the initial smoke diffusion rate. Finally, the smoke diffusion rate of the entire monitoring area was calculated by weighted averaging the smoke transmission rates between multiple sensors.

[0024] In a specific implementation, the data acquisition and filtering module receives the raw data stream uploaded by the sensing device. The raw data stream includes timestamps and corresponding fire parameter measurement values. For example, within a preset time window of sixty seconds, the data stream of a specific smoke sensor located in the clothing area of ​​a commercial complex is sampled, obtaining sixty smoke concentration measurement values, which are 0.05%, 0.06%, 0.07%, 0.08%, 0.09%, 0.11%... and 0.12%, respectively. The standard deviation of this set of measurement values ​​is calculated as an indicator of dispersion, and the standard deviation is found to be 0.025%. Within another preset time window of the same length, the sensor is affected by temporary decoration dust, and the measurement values ​​fluctuate drastically, which are 0.03%, 0.15%, 0.02%, 0.18%, 0.01%... and 0.20%, respectively, and the standard deviation is calculated to be 0.085%. The dispersion of the two sets of data was compared with a preset dispersion threshold of 0.050%. The dispersion of the former did not exceed the threshold, while the dispersion of the latter exceeded the threshold. Therefore, it was determined that the data within the time window of the latter contained noise interference, and all measurements within that time window were marked as invalid and removed from the original data stream. The measurements within the time window of the former were retained. From the retained measurements, outliers were identified using statistical intervals constructed based on a historical normal data model. If a measurement value was 0.14%, while the upper limit of the mean of the historical normal data model was 0.13%, exceeding the tolerance range, this measurement value was marked as data to be verified. Subsequently, this data to be verified was further removed from the retained set, finally obtaining the preliminary filtered fire parameters.

[0025] In some embodiments, a sequence of smoke concentration measurements is extracted from the initially screened fire parameters. Assuming the screened sequence is [0.052%, 0.054%, 0.058%, 0.063%, 0.068%, 0.074%], corresponding to time points respectively. to The time interval between adjacent time points is five seconds. The difference in smoke concentration between adjacent time points is calculated on the time series, for example... Time and The time difference is 0.002%. Time and The time difference is 0.004%, and so on, to obtain a series of instantaneous changes. The instantaneous changes over a total duration of thirty seconds for multiple consecutive time periods are summed to obtain a cumulative increase in smoke concentration of 0.022%. This cumulative increase is divided by the corresponding specified time length of thirty seconds to obtain an average increase in smoke concentration per unit time of 0.00073% per second, which is used as the preliminary smoke diffusion rate.

[0026] It is understandable that, considering the topological location of the sensing devices within the commercial complex space, if the location of the sensing devices with high concentrations is known... and neighboring locations Ten meters apart, location The initial smoke diffusion rate was 0.001% per second, at the location The initial smoke diffusion rate is 0.0005% per second. The smoke diffusion rate from location can be calculated based on the distance and rate difference between the two locations. To position The transmission rate is calculated using the following formula. : ; in: Indicates position The initial smoke diffusion rate, Indicates position The initial smoke diffusion rate, Indicates position With position The actual physical distance between them. Substituting the values, the transmission rate is 0.00005% per second per meter. This refers to the smoke transmission rate between multiple sensors within a comprehensive commercial complex, such as location. to The transmission rate is 0.00003% per second per meter, and the position is... to The transmission rate is 0.00004% per second per meter. The smoke diffusion rate of the entire monitoring area is finally obtained by weighted averaging with the coverage area of ​​each transmission path as the weight.

[0027] In one embodiment of the present invention, see [reference] Figure 3From the initially screened fire safety parameters, temperature measurements from sensors at different locations were extracted and grouped. Within the same time period, two adjacent locations within the commercial complex were selected, and their temperature measurements were obtained. The temperature difference between the two adjacent locations was calculated, and the difference was divided by the actual physical distance between them to obtain the local temperature change rate in that direction. Calculations were performed in multiple directions within the commercial complex to obtain the local temperature change rates in multiple directions. Based on the distribution density of people and goods within the commercial complex, different weights were assigned to the local temperature change rates in different directions, and a weighted average spatial temperature gradient was calculated.

