Intelligent building platform fire alarm monitoring system

Abstract

The invention discloses an intelligent building platform fire alarm monitoring system, and belongs to the technical field of building fire safety. The system comprises a composite sensing network module which is used for detecting smoke concentration, temperature, electrical parameters and combustible gas concentration and collecting and analyzing video images so as to achieve the effect of timely discovering fire behavior; the central intelligent processing unit is used for analyzing and processing the data of the various sensors so as to judge the fire occurrence probability and the danger level; the efficient communication module is used for realizing data and command interaction with other subsystems of the building platform; the communication monitoring module is used for monitoring a communication state and ensuring the stability and timeliness of data transmission; the multi-element early warning module is used for starting a multi-element early warning mechanism when the fire risk rises; and the remote monitoring and management terminal is used by property management departments and fire departments to check the fire alarm monitoring dynamic state and the equipment operation state in real time. According to the system, accurate, rapid and cooperative fire alarm monitoring and prevention and control can be realized.

Description

Technical Field

[0001] The present application relates to the field of building fire safety technology, and in particular to a fire alarm monitoring system for a smart building platform. Background Art

[0002] With the rapid development of urbanization, various complex buildings such as high-rise buildings and large commercial complexes have sprung up like mushrooms after a rain. The continuous expansion of building scale and the increasing complexity of functions have increased the potential for fire hazards. These places are densely populated and complex. Once a fire occurs, it is very easy to cause serious casualties and property losses if it is not discovered and handled in time.

[0003] Traditional fire alarm monitoring methods often rely on a single smoke detector or heat detector, which has certain limitations. For example, common smoke detectors are easily disturbed by environmental factors such as dust and humidity, and false alarms frequently occur, causing firefighters to be exhausted and wasting a lot of manpower and material resources; heat detectors often respond only after the fire develops to a certain stage and the temperature rises sharply, and the perception of the initial smoldering fire is delayed, which causes precious firefighting time to be lost. Moreover, most existing fire alarm monitoring systems operate in isolation, lacking effective coordination and linkage with other systems in the building, such as ventilation, lighting, and elevators, and unable to form an overall response mechanism at the moment of fire, greatly affecting the efficiency of personnel evacuation and fire extinguishing. Therefore, traditional fire alarm monitoring methods are difficult to meet the stringent needs of current building fire safety. Summary of the invention

[0004] Therefore, this application is proposed to address the problems and needs existing in the above-mentioned prior art.

[0005] The purpose of this application is to provide a fire alarm monitoring system for a smart building platform, which, relative to the defects of the prior art, can achieve accurate, rapid and coordinated fire alarm monitoring and prevention.

[0006] The purpose of this application is achieved through the following technical solutions:

[0007] The present application embodiment provides a fire alarm monitoring system for a smart building platform, the system comprising:

[0008] The composite sensor network module is composed of a variety of high-precision sensors, which are used to detect the smoke density, temperature, electrical parameters, and combustible gas concentration inside the building and collect and analyze video images to detect fires in a timely manner; the central intelligent processing unit is equipped with a high-performance processor, which runs a fire alarm judgment algorithm based on big data and artificial intelligence to analyze and process the data transmitted by various sensors to determine the probability of fire and the level of danger, and has a built-in self-learning module;

[0009] Efficient communication module, supporting multiple communication protocols, used to realize data and command interaction with other subsystems of the building platform based on the data processed by the central intelligent processing unit;

[0010] The communication monitoring module is used to monitor the communication status between various sensors, between each sensor and the central intelligent processing unit, and between the high-efficiency communication module and other subsystems of the building platform to ensure the stability and timeliness of data transmission;

[0011] A multi-warning module is used to activate the multi-warning mechanism when the fire risk increases;

[0012] Remote monitoring and management terminal, used for property and fire departments to remotely access through the Internet to view data such as fire alarm monitoring dynamics and equipment operating status in real time in the building.

[0013] In the above-mentioned smart building platform fire monitoring system, the multiple high-precision sensors include: intelligent optical smoke sensors, distributed fiber optic temperature detectors, combustible gas detectors, image-type fire detectors, and electrical fire detectors. The sensors in the composite sensor network module are reasonably installed according to the functional layout of the building and the fire risk assessment results. The multiple high-precision sensors are connected to the central intelligent processing unit through wired or wireless hybrid networking.

