A smoke exhaust method and system for fire protection engineering

Abstract

The present invention relates to the field of fire intelligent control technology, and in particular to a smoke exhaust method and system for fire engineering. The method comprises the following steps: obtaining fire monitoring data, and performing fire identification according to the fire monitoring data to obtain fire identification data; performing fire risk assessment according to the fire identification data to obtain fire risk assessment data; generating a smoke exhaust strategy according to the fire risk assessment data to obtain smoke exhaust strategy data; obtaining real-time fire monitoring data, and adjusting the smoke exhaust strategy data in real time according to the real-time fire monitoring data to obtain real-time adjustment data of the smoke exhaust strategy, so as to perform smoke exhaust operations for fire engineering. The present invention can effectively reduce smoke accumulation when a fire occurs, reduce the risk of poisoning and suffocation of personnel, and provide better protection for safe evacuation and rescue operations at the fire scene, thereby significantly reducing casualties and property losses caused by the fire.

Description

Technical Field

[0001] The present invention relates to the technical field of fire protection intelligent control, and in particular to a smoke exhaust method and system for fire protection engineering. Background Art

[0002] With the acceleration of urbanization in modern society, the continuous expansion of building scale and the increase in structural complexity, fire prevention and control has played a vital role in urban construction and security. Especially in complex places such as large commercial complexes, industrial parks, office buildings, underground buildings, etc., once a fire occurs, the spread of smoke will quickly threaten the safety of personnel and the integrity of the building structure. How to quickly and effectively exhaust smoke in the early stage of a fire and reduce the harm of smoke to evacuation passages and personnel is a core issue in fire emergency management.

[0003] Existing fire smoke exhaust systems rely on fixed smoke exhaust channels and preset fan power parameters for operation. Although they can perform basic smoke exhaust operations after a fire occurs, they have the following major limitations when dealing with complex fire scenarios: Traditional smoke exhaust systems rely on preset control logic and cannot dynamically adjust the smoke exhaust path or air volume according to the real-time fire situation. When the fire situation changes, such as the spread of flames and changes in the smoke diffusion path, the system cannot respond quickly, resulting in unsatisfactory smoke exhaust effects. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a smoke exhaust method and system for fire protection engineering to solve at least one of the above technical problems.

[0005] The present application provides a smoke exhaust method for fire protection engineering, the method comprising:

[0006] S1. Obtain fire monitoring data, and perform fire identification based on the fire monitoring data to obtain fire identification data;

[0007] S2. Perform fire risk assessment based on the fire identification data to obtain fire risk assessment data;

[0008] S3. Generate a smoke exhaust strategy based on the fire risk assessment data to obtain smoke exhaust strategy data;

[0009] S4. Acquire real-time fire monitoring data, and adjust the smoke exhaust strategy data in real time according to the real-time fire monitoring data to obtain real-time adjustment data of the smoke exhaust strategy to perform smoke exhaust operations for fire engineering.

[0010] In the present invention, fire monitoring data such as smoke concentration, temperature, gas composition and other information are obtained in real time through a sensor network to realize automatic identification of fire conditions, reducing the lag of manual monitoring and the possibility of subjective misjudgment. Fire risk assessment is not only an analysis of the current fire situation, but also includes a comprehensive assessment of the possible direction of fire spread, surrounding environmental risk factors (such as flammable substances, personnel density, etc.) and the potential damage of fire to structures. According to the fire risk assessment data, the smoke exhaust strategy is intelligently generated, including the opening sequence and opening time of the smoke exhaust equipment, the power adjustment of the smoke exhaust fan, etc. Different from traditional manual operation, this strategy can accurately control the smoke exhaust effect and minimize the harm of smoke to personnel and equipment. The smoke exhaust strategy can be adjusted according to the development of fire in different areas, such as giving priority to removing smoke in high-risk areas, while maintaining ventilation in low-risk areas to prevent the spread of fire, and can adapt to the complex situation of the fire scene, avoiding the untimely emergency response caused by traditional single strategies. After obtaining real-time fire monitoring data, the system can optimize and adjust the existing smoke exhaust strategy in real time to ensure that the smoke exhaust system can adapt to the dynamic changes of fire development. By monitoring the system feedback in real time, the smoke exhaust effect can be evaluated and, if necessary, the operating parameters of the smoke exhaust fan can be adjusted immediately or the smoke exhaust path can be changed.

[0011] Optionally, S1 includes:

[0012] Smoke data and temperature data are collected through smoke sensors and temperature sensors to obtain smoke data and temperature data respectively;

[0013] Fire judgment is performed according to smoke data and temperature data to obtain fire judgment data;

[0014] When it is determined that the fire judgment data includes suspected fire data, the image sensor is controlled to collect images to obtain scene image data;

[0015] Perform fire detection according to the scene image data to obtain fire detection data;

[0016] Integrate smoke data, temperature data and fire detection data to obtain fire monitoring data;

[0017] Fire identification is performed based on the fire monitoring data to obtain fire identification data.

[0018] In the present invention, multi-dimensional monitoring of fire is achieved through the collaborative work of smoke sensors, temperature sensors and image sensors. Smoke sensors and temperature sensors can provide physical characteristic data at the early stage of fire, such as rising smoke concentration, sudden temperature changes, etc., while image sensors can intuitively capture visual information such as flames and smoke spread. Through the preliminary judgment of smoke and temperature sensors, and then enabling image sensors based on suspected fire data, the fire status is confirmed in a hierarchical and progressive manner, which can quickly make preliminary judgments in the early stages of the fire and further confirm suspected fires, effectively improving the sensitivity and response speed of the system. By integrating smoke data, temperature data and image data, and using multi-source information for intelligent fire identification, the fused data can more comprehensively reflect the fire status compared to fire identification based on a single data source, and help accurately identify the type, location and scale of the fire.

[0019] Optionally, the fire determination includes:

[0020] Calculate the change rate according to the smoke data and the temperature data to obtain the smoke change rate data and the temperature change rate data;

[0021] Determine whether the smoke change rate data is greater than or equal to a preset smoke change rate threshold and whether the temperature change rate data is greater than or equal to a preset temperature change rate threshold;

[0022] When it is determined that the smoke change rate data is greater than or equal to a preset smoke change rate threshold and the temperature change rate data is greater than or equal to a preset temperature change rate threshold, feature extraction is performed on the smoke data and the temperature data to obtain smoke feature data and temperature feature data respectively;

[0023] The preset fire decision forest model is used to perform fire judgment on the smoke characteristic data and the temperature characteristic data to obtain fire judgment data.

[0024] In the present invention, when a fire occurs, the smoke concentration and temperature usually change rapidly at the same time. Using the smoke change rate and the temperature change rate as the basis for judgment at the same time can effectively reduce the occurrence of false alarms (for example, smoke generated by cooking but the temperature does not rise significantly) and missed alarms (the temperature rises slightly but is accompanied by obvious smoke). When the smoke change rate and the temperature change rate exceed the threshold, further feature extraction (such as average value, maximum value, fluctuation amplitude, fluctuation frequency, etc.) of the smoke data and temperature data can better describe the physical state of the fire and help distinguish different types of fire situations. The decision forest model is an integrated learning method that makes a comprehensive judgment by combining the judgment results of multiple decision tree models. Compared with a single model, the decision forest model has stronger generalization ability and robustness, can effectively cope with complex and changeable fire scenes, and improve the accuracy of fire judgment.

[0025] Optionally, the preset fire decision forest model is constructed in the following manner:

[0026] Obtain historical smoke data, historical temperature data, and corresponding historical fire label data;

[0027] Perform feature extraction on historical smoke data and historical temperature data to obtain historical smoke feature data and historical temperature feature data;

[0028] Perform time series fitting degree selection and space fitting degree selection on historical smoke characteristic data and historical temperature characteristic data, and obtain smoke temperature time fitting degree characteristic data and smoke temperature space fitting degree characteristic data respectively;

[0029] Data is divided according to the smoke temperature time fitting characteristic data and the smoke temperature space fitting characteristic data, to obtain smoke temperature time fitting division data and smoke temperature space fitting division data respectively, wherein the smoke temperature time fitting division data includes smoke temperature time fitting training data and smoke temperature time fitting test data, and the smoke temperature space fitting division data includes smoke temperature space fitting training data and smoke temperature space fitting test data;

[0030] Perform node splitting on the smoke temperature time fitting training data and the smoke temperature space fitting training data to obtain first smoke temperature node data and second smoke temperature node data respectively;

[0031] Generate a tree based on the first smoke temperature node data, the second smoke temperature node data and the corresponding historical fire label data to obtain primary first fire decision tree data and primary second decision tree data;

[0032] Post-pruning the primary first fire decision tree data and the primary second decision tree data to obtain secondary first fire decision tree data and secondary second decision tree data;

[0033] The secondary first fire decision tree data is iteratively optimized using the smoke temperature time fitting test data to obtain the first fire decision tree model, and the secondary second decision tree data is iteratively optimized using the smoke temperature space fitting test data to obtain the second fire decision tree model;

[0034] The first fire decision tree model and the second fire decision tree model are added with voting layers to obtain a fire decision tree forest model.