[0028] From the initially screened fire parameters, oxygen concentration measurement sequences from sensors at different locations were extracted. Based on a time reference, the oxygen concentration measurement sequences for each sensor were arranged chronologically. The difference between the current oxygen concentration measurement value and the preset historical normal baseline value for each sensor was calculated to obtain the absolute change in oxygen concentration at each sensor location. Within a preset fixed time interval, the direction and magnitude of the absolute change in oxygen concentration at each sensor location were statistically analyzed. Sensors with consistent directions of change and changes exceeding a set threshold were grouped into the same change region. For each defined change region, the sensor with the largest absolute change in oxygen concentration within that region was selected as a representative point, and the oxygen concentration measurement sequence for that representative point was obtained over the entire fixed time interval. A linear fit was performed on the oxygen concentration measurement sequence of the representative point to calculate the average rate of decrease in oxygen concentration over time within the change region. The average rate of decrease across all change regions was then combined and weighted by area to calculate the oxygen concentration decrease rate for the entire monitoring area.

[0029] In the specific implementation, temperature measurement values ​​from sensors at different locations are extracted from the initially screened fire safety parameters. These extracted temperature values ​​are then grouped according to a spatial grid. For example, the grid for the first-floor dining area of ​​the commercial complex contains four adjacent sensors with temperature measurements of 25.3℃, 25.8℃, 26.1℃, and 25.9℃, respectively. The grid for the second-floor clothing area of ​​the commercial complex contains four adjacent sensors with temperature measurements of 24.7℃, 24.9℃, 25.2℃, and 25.0℃, respectively. Within the same time period, two adjacent points within the commercial complex space are selected, such as horizontally adjacent points P1 and P2 in the dining area grid. The temperature measurements of these two points are obtained as 25.3℃ and 25.8℃, respectively. The temperature difference between the two adjacent points is calculated as ΔT = 25.8℃ - 25.3℃ = 0.5℃. This temperature difference is then divided by the actual physical distance d = 4.0m between the two points to obtain the local temperature change rate in the direction, which is 0.125℃ / m. Calculations are performed in multiple directions within the commercial complex space. For example, in the grid of the dining area, points P1 and P3 are selected vertically, and the calculated local temperature change rate is 0.075℃ / m. In the grid of the clothing area, points Q1 and Q2 are selected horizontally, and the calculated local temperature change rate is 0.100℃ / m. Based on the distribution density of people and goods within the commercial complex, for example, the dining area has a higher personnel density and is assigned a weight of 0.60, while the clothing area has a lower personnel density and is assigned a weight of 0.40. The average local temperature change rate of 0.100℃ / m measured in the two directions in the dining area is multiplied by the weight 0.60, and the local temperature change rate of 0.100℃ / m measured in the clothing area is multiplied by the weight 0.40. The weighted average spatial temperature gradient is calculated as 0.060 + 0.040 = 0.100℃ / m.

[0030] In some embodiments, oxygen concentration measurement sequences from sensors at different locations are extracted from the initially screened fire safety parameters. For example, the oxygen concentration measurement sequence for sensor S1 in the catering area is [20.91%, 20.88%, 20.84%, 20.80%, 20.75%], and the oxygen concentration measurement sequence for sensor S2 in the clothing area is [20.90%, 20.89%, 20.87%, 20.85%, 20.83%]. Based on a time reference, the oxygen concentration measurement sequences for each sensor are arranged chronologically, with a time interval of ten seconds. The difference between the oxygen concentration measurement value of each sensor at the current time point and the preset historical normal baseline value of 20.95% is calculated. For example, the absolute change of the current measurement value of S1 (20.75%) is -0.20%, and the absolute change of the current measurement value of S2 (20.83%) is -0.12%. Within a preset fixed time interval of fifty seconds, the absolute change in oxygen concentration at each sensor location was statistically analyzed. The direction of change was negative and the change amount exceeded the set threshold of 0.05%. These sensors were then divided into the same change area, namely the northwest area composed of the catering area and the clothing area.

[0031] Optionally, for each defined variation region, the sensor S1 with the largest absolute change in oxygen concentration within that region is selected as a representative point. The oxygen concentration measurement sequence of this representative point over the entire fixed time interval of fifty seconds is obtained [20.91%, 20.88%, 20.84%, 20.80%, 20.75%]. A linear fit is performed on the oxygen concentration measurement sequence of the representative point. Assuming the time variable is t (unit: seconds) and the oxygen concentration is C (unit: %), the linear relationship is obtained as C = -0.0032t + 20.912. The average rate of decrease in oxygen concentration over time within the variation region is calculated to be 0.0032% / s. By combining the average rate of decrease across all affected areas—for example, the average rate of decrease in the northwest region is 0.0032% / s, and the average rate of decrease in the southeast region is 0.0025% / s; the area of ​​the northwest region is 800 m², and the area of ​​the southeast region is 600 m²—the oxygen concentration decrease rate for the entire monitoring area is calculated using a weighted average of the area values. The result is (0.0032×800+0.0025×600) / (800+600)=0.0286 / 1400≈0.00002043% / s.