[0014] In the above-mentioned smart building platform fire monitoring system, the multiple high-precision sensors send collected data to the central intelligent processing unit at a first frequency, and the central intelligent processing unit processes and analyzes the various sensor data received at a second frequency, and the first frequency is greater than the second frequency.

[0015] In the above-mentioned smart building platform fire monitoring system, the central intelligent processing unit processes and analyzes the various sensor data received, including: data preprocessing, real-time analysis and fusion judgment, self-learning and optimization; the data preprocessing is to use a filtering algorithm to process a single sensor data to remove noise interference.

[0016] In the above-mentioned smart building platform fire monitoring system, the real-time analysis and fusion judgment include: single sensor dynamic threshold setting and multi-sensor data fusion judgment; the single sensor dynamic threshold setting is to dynamically calculate and set the threshold according to the building area function weight coefficient, environmental humidity influence coefficient, and historical data fluctuation adjustment coefficient; the multi-sensor data fusion judgment is to use a pre-constructed fire feature database to calculate the similarity between the current sensor data and the typical pattern using a similarity matching algorithm, and then calculate the comprehensive similarity through weighted average calculation, and then judge whether the comprehensive similarity is greater than the pre-set similarity threshold. If the judgment result is yes, the possibility of fire is extremely high, and then according to the pre-set risk level classification rules, a linear regression model or a decision tree model is used to locate the fire risk level.

[0017] In the above-mentioned fire alarm monitoring system of a smart building platform, the self-learning and optimization are specifically as follows: in daily operation, each sensor data fluctuation, alarm triggering situation, and whether it is finally confirmed as a fire event are continuously recorded as training samples, and a multi-layer perceptron neural network model in deep learning is used. The input layer receives sensor data features, the hidden layer abstracts and converts features, and the output layer predicts the probability of fire occurrence. Through the back propagation algorithm, based on the error between the predicted value and the actual confirmed fire event, the gradient descent method is used to continuously adjust the weights of each layer of the neural network, optimize the model parameters, and continuously reduce the false alarm rate.

[0018] In the above-mentioned smart building platform fire monitoring system, the communication methods adopted by the efficient communication module include 4G, 5G, WiFi, and industrial Ethernet.

[0019] In the above-mentioned fire alarm monitoring system of a smart building platform, the warning methods of the multi-warning module include sound and light alarms, mobile phone push notifications, and scrolling prompts on the display screen inside the building.

[0020] In the above-mentioned smart building platform fire alarm monitoring system, the system also includes: a fire protection facility management module, which is used for detecting and managing fire protection equipment; a fault display module, which is used to display the fault information detected by the communication monitoring module, so that the manager can perform maintenance on the faulty equipment or modules as soon as possible.

[0021] In the above-mentioned fire alarm monitoring system of a smart building platform, the fire protection facility management module includes: a fire extinguisher management unit, which is used for monitoring and managing the validity period, location, and real-time pressure of fire extinguishers; a fire hydrant management unit, which is used for monitoring and managing the water flow, water pressure, and location of fire hydrants; a fire detector management unit, which is used for managing various fire detection sensors, including the type, location, and status of the detectors.

[0022] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0023] The present application discloses a fire alarm monitoring system for a smart building platform. The composite sensor network module collects data such as smoke concentration, temperature, electrical parameters, combustible gas concentration, video images, etc. by setting single or multiple high-precision fire detection sensors at key locations of the building. For example, the intelligent optical smoke sensor can effectively distinguish between smoke and dust, greatly reducing the false alarm rate, and has a high-sensitivity detection capability, which can quickly and accurately detect the presence of smoke, and can promptly issue an alarm signal for a small amount of smoke produced by smoldering. The distributed fiber optic temperature detector utilizes the temperature sensitivity of the optical fiber to monitor temperature changes along the line in real time, accurately locate hot spots, and is not subject to electromagnetic interference, and is particularly suitable for electrical fire hazard inspection. The combustible gas detector can quickly detect the concentration of methane, propane and other combustible gases, and immediately issue an early warning once the concentration exceeds the standard. The image-based fire detector can identify flame characteristics as well as signs of faint smoke produced by smoldering, color changes, and subtle deformation of the surface of objects due to heat.

