Multivariable composite fire detection method, system, storage medium and device

Abstract

The invention relates to the field of fire detection, and discloses a multivariable composite fire detection method and system, a storage medium and a device, and the method comprises the steps: obtaining first and second sensing data based on a first sensor and a second sensor, and extracting sensing target data, a sensing position and a timestamp; calculating a first timestamp difference value between the first timestamp and the second timestamp and a first position data difference value between the first sensing position and the second sensing position; calculating a first process matching value of the first timestamp difference value and the first position data difference value; if the first process matching value is smaller than a preset reference matching value, a first target difference value of the first sensing target data and the second sensing target data is calculated, and if the first target difference value is larger than or equal to a preset reference target value, sensor fault early warning is conducted and data are output; if the first process matching value is greater than or equal to a preset reference matching value, calculating first induction fusion data, and if the first induction fusion data is greater than or equal to a preset fusion reference value, outputting a fire alarm prompt, otherwise, outputting a fire early warning prompt. Fire detection accuracy and reliability are improved, and timely early warning is realized.

Description

Technical Field

[0001] The present application relates to the technical field of fire detection, and in particular to a multivariable composite fire detection method, system, storage medium, and device. Background Art

[0002] Fire detectors are the detection components of automatic fire alarm systems. They convert the smoke, heat, and light generated in the early stages of a fire into electrical signals, which are input into the automatic fire alarm system. After processing, an alarm is issued or a corresponding action is taken. Depending on the combustion characteristics of a fire, fire detectors can be divided into smoke, temperature, light, composite, intelligent, and combustible gas fire detectors. Existing fire detector systems use a separate detection, control, and linkage system, and are composed of multiple point-type fire detectors, local fire alarm controllers, and local fire extinguishing systems. Existing fire detectors, depending on the application industry, have derived uses for industrial and civil buildings, industrial and commercial purposes, and electrical fires.

[0003] Existing fire detectors are point-type detectors, including smoke, temperature, infrared, ultraviolet, image, and gas detectors. Each detector has only one sensor type and can detect only one status parameter. This detection only applies to a specific environmental condition, making it unable to identify other potential risks. Deploying additional sensor types requires secondary wiring, creating an aesthetically pleasing design and requiring additional disassembly and installation, impacting construction timelines and efficiency. Furthermore, the detector has a high false alarm rate, requiring external manual fire alarm buttons and fire telephones for communication and confirmation with the fire alarm controller. Summary of the Invention

[0004] In order to enable a fire detector to detect multiple state parameters, the present application provides a multi-variable composite fire detection method, system, storage medium and device.

[0005] In a first aspect, the present application provides a multivariable composite fire detection method, which adopts the following technical solutions: A multivariable composite fire detection method comprises the following steps: acquiring first sensing data based on a first sensor, and extracting first sensing target data, a first sensing position, and a first timestamp from the first sensing data; acquiring second sensing data based on the second sensor, and extracting second sensing target data, a second sensing position, and a second timestamp from the second sensing data; Calculating a first timestamp difference between the first timestamp and the second timestamp; calculating a first position data difference between the first sensing position and the second sensing position; Calculating a first process matching value between the first timestamp difference and the first position data difference; If the first process matching value is less than a preset reference matching value, a first target difference between the first sensing target data and the second sensing target data is calculated; if the first target difference is greater than or equal to a preset reference target value, a sensor failure warning prompt is issued, and the first sensing target data and the second sensing target data are output; If the first process matching value is greater than or equal to the preset reference matching value, the first sensing fusion data of the first sensing target data and the second sensing target data is calculated; if the first sensing fusion data is greater than or equal to the preset fusion reference value, a fire alarm prompt is output, otherwise a fire warning prompt is output.

[0006] By adopting the above technical solution, by acquiring sensor data from different sensors (a first sensor and a second sensor) and extracting information such as the target data, sensing location, and timestamp, a single fire detector can detect multiple status parameters. This overcomes the limitation of existing fire detectors, which typically only have one sensor and can only detect one status parameter. This allows for a more comprehensive perception of multiple potential fire risks, improving the accuracy and reliability of fire detection. The timestamp difference, position data difference, and process matching value of the two sensors are calculated and compared with preset values. When the process matching value is less than a preset reference matching value, a target difference is further calculated between the first and second target data. If this target difference is greater than the preset reference matching value, a sensor failure warning is issued. This enables the system to promptly detect potential sensor failures, avoid false alarms or missed alarms caused by sensor failure, and ensure the normal operation of the fire detection system. When the process matching value is greater than the preset reference matching value, a fusion data of the first and second target data is calculated and compared with a preset fusion reference value. Based on the comparison result, a fire alarm or fire warning is output. This approach comprehensively considers data from different sensors to more accurately determine the presence and severity of a fire. Compared to existing technologies that only detect a single state parameter within a specific environmental condition, it can detect fires more promptly and accurately, enabling appropriate responses and reducing false alarm rates. Because a single fire detector can detect multiple state parameters, there's no need to deploy additional sensors and perform secondary wiring to detect different parameters, as is required with existing technologies. This avoids the unsightly effects of secondary wiring and the associated disassembly and installation issues associated with renovations, minimizing construction timelines and efficiency, and reducing construction costs and complexity.