[0032] It is understandable that the calculation of temperature gradient depends on the ratio of temperature difference to distance between adjacent points in space. By measuring in multiple directions and introducing personnel density weight, the heat transfer characteristics in densely populated areas can be more accurately reflected. The calculation of oxygen concentration decrease rate is achieved by identifying areas of change in the same direction and selecting representative points for linear fitting. Combined with area weighting, this can effectively characterize the oxygen consumption trend of the entire monitoring area.

[0033] In one embodiment of the present invention, the calculated smoke diffusion rate is normalized and mapped to a numerical range of zero to one; the calculated temperature gradient is normalized and mapped to a numerical range of zero to one; the calculated oxygen concentration decrease rate is normalized and mapped to a numerical range of zero to one; a weighting coefficient is assigned to the normalized smoke diffusion rate, the normalized temperature gradient, and the normalized oxygen concentration decrease rate; the normalized smoke diffusion rate is multiplied by its corresponding weighting coefficient, the normalized temperature gradient is multiplied by its corresponding weighting coefficient, and the normalized oxygen concentration decrease rate is multiplied by its corresponding weighting coefficient to obtain three weighted parameters; the three weighted parameters are linearly superimposed and summed to obtain a comprehensive metric value characterizing the current space fire hazard.

[0034] A set of incremental hazard level boundary values ​​is pre-defined, with each boundary value corresponding to a fire hazard level. A calculated comprehensive metric value is obtained, and starting from the lowest hazard level boundary value, it is compared sequentially with each of the pre-defined boundary values. If the comprehensive metric value is greater than or equal to a certain hazard level boundary value but less than the next higher boundary value, the current fire hazard level is determined to be the level corresponding to that boundary value. If the comprehensive metric value is less than the lowest hazard level boundary value, the current fire hazard level is determined to be a safe level. The determined current fire hazard level is output as input for subsequent resource allocation decisions.

[0035] In practice, the calculated smoke diffusion rate, temperature gradient, and oxygen concentration decrease rate are normalized to map them to a value range of zero to one. Assuming the smoke diffusion rate in a certain area of ​​a commercial complex is 0.008% per second at a given moment, and the system's maximum reference value for this type of parameter is known to be 0.010% per second and the minimum reference value is 0.001% per second, the normalized smoke diffusion rate, calculated through linear mapping, is approximately (0.008-0.001) / (0.010-0.001)≈0.778; the monitored temperature gradient is 0.150℃ / m, and the known maximum reference value is 0.200℃. / m, the minimum reference value is 0.050℃ / m, and the normalized temperature gradient is (0.150-0.050) / (0.200-0.050)≈0.667; the monitored oxygen concentration decrease rate is 0.005% per second, the known maximum reference value is 0.006% per second, the minimum reference value is 0.001% per second, and the normalized oxygen concentration decrease rate is (0.005-0.001) / (0.006-0.001)=0.800.

[0036] In some embodiments, weighting coefficients are assigned to the normalized smoke diffusion rate, the normalized temperature gradient, and the normalized oxygen concentration decrease rate. These weighting coefficients are set based on the parameters' contribution to fire risk. The smoke diffusion rate, reflecting the speed of smoke spread, is assigned a larger weight of 0.50; the temperature gradient, reflecting the degree of heat accumulation, is assigned a weight of 0.30; and the oxygen concentration decrease rate, reflecting the oxygen consumption during combustion, is assigned a weight of 0.20. Multiplying the normalized smoke diffusion rate (0.778) by its corresponding weighting coefficient (0.50) yields a weighted parameter of 0.389; multiplying the normalized temperature gradient (0.667) by its corresponding weighting coefficient (0.30) yields a weighted parameter of 0.200; and multiplying the normalized oxygen concentration decrease rate (0.800) by its corresponding weighting coefficient (0.20) yields a weighted parameter of 0.160.

[0037] Optionally, the three weighted parameters mentioned above can be linearly superimposed and summed to obtain a comprehensive metric value representing the current fire hazard of the space. The calculation formula is as follows: ; in: This represents a comprehensive metric. The weighting coefficients represent the normalized smoke diffusion rate. This represents the normalized smoke diffusion rate. These represent the weighting coefficients of the normalized temperature gradient. This represents the normalized temperature gradient. The weighting coefficients represent the normalized rate of decrease in oxygen concentration. This represents the rate of decrease in oxygen concentration after normalization. Substituting the values ​​into the calculation yields the comprehensive metric. .