[0035] In the present invention, the degree of fit is selected according to the data change trend in the time dimension, and the change characteristics of the smoke and temperature data on the time axis are extracted, which is helpful to capture the dynamic changes of the occurrence and development of the fire, and can better identify the time characteristics of the fire, such as the suddenness of the fire and the change speed of the fire. The degree of fit of the data is selected in the spatial dimension, and the distribution law of the fire in different locations can be analyzed, such as the possible location of the fire source, the direction of the fire spread, etc., which is of great significance for the location and prediction of the fire in complex scenes (such as multi-story buildings or vast industrial areas). In model training, node splitting divides the data through different feature dimensions (such as time series features and spatial features), and continuously refines the structure of the decision tree, which can fully mine the useful information in the data and improve the model's ability to judge complex fire scenes. The smoke temperature time fit feature data and the smoke temperature space fit feature data are divided into training data and test data respectively, so that the model training and testing processes can be ensured to be independent of each other and avoid the occurrence of overfitting problems. The first and second smoke temperature node data obtained by node splitting are used to generate a primary decision tree. This process is a process of recursively dividing the data according to different features, and gradually forming a decision path for fire judgment. During the generation process of the decision tree model, too many splits may lead to overfitting of the model. Post-pruning can effectively simplify the decision tree structure and reduce the complexity of the model by removing branches that perform poorly on the test data, thereby improving the generalization ability of the model.

[0036] Optionally, the timing fit selection includes:

[0037] Perform trend slope extraction according to the smoke characteristic data and the temperature characteristic data to obtain smoke characteristic trend slope data and temperature characteristic trend slope data;

[0038] Performing piecewise linear regression on the smoke characteristic data and the temperature characteristic data according to the smoke characteristic trend slope data and the temperature characteristic trend slope data to obtain first smoke temperature-time fitting characteristic data;

[0039] Cluster segmentation is performed according to the smoke characteristic data and the temperature characteristic data to obtain the smoke characteristic inflection point position data and the temperature characteristic inflection point position data;

[0040] Fitting and screening the smoke characteristic data and the temperature characteristic data according to the smoke characteristic inflection point position data and the temperature characteristic inflection point position data to obtain second smoke temperature-time fitting characteristic data;

[0041] Bayesian cross screening is performed based on the first smoke temperature time fitting characteristic data and the second smoke temperature time fitting characteristic data to obtain the smoke temperature time fitting characteristic data.

[0042] After extracting the trend slope in the present invention, the data is further fitted by piecewise linear regression, which can handle nonlinear changes and mutation points in the data. Compared with the traditional overall linear regression, the piecewise linear regression can better capture the local characteristics in the fire data, such as the sudden rise of sudden fires, the inflection point of the cooling process, etc., and improve the precision and flexibility of data fitting. Using cluster segmentation to analyze the smoke and temperature characteristic data can effectively identify the change areas of different patterns in the data. Compared with the traditional simple segmentation method, cluster segmentation can more intelligently capture the mutation of the pattern in the data, such as the difference in data distribution before and after the fire, so as to more accurately locate the inflection point. According to the cluster segmentation results, inflection point detection can identify the key turning points in the fire data (such as the sudden intensification of the fire, the temperature drop after the fire extinguishing measures are initiated, etc.). Compared with the traditional single threshold method, the inflection point detection after cluster segmentation can more flexibly adapt to the changes in fire in different scenarios and improve the response capability of sudden fire events. The Bayesian method allows the model parameters to be adjusted dynamically according to the actual data. When the data sample size is small, it can use prior knowledge to make reasonable inferences. When the data sample size increases, it can gradually rely on posterior data to make more accurate judgments. It can continuously optimize the fire judgment model in the early stages and during the development of the fire, thereby improving the judgment accuracy.

[0043] Optionally, the spatial fit selection includes:

[0044] Obtain smoke collection position data corresponding to the smoke characteristic data and temperature collection position data corresponding to the temperature characteristic data;

[0045] A smoke matrix is ​​constructed according to the smoke characteristic data and the smoke collection position data to obtain smoke characteristic position matrix data, and a temperature matrix is ​​constructed according to the temperature characteristic data and the temperature collection position data to obtain temperature characteristic position matrix data;

[0046] Calculating the spatial distance variation degree of the smoke characteristic position matrix data and the temperature characteristic position matrix data to obtain the first smoke temperature spatial correlation data;

[0047] Calculating the spatial correlation coefficient of the smoke characteristic position matrix data and the temperature characteristic position matrix data to obtain the second smoke temperature spatial correlation data;

[0048] The smoke characteristic data and the temperature characteristic data are spatially fitted and screened according to the first smoke temperature spatial correlation data and the second smoke temperature spatial correlation data to obtain smoke temperature spatial fitting degree characteristic data.

[0049] In the present invention, by associating smoke feature data and temperature feature data with their corresponding collection position data, a smoke feature position matrix and a temperature feature position matrix are constructed, which can better express the distribution characteristics of spatial data and provide rich spatial position information and feature correlation for spatial analysis. By calculating the spatial distance variation between data points at different positions in the smoke feature position matrix and the temperature feature position matrix, the volatility and uncertainty of the data at different spatial positions can be evaluated, the local variation characteristics of spatial data can be effectively captured, and data anomalies and areas with drastic changes can be identified. In the prior art, spatial correlation analysis only uses a single spatial correlation coefficient (such as the Moran index). The present invention uses both spatial distance variation and spatial correlation coefficient as two indicators, which can analyze the spatial correlation of data from different angles and provide a more comprehensive reference for data feature screening. By analyzing the first and second spatial correlation data, smoke and temperature feature data with higher fitting degree are screened out, noise data and anomalies can be filtered out, data with stronger discrimination of fire characteristics can be retained, and the effectiveness of model training can be improved.

[0050] Optionally, S2 includes:

[0051] A fire state model is constructed according to the fire identification data to obtain a fire state model;

[0052] Conduct fire risk factor analysis based on the fire state model to obtain fire risk factor data;

[0053] Fire damage estimation is performed based on fire risk factor data to obtain fire damage estimation data;

[0054] Construct a fire risk matrix based on fire damage estimation data to obtain fire risk matrix data;

[0055] Risk areas are divided according to the fire risk matrix data to obtain fire risk assessment data.

[0056] In the present invention, by systematically constructing a model for fire identification data, various types of fire status information (such as fire occurrence time, fire source location, fire development speed, spread direction, etc.) are incorporated into the model to construct a comprehensive fire status model. Fire risk factor analysis not only considers traditional fire risk factors (such as fire source location, flame intensity, etc.), but also introduces environmental factors (such as building structure, combustible distribution, personnel density, etc.), making the analysis of risk factors more comprehensive. By analyzing the fire risk factor data, the damage that may be caused by the fire can be quantitatively estimated, including casualties, property losses, and the degree of damage to the building structure. The fire risk matrix processes different fire risk factors and damage estimation data in a matrix to form structured risk data. According to the fire risk matrix data, the risk area can be divided, and the fire scene can be divided into different risk level areas (such as high risk, medium risk, and low risk areas), which is helpful to formulate more targeted emergency response strategies, such as giving priority to evacuating people in high-risk areas and giving priority to extinguishing areas with larger fires.

[0057] Optionally, S3 includes:

[0058] Obtain smoke exhaust duct parameter data;

[0059] Prioritize smoke exhaust requirements based on fire risk assessment data to obtain fire risk smoke exhaust requirement priority data;

[0060] Generate a smoke exhaust path based on smoke exhaust pipe parameter data and fire risk smoke exhaust demand priority data to obtain smoke exhaust path data;

[0061] A smoke exhaust strategy is generated according to the smoke exhaust path data and the smoke exhaust duct parameter data to obtain the smoke exhaust strategy data.

[0062] According to the smoke exhaust demand priority data, the smoke exhaust demand in high-risk areas can be prioritized to prevent smoke from accumulating in high-risk areas during fires, causing greater danger. Targeted smoke exhaust strategies can effectively reduce the impact of fires on personnel evacuation and firefighting operations, and improve the overall smoke exhaust effect. By obtaining smoke exhaust pipe parameter data (such as pipe diameter, length, resistance coefficient, etc.), the physical characteristics of the smoke exhaust system can be better understood. Compared with the analysis of traditional single fire data, the present invention takes into account the adaptability and transmission capacity of the smoke exhaust pipe, providing a solid foundation for the rational planning of the smoke exhaust path. By analyzing the smoke exhaust pipe parameter data and the fire risk smoke exhaust demand priority data, the optimal smoke exhaust path is intelligently generated. Compared with the traditional smoke exhaust method that relies on experience or preset paths, the intelligent path generation can better adapt to complex building structures and changeable fire scenes, and improve the smoke exhaust efficiency. According to the generated smoke exhaust path data and smoke exhaust pipe parameter data, the opening and closing time of the smoke exhaust fan, the air volume adjustment and other operations can be accurately controlled to achieve precise control of the smoke exhaust equipment, avoiding the situation of untimely or insufficient smoke exhaust due to improper control.

[0063] Optionally, S4 includes:

[0064] Obtain real-time fire monitoring data;

[0065] Fire identification is performed according to real-time fire monitoring data to obtain real-time fire identification data;

[0066] Real-time adjustments are made based on real-time fire identification data and smoke exhaust strategy data to obtain real-time adjustment data for the smoke exhaust strategy in order to perform smoke exhaust operations for fire protection engineering.

