A fire alarm method and system based on Internet of Things

By combining physical and chemical, time and position data in the Internet of Things fire alarm system, and using clustering algorithms to judge the fire status, the problem of inaccurate fire warning in the existing technology is solved, and efficient and accurate fire identification and classification is achieved.

CN118736799BActive Publication Date: 2024-11-15SHANDONG AOSHEN INTELLIGENT ENG CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411224519.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-03
Publication Date
2024-11-15
Estimated Expiration
2044-09-03

AI Technical Summary

Technical Problem

The Internet of Things-based fire alarm system fails to optimize algorithms based on the characteristics of fire field data when analyzing fire sensor data using big data technology, resulting in inaccurate or inefficient fire warnings.

Method used

A fire alarm system based on the Internet of Things is designed to collect physical and chemical information through the perception layer, and the data transmission layer transmits and processes data. The management center layer uses physical and chemical, time and position data to judge the fire state, and combines the clustering algorithm to perform fire state. Classify to improve judgment accuracy and efficiency.

Benefits of technology

Accurate and efficient judgment of the fire status is achieved, misjudgment is reduced, the accuracy and efficiency of fire warning is improved, and the consumption and cost of computing resources is reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118736799B_ABST
    Figure CN118736799B_ABST
Patent Text Reader

Abstract

The present application discloses a fire alarm method and system based on the Internet of Things, which belongs to the technical field of fire signal alarm. The perception layer is used to collect physical and chemical information in the room; the data transmission layer is used to receive the physical and chemical information from the perception layer, and package the physical and chemical information and the room ID for transmission; the management center layer is used to receive the physical and chemical, time, and location data from the data transmission layer to determine the fire state of the room; the application layer is used to respond to the control instructions and / or fire information of the management center layer; wherein the time data is used to determine the self-extinguishing fire state, and the location data is used to determine the continuous spreading fire state. The method and system in the present application combine the physical and chemical, time and location characteristics of the fire state, and can accurately, timely and efficiently detect fires through clustering algorithms.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of fire signal alarm, and in particular to a fire alarm system and method based on the Internet of Things. Background Art

[0002] Driven by information and communication technology (ICT), the Internet of Things (IoT) and artificial intelligence are being used more and more widely in the construction of smart cities. As an indispensable and important part of the smart city system, the smart fire protection system undertakes multiple responsibilities such as disaster perception, intelligent disaster prevention and emergency rescue.

[0003] Smart fire protection is based on the fire alarm system and relies on technologies such as fire big data, cloud computing and the Internet of Things. It is a modern fire safety management system that integrates fire prevention and firefighting. The construction of the fire Internet of Things comprehensively utilizes technologies such as wireless sensor networks (WSN), LoRa, and NB-IoT to provide real-time monitoring of building fire protection facilities. Sensors are installed at key parts of fire protection facilities to achieve intelligent monitoring of the working status of fire protection facilities and equipment; fire warning monitoring deploys sensors that monitor fire parameters in the environment, such as temperature, humidity, smoke, gas, and infrared sensors, to sense and collect environmental data in real time. These sensors are connected to the Internet of Things through appropriate access technologies, and are connected to the fire cloud platform through wireless or wired transmission methods. The cloud platform is equipped with a database to store data, which is convenient for subsequent use of artificial intelligence or machine learning technologies to analyze and learn data, warn of abnormal data, and identify fires in time in the early stages of a fire, thereby buying time for subsequent fire extinguishing work. Compared with the traditional fire protection model, the smart fire protection system saves manpower, makes comprehensive judgments based on data, effectively reduces the occurrence of fires, and realizes "automation", "intelligence" and "systematization" of prevention and control.

[0004] In related technologies, the fire alarm system based on the Internet of Things, when applying big data technology to analyze the data collected by fire sensors, fails to deeply optimize the algorithm based on the characteristics of the fire scene data, which may lead to inaccurate fire warnings or inefficiencies. Summary of the invention

[0005] In order to solve the above technical problems, the present application provides an efficient and accurate fire alarm system and method based on the Internet of Things.