[0038] It is understandable that a set of incremental hazard level boundary values ​​is pre-defined, with each boundary value corresponding to a fire hazard level. For specific boundary value settings, please refer to Table 1. Table 1: Hazard Level Boundary Values ​​and Corresponding Levels

[0039] The calculated composite metric value of 0.749 is obtained. Starting from the lowest hazard level boundary value of 0.300, the composite metric value is compared sequentially with the preset hazard level boundary values. If the composite metric value of 0.749 is greater than or equal to the hazard level boundary value of 0.700, but less than the next higher hazard level boundary value of 0.900, the current fire hazard level is determined to be Level 3 (high risk). If the composite metric value is less than the lowest hazard level boundary value of 0.300, the current fire hazard level is determined to be a safe level. The determined current fire hazard level is output as input for subsequent resource allocation decisions.

[0040] In one embodiment of the present invention, fire-fighting resources include smoke exhaust fans, fire pumps, and emergency lighting and evacuation guidance systems. A fire-fighting resource list is established, which includes all available fire-fighting equipment and their basic attributes within the commercial complex. Corresponding resource allocation strategies are preset for different fire hazard levels, and each resource allocation strategy defines the expected activation intensity of various types of fire-fighting equipment at each level. Based on the current fire hazard level, the corresponding resource allocation strategy is queried to obtain the expected activation intensity requirement for each type of fire-fighting equipment at the current level. Based on the current availability and physical location of each type of fire-fighting equipment, its response efficiency coefficient for reaching the designated risk area is calculated. The expected activation intensity requirement and the corresponding response efficiency coefficient are fused together to obtain the theoretical activation weight of each type of fire-fighting equipment in the current situation. The theoretical activation weights of all fire-fighting equipment of the same type within the commercial complex are summarized and standardized to obtain the final activation weight to be allocated to each type of fire-fighting resource.

[0041] The calculated final activation weights for each type of fire-fighting resource are converted into instruction formats recognizable by the corresponding fire equipment control units. For smoke exhaust fans, control signals containing the number of fans to be activated and their speed settings are generated based on their assigned activation weights. For fire pumps, control signals containing the number of pumps to be activated and their outlet pressure setpoints are generated based on their assigned activation weights. For emergency lighting and evacuation guidance systems, control signals containing lighting brightness levels and guidance paths are generated based on their assigned activation weights. These different control signals are distributed to the corresponding fire equipment control units through the dedicated fire protection network within the commercial complex. The fire equipment control units receive and parse the control signals, drive the actuators of the fire equipment, and adjust the operating status of the fire equipment to the target state specified by the control signals.

[0042] In specific implementation, fire-fighting resources include smoke exhaust fans, fire pumps, and emergency lighting and evacuation guidance systems. A fire-fighting resource list is established, which includes all available fire-fighting equipment and their basic attributes within the commercial complex. For example, the first floor of the commercial complex may have three smoke exhaust fans (numbered FAN-01 to FAN-03), two fire pumps (numbered PUMP-A1 to PUMP-A2), and one emergency lighting and evacuation guidance system (numbered LIGHT-SYS-Z1); the second floor may have two smoke exhaust fans (numbered FAN-04 to FAN-05), one fire pump (numbered PUMP-B1), and one emergency lighting and evacuation guidance system (numbered LIGHT-SYS-Z2). In some embodiments, corresponding resource allocation strategies are preset for different fire hazard levels. The resource allocation strategy defines the expected activation intensity of various fire-fighting equipment at each level. For specific rules, please refer to Table 2. Table 2: Strategies for Expected Activation Intensity of Firefighting Equipment under Different Hazard Levels

[0043] Based on the current level of fire hazard (e.g., Level 3 high risk), query the corresponding resource allocation strategy to obtain the expected activation intensity requirements for each type of fire-fighting equipment under the current level: the activation intensity requirement for smoke exhaust fans is 70% of the total, the activation intensity requirement for fire pumps is 65% of the total, and the activation intensity requirement for emergency lighting and evacuation guidance systems is Level 5 brightness and emergency path mode.