[0067] In the present invention, through real-time data collection, the changing trend and development status of the fire can be continuously perceived at different stages of the fire. Fire identification can be performed based on real-time monitoring data, and the current state of the fire can be judged in real time, including the severity of the fire, the speed of spread, the direction of smoke diffusion, etc., which can avoid the recognition error caused by environmental changes in the traditional static judgment method and improve the accuracy of fire state judgment. The system can dynamically adjust the operating parameters of the smoke exhaust fan, adjust the smoke exhaust path, change the smoke exhaust priority, etc. to cope with sudden changes in the fire situation. The adaptability and flexibility of the smoke exhaust system in complex fire scenes are significantly improved. By adjusting the smoke exhaust strategy in real time, the smoke exhaust effect can be continuously optimized according to the changes in the fire situation, avoiding the situation where the smoke exhaust effect is poor due to the untimely adjustment of the strategy. For example, when the fire spreads to a new area, the smoke exhaust intensity of the area can be quickly increased, or the original smoke exhaust path can be adjusted to reduce the retention of smoke in key channels.

[0068] Optionally, the present application further provides a smoke exhaust system for fire engineering, which is used to perform the smoke exhaust method for fire engineering as described above, and the smoke exhaust system for fire engineering comprises:

[0069] A fire identification module is used to obtain fire monitoring data and perform fire identification based on the fire monitoring data to obtain fire identification data;

[0070] A fire risk assessment module, used to perform fire risk assessment based on fire identification data to obtain fire risk assessment data;

[0071] A smoke exhaust strategy generation module is used to generate a smoke exhaust strategy according to the fire risk assessment data to obtain smoke exhaust strategy data;

[0072] The smoke exhaust strategy real-time adjustment module is used to obtain real-time fire monitoring data and adjust the smoke exhaust strategy data in real time according to the real-time fire monitoring data to obtain the real-time adjustment data of the smoke exhaust strategy for performing smoke exhaust operations for fire protection engineering.

[0073] The purpose of the present invention is to analyze the dynamic changes of fire monitoring data in real time, and the present invention can make an early warning response in the early stage of the fire (such as small abnormal changes in smoke concentration and temperature). Based on the fire risk factor data, the present invention can quantitatively estimate the damage that may be caused by the fire, including casualties, property losses, and the degree of damage to the building structure. Fire risk assessment is not a static process, but a system that changes dynamically as the fire develops. The present invention can continuously update the risk matrix according to real-time fire identification data, and map the risk level, fire change trend, smoke diffusion path and other information of different regions into the risk matrix in real time, and provide the latest and most accurate risk assessment results. On the basis of fire risk assessment, the smoke exhaust demand priority of each region is intelligently analyzed. The system can automatically identify high-risk areas for fire and areas that need to be focused on smoke exhaust, and generate the optimal smoke exhaust strategy through an intelligent matching mechanism to ensure that limited smoke exhaust resources (such as smoke exhaust fans and pipes) are used in the most needed areas first, maximizing the smoke exhaust effect. According to real-time fire monitoring data and system feedback results, the present invention can automatically adjust the power of the smoke exhaust fan, the opening order of the smoke exhaust pipe, the switching of the smoke exhaust path and other operations to improve the intelligence level of the smoke exhaust strategy. The system can respond quickly to changes in the fire situation (such as sudden intensification of fire and change in smoke diffusion path), and reduce the impact of fire on building structures and personnel through real-time optimization strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting implementations made with reference to the following drawings:

[0075] Figure 1A flowchart showing a smoke exhaust method for fire protection engineering according to an embodiment of the present invention is shown;

[0076] Figure 2 A flowchart showing the steps of a fire identification method according to an embodiment is shown;

[0077] Figure 3 A flowchart showing a method for fire risk assessment according to an embodiment is shown;

[0078] Figure 4 A flowchart showing a method for generating a smoke exhaust strategy according to an embodiment of the present invention is shown;

[0079] Figure 5 A flowchart showing a method for real-time adjustment of a smoke exhaust strategy according to an embodiment of the present invention is provided;

[0080] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0081] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0082] In addition, the drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0083] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0084] In a large shopping mall with a total area of ​​30,000 square meters and 4 floors, the fire monitoring system deploys the following sensors: Smoke sensors: 100, distributed in key locations on each floor. Temperature sensors: 100, deployed on the ceiling and around flammable objects. CO (gas) sensors: 50, distributed in main passages and flammable areas. Video surveillance: 50, distributed at main entrances and high-risk areas on each floor.

[0085] In a fire situation, a fire was detected in an area in the middle of the third floor. The fire monitoring data is as follows: The smoke sensor data shows that the smoke concentration in the area has risen to 300ppm (increasing by 30ppm per minute). The temperature sensor shows that the temperature in the area has risen to 180°C, with a rising rate of 15°C per minute. The CO sensor detected that the CO content reached 400ppm, and there was a risk of medium-concentration toxic gas leakage. Video surveillance identified that flames appeared in the area and smoke spread rapidly. According to the fire identification algorithm, the system combined smoke, temperature, CO content and video images to determine that a medium-sized fire had occurred in the area. The fire was identified as a Class C fire (a fire caused by electrical equipment) with a flame spread rate of 5 meters per minute.

[0086] The system conducts a fire risk assessment based on the fire identification data and analyzes the following key risk factors: Population density: There are currently about 500 people on the third floor, which is a high density, and the area is close to the evacuation passage. Fire spread speed: The spread speed is 5 meters per minute, and there is a risk of spreading to other areas; Building structure: There are multiple electrical equipment and combustible materials on the ceiling above the middle area of ​​the third floor, which is a high risk; Smoke exhaust path complexity: The smoke exhaust system on the third floor is complex, mainly composed of two main smoke exhaust pipes with a diameter of 80cm and other branch pipes; The fire risk assessment results show that the fire risk in the middle of the third floor is assessed as a high-risk area, and the main risk factors include: dense population, rapid fire spread, and severe smoke diffusion. The fire damage is estimated as follows: If smoke is not exhausted and the fire is not controlled in time, it will cause poisoning or suffocation to people, and the economic loss is estimated to be 50 million yuan.

[0087] According to the fire risk assessment data, the system generated the following smoke exhaust strategy: According to the risk assessment, the middle area of ​​the third floor is classified as smoke exhaust priority level 1, that is, this area requires the most urgent smoke exhaust operation. The surrounding area (the area close to the evacuation passage) is classified as smoke exhaust priority level 2 to ensure the safety of personnel evacuation. The system analyzed the smoke exhaust duct parameters. The smoke exhaust duct on the third floor mainly has two main smoke exhaust ducts, which lead to different exhaust ports. The main smoke exhaust duct is 100 meters long, 80 cm in diameter, and has a resistance coefficient of 0.5. After comprehensive analysis, the system decided to optimize the fan speed to 1000 cubic meters / minute through the main smoke exhaust duct 1 (the duct closer to the fire source area) to ensure rapid smoke exhaust, and at the same time open the branch smoke exhaust duct for auxiliary smoke exhaust. The system will adjust the air volume of the smoke exhaust fan according to the fire risk data, increase the smoke exhaust intensity in high-risk areas, and reduce the fan operation time in low-risk areas. The specific smoke exhaust strategy is to start the main smoke exhaust duct fan on the third floor, set the air volume to 1,000 cubic meters per minute, and enable 5 branch smoke exhaust fans, each with an air volume of 200 cubic meters per minute. At the same time, the smoke exhaust fan on the fourth floor is turned on to prevent smoke from spreading to the upper floors. The smoke exhaust system in the stairwell on the third floor is set as the second priority to assist personnel evacuation.

[0088] During the smoke exhaust operation, the system continued to obtain real-time fire monitoring data. After 10 minutes, the fire monitoring data showed that the smoke concentration continued to rise to 600ppm, and the CO content increased to 500ppm, indicating that the fire had intensified and the smoke spread rapidly. The system adjusted the smoke exhaust strategy based on the new monitoring data, increased the air volume of the main smoke exhaust duct 1 to 1200 cubic meters / minute, and simultaneously started another main smoke exhaust duct on the third floor for auxiliary smoke exhaust. The air volume of the smoke exhaust fan on the fourth floor was increased to 600 cubic meters / minute to prevent the smoke from continuing to spread upward. The air volume of the branch smoke exhaust fan was adjusted to 300 cubic meters / minute in real time to ensure that the smoke in the high-risk area was quickly discharged. Through adjustment, the system successfully controlled the smoke concentration on the third floor to 400ppm and the CO content to 300ppm within the first 20 minutes of the fire. The smoke concentration in the personnel evacuation channel was less than 100ppm, effectively ensuring the safe evacuation of 500 people. The fire was finally successfully extinguished, the property loss was controlled within RMB 20 million, and no poisoning or suffocation incidents occurred.

[0089] See also Figures 1 to 5 The present application provides a smoke exhaust method for fire protection engineering, the method comprising:

[0090] S1. Obtain fire monitoring data, and perform fire identification based on the fire monitoring data to obtain fire identification data;

[0091] Specifically, real-time monitoring data is obtained through fire detectors (such as smoke sensors, temperature sensors or infrared sensors) installed in buildings or specific areas. Key signal features are extracted from the real-time monitoring data, such as identifying the rate of temperature change, smoke concentration change pattern, etc. Based on the extracted features, it is compared with the preset fire feature threshold or through pattern recognition to determine whether a fire has occurred. If the features meet the fire conditions, it is identified as a fire and fire identification data is generated.

[0092] S2. Perform fire risk assessment based on the fire identification data to obtain fire risk assessment data;

[0093] Specifically, combined with fire identification data, multiple risk factors such as the location, scale, speed of fire spread, and smoke concentration are analyzed to determine the severity of the fire. The potential risk of fire spread is assessed by obtaining information such as the layout of the building, ventilation conditions, and the types and quantities of surrounding combustibles. For example, whether there are walls or air ducts in the building that block the spread of smoke. Using existing fire spread data or models, the possible direction and speed of fire spread in a short period of time are predicted, and the range of smoke spread in the building is assessed.