[0006] The technical solution of the present application to solve the above problems is: a fire alarm system based on the Internet of Things, comprising:

[0007] The perception layer is used to collect the physical and chemical information in the room; the data transmission layer is connected to the perception layer, used to receive the physical and chemical information from the perception layer, and package the physical and chemical information and the room ID for transmission; the management center layer is connected to the data transmission layer, used to receive the data from the data transmission layer, and process the data of the data transmission layer into a multi-element array containing physical and chemical data, time data, location data and room ID, and judge the fire state of the room according to the multi-element array, wherein the fire state of the room includes no fire, self-extinguishing fire, continuous limited fire and continuous spreading fire; the application layer is connected to the data transmission layer, used to respond to the control instructions and / or fire information of the management center layer; wherein the time data is used to judge the self-extinguishing fire state, and the location data is used to judge the continuous spreading fire state. The fire alarm system in the present application makes comprehensive use of physical and chemical data, time data, location data and room ID according to the characteristics of different fire states, so it can judge the specific state of the fire more accurately and efficiently.

[0008] In some embodiments of the present application, the physical and chemical data include but are not limited to temperature, and / or humidity, and / or infrared intensity. According to the specific conditions of the room, the composition of the physical and chemical data can be selectively selected, so it is more flexible and efficient.

[0009] In some embodiments of the present application, the time data used to determine the self-extinguishing fire belongs to the same room, and the location data belongs to multiple adjacent rooms. The duration of the fire can be determined by the time data of the fire event in the same room, and if the fire occurs in multiple adjacent rooms, it can be determined whether the fire spreads, so the fire status judgment of the present application is more accurate.

[0010] In some embodiments of the present application, the management center layer includes a fire situation database and multiple cluster centers, and the cluster centers include: a no-fire cluster center, a fire cluster center, a self-extinguishing fire center, a continuous fire center, a continuous localized fire center, and a continuous spreading fire center; the multiple cluster centers can be obtained by clustering data in the fire situation database. The square distance from the cluster center to the real-time data can be simply calculated through the cluster center, which is more efficient and timely.

[0011] In some embodiments of the present application, the management center layer periodically receives data from the data transmission layer and determines the fire status. Since the fire status is determined by the square distance, the computing power resources used are relatively small, so a smaller detection cycle can be appropriately set as needed to detect the fire in time.

[0012] In some embodiments of the present application, the cluster center periodically updates the cluster center, and the update period of the cluster center is greater than the period of the management center layer receiving the data of the data transmission layer. The update period of the cluster center is longer, so more computing power of the management center can be saved, which is more efficient and reduces costs.

[0013] The present application also provides a fire alarm method based on the Internet of Things, which includes:

[0014] In the first step, the management center layer uses a clustering algorithm to determine whether the room is in a fire state or a non-fire state based on the physical and chemical data of the room; in the second step, if the room is in a fire state, the management center layer uses a clustering algorithm to determine whether the room is in a self-extinguishing fire state or a continuous fire state based on the physical and chemical data and time data of the room; in the third step, if the room is in a continuous fire state, the management center layer uses a clustering algorithm to determine whether the room is in a continuous localized fire state or a continuous spreading fire state based on the physical and chemical data and location data of the room.

[0015] In some embodiments of the present application, in the second step, the fire state is first classified preliminarily through physical and chemical data, and then the self-extinguishing fire state and the continuous fire state in the room are confirmed through time data. The duration of the fire can be determined through the time data of the fire event in the same room, and if multiple adjacent rooms have fires, it can be determined whether the fire has spread. Therefore, the fire state judgment of the present application is more accurate. Preliminary judgment is made first and then confirmed through time data, which is more accurate to avoid misjudgment.

[0016] In some embodiments of the present application, in the third step, the fire state is first classified preliminarily through physical and chemical data, and then the position data is used to confirm whether the room is in a state of continuous localized fire or a state of continuous spreading fire. The duration of the fire can be determined through the time data of the fire event in the same room, and if a fire occurs in multiple adjacent rooms, it can be determined whether the fire spreads. Therefore, the fire state judgment of the present application is more accurate. Preliminary judgment is made first, and then confirmed through physical data, which is more accurate and avoids misjudgment.

[0017] In some embodiments of the present application, the clustering algorithm includes a binary k-means algorithm. The fire alarm device of the present application performs accurate fire classification based on a comprehensive analysis of fire scene fire data, and based on the characteristics of fire conditions in different states, combines clustering algorithms, time, and location data to perform accurate, timely, and efficient fire identification and classification;

[0018] At the same time, this application comprehensively uses sensor acquisition modules to collect time, location, and room ID information, and adopts IoT RFID technology with stable transmission and extremely high network reliability for wireless self-organizing networking, thereby improving the reliability of signal transmission and the utilization efficiency of data. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a schematic diagram of a fire alarm system based on the Internet of Things in some embodiments of the present application;

[0020] Figure 2 This is a flow chart of a fire alarm method based on the Internet of Things in some embodiments of the present application;

[0021] Figure 3 A flow chart of a method for determining whether a fire has occurred in some embodiments of the present application;

[0022] Figure 4 A flow chart of a method for determining whether a fire is self-extinguishing or persistent in some embodiments of the present application;

[0023] Figure 5 This is a flow chart of a method for determining whether a fire is continuing to be localized or spreading in some embodiments of the present application. DETAILED DESCRIPTION

[0024] The present application is further described below in conjunction with the accompanying drawings and embodiments.