[0044] Optionally, the response efficiency coefficient for reaching a designated risk area is calculated based on the current availability and physical location of each type of fire-fighting equipment. For example, if the risk area is located in the center of the catering area on the first floor of a commercial complex, 15 meters from smoke exhaust fan FAN-01, 20 meters from smoke exhaust fan FAN-02, and 25 meters from smoke exhaust fan FAN-03; and if the system defines the response efficiency coefficient as inversely proportional to distance, with a coefficient of 1.0 at a baseline distance of 20 meters, then the response efficiency coefficient for FAN-01 is 1.333, for FAN-02 it is 1.000, and for FAN-03 it is 0.800. The expected activation intensity requirement is then fused with the corresponding response efficiency coefficient to obtain the theoretical activation weight of each type of fire-fighting equipment in the current situation. The fusion calculation formula is as follows: ; in: This represents the theoretical activation weight of the i-th fire-fighting device. This indicates the expected activation strength requirement for this type of fire-fighting equipment under the current hazard level. Let represent the response efficiency coefficient of the i-th fire-fighting equipment. For the smoke exhaust fan FAN-01, substituting the numerical values, the theoretical start-up weight is approximately 70% × 1.333 ≈ 93.31%.

[0045] It is understandable that the theoretical activation weights of all fire-fighting equipment of the same type within a commercial complex are aggregated and standardized to obtain the final activation weights required for each type of fire-fighting resource. For example, the theoretical activation weights of the three smoke exhaust fans on the first floor are 93.31%, 70.00%, and 56.00%, respectively, with a total of 219.31%. After standardization, each weight is divided by the sum, resulting in a final activation weight of approximately 42.55% for FAN-01, approximately 31.92% for FAN-02, and approximately 25.53% for FAN-03.

[0046] In practical implementation, the calculated final activation weight of each type of fire-fighting resource is converted into an instruction format recognizable by the corresponding fire-fighting equipment control unit. The conversion rule is to map the percentage weight to the inherent control code range of the equipment. For example, the speed range of smoke exhaust fans is divided into ten levels, and the final activation weight of 42.55% corresponds to the fourth speed level. For smoke exhaust fans, based on the assigned activation weight, a control signal containing the number of fans to be activated and the speed range is generated, for example, the control signal code is {FAN-01:START,SPEED=4;FAN-02:START,SPEED=3}. For fire pumps, based on the assigned activation weight, a control signal containing the number of pumps to be activated and the set value of the outlet water pressure is generated, for example, the control signal code is {PUMP-A1:START,PRESSURE=0.8MPa}. For emergency lighting and evacuation guidance systems, based on the assigned activation weight, a control signal containing the lighting brightness level and the indication path is generated, for example, the control signal code is {LIGHT-SYS-Z1:LEVEL=5,PATH_MODE=EMERGENCY}. The generated control signals are distributed to the corresponding fire equipment control units through the fire protection network within the commercial complex. The fire equipment control units receive and parse the control signals, drive the actuators of the fire equipment, and adjust the operating status of the fire equipment to the target state specified by the control signals.

[0047] In one embodiment of the present invention, after the fire-fighting equipment adjusts its operating status according to the control signal, at a set time interval, new fire-fighting parameters are re-acquired from sensors deployed in different functional areas within the commercial complex. A complete process is then executed for the new fire-fighting parameters, including preliminary screening, feature identification, comprehensive metric construction, hazard level determination, activation weight calculation, and control signal generation. The activation weights of the fire-fighting resources generated in the new process are compared and analyzed with the activation weights used in the previous adjustment round to calculate the direction and magnitude of the weight changes. If the weight changes indicate an increasing trend in fire hazard, the newly generated activation weights and control signals are immediately used to strengthen the adjustment of the operating status of the fire-fighting resources. If the weight changes indicate a decreasing trend in fire hazard, the newly generated activation weights are compared with the current operating status to gradually fine-tune the operating status of the fire-fighting resources. This process of acquisition, analysis, decision-making, and adjustment is continuously cyclically executed until all fire-fighting parameters acquired from the sensors fall back to within the preset safety threshold range.

[0048] In practice, after the fire-fighting equipment adjusts its operating status according to the control signal, at a set time interval, such as sixty seconds, new fire parameters are obtained again from the sensors deployed in different functional areas inside the commercial complex. Assume that a set of sensors in the catering area on the first floor of the commercial complex returns the updated smoke concentration measurement sequence as [0.102%, 0.115%, 0.128%, 0.142%, 0.155%], the temperature measurement sequence as [28.5℃, 29.1℃, 29.8℃, 30.5℃, 31.2℃], and the oxygen concentration measurement sequence as [20.68%, 20.62%, 20.57%, 20.51%, 20.46%]. The process involves initial screening, feature identification, construction of comprehensive metrics, determination of hazard level, calculation of activation weights, and generation of control signals for new fire protection parameters. For example, the calculated activation weights for the new round of fire protection resources are 0.850 for smoke exhaust fans, 0.720 for fire pumps, and 0.780 for emergency lighting and evacuation guidance systems.