[0094] S3. Generate a smoke exhaust strategy based on the fire risk assessment data to obtain smoke exhaust strategy data;

[0095] Specifically, according to the results of the fire risk assessment, different smoke exhaust areas are divided, and key areas that need to be given priority for smoke exhaust are determined to prevent smoke from spreading to areas where people are concentrated. According to information such as the building layout, the location of the smoke exhaust channel, and the configuration of the ventilation system, a suitable smoke exhaust path is selected. The selection of the smoke exhaust path needs to consider the direction of smoke flow and prevent high-temperature smoke from spreading to areas without fire. Generate instructions to start or shut down the smoke exhaust fan in a specific area, dispatch the ventilation equipment in the building, and form an effective smoke exhaust path. By calculating the opening and closing time of each smoke exhaust port, the air volume adjustment, and the power of the smoke exhaust fan, ensure that the smoke is removed to the greatest extent in the shortest possible time and protect the escape route.

[0096] S4. Acquire real-time fire monitoring data, and adjust the smoke exhaust strategy data in real time according to the real-time fire monitoring data to obtain real-time adjustment data of the smoke exhaust strategy to perform smoke exhaust operations for fire engineering.

[0097] Specifically, the fire detection system continuously obtains the development of the fire in the building, including the fire intensity, smoke concentration and airflow changes. According to the data of the sensors in the smoke exhaust channel, the effectiveness of the current smoke exhaust strategy is judged to determine whether the smoke is discharged smoothly along the predetermined route and whether the smoke concentration in the fire area has dropped significantly. If it is found that the smoke exhaust effect is not ideal (for example, the smoke concentration in some areas has not been effectively reduced), the fire risk is reassessed in real time and the strategy is adjusted. For example, it is necessary to change the smoke exhaust path, increase the opening of the smoke exhaust port, or adjust the fan power. In the case of rapid changes in the fire environment, the operating status of the smoke exhaust equipment is dynamically adjusted, such as immediately closing unnecessary smoke exhaust ports or adjusting the speed of the fan to ensure flexible response of the entire system.

[0098] Optionally, S1 includes:

[0099] S11, collecting smoke data and temperature data through a smoke sensor and a temperature sensor to obtain smoke data and temperature data respectively;

[0100] Specifically, smoke sensors installed in buildings collect smoke concentration data by detecting tiny particles in the air or changes in light scattering. Sensors usually continuously collect and upload real-time data at fixed time intervals. Temperature sensors monitor changes in ambient temperature in real time. In the early stages of a fire, the temperature usually rises rapidly. Temperature sensors record these temperature fluctuations through high-frequency sampling and generate ambient temperature data.

[0101] S12, performing fire judgment according to the smoke data and the temperature data to obtain fire judgment data;

[0102] Specifically, the collected smoke data and temperature data are preprocessed to remove noise and errors in the sampling process to ensure the accuracy of the data. According to the preset threshold range of smoke concentration and temperature, it is judged whether there is an abnormality. If the smoke concentration exceeds the set threshold and the temperature rises rapidly within a certain period of time, the system determines that a fire may have occurred and generates suspected fire data.

[0103] S13, when it is determined that the fire judgment data includes suspected fire data, the image sensor is controlled to collect images to obtain scene image data;

[0104] Specifically, when the fire judgment data contains suspected fire, the system will immediately trigger the image sensor to work. The image sensor is installed in the high-risk area of ​​fire to collect real-time images of the fire area. The image sensor captures the image data of the current scene, including visual images or infrared images.

[0105] S14, performing fire detection according to the scene image data to obtain fire detection data;

[0106] Specifically, by analyzing the fire features in the image (such as smoke shape, flame brightness and color), visual features related to fire are extracted. If flames or abnormal smoke patterns are identified, it is determined that a fire has occurred and fire detection data is generated.

[0107] S15, integrating the smoke data, temperature data and fire detection data to obtain fire monitoring data;

[0108] Specifically, smoke data, temperature data and fire detection data are integrated to establish a complete fire monitoring data.

[0109] S16. Perform fire identification based on the fire monitoring data to obtain fire identification data.

[0110] Specifically, the integrated fire monitoring data is used for fire determination, and based on preset rules or models (such as the weights of smoke, temperature, and image detection results), it is determined whether a fire has actually occurred. When all data sources indicate the existence of a fire, the system generates fire identification data, confirms the fire status, and triggers subsequent alarm and smoke exhaust operations. If there is no clear indication of a fire in the data, the monitoring status is maintained and data collection continues to prevent the possibility of subsequent fires.

[0111] Optionally, the fire determination includes:

[0112] Calculate the change rate according to the smoke data and the temperature data to obtain the smoke change rate data and the temperature change rate data;

[0113] Specifically, by performing time differentiation on the continuous smoke density data and temperature data, the change rate of smoke density and the change rate of temperature are calculated. The formula is: smoke change rate = (current smoke value - smoke value at the previous moment) / time interval. Temperature change rate = (current temperature value - temperature value at the previous moment) / time interval.

[0114] Determine whether the smoke change rate data is greater than or equal to a preset smoke change rate threshold and whether the temperature change rate data is greater than or equal to a preset temperature change rate threshold;

[0115] Specifically, the calculated smoke change rate data is compared with the preset smoke rate threshold, and similarly, the temperature change rate data is compared with the temperature rate threshold. If the smoke change rate is greater than or equal to the smoke rate threshold, and the temperature change rate is greater than or equal to the temperature rate threshold, it is considered that a fire has occurred. If both conditions are met at the same time, the system generates a suspected fire sign and enters the next feature extraction stage. If any rate does not meet the threshold condition, monitoring continues without entering the next stage.

[0116] When it is determined that the smoke change rate data is greater than or equal to a preset smoke change rate threshold and the temperature change rate data is greater than or equal to a preset temperature change rate threshold, feature extraction is performed on the smoke data and the temperature data to obtain smoke feature data and temperature feature data respectively;

[0117] Specifically, features that are helpful for fire judgment are extracted from historical smoke data and temperature data. Features are data that reflect data characteristics or specific patterns, such as peak value, rising speed, cumulative change, etc. The features extracted from smoke features include: average smoke concentration, volatility of smoke concentration, and maximum rising rate in a short period of time. The features extracted from temperature features include: acceleration of temperature change (i.e., second-order derivative), high temperature duration, and cumulative time for temperature to reach a certain threshold.

[0118] The preset fire decision forest model is used to perform fire judgment on the smoke characteristic data and the temperature characteristic data to obtain fire judgment data.

[0119] Specifically, the fire decision tree model is a data classification tool based on multidimensional features. It makes fire judgments according to the preset decision path based on the smoke and temperature feature data. The decision tree consists of multiple nodes, each of which represents a decision rule (such as whether a certain feature value exceeds a threshold). The root node of the decision tree makes the first judgment based on a certain feature (such as smoke density). If the feature value is higher than a certain threshold, it proceeds along the "suspected fire" branch, otherwise it continues along the "no fire" branch. Next, the next node of the decision tree will further refine the judgment based on another feature (such as the temperature change rate). This process is carried out in sequence until the leaf node is reached, and a fire (suspected fire data) or non-fire judgment is made. When the model calculates all the features, it makes a fire judgment and generates fire judgment data. If it is judged to be a fire, the system will trigger an alarm or other subsequent emergency response. If no fire is identified, monitoring continues.

[0120] Optionally, the preset fire decision forest model is constructed in the following manner:

[0121] Obtain historical smoke data, historical temperature data, and corresponding historical fire label data;

[0122] Specifically, historical smoke data, historical temperature data, and corresponding fire label data are obtained from the historical fire monitoring system. The fire label data is used to mark whether a fire occurred in each time period. The historical smoke data records the changes in smoke concentration at each time point in history. The historical temperature data records the temperature changes at each time point. The fire label indicates whether a fire occurred in each time period (0 means no fire, 1 means fire).

[0123] Perform feature extraction on historical smoke data and historical temperature data to obtain historical smoke feature data and historical temperature feature data;

[0124] Specifically, useful features are extracted from the original smoke and temperature data: smoke features: maximum smoke value, minimum smoke value, average smoke value, smoke change rate, etc. Temperature features: average temperature, maximum temperature change, rising rate, etc.

[0125] Perform time series fitting degree selection and space fitting degree selection on historical smoke characteristic data and historical temperature characteristic data, and obtain smoke temperature time fitting degree characteristic data and smoke temperature space fitting degree characteristic data respectively;

[0126] Specifically, the time series fit is selected to analyze the trend of smoke and temperature changes over time, extract characteristic slopes and trend changes. The spatial fit is selected to extract fit features based on the spatial differences of sensor data at different locations. The time series fit is selected to extract time fit features by analyzing the trend of smoke concentration and temperature changes over time, and describe the change of data over time during the fire process through trend analysis (such as slope, fluctuation, etc.), such as performing linear fitting on the temperature data in the past 1 hour, calculating the rising slope, and obtaining the time fit features. The spatial fit is selected to obtain spatial fit features by analyzing the data distribution of smoke and temperature sensors at different locations. For example, if the smoke concentration in one area is significantly different from that in other areas, the spatial fit will be lower. The smoke data of multiple sensors in different areas of the building are subjected to spatial correlation analysis to obtain the spatial fit of each location.