[0025] Figure 1 This is a schematic diagram of an Internet of Things-based fire alarm system in some embodiments of the present application.

[0026] like Figure 1 As shown, a fire alarm system based on the Internet of Things includes a perception layer, a data transmission layer, a management center layer, and an application layer.

[0027] The perception layer is used to obtain physical and chemical data related to the fire in the room. The perception layer equipment includes one or more sensor acquisition modules installed in each room for collecting room status data. The perception layer is mainly equivalent to the human body's sensory organs, and is used to collect physical and chemical information (referred to as physical and chemical information) that reflects the fire information or fire status in the room. The above physical and chemical information includes temperature, humidity, smoke, carbon dioxide concentration, current, voltage, infrared, etc.

[0028] The physical implementation of the perception layer is mainly achieved through fire sensors and smart meters based on the Internet of Things. However, it should be noted that the above-mentioned devices do not necessarily belong only to the perception layer. For example, the fire sensor may also include a communication module, which is used to transmit the fire signal or information collected by the fire sensor. In this case, the communication module in the sensor belongs to the data transmission layer. At the same time, some types of fire sensors may also have an alarm module. When a fire occurs, the alarm module can issue a fire alarm or a voice alarm. At this time, the alarm module belongs to the application layer. In some cases, the fire sensor based on the Internet of Things may also include an intelligent recognition system that can directly determine whether a fire has occurred based on local data and directly alarm. However, the intelligent recognition system does not belong to the management center layer. The management center layer in this application can communicate with terminals such as fire sensors and smart meters based on the Internet of Things through wired or wireless communication methods, but the management center is not located in terminals such as fire sensors or smart meters, because the management center needs to rely on hardware facilities with certain computing power.

[0029] In general, the physical and chemical information collected by the perception layer is mainly discrete data, such as temperature and humidity data collected periodically. However, it may also contain continuous time series information, such as current and voltage information. It should be further explained that the range of information collected by sensors of different types of places may be different. For example, in some spaces, only temperature information needs to be collected, while in another space scenario, temperature, humidity, and infrared information may need to be collected. However, whether it is discrete data or continuous data, the data finally processed by the management center is processed discrete data. For continuous data, it can usually be processed into discrete data through periodic sampling and processed into discrete data through normalization, which is convenient for the intelligent computing system of the management center to identify and classify. It should be further explained that for continuous data such as current and voltage, when a fire accident occurs, its instantaneous value may contain periodic components and non-periodic components, but when discretizing, its effective value (the sum of the effective value of the periodic component and the effective value of the non-periodic component) is taken.

[0030] It should be noted that the term "room" in this application is not limited to residential places. Rooms for work, entertainment, study, etc. are all within the scope of the rooms in this application. For example, rooms in substations and shopping malls are also referred to as rooms in this application.

[0031] The sensor acquisition module periodically collects the status data of the room and uploads the collected data to the communication terminal of the data transmission layer in real time.

[0032] In some embodiments, the sensor acquisition module includes one or more sensors installed in one place, or multiple sensors installed in multiple places in the room. For example, a fire smoke sensor attached to the inside of the room for real-time monitoring of the temperature, humidity and smoke concentration information in the room; a temperature sensor attached to the outer wall of the room for real-time monitoring of the room temperature information; a water immersion sensor attached to the inside of the room for real-time monitoring of the water level information in the room; a combustible gas sensor attached to the inside of the room for real-time monitoring of the combustible gas concentration information in the room; the fire smoke sensor, temperature sensor, water immersion sensor, and combustible gas sensor communicate wirelessly with the communication terminal to send the collected signals to the communication terminal. Each sensor has a unique ID value. In some embodiments, some digits of the ID value of the sensors located in the same room are related to or the same as the room ID, so that they can be identified as coming from the same room by the data center.

[0033] In some embodiments, the communication terminal is a communication device independent of the sensor, which is used to collect and aggregate the data collected from each sensor or each acquisition module of each sensor by wired or wireless means, and merge the data into a data packet, and add a room ID to identify the room in the data packet. Therefore, in this way, the independent communication module is less affected by the working environment and working position of the sensor, and has higher communication reliability.