[0049] In some embodiments, the activation weights of fire-fighting resources generated by the new process are compared and analyzed with the activation weights used in the previous round of adjustments. In the previous round, the activation weights for smoke exhaust fans were 0.750, fire pumps were 0.680, and emergency lighting and evacuation guidance systems were 0.730. The direction and magnitude of the weight changes are calculated, for example, using the following formula to calculate the magnitude of the weight changes. : ; in: This indicates the magnitude of the change in total weight. This indicates the activation weight of the j-th type of fire-fighting resource in the new round. This represents the activation weight of the j-th type of fire-fighting resource in the previous round. This indicates the type and quantity of fire-fighting resources. Substituting the values ​​yields the calculated results. =|0.850-0.750|+|0.720-0.680|+|0.780-0.730|=0.100+0.040+0.050=0.190, and all weights show an upward trend.

[0050] Optionally, if the weight changes indicate an increasing trend in fire hazard, such as a total weight change exceeding the set trend threshold of 0.120 (0.190), and an increase in the activation weights of all resource types, the newly generated activation weights and control signals are immediately used to strengthen the adjustment of the operational status of fire-fighting resources. For example, the number of smoke exhaust fans started is increased from three to four, the speed is increased from level five to level seven, the number of fire pumps started is increased from two to three, the outlet pressure setpoint is increased from 0.9 MPa to 1.1 MPa, the brightness level of the emergency lighting and evacuation guidance system is increased from level five to level six, and the indicated path is switched from emergency path to extreme path. If the weight changes indicate a decreasing trend in fire hazard, such as a newly generated activation weight lower than the weight corresponding to the current operational status in the next monitoring round, the newly generated activation weight is compared with the current operational status, and the operational status of fire-fighting resources is fine-tuned gradually. For example, the speed of the smoke exhaust fans is decreased by one level each time, adjusted every thirty seconds, until it matches the new weight.

[0051] It is understandable that the process of continuously acquiring, analyzing, making decisions, and adjusting is executed in a loop. For example, every sixty seconds, the entire process from acquiring data from the sensor to generating a control signal is completed. The system continuously compares the changes in old and new weights and dynamically adjusts the operating status of fire-fighting resources until all the fire parameters acquired from the sensor fall back to the preset safety threshold range, such as smoke concentration below 0.080%, temperature below 27.0℃, and oxygen concentration above 20.80%. At this point, the system stops adjusting and maintains the basic monitoring status.

[0052] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A fire hazard early warning system for commercial complexes based on sensor networks, characterized in that, include: The data acquisition and filtering module obtains fire protection parameters from sensors deployed in different functional areas within the commercial complex, performs preliminary filtering of the fire protection parameters, and removes invalid data. The risk characteristic calculation module identifies features reflecting the trend of smoke accumulation and calculates the smoke diffusion rate from the fire parameters after preliminary screening; it identifies features reflecting the temperature change process and calculates the temperature gradient; and it identifies features reflecting the degree of oxygen consumption and calculates the oxygen concentration decrease rate. The hazard level assessment module constructs a comprehensive metric value characterizing the current fire hazard level of the space based on the smoke diffusion rate, the temperature gradient, and the oxygen concentration decrease rate. The comprehensive metric value is compared with multiple preset hazard level boundaries to determine the current fire hazard level. The resource dynamic control module calculates the activation weights to be allocated to different fire-fighting resources within the commercial complex based on the current level of fire hazard, and generates control signals based on the activation weights. The control signals are used to adjust the operating status of the corresponding fire-fighting resources within the commercial complex. Among them, after adjusting the operational status of fire-fighting resources, the system continuously acquires updated fire-fighting parameters and repeatedly executes the module's functions to achieve continuous early warning of fire hazards.