[0127] Data is divided according to the smoke temperature time fitting characteristic data and the smoke temperature space fitting characteristic data, to obtain smoke temperature time fitting division data and smoke temperature space fitting division data respectively, wherein the smoke temperature time fitting division data includes smoke temperature time fitting training data and smoke temperature time fitting test data, and the smoke temperature space fitting division data includes smoke temperature space fitting training data and smoke temperature space fitting test data;

[0128] Specifically, the feature data is divided into a training set and a test set. Usually 70% of the data is used for training and 30% for testing. The temporal fit data is divided to obtain the training data and test data of the time dimension. The spatial fit data is divided to obtain the training data and test data of the spatial dimension.

[0129] Perform node splitting on the smoke temperature time fitting training data and the smoke temperature space fitting training data to obtain first smoke temperature node data and second smoke temperature node data respectively;

[0130] Specifically, conditional splitting is performed based on features, and each node represents a condition, such as "smoke concentration > 300ppm". The purpose of splitting is to make the data set purer. The first smoke temperature node data is based on the time fit feature as the main splitting basis. The second smoke temperature node data is based on the space fit feature as the main splitting basis.

[0131] Specifically, when constructing a decision tree, the data is conditionally split based on a certain feature (such as the time or space fit of smoke concentration or temperature). Each node represents a decision condition (such as "whether the smoke concentration is greater than a certain threshold"). The criterion for each split is to make the subset after the split purer (i.e., samples of the same type are concentrated together), such as by minimizing the Gini coefficient or maximizing information gain. The split data is used to generate the corresponding node.

[0132] Generate a tree based on the first smoke temperature node data, the second smoke temperature node data and the corresponding historical fire label data to obtain primary first fire decision tree data and primary second decision tree data;

[0133] Specifically, a primary decision tree is constructed based on the split node data, and the decision tree determines the final output of fire or non-fire results by splitting layer by layer. The primary first decision tree is constructed based on time fit. The primary second decision tree is constructed based on spatial fit. The primary first fire decision tree uses smoke characteristics as the main splitting basis, and the nodes are split step by step from "smoke concentration > 150 ppm" to "temperature rise rate > 0.5°C / minute", and the output fire risk is high or low. The primary second fire decision tree uses spatial fit as the main feature, analyzes the smoke and temperature data at each location step by step, and derives the possibility of fire.

[0134] Post-pruning the primary first fire decision tree data and the primary second decision tree data to obtain secondary first fire decision tree data and secondary second decision tree data;

[0135] Specifically, by pruning redundant branches, the decision tree is simplified to prevent model overfitting. After pruning, the decision tree is more general and suitable for different data sets. Nodes that perform poorly on the test set are pruned, and the most important split nodes are retained. In the primary first decision tree, some feature branches (such as "smoke concentration > 120 ppm") contribute less to subsequent classification and are deleted after pruning. Irrelevant feature branches in the second decision tree are also pruned, and branches that mainly affect fire judgment are retained.

[0136] The secondary first fire decision tree data is iteratively optimized using the smoke temperature time fitting test data to obtain the first fire decision tree model, and the secondary second decision tree data is iteratively optimized using the smoke temperature space fitting test data to obtain the second fire decision tree model;

[0137] Specifically, the secondary decision tree was verified using the test data, and the node splitting conditions and pruning strategies were adjusted to make the decision tree perform best on the test set. The first fire decision tree model was optimized, and the feature splitting conditions were adjusted using the test data, and "smoke concentration > 140 ppm" was changed to "smoke concentration > 130 ppm" to improve the fire recognition rate. The second fire decision tree model was similarly optimized to improve the spatial correlation judgment threshold and improve the spatial fit performance of the model.

[0138] The first fire decision tree model and the second fire decision tree model are added with voting layers to obtain a fire decision tree forest model.

[0139] Specifically, multiple decision tree models are integrated and the fire prediction result is determined by a voting mechanism. If the majority of decision trees predict a fire, the "fire" label is output. Each tree independently judges a fire, and the final result is determined by the majority vote. Fire prediction: The first decision tree judges that a fire has occurred, and the second decision tree also judges that a fire has occurred. The final fire decision tree forest model outputs the result "fire".

[0140] Optionally, the timing fit selection includes:

[0141] Perform trend slope extraction according to the smoke characteristic data and the temperature characteristic data to obtain smoke characteristic trend slope data and temperature characteristic trend slope data;

[0142] Specifically, the smoke concentration and temperature data that change over time are analyzed to calculate their trend slope, that is, the rate of change. The slope is an indicator that measures the direction and speed of change, and can reflect the potential development trend of the fire. The trend slope of the smoke characteristic is to calculate the rate of change based on the smoke data at multiple time points. The trend slope of the temperature characteristic is to calculate the slope of the temperature change. The initial smoke concentration is 50 ppm, which rises to 150 ppm within 5 minutes. The initial temperature is 22°C, which rises to 30°C within 5 minutes. Calculation: Smoke slope = 150-505=20 ppm / minute. Temperature slope = 30-225=1.6 °C / minute.

[0143] Performing piecewise linear regression on the smoke characteristic data and the temperature characteristic data according to the smoke characteristic trend slope data and the temperature characteristic trend slope data to obtain first smoke temperature-time fitting characteristic data;

[0144] Specifically, the time series data are divided into different intervals according to time periods, and linear regression is performed on each interval to fit its changing trend. , where a is the intercept and b is the slope, which represents the rate of change of the data (obtained through fitting). In each time period, the slope of the fitted line is calculated using the smoke data and the intercept Similarly, regressing the temperature data yields the slope and the intercept . Through regression, the slope and fit of each time period are calculated. Time periods with high fit indicate that the data change trend is relatively stable. The first smoke temperature time fit characteristic data is based on the regression results to obtain the fitting slope and correlation coefficient of each interval, indicating the time fit of the interval. For each time period, the fit of the linear regression is calculated, indicating that the data changes in this time period are relatively stable, and the regression model can well describe the trend .in, is the actual smoke / temperature value, is the smoke / temperature value predicted by regression, is the average of the actual values. A time period with a higher degree of fit indicates that the data changes during this time period are relatively stable, and the regression model can describe the trend well.

[0145] The smoke concentration in a 5-minute interval changes from 300 ppm to 500 ppm, and the temperature changes from 25°C to 40°C. The slope obtained by linear regression fitting is: Smoke slope Minutes, temperature slope Minutes. The historical trend slope is: Smoke characteristic trend slope Minutes, temperature characteristic trend slope Minutes. Slope Deviation Calculation Results: Slope Deviation Smoke Min. Slope Deviation Temperature Minutes. Through these deviation data, it can be evaluated whether the smoke and temperature changes in this period are in line with the normal trend. If the deviation is too large, it means that the fire situation is getting worse.

[0146] Cluster segmentation is performed according to the smoke characteristic data and the temperature characteristic data to obtain the smoke characteristic inflection point position data and the temperature characteristic inflection point position data;

[0147] Specifically, the time series data is segmented through a clustering algorithm to automatically identify important turning points or inflection point locations in the data. Inflection points usually correspond to significant moments of data change, such as when the fire suddenly increases or the smoke spreads rapidly. For example, by using K-means clustering, similar time points are clustered into one category to identify nodes with significant changes (i.e., inflection points). According to the clustering results, the moments when the smoke concentration or temperature changes dramatically are found and recorded as inflection points. Through clustering, the system identifies that the smoke changes most significantly in the 3rd and 4th minutes: Smoke inflection point: At the 3rd minute, the smoke concentration rises rapidly from 120 ppm to 150 ppm. Temperature inflection point: At the 4th minute, the temperature suddenly rises from 26°C to 30°C.

[0148] Specifically, the collected smoke concentration data range is [50, 500 ppm], and the temperature data range is [20°C, 100°C]. After standardization, the data values ​​are adjusted to the range of [0, 1]. Initialize the centroid positions of K clusters. Calculate the distance between each time point and the centroid, and assign each point to the nearest cluster. Update the centroid to the mean of all points assigned to the cluster. Repeat the steps until the cluster assignment no longer changes significantly. When processing time series data, time points (such as changes in smoke concentration and temperature per minute) are used as input data, and similar time points are divided into the same cluster through K-means clustering. Points with significant changes will be clustered into a cluster alone, or form a relatively independent cluster with surrounding points. In this clustering, the inflection point can be defined as the point at the cluster boundary, that is, the turning point from one cluster to another. Because K-means segments according to the severity of data changes, when a certain segment of data changes drastically, it will form a significant difference with the previous and next data, thus generating a cluster boundary. In the clustering process, assuming K = 3, the time series data is divided into three clusters. The first cluster: time period 0-2 minutes, the smoke concentration changes slowly. The second cluster: time period 3-4 minutes, the smoke and temperature change rapidly. The third cluster: time period 5-6 minutes, the fire gradually stabilizes. The system recognizes that the smoke suddenly rises at the 3rd minute, forming a significant change cluster. This time point is the inflection point of the smoke.

[0149] Specifically, the goal of K-means clustering is to minimize the variance of data points within a cluster while maximizing the differences between different clusters. Therefore, time points with drastic changes are usually divided into new clusters because these points are very different from the data in the previous time period. The sum of the squares of the distances between the data points in each cluster and the cluster center is used to measure the similarity of the data points within the cluster. Points with drastic changes will be very different from the previous and subsequent clusters, which will increase the distance between clusters and form an inflection point. K-means automatically divides those points with significant changes into new cluster boundaries through the change of the intra-cluster variance SSE. The formation of a new cluster means that the data in the previous and subsequent time periods have changed significantly. In the time series data, the system recognizes that the smoke concentration increased from 120 ppm to 150 ppm in the 3rd minute. This change is so drastic that the 3rd minute is divided into a new cluster. Therefore, the 3rd minute becomes the inflection point of the smoke. Similarly, the temperature suddenly rises from 26°C to 30°C in the 4th minute, and the temperature data is also divided into a new cluster separately, forming a significant inflection point.