[0034] In some embodiments, the communication terminal is a communication module set in a certain sensor, which is used to aggregate the fire-related physical and chemical data from the sensor or even from other sensors through wired or wireless means, and merge these data into a data packet, and add a room ID to identify the room in the data packet. In this way, if there are multiple sensors in the same room, and multiple sensors have communication modules, only one of the multiple communication modules will be used to communicate with the management center, and the other communication modules will be used to communicate with the communication module; at the same time, the other communication modules are used as backup modules for the communication modules that communicate with the management center, so that the communication reliability between the management center and the room is stronger.

[0035] The signals transmitted to the communication terminal by the fire smoke sensor, temperature sensor, water immersion sensor and combustible gas sensor in the fire sensor acquisition module also include installation address signals corresponding to the IDs of the sensors.

[0036] The fire smoke sensors, temperature sensors, water immersion sensors, and combustible gas sensors in the room first communicate wirelessly with the communication terminal in the same room. If the communication terminal in the same room communicates abnormally, it searches for other communication terminals within the communication range and transmits the data to the background monitoring host through other communication terminals.

[0037] The sensor acquisition module and the room current acquisition module 2 installed inside the room can jointly collect the operating status information in the room, and connect to the wireless ad hoc network communication terminal through RFID (radio frequency identification technology). After receiving the above-mentioned collected data, the wireless ad hoc network communication terminal summarizes it and periodically transmits it to the data center of the main station system through wireless ad hoc network technology.

[0038] The data transmission layer includes communication terminals 3 installed in each room, and the management center layer includes a data center. Each communication terminal 3 wirelessly communicates with the sensor acquisition module and the room current acquisition module 2 located in the same room, receives the data collected by the sensor acquisition module and the room current acquisition module 2, and wirelessly transmits the collected data to the data center after summarizing the collected data. The wireless ad hoc network communication terminal supports the ad hoc network protocol and USB interface, and its own design meets the requirements of low power consumption. It is powered by a built-in battery, which can be automatically replaced when the power is insufficient.

[0039] The data center is located in the centralized control center, receives data from the wireless ad hoc network communication terminal in real time, uses various data processing and analysis methods to process and analyze the received data in real time, displays the detailed installation address of the relevant alarm information equipment, and edits the corresponding text message (with detailed address) according to the type of alarm information and sends it to the mobile terminal of the relevant staff, and sends the information to the intelligent alarm unit at the room site. Here, various data processing and analysis methods can be implemented using existing technologies. The data center is equipped with a memory for storing various monitoring and early warning data.

[0040] The data center also has a statistical query module and a real-time operating information query module. The statistical query module provides historical information of various variables, including fire smoke conditions, combustible gas concentration, room temperature, room current, displacement abnormalities, etc.; the operating information query provides real-time data of various variables.

[0041] The data center has the functions of trend analysis and early warning, alarm record analysis, historical data storage, and data report generation. It can also provide an intuitive and concise data display and system control interface for operation and management.

[0042] Wireless ad hoc network technology specifically refers to that the wireless ad hoc network communication terminal communicates with the data center through the RFID wireless communication module, wherein the wireless ad hoc network communication terminal refers to a communication terminal closest to the sensor acquisition module and the room current acquisition module 2, that is, data is uploaded to the data center via the communication terminal closest to the sensor acquisition module. If the communication terminal closest to the sensor acquisition module has communication abnormalities (such as insufficient power in the communication terminal and cannot work normally), other communication terminals within the communication range (for example, a visible distance of 1000 meters) are searched, and a new transmission path is established again through the RFID wireless communication module for data transmission.

[0043] The RFID wireless communication module adopts a communication main control board card, which can be embedded or other products. The RFID wireless communication module belongs to the Internet of Things communication and has a self-organizing network function.

[0044] The application layer includes a mobile terminal 5 worn by the staff and an intelligent alarm unit 6 located in the room. The data center outputs corresponding control signals based on the data received from each communication terminal 3 and transmits them to the mobile terminal 5 and the intelligent alarm unit 6 via wireless. The mobile terminal 5 and the intelligent alarm unit 6 perform corresponding operations based on the received control signals.

[0045] The mobile terminal 5 has 4G / 5G or wireless communication function, and can receive various detailed alarm text messages sent by the data center of the master station system, realize remote scheduling operation management and intelligent operation inspection functions, and facilitate multi-terminal remote interaction and on-site tracking.