2. The fire hazard early warning system for commercial complexes based on sensor networks according to claim 1, characterized in that, The preliminary screening of fire protection parameters, eliminating invalid data, specifically includes: Receive raw data stream uploaded by the sensing device, the raw data stream containing timestamps and corresponding fire parameter measurement values; Within a continuous preset time window, the data stream of each sensor is sampled, and the dispersion of the measured values ​​within the time window is calculated. The degree of dispersion of the measured values ​​is compared with the dispersion threshold. If the degree of dispersion exceeds the dispersion threshold, it is determined that there is noise interference in the data within the time window. Measurements within the time window where noise interference is identified are marked as invalid and removed from the original data stream; measurements within the time window where the dispersion is below the threshold are retained. From the retained measurements, outliers that deviate from the historical normal data model by more than the tolerance range are identified, and these outliers are marked as data to be verified. From the retained measurements, data marked as pending verification are further removed to obtain the preliminary filtered fire protection parameters.

3. The fire hazard early warning system for commercial complexes based on sensor networks according to claim 2, characterized in that, The process of identifying features reflecting smoke accumulation trends and calculating smoke diffusion rates from the initially screened fire parameters specifically includes: From the initially screened fire parameters, the smoke concentration measurement value sequence was extracted and arranged in chronological order. In the time series, the difference in smoke concentration between adjacent time points is calculated to obtain the instantaneous change in smoke concentration; The instantaneous changes over multiple consecutive time periods are summed to calculate the cumulative increase in smoke concentration over a specified time period. The cumulative increase in smoke concentration is divided by the corresponding specified time length to obtain the average increase in smoke concentration per unit time, and the average increase is used as the initial smoke diffusion rate. Based on the topological location of the sensing device within the commercial complex space and the preliminary smoke diffusion rate, the rate at which smoke travels from the location of the sensing device with high concentration to a nearby location is calculated. The smoke diffusion rate of the entire monitoring area is calculated by combining the smoke transfer rates of multiple sensing devices and using a weighted average.

4. The fire hazard early warning system for commercial complexes based on sensor networks according to claim 3, characterized in that, The process of identifying features reflecting temperature changes and calculating temperature gradients specifically includes: From the initial screening of fire parameters, temperature measurement values ​​from sensors at different locations were extracted and grouped. Within the same time period, select two adjacent locations within the commercial complex space and obtain the temperature measurements at these two locations. Calculate the temperature difference between two adjacent locations and divide the temperature difference by the actual physical distance between the two locations to obtain the local temperature change rate in the direction. Calculations were performed in multiple different directions within the commercial complex space to obtain the local temperature change rate in multiple different directions; Based on the distribution density of people and goods within the commercial complex, different weights are assigned to the local temperature change rate in different directions, and the weighted average spatial temperature gradient is calculated.

5. The fire hazard early warning system for commercial complexes based on sensor networks according to claim 1, characterized in that, The process of identifying features reflecting the degree of oxygen consumption and calculating the rate of decrease in oxygen concentration includes: From the initially screened fire parameters, the oxygen concentration measurement value sequence of sensors at different locations was extracted; Based on the time reference, the oxygen concentration measurement sequence for each sensor is arranged in chronological order; Calculate the difference between the oxygen concentration measurement value of each sensor at the current time point and the preset historical normal baseline value to obtain the absolute change in oxygen concentration at the location of each sensor. Within a preset fixed time interval, the direction and magnitude of the absolute change in oxygen concentration at each sensor location are statistically analyzed, and sensors with the same direction of change and a change exceeding a set threshold are classified into the same change region. For each defined change region, the sensor with the largest absolute change in oxygen concentration within the change region is selected as a representative point, and the oxygen concentration measurement sequence of the representative point is obtained over the entire fixed time interval. By performing linear fitting on the oxygen concentration measurement sequence of representative points, the average rate of decrease of oxygen concentration over time in the changing region can be calculated. By combining the average rate of decrease across all affected areas, and using a weighted average of the area values, the oxygen concentration decrease rate for the entire monitoring area is obtained.

6. The fire hazard early warning system for commercial complexes based on sensor networks according to claim 1, characterized in that, Based on the smoke diffusion rate, the temperature gradient, and the oxygen concentration decrease rate, a comprehensive metric is constructed to characterize the current fire hazard level in the space, specifically including: The calculated smoke diffusion rate is normalized and mapped to a numerical range of zero to one. The calculated temperature gradient is normalized and mapped to a numerical range of zero to one. The calculated rate of decrease in oxygen concentration was normalized and mapped to a numerical range of zero to one. Assign a weighting factor to the normalized smoke diffusion rate, a weighting factor to the normalized temperature gradient, and a weighting factor to the normalized oxygen concentration decrease rate. The normalized smoke diffusion rate is multiplied by its corresponding weighting coefficient, the normalized temperature gradient is multiplied by its corresponding weighting coefficient, and the normalized oxygen concentration decrease rate is multiplied by its corresponding weighting coefficient to obtain three weighted parameters. The three weighted parameters are linearly superimposed and summed to obtain a comprehensive metric value that characterizes the current fire hazard in the space.