[0150] Fitting and screening the smoke characteristic data and the temperature characteristic data according to the smoke characteristic inflection point position data and the temperature characteristic inflection point position data to obtain second smoke temperature-time fitting characteristic data;

[0151] Specifically, based on the identified inflection point data, the data is rescreened for a more refined fit. The focus is on fitting the data occurring near the inflection point, and extracting the second-layer time fitting characteristics. The goal of this step is to optimize the previous fitting results and capture the characteristics of the rapidly changing stage of the fire. The fitting method is similar to the previous linear regression, but focuses on the local fitting of the data before and after the inflection point. For the inflection points at the 3rd and 4th minutes, the fitting results before and after the inflection point are that the smoke rise rate after the 3rd minute reaches 30 ppm / minute, and the temperature rise rate after the 4th minute reaches 2°C / minute.

[0152] Bayesian cross screening is performed based on the first smoke temperature time fitting characteristic data and the second smoke temperature time fitting characteristic data to obtain the smoke temperature time fitting characteristic data.

[0153] Specifically, the time fit features of the first and second layers are cross-screened by the Bayesian method. Bayesian screening uses a probability model to combine prior probability and posterior probability to screen out the feature data that best represents the fire trend. The prior probability is to preliminarily infer the development mode of the fire based on the previous fit features. The posterior probability is to correct the probability of the fire mode after combining new data (such as the fit features near the inflection point). After screening, the most representative time fit feature data is finally output.

[0154] The Bayesian model initially estimated that the fire developed relatively smoothly in the first 3 minutes, but developed rapidly after the inflection point (starting from the 3rd minute). The Bayesian model finally confirmed that the data from the 3rd to the 4th minute were of great significance for fire judgment, and the data during this period were selected as key features. Initial conditions: smoke concentration was 50 ppm in the 1st minute, 120 ppm in the 3rd minute, and 150 ppm in the 5th minute. The temperature was 22°C in the 1st minute, 26°C in the 3rd minute, and 30°C in the 5th minute. The trend slope was 20 ppm / minute for smoke and 1.6°C / minute for temperature. The segmented regression had a fit of 0.95 for the first 3 minutes and 0.92 for the last 2 minutes. The inflection point was that the smoke rose significantly in the 3rd minute and the temperature changed sharply in the 4th minute. The Bayesian screening result confirmed that the smoke and temperature changes from the 3rd to the 4th minute were the key moments of the fire, as the final time fit feature.

[0155] Optionally, the spatial fit selection includes:

[0156] Obtain smoke collection position data corresponding to the smoke characteristic data and temperature collection position data corresponding to the temperature characteristic data;

[0157] Specifically, smoke and temperature sensors are installed at different locations in the building, and the collected data must be associated with the spatial location of the sensor. Each data point contains the collected smoke or temperature value and the specific location of the sensor in the building (such as floor, room). For example, sensor A is located in the north of the first floor, sensor B is located in the south of the first floor, sensor C is located on the second floor, and so on. If the building has three floors, and multiple smoke and temperature sensors are installed on each floor: Sensor A (location: north of the first floor) collects a smoke concentration of 100 ppm and a temperature of 25°C. Sensor B (location: south of the first floor) collects a smoke concentration of 80ppm and a temperature of 23°C. Sensor C (location: east of the second floor) collects a smoke concentration of 150 ppm and a temperature of 28°C.

[0158] A smoke matrix is ​​constructed according to the smoke characteristic data and the smoke collection position data to obtain smoke characteristic position matrix data, and a temperature matrix is ​​constructed according to the temperature characteristic data and the temperature collection position data to obtain temperature characteristic position matrix data;

[0159] Specifically, based on the smoke data and the corresponding positions, a two-dimensional smoke feature position matrix is ​​constructed. Each element in the matrix represents the smoke concentration at a specific location. For example, rows can represent different floors, and columns can represent sensors at different locations within a floor. Temperature matrix construction constructs a temperature feature position matrix based on temperature data, reflecting the temperature distribution at each spatial location.

[0160] Calculating the spatial distance variation degree of the smoke characteristic position matrix data and the temperature characteristic position matrix data to obtain the first smoke temperature spatial correlation data;

[0161] Specifically, by calculating the physical distance between different sensor locations (e.g., using Euclidean distance), the relative positions of each sensor are determined, and the distance can be used as the basis for analyzing the data differences between each location. The smoke concentration and temperature changes between adjacent sensors are calculated to evaluate the spatial variation of the data. The higher the variation, the greater the change in the fire situation in the area.

[0162] , is the spatial correlation data of the first smoke temperature, is the number of the smoke feature position matrix data / temperature feature position matrix data numerical values, is the number of the smoke feature position matrix data / temperature feature position matrix data numerical values, For the The value and The physical distance between the values. The distance between sensors A and B is 50 meters, and the distance between A and C is 100 meters: Smoke variability: The smoke variability between A and B Meters, smoke variability between A and C Meters. Temperature variability: The temperature variability between A and B Meters, temperature variation between A and C rice.

[0163] Calculating the spatial correlation coefficient of the smoke characteristic position matrix data and the temperature characteristic position matrix data to obtain the second smoke temperature spatial correlation data;

[0164] Specifically, the correlation between smoke and temperature data at different locations in the space is calculated. For example, the Pearson correlation coefficient is calculated to quantify whether the smoke concentration or temperature at each location is correlated with each other. A correlation coefficient of 1 indicates a high correlation, 0 indicates no correlation, and a negative value indicates a negative correlation. The correlation coefficients are calculated using the smoke and temperature data of sensors A, B, and C: Correlation coefficients of smoke concentration: The correlation coefficient between A and B = 0.9 (high correlation), and the correlation coefficient between A and C = 0.7 (moderate correlation). Correlation coefficients of temperature: The correlation coefficient between A and B = 0.95 (high correlation), and the correlation coefficient between A and C = 0.6 (moderate correlation).

[0165] The smoke characteristic data and the temperature characteristic data are spatially fitted and screened according to the first smoke temperature spatial correlation data and the second smoke temperature spatial correlation data to obtain smoke temperature spatial fitting degree characteristic data.

[0166] Specifically, combining the variability and correlation coefficient, the location data with dramatic changes and high correlation with surrounding locations are screened out. Focus on areas with large variability but weak correlation, because these areas are at risk of rapid spread of fire. Areas with high spatial fit indicate smooth data changes, and areas with low fit have abnormal fire characteristics. The screened spatial fit feature data reflects the distribution trend and change law of fire in space. Through screening, the system identified that: the smoke variability between A and B is small and the correlation is high, which belongs to the high fit area. The smoke variability between A and C is large, but the correlation is medium, indicating that the fire in the east of the second floor is developing rapidly, which belongs to the low fit area.

[0167] Optionally, S2 includes:

[0168] S21, constructing a fire state model according to the fire identification data to obtain a fire state model;

[0169] Specifically, key data is obtained from the fire identification system, including the location of the fire, smoke concentration, temperature change rate, fire spread speed, etc. These data can help describe the current state of the fire. Fire state model construction Based on these data, a fire state model is constructed to simulate the spread process and change trend of the fire. The model includes the size of the fire, the direction of flame spread, and the temperature dynamics of the fire. Model elements include the location of the fire source, the fire spread path, and the risk level of different areas. If a fire occurs in a building, the fire identification system provides the following data: The fire occurred in the northern area of ​​the second floor, and the smoke concentration was 200 ppm. The temperature rise rate was 3°C / minute. The fire spread speed was 5 meters / minute. Based on these data, the fire state model simulates the spread of the fire from the northern part of the second floor to the surrounding areas.

[0170] S22, performing fire risk factor analysis according to the fire state model to obtain fire risk factor data;

[0171] Specifically, according to the fire state model, the risk factors that affect the spread of fire and damage are identified. For example: the location of the fire source, the fire is close to flammable materials or crowded areas, the risk is higher. Environmental conditions, such as ventilation conditions, building materials, and the openness of the fire spread path. The size and spread speed of the fire, the faster the fire develops, the greater the risk. Building structure and evacuation difficulty, complex building structure may make the fire difficult to control. Generate risk factor data and conduct quantitative analysis on each risk factor, such as assigning different weights to the location of the fire source, the size of the fire, and the building structure. The results of the fire risk factor analysis are as follows: the fire source is close to flammable materials, and a large amount of paper materials are piled up in the northern area of ​​the second floor, which has a high fire risk; the size of the fire, the fire spread speed is 5 meters / minute, which is a relatively fast spread; the building structure, the second floor area has complex partitions, which will hinder evacuation. Based on these factors, the system scores each factor, such as the location of the fire source: 8 / 10, the size of the fire: 7 / 10, and the building structure: 6 / 10.