[0046] The intelligent alarm unit 6 has a self-organizing network protocol, and can receive alarm information sent by the data center of the master station system through a wireless self-organizing network communication terminal, and perform sound and light alarms in the room.

[0047] In some embodiments of the present application, the physical and chemical data include but are not limited to temperature, and / or humidity, and / or infrared intensity. According to the specific conditions of the room, the composition of the physical and chemical data can be selectively selected, so it is more flexible and efficient.

[0048] In some embodiments of the present application, the time data used to determine the self-extinguishing fire belongs to the same room, and the location data belongs to multiple adjacent rooms. The duration of the fire can be determined by the time data of the fire event in the same room, and if the fire occurs in multiple adjacent rooms, it can be determined whether the fire spreads, so the fire status judgment of the present application is more accurate.

[0049] In some embodiments of the present application, the management center layer includes a fire situation database and multiple cluster centers, and the cluster centers include: a no-fire cluster center, a fire cluster center, a self-extinguishing fire center, a continuous fire center, a continuous localized fire center, and a continuous spreading fire center; the multiple cluster centers can be obtained by clustering data in the fire situation database. The square distance from the cluster center to the real-time data can be simply calculated through the cluster center, which is more efficient and timely.

[0050] In some embodiments of the present application, the management center layer periodically receives data from the data transmission layer and determines the fire status. Since the fire status is determined by the square distance, the computing power resources used are relatively small, so a smaller detection cycle can be appropriately set as needed to detect the fire in time.

[0051] In some embodiments of the present application, the cluster center periodically updates the cluster center, and the update period of the cluster center is greater than the period of the management center layer receiving the data of the data transmission layer. The update period of the cluster center is longer, so more computing power of the management center can be saved, which is more efficient and reduces costs.

[0052] The data transmission layer device periodically sends the data collected by all the perception layer devices in the same room to the data center. The data in all rooms will be aggregated to the data center. There can be multiple data centers or one data center. Some or all of the data can be shared between data centers. The specific network structure can be set according to the required scenario, and no specific limitation is made here.

[0053] The above data usually includes the physical and chemical data of the fire in the same room, such as temperature, humidity, infrared intensity, carbon dioxide concentration, etc. In addition to the above physical and chemical data, it usually also includes an ID data representing the room where the sensor is located, and the ID data of each room is different. The data transmission layer usually packages the data detected by all sensors in a room and sends it to the data center. In the process of packaging, the data transmission layer not only receives the physical and chemical information in each room, but also adds the room ID information to the data packet. After receiving the data packet transmitted by the data transmission layer, the data center will normalize the data packet. The normalization of the data packet mainly refers to the processing of the relevant physical and chemical signal information into unified and discrete information, and the offline information still exists in a relatively independent manner. In addition, the data center will query the location information of the data packet according to the room ID, that is, the information of the address of the room. In addition, the data center will also add time information to the data packet according to the time when the data packet is received. The time information is generally based on the information when the data center receives the data. In some embodiments, the time information can also be added by the data transmission layer, and the time information can be the time when the data transmission layer sends the data packet.

[0054] In short, after a series of data processing processes, the data in the data packet contains the following information: physicochemical information, location information, time information and ID. It should be noted that the physicochemical information may only contain physical information, or only contain chemical information, or may contain physicochemical information at the same time. The physicochemical information may exist in the form of a multivariate array of fixed size, the elements of which are a single physical or chemical information, such as temperature information as an element, infrared intensity information as an element, etc. For example, a multivariate array of physicochemical information in a room (x1, x2, x3...xn), where n is a positive integer, xn is the nth element of the array, representing a certain physicochemical information in the room, for example, x1 represents temperature, x2 represents humidity, etc. In some embodiments, the elements of the multivariate array have a uniform numerical range, such as decimals between [0-1], which facilitates unified calculation and processing. At the same time, the location information, time information and ID information are also combined with the multi-element array containing physical and chemical information in the form of elements to form an extended array (x1, x2, x3...xn, t, p, ID), where t represents time, p represents location, and ID is room ID (number).

[0055] Figure 2 This is a flow chart of a fire alarm method based on the Internet of Things in some embodiments of the present application; Figure 2 As shown, a fire alarm method based on the Internet of Things includes:

[0056] Step 1: Based on the physical and chemical information, determine whether the room is in a fire state or a non-fire state.