[0024] The central intelligent processing unit runs a fire alarm judgment algorithm based on big data and artificial intelligence to analyze and process the collected data to determine the probability of fire and the level of danger. When a fire is determined to have occurred, the subsequent emergency response mechanism is initiated according to the level of danger to avoid losses caused by false alarms and late alarms generated by a single fire detection sensor. The central intelligent processing unit also has a built-in self-learning module that uses the multi-layer perceptron neural network model in deep learning to continuously optimize the model parameters and continuously reduce the false alarm rate.

[0025] Through the efficient communication module, data and command interaction with other subsystems of the building platform (such as ventilation, lighting, elevators, and fire sprinkler systems) can be achieved, thereby realizing effective coordination and linkage with other systems and an overall response mechanism. When a fire occurs, a variety of warning methods are generated through the multi-warning module to promptly and effectively notify personnel to respond and evacuate, ensure personal safety and avoid property losses. Through the communication monitoring module, the communication status between various sensors, between each sensor and the central intelligent processing unit, and between the efficient communication module and other subsystems of the building platform is monitored to promptly discover fault information of equipment or modules, thereby ensuring the stability and timeliness of data transmission.

[0026] The fire-fighting facility management module detects and manages fire-fighting equipment so that it can be used to extinguish fires in a timely and effective manner when a fire occurs. The fault display module displays the fault information detected by the communication monitoring module so that managers can maintain the faulty equipment or modules as soon as possible, thereby ensuring the stable and reliable operation of the system. Finally, through the remote monitoring and management terminal, the property and fire departments can remotely view the fire alarm monitoring dynamics and equipment operation status data in the building in real time, which facilitates remote management and timely detection of problems and early response, ultimately achieving accurate, rapid and coordinated fire alarm monitoring and prevention. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The above and other purposes, features and advantages of the present invention will become more apparent by describing the embodiments of the present invention in more detail in conjunction with the accompanying drawings. The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention.

[0028] Figure 1 It is a structural framework diagram of a smart building platform fire alarm monitoring system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be understood that the specific embodiments described here are only part of the embodiments of the present application, and the present application is not limited to the exemplary embodiments described here.

[0030] Example 1

[0031] like Figure 1 As shown, the present application provides a fire alarm monitoring system for a smart building platform, which is implemented by including the following modules:

[0032] The composite sensor network module is composed of a variety of high-precision sensors, which are used to detect the smoke density, temperature, electrical parameters, and combustible gas concentration inside the building and collect and analyze video images to detect fires in a timely manner;

[0033] The central intelligent processing unit is equipped with a high-performance processor, which runs a fire alarm judgment algorithm based on big data and artificial intelligence to analyze and process the data transmitted by various sensors to determine the probability of fire and the level of danger, and has a built-in self-learning module;

[0034] Efficient communication module, supporting multiple communication protocols, used to realize data and command interaction with other subsystems of the building platform based on the data processed by the central intelligent processing unit;

[0035] The communication monitoring module is used to monitor the communication status between various sensors, between each sensor and the central intelligent processing unit, and between the high-efficiency communication module and other subsystems of the building platform to ensure the stability and timeliness of data transmission;

[0036] A multi-warning module is used to activate the multi-warning mechanism when the fire risk increases;

[0037] Remote monitoring and management terminal, used for property management and fire departments to access remotely through the Internet to view real-time data such as fire alarm monitoring dynamics and equipment operation status in the building;