[0172] S23, estimating fire damage according to the fire risk factor data to obtain fire damage estimation data;

[0173] Specifically, based on the fire risk factors, combined with historical data or cases of similar fire events (conducting numerical data queries or data processing based on knowledge engines based on expert knowledge and experience), the damage that may be caused by the fire is estimated. Damage assessment is divided into three categories: Casualty estimation includes assessing possible casualties based on the speed of fire spread, density of personnel, difficulty of evacuation, etc. Property loss estimation includes assessing the value of buildings, equipment and other materials damaged by the fire. Environmental damage estimation includes assessing environmental pollution or damage caused by the fire, such as the spread of toxic smoke. Damage estimation data generation includes quantifying the damage caused by the fire to different aspects. For example, there are 30 employees in the fire area, the evacuation difficulty is medium, and the value of buildings and materials is 2 million yuan: Casualty estimation, if the fire is not controlled in time, it is expected that 5 people may be injured, and serious evacuation difficulties will lead to greater casualties. Property loss estimation, the fire may cause property losses of about 1 million yuan, involving equipment damage and building damage. Environmental damage estimation, due to the large amount of toxic smoke generated by the fire, it will have a serious impact on the air quality nearby.

[0174] S24, constructing a fire risk matrix according to the fire damage estimation data to obtain fire risk matrix data;

[0175] Specifically, a fire risk matrix is ​​constructed based on the fire damage estimation data. The rows of the matrix represent different types of damage (such as casualties, property losses, environmental damage, etc.), and the columns represent the areas affected by the fire (such as floors, rooms). Each matrix element represents the risk level of the area for a specific type of damage. For example, the matrix can show that the risk of property loss in the northern area of ​​the second floor is high, while the risk of casualties on the first floor is medium. Matrix data is generated, and the quantified matrix data is used for risk area division. The constructed fire risk matrix is ​​as follows: Risk of casualties: northern area of ​​the second floor: high, other areas of the second floor: medium, first floor: low. Risk of property loss: northern area of ​​the second floor: high, other areas of the second floor: medium, first floor: low. Risk of environmental damage: northern area of ​​the second floor: medium, other areas of the second floor: medium, first floor: low.

[0176] S25. Divide the risk areas according to the fire risk matrix data to obtain fire risk assessment data.

[0177] Specifically, according to the data in the fire risk matrix, the risk area of ​​the building or fire scene is divided. High-risk areas should give priority to personnel evacuation and fire-fighting measures, while low-risk areas can be used as secondary protection. The system divides the building into high-, medium-, and low-risk areas based on the matrix data, and generates corresponding fire risk assessment data as a guide for firefighting work. According to the fire risk matrix, the building is divided into the following risk areas: High-risk area: the northern part of the second floor, due to the concentration of fire sources and flammable materials, the fire risk is high. Medium-risk area: other areas on the second floor, the fire may spread to these areas, and there is a certain risk. Low-risk area: the first floor, the fire has not yet spread here, and the risk is relatively low.

[0178] A fire broke out in the building. The system constructed the following fire status model through fire identification data: Fire location: North of the second floor, Fire spread speed: 5 meters / minute, Smoke concentration: 200 ppm, Temperature rise rate: 3°C / minute. The fire risk factor analysis scores are as follows: Fire source location close to flammable materials: 8 / 10, Fire size: 7 / 10, Building structure complexity: 6 / 10. Based on these risk factors, the fire damage is estimated as follows: Casualties: 5 people are expected to be injured, Property loss: about 1 million yuan, Environmental damage: High air pollution, toxic smoke spread. After constructing the fire risk matrix, the building is divided into high, medium and low risk areas. The system generates fire risk assessment data and prioritizes rescue and smoke exhaust operations in the high-risk area in the north of the second floor.

[0179] Optionally, S3 includes:

[0180] S31, obtaining smoke exhaust pipe parameter data;

[0181] Specifically, relevant pipe parameters are obtained from the smoke exhaust system of the building. These parameters are very important for the formulation of smoke exhaust strategies. Specific parameters include pipe diameter. The diameter of the smoke exhaust pipe directly affects the smoke exhaust capacity. Larger pipes can exhaust smoke quickly, while smaller pipes have flow restrictions; fan power. The power of the smoke exhaust fan determines the speed of the airflow and the smoke exhaust efficiency. A larger fan power can cover a larger area and exhaust more smoke; smoke exhaust port location. The distribution and number of smoke exhaust ports in the building will affect the choice of smoke exhaust path. Smoke exhaust ports close to high-risk areas are used first; ventilation resistance. The resistance in the pipe affects the flow rate of smoke, and its weakening effect on smoke exhaust efficiency needs to be considered. The parameters of the building's smoke exhaust system are pipe diameter of 500 mm. The maximum smoke exhaust capacity of each fan is 4000 m³ / h. There are 4 smoke exhaust ports on the first and second floors respectively. The total resistance of the smoke exhaust pipe is calculated to be 200 Pa.

[0182] S32, prioritizing smoke exhaust requirements according to the fire risk assessment data to obtain fire risk smoke exhaust requirement priority data;

[0183] Specifically, according to the fire risk assessment data generated in the previous step, the fire risks in different areas of the building are divided. In high-risk areas, the fire spreads quickly and is densely populated, so smoke exhaust operations are given priority; medium and low-risk areas are handled in turn according to specific circumstances. Determination of smoke exhaust demand priority: According to the fire risk level, the system prioritizes the smoke exhaust demand of each area. The higher the priority, the more urgent the smoke exhaust demand. High-risk areas (such as those near the fire source) are given priority for smoke exhaust. Medium-risk areas (such as areas where the fire may spread) are second-best for smoke exhaust. Low-risk areas are smoke exhausted last. According to the fire risk assessment, the northern part of the second floor of the building is a high-risk area, the first floor is a medium-risk area, and the southern part of the second floor is a low-risk area. The system prioritizes: High-risk area (north of the second floor): priority 1, emergency smoke exhaust. Medium-risk area (first floor): priority 2, second-best smoke exhaust. Low-risk area (south of the second floor): priority 3, last smoke exhaust.

[0184] S33, generating a smoke exhaust path according to the smoke exhaust pipe parameter data and the fire risk smoke exhaust demand priority data to obtain smoke exhaust path data;

[0185] Specifically, through path algorithms, such as the shortest planning, priority is given to paths with the least smoke exhaust resistance and sufficient fan power to ensure the highest smoke exhaust efficiency. Smoke exhaust paths are planned in high-risk areas to ensure that smoke in the area can be exhausted as soon as possible. When planning the smoke exhaust path, high-temperature smoke should be prevented from spreading to other safe areas. Combined with the layout of the building's smoke exhaust ducts, the best path for exhausting smoke from high-risk areas is calculated. The path should minimize bends and resistance. The high-risk area (north of the second floor) selects the nearest smoke exhaust port on the north of the second floor, with a calculated path length of 30 meters, a fan smoke exhaust capacity of 4000 m³ / h, and an estimated smoke exhaust time of 15 minutes. The medium-risk area (first floor) selects the smoke exhaust port in the middle of the first floor, with a path length of 25 meters and a smoke exhaust time of 12 minutes. The low-risk area (south of the second floor) selects the smoke exhaust port on the south of the second floor, with a path length of 40 meters and a smoke exhaust time of 20 minutes.

[0186] S34. Generate a smoke exhaust strategy according to the smoke exhaust path data and the smoke exhaust pipe parameter data to obtain smoke exhaust strategy data.

[0187] Specifically, the order of fan opening is scheduled according to the smoke exhaust path data. The fans in the high-risk area are started first to ensure that the smoke in this area is discharged first. The optimal smoke exhaust port on the path is selected to open to avoid airflow interference caused by the simultaneous opening of multiple smoke exhaust ports. The fan power is dynamically adjusted according to the smoke concentration and smoke exhaust demand. In high-risk areas, the fan should run at full power; in low-risk areas, lower power can be used to save energy. During the smoke exhaust process, the changes in smoke concentration and temperature are monitored in real time, and the smoke exhaust strategy is adjusted according to real-time data to ensure efficient smoke exhaust and avoid resource waste. Phase 1 (high-risk area, north of the second floor): Start the smoke exhaust port on the north of the second floor, turn on the main smoke exhaust fan, the fan power is 4000 m³ / h, and exhaust smoke at full capacity. Smoke exhaust time: 15 minutes. Phase 2 (medium-risk area, first floor): Start the smoke exhaust port in the middle of the first floor, the fan power is 3000 m³ / h. Smoke exhaust time: 12 minutes. Phase 3 (low-risk area, south of the second floor): Start the smoke exhaust port on the south of the second floor, the fan power is 2000 m³ / h. Smoke exhaust time: 20 minutes.

[0188] Optionally, S4 includes:

[0189] S41, obtaining real-time fire monitoring data;

[0190] Specifically, various sensors installed in the building (such as smoke sensors, temperature sensors, airflow sensors, image sensors, etc.) continuously monitor the dynamic changes in the fire area, and the data is transmitted to the fire control center in real time through the Internet of Things technology. Monitor the concentration changes of smoke and obtain smoke concentration data in ppm (per million concentration). Detect temperature changes in the fire area and obtain temperature data to reflect the size and location of the fire. Monitor the air flow speed and direction to obtain airflow data to help determine the spread of smoke. Obtain image data through infrared or visible light cameras to monitor the development of the fire.

[0191] S42, performing fire identification according to the real-time fire monitoring data to obtain real-time fire identification data;

[0192] Specifically, the real-time data provided by the sensors is analyzed to determine whether the fire has spread, whether the fire has intensified, and whether the smoke concentration has changed, etc., to conduct real-time fire identification. By continuously monitoring changes in temperature and smoke concentration and comparing fire models, the dynamic development of the fire can be determined.