[0057] Figure 3 Flowchart of a method for determining whether a fire has occurred in some embodiments of the present application; Figure 3 As shown in the figure, in this step, the physical and chemical information of all rooms is taken as one point, and the time and location information are ignored, and the classification is performed based on the binary k-means algorithm. In most cases, the room is in a state of no fire, and only in a few cases will the room have a fire. Most of the points are points in a state of no fire, and a few points are points in a state of fire. When a fire occurs, the physical and chemical data collected by the fire sensor in the room will be obviously abnormal. Therefore, the state of fire and the state of no fire can be easily distinguished based on the physical and chemical information of the room alone.

[0058] The clustering algorithm in the technical solution of the present application is a k-means clustering algorithm, in particular a binary k-means clustering algorithm. The main steps of the binary k-means algorithm include: The main steps of the binary K-means algorithm are as follows:

[0059] Data preparation, collecting the data that needs to be classified and preprocessing the data, such as standardization or normalization;

[0060] Initialize the cluster centers and randomly select two data points as the initial two cluster centers;

[0061] Assign data points, calculate the distance between each data point and the two cluster centers, and assign the data point to the category to which the cluster center with the closer distance belongs;

[0062] Update the cluster center, calculate the mean of all data points in each cluster, and use it as the new cluster center;

[0063] Repeat the above two steps until the cluster center no longer changes or the preset number of iterations is reached.

[0064] Specifically, in the technical solution of this application, the conventional binary classification k-means algorithm is further improved. At the beginning, a data point is artificially set as the cluster center of the fire, and the cluster center of the non-fire can be artificially set or randomly selected. This solution is only applicable to the situation where no data is collected at the beginning, no virtual data is artificially generated, and no real data from other rooms is imported, so it is not commonly used.

[0065] In some other implementations, the data center may first import data collected by other means, or artificially generated virtual data as a starting point.

[0066] When the data center is running, there are first two cluster centers, one is the no-fire cluster center, and the other is the fire cluster center. At this time, the data center continuously receives fire data from the fire sensor. For the real-time collected fire data, the data center first calculates the "distance" from the fire data at a certain moment to the two cluster centers, and determines whether a fire has occurred based on the distance to the cluster center: if the real-time fire data is closer to the no-fire cluster center, then there is no fire, otherwise there is a fire.

[0067] Here, as a special case, if the distance between the real-time data and the two centers is exactly the same, it can be considered as whether there is a fire or not, or the data can be simply abandoned and the next real-time data can be used for judgment.

[0068] In order to improve the accuracy of judgment, multiple continuous real-time data can be used for comparison and a comprehensive judgment can be made as to whether a fire has occurred.

[0069] In this step, the data center not only determines whether a fire has occurred based on the real-time fire data, but also records the real-time fire data in real time. The recorded real-time data will be updated to the fire database together with the historical data. The data center will regularly cluster the data recorded in the fire database and update the cluster center, which means that the cluster center without fire and the cluster center with fire are not two fixed points.

[0070] It should be noted that the updated cluster center at this time is to update the point of the cluster center used to judge the fire situation, and it is not the operation of updating the cluster center during the operation of the binary k-means clustering algorithm. In other words, the updated cluster center at this time is the result of the operation used in the clustering algorithm.

[0071] It should be noted that the fire data used by the clustering algorithm in this step are all physical and chemical data, and do not involve clustering calculations of time and location information.

[0072] Since it is only necessary to calculate the distance between the real-time fire data and the two cluster centers in real time and judge the fire situation according to their size relationship, the method operates promptly and quickly. Since the real-time fire data will be added to the fire database and used to regularly update the cluster centers, the cluster centers of the present application can be "automatically" optimized adaptively according to the operation situation, so the selection of cluster centers is more accurate and the judgment of the fire situation is more accurate.

[0073] Step 2: If there is a fire, determine whether the room is in a self-extinguishing fire state or a continuous fire state based on physical, chemical and time information.

[0074] Figure 4 This is a flow chart of a method for determining a self-extinguishing fire or a persistent fire in some embodiments of the present application; Figure 4 As shown, if the fire is in a self-extinguishing state, the data center will send relevant information to the application layer, but will not trigger an alarm to the application layer;

[0075] If the fire is ongoing, the data center will also send relevant information to the application and further initiate a judgment on the fire status.

[0076] A self-extinguishing fire is a temporary fire that occurs due to accidental reasons, such as a kitchen pot with too high a temperature, which causes a short-term combustion, or an accidental short circuit that causes a fire. The characteristics of a self-extinguishing fire are that it lasts for a short time and is confined to one place (usually a part of a room). Even if people do not intervene, the fire will go out automatically.