[0038] Specifically, the inventors found that traditional fire monitoring methods often rely on single smoke detectors or temperature detectors, which have certain limitations. For example, common smoke detectors are easily disturbed by environmental factors such as dust and humidity, and false alarms frequently occur, causing firefighters to be exhausted and wasting a lot of manpower and material resources; temperature detectors often respond only after the fire develops to a certain stage and the temperature rises sharply, and the perception of the initial smoldering fire is delayed, which causes precious firefighting time to be lost; moreover, most existing fire monitoring systems operate in isolation, lacking effective coordination and linkage with other systems in the building, such as ventilation, lighting, and elevators, and unable to form an overall response mechanism at the moment of a fire. The traditional fire alarm monitoring methods are difficult to meet the stringent requirements of current building fire safety. In order to solve these technical problems, the inventors have proposed a fire alarm monitoring system for a smart building platform, namely, the fire alarm monitoring system comprising a composite sensor network module, a central intelligent processing unit, an efficient communication module, a communication monitoring module, a multi-warning module, and a remote monitoring and management terminal. The composite sensor network module collects data such as smoke concentration, temperature, electrical parameters, combustible gas concentration, and video images by setting a single or multiple high-precision fire detection sensors at key locations of the building. The collected data is processed by the central intelligent processing unit based on big data and human intelligence. The fire alarm judgment algorithm of artificial intelligence is used for analysis and processing to determine the probability of fire and the danger level. When a fire is judged to occur, the subsequent emergency response mechanism is activated according to the danger level to avoid losses caused by false alarms and late alarms generated by a single fire detection sensor. The central intelligent processing unit also has a built-in self-learning module, which uses the multi-layer perceptron neural network model in deep learning to continuously optimize the model parameters and continuously reduce the false alarm rate; through the efficient communication module, it realizes data and command interaction with other subsystems of the building platform (such as ventilation, lighting, elevators, and fire sprinkler systems), thereby realizing effective collaborative linkage and overall response mechanism with other systems. When a fire is judged to occur, the multi-warning model is used to The module generates a variety of early warning methods to timely and effectively notify personnel to respond and evacuate, ensure personal safety and avoid property losses; in addition, the communication monitoring module monitors the communication status between various sensors, between each sensor and the central intelligent processing unit, and between the high-efficiency communication module and other subsystems of the building platform, so as to timely discover the fault information of equipment or modules, thereby ensuring the stability and timeliness of data transmission. Finally, through the remote monitoring and management terminal, the property and fire departments can remotely view the fire alarm monitoring dynamics and equipment operation status data in the building in real time, which is convenient for remote management and timely discovery of problems and early response, and ultimately achieve accurate, fast and coordinated fire alarm monitoring and prevention.

[0039] Furthermore, the various high-precision sensors in the composite sensor network module include: intelligent optical smoke sensors, distributed fiber optic temperature detectors, combustible gas detectors, image-based fire detectors, and electrical fire detectors. The intelligent optical smoke sensors use advanced scattered light or dimming optical principles, and are combined with an intelligent algorithm that adaptively adjusts dynamic thresholds based on real-time environmental conditions such as humidity, temperature, and light intensity. It can effectively distinguish between smoke and dust, greatly reducing the false alarm rate, and has high-sensitivity detection capabilities. It can quickly and accurately detect the presence of smoke, whether it is a small amount of smoke produced by smoldering or a large amount of smoke produced by open flames. It can send out alarm signals in a timely manner; distributed fiber optic temperature detectors are laid along key parts of buildings such as cable trays, ventilation ducts, and other areas prone to smoldering. The temperature-sensitive characteristics of optical fibers can be used to monitor temperature changes along the line in real time, accurately locate hot spots, and are not subject to electromagnetic interference. They are particularly suitable for electrical fire hazard inspections; combustible gas detectors are deployed in key areas such as kitchens and boiler rooms that are prone to the accumulation of combustible gases, and can quickly detect methane, propane, and other gases. The image-based fire detector uses a high-definition camera combined with intelligent image recognition technology to collect and analyze real-time images of the monitored area. It can not only identify the characteristics of flames and the signs of faint smoke produced by smoldering, color changes, and subtle deformation of the surface of objects due to heat, but also determine the direction of fire spread, providing an intuitive basis for fire-fighting decisions. The electrical fire detector can detect residual current and send out an audible and visual alarm signal when the residual current reaches the set value. These sensors are reasonably installed in various key areas of the building based on the functional layout of the building and the results of fire risk assessment. For example, intelligent optical smoke sensors are installed every three floors in the stairwells of high-rise residential buildings, distributed optical fiber temperature detectors and electrical fire detectors are densely laid in the distribution room and electrical shaft, and the kitchen area is equipped with combustible gas detectors and intelligent optical smoke sensors to ensure all-round monitoring of key areas. The above multiple high-precision sensors are connected to the central intelligent processing unit through wired or wireless hybrid networking to ensure stable and real-time transmission of the collected data to the central intelligent processing unit.