7. The fire hazard early warning system for commercial complexes based on sensor networks according to claim 6, characterized in that, The comprehensive metric value is compared with multiple preset hazard level boundaries to determine the current level of fire hazard, specifically including: A set of incremental hazard level boundary values ​​is pre-defined, with each hazard level boundary value corresponding to a fire hazard level; Obtain the calculated comprehensive metric value, starting from the lowest hazard level boundary value, and compare the comprehensive metric value with the preset hazard level boundary value in sequence; When the comprehensive metric value is greater than or equal to a certain hazard level boundary value, but less than the next higher hazard level boundary value, the current level of fire hazard is determined to be the level corresponding to the hazard level boundary value. If the comprehensive metric value is less than the lowest hazard level boundary value, then the current fire hazard level is determined to be a safe level; The current level of fire hazard is output as input for subsequent resource allocation decisions.

8. The fire hazard early warning system for commercial complexes based on sensor networks according to claim 7, characterized in that, Based on the current level of fire hazard, calculate the activation weights required for different fire response resources within the commercial complex, specifically including: The fire-fighting resources include smoke exhaust fans, fire pumps, and emergency lighting and evacuation guidance systems; Establish a fire response resource list, which includes all available fire-fighting equipment and their basic attributes within the commercial complex; For different fire hazard levels, corresponding resource allocation strategies are preset, and the resource allocation strategies define the expected activation intensity of various fire-fighting equipment under each level; Based on the current level of fire hazard, query the corresponding resource allocation strategy and obtain the expected activation intensity requirements for each type of fire-fighting equipment under the current level; Calculate the response efficiency coefficient for each type of fire-fighting equipment to reach the designated risk area based on its current availability and physical location; The expected activation intensity requirement is combined with the corresponding response efficiency coefficient to calculate the theoretical activation weight of each type of fire-fighting equipment in the current situation. The theoretical activation weights of all fire-fighting equipment of the same type within the commercial complex are summarized and standardized to obtain the final activation weights required for each type of fire-fighting resource.

9. The fire hazard early warning system for commercial complexes based on sensor networks according to claim 8, characterized in that, A control signal is generated based on the aforementioned activation weight. This control signal is used to adjust the operational status of the corresponding fire-fighting resources within the commercial complex, specifically including: The calculated final activation weight of each type of fire-fighting resource is converted into an instruction format that can be recognized by the corresponding fire-fighting equipment control unit. For exhaust fans, control signals containing the number of fans to be started and the speed level are generated based on the start weights assigned to them. For fire pumps, a control signal containing the number of pumps to be started and the set value of the outlet pressure is generated according to the start weight assigned to them. For emergency lighting and evacuation guidance systems, control signals containing lighting brightness levels and guidance paths are generated based on the activation weights assigned to them. The generated control signals are distributed to the corresponding fire equipment control units through the fire protection network within the commercial complex. The fire equipment control unit receives and parses control signals, drives the actuators of the fire equipment, and adjusts the operating status of the fire equipment to the target state specified by the control signal.

10. The fire hazard early warning system for commercial complexes based on sensor networks according to claim 1, characterized in that, After adjusting the operational status of fire-fighting resources, the module continuously acquires updated fire-fighting parameters and repeatedly executes its functions to achieve continuous early warning of fire hazards. Specifically, this includes: At a set time interval after the fire-fighting equipment adjusts its operating status according to the control signal, new fire-fighting parameters are re-obtained from the sensing devices deployed in different functional areas inside the commercial complex. The complete process of performing preliminary screening, feature identification, comprehensive metric construction, hazard level determination, activation weight calculation and control signal generation for new fire protection parameters; The activation weights of fire response resources generated by the new process are compared and analyzed with the activation weights used in the previous round of adjustments to calculate the direction and magnitude of the weight changes. If the weight changes indicate an increasing trend in fire hazard, the newly generated activation weights and control signals will be immediately used to strengthen the adjustment of the operational status of fire response resources. If the weight changes indicate that the fire hazard is decreasing, the newly generated activation weights will be compared with the current operating status to fine-tune the operating status of fire response resources in a gradual manner. The process of acquiring, analyzing, making decisions, and adjusting is continuously repeated until all fire parameters acquired from the sensing devices fall back to within the preset safety threshold range.

Images