[0193] S43. Perform real-time adjustment according to the real-time fire identification data and the smoke exhaust strategy data to obtain real-time adjustment data of the smoke exhaust strategy to perform smoke exhaust operations for fire protection engineering.

[0194] Specifically, based on the fire identification results, smoke from high-risk areas is discharged first. If the fire intensifies in the northern area of ​​the second floor, adjust the fan power and the opening status of the smoke exhaust port to speed up the smoke exhaust. When the smoke spreads to new areas (such as the southern part of the second floor and the first floor), the smoke exhaust path needs to be replanned to prevent the smoke from entering new unaffected areas. The power of the smoke exhaust fan is dynamically adjusted according to the real-time smoke concentration and temperature changes. If the smoke concentration rises too quickly, the fan power should be increased to the maximum. According to the direction of smoke diffusion, specific smoke exhaust ports are opened or closed in real time to ensure that the smoke is directed to the correct exhaust channel. The system continuously optimizes the smoke exhaust strategy based on real-time data to ensure that high-concentration smoke is removed in the shortest time and avoid affecting other areas.

[0195] Optionally, the present application further provides a smoke exhaust system for fire engineering, which is used to perform the smoke exhaust method for fire engineering as described above, and the smoke exhaust system for fire engineering comprises:

[0196] A fire identification module is used to obtain fire monitoring data and perform fire identification based on the fire monitoring data to obtain fire identification data;

[0197] A fire risk assessment module, used to perform fire risk assessment based on fire identification data to obtain fire risk assessment data;

[0198] A smoke exhaust strategy generation module is used to generate a smoke exhaust strategy according to the fire risk assessment data to obtain smoke exhaust strategy data;

[0199] The smoke exhaust strategy real-time adjustment module is used to obtain real-time fire monitoring data and adjust the smoke exhaust strategy data in real time according to the real-time fire monitoring data to obtain the real-time adjustment data of the smoke exhaust strategy for performing smoke exhaust operations for fire protection engineering.

[0200] Therefore, from any point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the attached application documents rather than the above description, and it is intended that all changes falling within the meaning and scope of equivalent elements of the application documents are included in the present invention.

[0201] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A smoke exhaust method for fire protection engineering, characterized in that: The method comprises: S1. Obtain fire monitoring data, and perform fire identification based on the fire monitoring data to obtain fire identification data; S2. Perform fire risk assessment based on the fire identification data to obtain fire risk assessment data; S3. Generate a smoke exhaust strategy based on the fire risk assessment data to obtain smoke exhaust strategy data; S4, obtaining real-time fire monitoring data, and adjusting the smoke exhaust strategy data in real time according to the real-time fire monitoring data, to obtain real-time adjustment data of the smoke exhaust strategy, so as to perform smoke exhaust operations for fire engineering; S1 includes: Smoke data and temperature data are collected through smoke sensors and temperature sensors to obtain smoke data and temperature data respectively; Fire judgment is performed according to smoke data and temperature data to obtain fire judgment data; When it is determined that the fire judgment data includes suspected fire data, the image sensor is controlled to collect images to obtain scene image data; Perform fire detection according to the scene image data to obtain fire detection data; Integrate smoke data, temperature data and fire detection data to obtain fire monitoring data; Perform fire identification according to the fire monitoring data to obtain fire identification data; Fire assessment includes: Calculate the change rate according to the smoke data and the temperature data to obtain the smoke change rate data and the temperature change rate data; Determine whether the smoke change rate data is greater than or equal to a preset smoke change rate threshold and whether the temperature change rate data is greater than or equal to a preset temperature change rate threshold; When it is determined that the smoke change rate data is greater than or equal to a preset smoke change rate threshold and the temperature change rate data is greater than or equal to a preset temperature change rate threshold, feature extraction is performed on the smoke data and the temperature data to obtain smoke feature data and temperature feature data respectively; Using the preset fire decision forest model to judge the smoke characteristic data and the temperature characteristic data, fire judgment data is obtained; The preset fire decision forest model is constructed in the following ways: Obtain historical smoke data, historical temperature data, and corresponding historical fire label data; Perform feature extraction on historical smoke data and historical temperature data to obtain historical smoke feature data and historical temperature feature data; Perform time series fitting degree selection and space fitting degree selection on historical smoke characteristic data and historical temperature characteristic data, and obtain smoke temperature time fitting degree characteristic data and smoke temperature space fitting degree characteristic data respectively; Data is divided according to the smoke temperature time fitting characteristic data and the smoke temperature space fitting characteristic data, to obtain smoke temperature time fitting division data and smoke temperature space fitting division data respectively, wherein the smoke temperature time fitting division data includes smoke temperature time fitting training data and smoke temperature time fitting test data, and the smoke temperature space fitting division data includes smoke temperature space fitting training data and smoke temperature space fitting test data; Perform node splitting on the smoke temperature time fitting training data and the smoke temperature space fitting training data to obtain first smoke temperature node data and second smoke temperature node data respectively; Generate a tree based on the first smoke temperature node data, the second smoke temperature node data and the corresponding historical fire label data to obtain primary first fire decision tree data and primary second decision tree data; Post-pruning the primary first fire decision tree data and the primary second decision tree data to obtain secondary first fire decision tree data and secondary second decision tree data; The secondary first fire decision tree data is iteratively optimized using the smoke temperature time fitting test data to obtain the first fire decision tree model, and the secondary second decision tree data is iteratively optimized using the smoke temperature space fitting test data to obtain the second fire decision tree model; The first fire decision tree model and the second fire decision tree model are added with a voting layer to obtain a fire decision tree forest model; The timing fit options include: Perform trend slope extraction according to the smoke characteristic data and the temperature characteristic data to obtain smoke characteristic trend slope data and temperature characteristic trend slope data; Performing piecewise linear regression on the smoke characteristic data and the temperature characteristic data according to the smoke characteristic trend slope data and the temperature characteristic trend slope data to obtain first smoke temperature-time fitting characteristic data; Cluster segmentation is performed according to the smoke characteristic data and the temperature characteristic data to obtain the smoke characteristic inflection point position data and the temperature characteristic inflection point position data; Fitting and screening the smoke characteristic data and the temperature characteristic data according to the smoke characteristic inflection point position data and the temperature characteristic inflection point position data to obtain second smoke temperature-time fitting characteristic data; Bayesian cross screening is performed based on the first smoke temperature time fitting characteristic data and the second smoke temperature time fitting characteristic data to obtain the smoke temperature time fitting characteristic data.

2. The method according to claim 1, characterized in that The spatial fit options include: Obtain smoke collection position data corresponding to the smoke characteristic data and temperature collection position data corresponding to the temperature characteristic data; A smoke matrix is ​​constructed according to the smoke characteristic data and the smoke collection position data to obtain smoke characteristic position matrix data, and a temperature matrix is ​​constructed according to the temperature characteristic data and the temperature collection position data to obtain temperature characteristic position matrix data; Calculating the spatial distance variation degree of the smoke characteristic position matrix data and the temperature characteristic position matrix data to obtain the first smoke temperature spatial correlation data; Calculating the spatial correlation coefficient of the smoke characteristic position matrix data and the temperature characteristic position matrix data to obtain the second smoke temperature spatial correlation data; The smoke characteristic data and the temperature characteristic data are spatially fitted and screened according to the first smoke temperature spatial correlation data and the second smoke temperature spatial correlation data to obtain smoke temperature spatial fitting degree characteristic data.

3. The method according to claim 1, characterized in that S2 include: A fire state model is constructed according to the fire identification data to obtain a fire state model; Conduct fire risk factor analysis based on the fire state model to obtain fire risk factor data; Fire damage estimation is performed based on fire risk factor data to obtain fire damage estimation data; Construct a fire risk matrix based on fire damage estimation data to obtain fire risk matrix data; Risk areas are divided according to the fire risk matrix data to obtain fire risk assessment data.

4. The method according to claim 1, characterized in that S3 include: Obtain exhaust pipe parameter data; Prioritize smoke exhaust requirements based on fire risk assessment data to obtain fire risk smoke exhaust requirement priority data; Generate a smoke exhaust path based on smoke exhaust pipe parameter data and fire risk smoke exhaust demand priority data to obtain smoke exhaust path data; A smoke exhaust strategy is generated according to the smoke exhaust path data and the smoke exhaust duct parameter data to obtain the smoke exhaust strategy data.

5. The method according to claim 1, characterized in that S4 includes: Obtain real-time fire monitoring data; Fire identification is performed according to real-time fire monitoring data to obtain real-time fire identification data; Real-time adjustments are made based on real-time fire identification data and smoke exhaust strategy data to obtain real-time adjustment data for the smoke exhaust strategy in order to perform smoke exhaust operations for fire protection engineering.

6. A smoke exhaust system for fire protection engineering, characterized in that: Used to perform the smoke exhaust method for fire engineering according to claim 1, the smoke exhaust system for fire engineering comprises: A fire identification module is used to obtain fire monitoring data and perform fire identification based on the fire monitoring data to obtain fire identification data; A fire risk assessment module, used to perform fire risk assessment based on fire identification data to obtain fire risk assessment data; A smoke exhaust strategy generation module is used to generate a smoke exhaust strategy according to the fire risk assessment data to obtain smoke exhaust strategy data; The smoke exhaust strategy real-time adjustment module is used to obtain real-time fire monitoring data and adjust the smoke exhaust strategy data in real time according to the real-time fire monitoring data to obtain the real-time adjustment data of the smoke exhaust strategy for performing smoke exhaust operations for fire protection engineering.

Images