[0040] Furthermore, the above-mentioned multiple high-precision sensors send collected data to the central intelligent processing unit at a first frequency, such as continuously collecting data once per second and transmitting it to the central intelligent processing unit in real time through a hybrid networking method. The central intelligent processing unit processes and analyzes the various sensor data received at a second frequency. The first frequency is greater than the second frequency, such as the central intelligent processing unit processes and analyzes the received data every 2 seconds. No specific limitation is made here.

[0041] Furthermore, the communication methods adopted by the efficient communication module include 4G, 5G, WiFi, industrial Ethernet, etc., to achieve seamless connection with various subsystems within the building platform.

[0042] Furthermore, the warning methods of the multi-warning module include sound and light alarms, mobile phone push notifications, and scrolling prompts on display screens inside buildings. Once the system detects an increased fire risk, the multi-warning mechanism is immediately activated. For example, in public areas inside the building (such as corridors, lobbies, etc.), an alarm is sounded through a high-decibel sound and light alarm, and fire warning information is scrolled on the display screen inside the building; customized fire alarm notifications are pushed to the smartphones of people in the building, including key information such as the fire location and evacuation routes, to ensure that people are aware of the danger at the first time and take timely and effective actions to ensure personal safety and avoid property losses.

[0043] Example 2

[0044] This embodiment is an explanation of the embodiment 1. The central intelligent processing unit receives various types of data from the composite sensor network module in real time at an extremely high frequency, and then processes and analyzes the received various sensor data, including: data preprocessing, real-time analysis and fusion judgment, self-learning and optimization; among which, data preprocessing is to use a filtering algorithm to process a single sensor data to remove noise interference, ensure the relative smoothness and reliability of the input data, and lay the foundation for subsequent accurate analysis.

[0045] Real-time analysis and fusion judgment, including: single sensor dynamic threshold setting, multi-sensor data fusion judgment;

[0046] The dynamic threshold of a single sensor is set by dynamically calculating the threshold based on the building area function weight coefficient, environmental humidity influence coefficient, and historical data fluctuation adjustment coefficient;

[0047] The multi-sensor data fusion judgment is to use a pre-constructed fire feature database, use a similarity matching algorithm to calculate the similarity between the current sensor data and the typical pattern, and then obtain the comprehensive similarity through weighted average calculation, and then judge whether the comprehensive similarity is greater than the pre-set similarity threshold. If the judgment result is yes, the possibility of fire is extremely high, and then according to the pre-set risk level classification rules, use a linear regression model or a decision tree model to locate the fire risk level.

[0048] Specifically, the central intelligent processing unit receives data from the composite sensor network module in real time at a frequency of f (millisecond level, assuming f = 10 milliseconds, that is, data is received every 10 milliseconds). Assume that the smoke concentration value sequence transmitted by the smoke sensor is C(t), the temperature value sequence fed back by the temperature sensor is T(t), the combustible gas concentration value sequence detected by the combustible gas sensor is G(t), and the electrical parameter value sequence monitored by the electrical fire detector is E(t), where t represents the time series. At the moment of receiving, a filtering algorithm is used to remove noise interference. Taking the sliding average filtering algorithm as an example, for the smoke concentration value sequence C(t), its filtered value Cf(t) is calculated and output by the following calculation formula:

[0049]

[0050] Where n is the sliding window size. Assume n=5, that is, the smoke concentration values ​​at the current moment and the previous four moments are averaged. This can effectively smooth the peak data caused by the instantaneous fluctuation of the sensor and ensure that the input data is relatively stable and reliable. Similarly, similar filtering processing can be performed on other sensor data.