[0007] Optionally, the method further comprises the following steps: acquiring third sensing data based on a third sensor, and extracting third sensing target data, a third sensing position, and a third timestamp from the third sensing data; Calculating a second timestamp difference between the third timestamp and the second timestamp; calculating a second position data difference between the third sensing position and the second sensing position; Calculating a second process matching value of the second timestamp difference and the second position data difference; If the second process matching value is less than a preset reference matching value, a second target difference between the third sensing target data and the second sensing target data is calculated, and a comprehensive target difference is calculated based on the first target difference and the second target difference. If the comprehensive target difference is greater than or equal to a preset reference target value, a sensor failure warning prompt is issued, and the first sensing target data, the second sensing target data, and the third sensing target data are output; If the second process matching value is greater than or equal to the preset reference matching value, the second sensing fusion data of the third sensing target data and the second sensing target data is calculated; the comprehensive sensing fusion data is calculated based on the first sensing fusion data and the second sensing fusion data. If the comprehensive sensing fusion data is greater than the preset fusion reference value, a fire alarm prompt is output, otherwise a fire warning prompt is output.

[0008] By adopting the above technical solution, introducing a third sensor and performing calculations and analysis, the accuracy and reliability of fire detection are further improved. It can comprehensively perceive potential fire-related risks from more dimensions, providing stronger support for protecting people's lives and property.

[0009] Optionally, the first sensor, the second sensor and the third sensor are all one of a smoke sensor, a temperature sensor, an infrared sensor, an ultraviolet sensor, an image sensor and a gas sensor, and the first sensor, the second sensor and the third sensor are different sensors.

[0010] Optionally, the step of acquiring first sensing data based on the first sensor and extracting first sensing target data, a first sensing position, and a first timestamp from the first sensing data further includes the following sub-steps: The first sensor is an image sensor, and the first sensor data is picture data captured by the image sensor; Extracting open flame, temperature, and smoke targets from the image data according to a preset first recognition algorithm, wherein the first sensing target data is image confidence values ​​of the open flame, temperature, and smoke targets; Extracting the time of the open flame, temperature and smoke targets as a first timestamp; The image position of the open flame, temperature and smoke characteristic targets in the image data is calculated, and a physical position is matched from a preset position database as the first sensing position according to the image position.

[0011] By employing the above technical solution, a preset first recognition algorithm is used to accurately extract open flame, temperature, and smoke targets from image data captured by the image sensor and obtain their image confidence values. This value can assess the severity of open flame, temperature, and smoke, providing intuitive and critical data for fire detection, significantly improving fire detection accuracy. The time of the open flame, temperature, and smoke target's appearance is extracted as the first timestamp, providing a precise reference for subsequent comprehensive analysis of the temporal correlation between different sensor data. This timestamp clearly illustrates the temporal context of the fire's development, helping to fully understand the process of fire occurrence and development, and providing strong temporal support for timely response measures. The location of the open flame, temperature, and smoke targets in the image data is calculated, and then the corresponding physical location is matched based on a preset location database as the first sensing location. This method achieves a precise mapping from image space to actual physical space, accurately locating the fire's location. Accurate location information is crucial in fire rescue, providing strong support for rapid response and effective rescue operations. It greatly improves the reliability and practicality of fire detection systems in practical applications, enhancing their overall effectiveness.

[0012] Optionally, the step of acquiring second sensing data based on the second sensor and extracting second sensing target data, a second sensing position, and a second timestamp from the second sensing data further includes the following sub-steps: The second sensor is a non-image sensor, and the second sensor data is electrical signal data collected by the non-image sensor; extracting open flame, temperature, and smoke characteristics from the electrical signal data according to a preset second recognition algorithm, wherein the second sensing target data is the value of the electrical signal data of the open flame, temperature, and smoke characteristics; extracting the time of the open flame, temperature and smoke characteristics as a second timestamp; A physical position of the non-image sensor is acquired as the second sensing position.