[0077] Specifically, in this step, similar to the first step, there are two cluster centers in the data center - the self-extinguishing fire center and the continuous fire center. The self-extinguishing fire center and the continuous fire center can be initially obtained through artificial settings or clustering methods. The data center receives the fire data from the room perception layer in real time. The data center calculates and compares the distance from the real-time fire data to the self-extinguishing fire center and the continuous fire center, and uses this to preliminarily judge the fire situation. It should be noted that the fire data used at this time is still the physical and chemical data from the sensor. It can be understood that this is only a preliminary judgment of the fire situation, and the data center will not judge the occurrence of self-extinguishing fires and continuous fires based solely on the data at a certain moment.

[0078] Next, the data center will introduce time data to further determine whether a self-extinguishing fire or a continuous fire has occurred. The data center is set with a time threshold, such as 5s, 10s or 1min. Within the time range that does not exceed the time threshold, if the fire data at most moments are close to the center of the self-extinguishing fire, it is determined that a self-extinguishing fire has occurred, otherwise it is a continuous fire. It should be noted that in this step, the fire data received by the data center at a certain moment must be the fire data determined in the first step to have caused a fire. That is to say, if the fire data at a certain time is determined to have not caused a fire, there is no need to perform the judgment in this step.

[0079] Similar to the first step, points that are judged to be self-extinguishing fires or persistent fires will also be added to the fire situation database of the data center. Through clustering, the positions of the self-extinguishing fire centers and the persistent fire centers will be recalculated and updated periodically or conditionally.

[0080] Step 3: If the fire is continuous, determine whether the room is in a localized fire state or a spreading fire state based on the physical, chemical and location information.

[0081] Figure 5 This is a flow chart of a method for determining whether a fire is continuously localized or continuously spreading in some embodiments of the present application; Figure 5 As shown, if there is a localized fire, the data center will send an alarm message to the application layer; for example, notify relevant personnel to rush to the fire scene as soon as possible; start the fire extinguishing device and sound an alarm.

[0082] If there is a spreading fire, the data center will send an alarm message to the application layer; for example, it will activate the fire extinguishing device at the application layer, inform relevant personnel to evacuate as soon as possible, and activate the alarm.

[0083] Specifically, in this step, similar to the first and second steps, there are two cluster centers in the data center - the continuous localized fire center and the continuous spreading fire center. The continuous localized fire center and the continuous spreading fire center can be initially obtained through artificial settings or clustering methods. The data center receives the fire data from the room perception layer in real time. The data center calculates and compares the distance from the real-time fire data to the continuous localized fire center and the continuous spreading fire center, and uses this to preliminarily judge the fire situation. It should be noted that the fire data used at this time is still the physical and chemical data from the sensor. It can be understood that this is only a preliminary judgment of the fire situation, and the data center will not judge the occurrence of continuous localized fire and continuous self-extinguishing fire based solely on the data at a certain moment.

[0084] Next, the data center will introduce location data to further determine whether a continuous localized fire or a continuous spreading fire has occurred. When multiple adjacent rooms are initially determined to have a continuous spreading fire at a certain moment, or for several consecutive moments, or within a short period of time (such as 2 seconds), it can be determined that a continuous spreading fire has occurred in these rooms.

[0085] Similar to the first and second steps, points that are determined to be the site of a persistent localized fire or a persistent spreading fire will also be added to the fire situation database of the data center. Through clustering, the locations of the centers of the persistent localized fire and the centers of the persistent spreading fire will be recalculated and updated periodically or conditionally.

[0086] It can be understood that the fire data in the second step is the fire data of one location (such as a room) within a period of time; and the fire data in the third step is the fire data of multiple locations at a certain time (or a short period of time, or several consecutive moments).

[0087] The fire alarm device of the present application performs accurate fire classification based on a comprehensive analysis of fire scene fire condition data, and performs accurate, timely and efficient fire identification and classification based on the characteristics of fire conditions in different states, combined with clustering algorithms, time and location data.

[0088] In some embodiments of the present application, in the second step, the fire state is first classified preliminarily through physical and chemical data, and then the self-extinguishing fire state and the continuous fire state in the room are confirmed through time data. The duration of the fire can be determined through the time data of the fire event in the same room, and if multiple adjacent rooms have fires, it can be determined whether the fire has spread. Therefore, the fire state judgment of the present application is more accurate. Preliminary judgment is made first and then confirmed through time data, which is more accurate to avoid misjudgment.