[0051] Then the central intelligent processing unit performs real-time analysis and fusion judgment on the data, and sets dynamic thresholds for single sensors to improve their accuracy. Taking the smoke sensor as an example, assuming that the functional weight coefficient of the building area is wfunc (such as wfunc=0.8 in the machine room area, wfunc=0.6 in the warehouse area, etc.), the environmental humidity influence coefficient is whum (when the humidity is greater than 70%, whum=1.2, when the humidity is less than 30%, whum=0.8), and the historical data fluctuation adjustment coefficient is whist (calculated based on the standard deviation of the past data in the area, if the standard deviation is small, whist=0.9, and if the standard deviation is large, whist=1.1), the initial default smoke alarm threshold is C 0 , then the dynamic smoke alarm threshold Cth(t) calculation formula is:

[0052] Cth(t)=C 0 ×wfunc×whum×whist;

[0053] When the smoke sensor and the temperature sensor send out abnormal signals at the same time (i.e. Cf(t)>Cth(t) and the temperature change rate ΔT(t)>ΔT 0 , ΔT 0 is the set temperature change initial threshold), and the combustible gas sensor detects an increase in the combustible gas concentration (G(t)>G 0 , G 0 When the combustible gas alarm threshold is reached, multi-sensor data fusion analysis is started.

[0054] The current multi-sensor data combination mode is compared through the pre-built fire feature database. Assuming that the typical smoke concentration change model from initial smoldering to open flame outbreak stored in the fire feature database is Cm(t), the temperature change model is Tm(t), and the combustible gas concentration change model is Gm(t), a similarity matching algorithm such as the Euclidean distance method is used to calculate the similarity between the current data and the typical mode. Taking smoke concentration as an example, the Euclidean distance calculation formula is:

[0055]

[0056] Among them, k is the length of the matching time window. Assume that k = 10, that is, compare the difference between the current 10 time data and the typical mode. Cf(i) is the smoke concentration value after filtering at time i, and Cm(i) is the typical smoke concentration change model value in the fire characteristic database at time i. Similarly, the Euclidean distance d of temperature and combustible gas concentration is calculated. T (t), d G (t), by calculating the Euclidean distance, the algorithm can quantitatively understand how close the current smoke density change trend is to the known smoke density change patterns of typical fires from initial smoldering to open flames, and then assist in judging whether a fire has occurred and the development trend of the fire, providing a key basis for subsequent comprehensive decision-making;

[0057] The comprehensive similarity S(t) can be obtained by weighted average, and the calculation formula is:

[0058]

[0059] Among them, wC, w T , wG is the weight coefficient of the corresponding sensor data, which is allocated according to the importance of the sensor to fire judgment, and wC+w T +wG=1. When S(t)>S 0 (S 0 is the set similarity threshold), combined with the current building environment information, the algorithm quickly determines that the possibility of fire is extremely high, and according to the pre-set risk level classification rules, based on S(t), Cf(t), ΔT(t), G(t) and other parameters, the linear regression model or decision tree model is used to locate the fire risk level R(t). Taking linear regression as an example, its calculation formula is:

[0060] R(t)=a×S(t)+b×Cf(t)+c×ΔT(t)+d×G(t)+e;

[0061] Among them, a, b, c, d, and e are the training parameters of the linear regression model, which are obtained through training with a large amount of historical fire data and simulated experimental data.

[0062] Example 3

[0063] This embodiment is an explanation of the embodiment 1, self-learning and optimization, specifically: in daily operation, each sensor data fluctuation, alarm triggering situation and whether it is finally confirmed as a fire event are continuously recorded as training samples, and a multi-layer perceptron neural network model in deep learning is used. The input layer receives sensor data features, the hidden layer abstracts and converts features, and the output layer predicts the probability of fire occurrence. Through the back propagation algorithm, based on the error between the predicted value and the actual confirmed fire event, the gradient descent method is used to continuously adjust the weights of each layer of the neural network, optimize the model parameters, and continuously reduce the false alarm rate.

[0064] Specifically, in daily operation, the algorithm continuously records every sensor data fluctuation, alarm triggering situation, and whether it is finally confirmed as a fire event. Assuming that the confirmed fire event is marked as F(t) (1 if a fire occurs, 0 if no fire occurs), each time Cf(t), ΔT(t), G(t), F(t) and other data are stored as training samples.

[0065] Using machine learning algorithms, such as the neural network model in deep learning, taking the multi-layer perceptron (MLP) as an example, the input layer receives sensor data features such as Cf(t), ΔT(t), G(t), and the hidden layer performs feature abstraction and conversion. Assuming that the output of the hidden layer neuron is hj(t) (j represents the number of the hidden layer neuron), the weight between the hidden layer and the input layer is wij (i represents the number of the input layer neuron), and the activation function of the hidden layer neuron is f (such as the commonly used sigmoid function), the hidden layer neuron output calculation formula is:

[0066]

[0067] Among them, xi(t) represents the sensor data feature of the input layer at time t (such as Cf(t) corresponding to x 1 (t)).

[0068] The output layer predicts the probability of fire occurrence P(t). Assuming the weight between the output layer and the hidden layer is wjk, the calculation formula of P(t) is:

[0069]

[0070] Through the back propagation algorithm, based on the error between the predicted value P(t) and the actual fire event mark F(t), the gradient descent method is used to continuously adjust the weights wij and wjk of each layer of the neural network. Taking the update of wij as an example, the calculation formula for the updated weight is:

[0071]

[0072] Among them, η is the learning rate, Loss is the loss function, such as the cross entropy loss function, is the partial derivative of the loss function with respect to the weight wij; this method is used to continuously optimize the model parameters. For example, if false alarms of smoke sensors are frequently caused by decoration dust in a certain area, the algorithm automatically learns the smoke data characteristics in the dust environment and dynamically adjusts the smoke judgment threshold and logic in the area. In the event of similar situations in the future, it can accurately distinguish between smoke and dust interference, continuously reduce the false alarm rate, improve the accuracy of fire alarm judgment, and adapt to the complex and changing environment of the building.

[0073] In summary, the intelligent algorithm of the central intelligent processing unit achieves accurate and intelligent monitoring of building fire alarms through sophisticated data reception and processing, dynamic threshold setting, multi-sensor fusion analysis, and continuous self-learning optimization.

[0074] Example 4

[0075] The fire alarm monitoring system of the present application also includes:

[0076] Firefighting facility management module, used for detection and management of firefighting equipment;

[0077] The fault display module is used to display the fault information detected by the communication monitoring module, so that the administrator can perform maintenance on the faulty equipment or module as soon as possible, thereby ensuring the stable and reliable operation of the system.

[0078] Fire protection facility management module, including:

[0079] Fire extinguisher management unit, which is used to monitor and manage the validity period, location, and real-time pressure of fire extinguishers, so that personnel can use fire extinguishers to put out fires in a timely and effective manner when fires occur;

[0080] Fire hydrant management unit, used for monitoring and managing the water flow, water pressure and location of fire hydrants, so as to facilitate early handling when the water flow and water pressure are found to be below the threshold and to find available fire hydrants as soon as possible when a fire occurs;

[0081] The fire detector management unit is used to manage various fire detection sensors, including the type, location and status of the detectors.

[0082] The basic principles of the present application are described above in conjunction with specific embodiments. It should be understood that the specific details disclosed above are only for illustration and ease of understanding, and are not restrictive, and are not used to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A fire alarm monitoring system for a smart building platform, characterized in that: The system comprises: The composite sensor network module is composed of a variety of high-precision sensors, which are used to detect the smoke density, temperature, electrical parameters, and combustible gas concentration inside the building and collect and analyze video images to detect fires in a timely manner; The central intelligent processing unit is equipped with a high-performance processor, which runs a fire alarm judgment algorithm based on big data and artificial intelligence to analyze and process the data transmitted by various sensors to determine the probability of fire and the level of danger, and has a built-in self-learning module; Efficient communication module, supporting multiple communication protocols, used to realize data and command interaction with other subsystems of the building platform based on the data processed by the central intelligent processing unit; The communication monitoring module is used to monitor the communication status between various sensors, between each sensor and the central intelligent processing unit, and between the high-efficiency communication module and other subsystems of the building platform to ensure the stability and timeliness of data transmission; A multi-warning module is used to activate the multi-warning mechanism when the fire risk increases; Remote monitoring and management terminal, used for property and fire departments to remotely access through the Internet to view data such as fire alarm monitoring dynamics and equipment operating status in real time in the building.