[0089] In some embodiments of the present application, in the third step, the fire state is first classified preliminarily through physical and chemical data, and then the position data is used to confirm whether the room is in a state of continuous localized fire or a state of continuous spreading fire. The duration of the fire can be determined through the time data of the fire event in the same room, and if a fire occurs in multiple adjacent rooms, it can be determined whether the fire spreads. Therefore, the fire state judgment of the present application is more accurate. Preliminary judgment is made first, and then confirmed through physical data, which is more accurate and avoids misjudgment.

[0090] In some embodiments of the present application, the clustering algorithm includes a binary k-means algorithm. The fire alarm device of the present application performs accurate fire classification based on a comprehensive analysis of fire scene fire data, and based on the characteristics of fire conditions in different states, combines clustering algorithms, time, and location data to perform accurate, timely, and efficient fire identification and classification.

[0091] At the same time, this application comprehensively uses sensor acquisition modules to collect time, location, and room ID information, and adopts IoT RFID technology with stable transmission and extremely high network reliability for wireless self-organizing networking, thereby improving the reliability of signal transmission and the utilization efficiency of data.

[0092] Specific examples are used herein to illustrate the principles and implementation methods of the present application, and the description of the above embodiments is only used to help understand the method and core ideas of the present application. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of the present application, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the scope of protection of the claims of the present application.

Claims

1. A fire alarm system based on the Internet of Things, comprising: The perception layer is used to collect physical and chemical information in the room; A data transmission layer, connected to the perception layer, for receiving the physical and chemical information from the perception layer, and packaging and sending the physical and chemical information and the room ID; a management center layer connected to the data transmission layer, configured to receive data from the data transmission layer, process the data from the data transmission layer into a multi-element array including physical and chemical data, time data, location data and room ID, and determine the fire status of the room according to the multi-element array, wherein the fire status of the room includes no fire, self-extinguishing fire, continuous localized fire and continuous spreading fire; An application layer, connected to the data transmission layer, for responding to control instructions and / or fire information of the management center layer; The time data is used to determine the self-extinguishing fire state, and the position data is used to determine the continuously spreading fire state.

2. The fire alarm system based on the Internet of Things according to claim 1, characterized in that: The physicochemical data include but are not limited to temperature, and / or humidity, and / or infrared intensity.

3. The fire alarm system based on the Internet of Things according to claim 1, characterized in that: The time data used to determine the self-extinguishing fire belongs to the same room, and the location data belongs to a plurality of adjacent rooms.

4. The fire alarm system based on the Internet of Things according to claim 3 is characterized in that: The management center layer includes a fire situation database and multiple cluster centers, and the cluster centers include: a no-fire cluster center, a fire cluster center, a self-extinguishing fire center, a continuous fire center, a continuous localized fire center, and a continuous spreading fire center; the multiple cluster centers can be obtained by clustering data in the fire situation database.

5. The fire alarm system based on the Internet of Things according to claim 4 is characterized in that: The management center layer periodically receives data from the data transmission layer and determines the fire status.

6. The fire alarm system based on the Internet of Things according to claim 5 is characterized in that: The cluster center is updated periodically, and the update period of the cluster center is greater than the period at which the management center layer receives the data of the data transmission layer.

7. A fire alarm method based on the Internet of Things, characterized in that: The Internet of Things-based fire alarm system applied to any one of claims 1-6, the method comprising: In the first step, the management center layer determines whether the room is in a fire state or not in a fire state according to the physical and chemical data of the room through a clustering algorithm; In the second step, if there is a fire in the room, the management center layer determines whether the room is in a self-extinguishing fire state or a continuous fire state according to the physical and chemical data and time data of the room through a clustering algorithm; In the third step, if the room is in a continuous fire state, the management center layer determines whether the room is in a continuous localized fire state or a continuous spreading fire state based on the physical and chemical data and location data of the room through a clustering algorithm.

8. The fire alarm method based on the Internet of Things according to claim 7 is characterized in that: In the second step, a preliminary fire state classification is first performed using physical and chemical data, and then the self-extinguishing fire state and the continuous fire state in the room are confirmed using time data.

9. The fire alarm method based on the Internet of Things according to claim 7, characterized in that: In the third step, a preliminary fire state classification is first performed using physical and chemical data, and then the position data is used to confirm whether the room is in a continuous localized fire state or a continuous spreading fire state.

10. The fire alarm method based on the Internet of Things according to any one of claims 7 to 9, characterized in that: The clustering algorithm includes a binary classification k-means algorithm.