2. A fire alarm monitoring system for a smart building platform as claimed in claim 1, characterized in that: The multiple high-precision sensors include: intelligent optical smoke sensors, distributed optical fiber temperature detectors, combustible gas detectors, image-type fire detectors, and electrical fire detectors. The sensors in the composite sensor network module are reasonably installed according to the functional layout of the building and the results of fire risk assessment. The multiple high-precision sensors are connected to the central intelligent processing unit through wired or wireless hybrid networking.

3. The fire alarm monitoring system for a smart building platform as claimed in claim 1, characterized in that: The multiple high-precision sensors send collected data to the central intelligent processing unit at a first frequency, and the central intelligent processing unit processes and analyzes the received various sensor data at a second frequency, and the first frequency is greater than the second frequency.

4. A fire alarm monitoring system for a smart building platform as claimed in claim 1 or 3, characterized in that: The central intelligent processing unit processes and analyzes the various sensor data received, including: Data preprocessing, real-time analysis and fusion judgment, self-learning and optimization; The data preprocessing is to use a filtering algorithm to process the single sensor data to remove noise interference.

5. A fire alarm monitoring system for a smart building platform as claimed in claim 4, characterized in that: The real-time analysis and fusion judgment include: single sensor dynamic threshold setting, multi-sensor data fusion judgment; The single sensor dynamic threshold is set by dynamically calculating the threshold based on the building area function weight coefficient, the environmental humidity influence coefficient, and the historical data fluctuation adjustment coefficient; The multi-sensor data fusion judgment is to use a pre-constructed fire feature database, use a similarity matching algorithm to calculate the similarity between the current sensor data and the typical pattern, and then obtain the comprehensive similarity through weighted average calculation, and then judge whether the comprehensive similarity is greater than a pre-set similarity threshold. If the judgment result is yes, the possibility of fire is very high, and then according to the pre-set risk level classification rules, use a linear regression model or a decision tree model to locate the fire risk level.

6. A fire alarm monitoring system for a smart building platform as claimed in claim 4, characterized in that: The self-learning and optimization are specifically as follows: in daily operation, each sensor data fluctuation, alarm triggering situation, and whether it is finally confirmed as a fire event are continuously recorded as training samples, and a multi-layer perceptron neural network model in deep learning is used. The input layer receives sensor data features, the hidden layer abstracts and converts features, and the output layer predicts the probability of fire occurrence. Through the back propagation algorithm, based on the error between the predicted value and the actual confirmed fire event, the gradient descent method is used to continuously adjust the weights of each layer of the neural network, optimize the model parameters, and continuously reduce the false alarm rate.

7. The fire alarm monitoring system for a smart building platform as claimed in claim 1, characterized in that: The communication methods adopted by the efficient communication module include 4G, 5G, WiFi, and industrial Ethernet.

8. The fire alarm monitoring system for a smart building platform as claimed in claim 1, characterized in that: The warning modes of the multi-dimensional warning module include sound and light alarms, mobile phone push notifications, and scrolling prompts on display screens inside buildings.

9. The fire alarm monitoring system for a smart building platform as claimed in claim 1, characterized in that: The system further comprises: Firefighting facility management module, used for detection and management of firefighting equipment; The fault display module is used to display the fault information detected by the communication monitoring module so that the administrator can perform maintenance on the faulty equipment or module as soon as possible.

10. A fire alarm monitoring system for a smart building platform as claimed in claim 9, characterized in that: The fire protection facility management module comprises: Fire extinguisher management unit, used for monitoring and management of the validity period, location, and real-time pressure of fire extinguishers; Fire hydrant management unit, used for monitoring and management of water flow, water pressure and location of fire hydrants; The fire detector management unit is used to manage various fire detection sensors, including the type, location and status of the detectors.

Images