[0013] By employing the above technical solution, a pre-set second recognition algorithm extracts flame, temperature, and smoke characteristics from the electrical signal data collected by the non-image sensor. The values ​​of these electrical signal data are used as the second sensing target data, capturing fire-related information from a dimension distinct from the image. This complements the visual information acquired by the image sensor, providing comprehensive perception of the fire scene, improving the comprehensiveness and accuracy of fire detection, more accurately assessing the fire situation, and reducing potential misjudgments caused by a single sensor. The time of occurrence of the flame, temperature, and smoke characteristics is extracted as a second timestamp, which is cross-referenced with the timestamp of the first sensor, providing precise time synchronization for multi-sensor data fusion analysis. This facilitates more accurate analysis of the temporal correlation between different physical phenomena during a fire, further untangling the temporal context of a fire's development, and providing robust temporal clues for early warning and rapid response. In fixed scenarios, such as building interiors, the installation location of each non-image sensor is predetermined and recorded in a pre-set location database. When determining the second sensing location, simply match the sensor's identification information to the corresponding physical location in the database. This approach provides reliable localization of the fire's origin. Accurately determine the location of fire-related materials or sensors, and combine it with the location information determined by the first sensor to achieve double confirmation of the fire location and more accurate positioning, providing key location information support for fire rescue personnel to quickly arrive at the fire scene and implement effective rescue, thereby improving the efficiency and success rate of fire rescue.

[0014] Optionally, the method further comprises the following steps: With time as the horizontal axis and the first sensing target data as the vertical axis, a first characteristic change curve of the first sensor is constructed; With time as the horizontal axis and the second sensing target data as the vertical axis, construct a second characteristic change curve of the second sensor; identifying, based on the first characteristic change curve and the second characteristic change curve, change trends of the first characteristic change curve and the second characteristic change curve, and calculating consistency of the change trends as a curve matching value; Calculating a reasonable time difference and a reasonable position data difference respectively according to the first timestamp difference and the first position data difference; Reasonable time difference = first position data difference / first sensing target data; Reasonable position data difference = first timestamp difference × first sensing target data; Calculate the difference between the reasonable time difference and the first timestamp difference as a comprehensive time difference; Calculate the difference between the reasonable position data difference and the first position data difference as the comprehensive position difference; Calculating a reasonable time-distance value based on the integrated time difference and the integrated position difference; Reasonable value of time-distance = k / (comprehensive position difference 2 +Comprehensive time difference 2 ), where k is a tuning parameter; According to the curve matching value and the reasonable time-distance value, the first process matching value is calculated: The first process matching value = α × curve matching value + β × (time-distance reasonable value), where α + β = 1.

[0015] By employing the above technical solution, by constructing a first characteristic change curve for the first sensor (image sensor) and a second characteristic change curve for the second sensor (smoke sensor), and calculating the curve matching value, the consistency of the changing trends of the two sensor data can be accurately quantified. This allows for correlation analysis of the fire-related characteristics in data from different sensor types. Compared to single-sensor data analysis, this significantly improves the accuracy and comprehensiveness of fire characteristic identification and reduces the possibility of misjudgment. For example, if the trend of smoke area change in the image is highly consistent with the trend of smoke concentration change detected by the smoke sensor, the curve matching value is high, further confirming the occurrence and development of the fire. Based on the difference between the timestamp and location data, a reasonable time difference and a reasonable location data difference are calculated, which in turn leads to a comprehensive time difference and comprehensive location difference, and a reasonable time-distance value is calculated. This process fully considers the logical relationship between the time and space of smoke diffusion after a fire occurs. In real-world scenarios, the time and distance of smoke spreading from the fire location to the smoke sensor location have a certain reasonable correspondence. By quantifying this relationship, the rationality of sensor detection data in both time and space can be effectively verified. If the actual timestamp difference and location data difference are close to the calculated reasonable values, it indicates that the sensor detection results conform to the normal pattern of fire smoke diffusion, further enhancing the reliability of fire diagnosis. The process matching value is calculated by combining the curve matching value and the reasonable time-distance value, taking into account the consistency of sensor data feature changes and the temporal and spatial rationality. By setting the weights α and β, the importance of each can be flexibly adjusted according to the needs of different application scenarios. For example, in scenarios where timely fire response is crucial, the weight α of the curve matching value can be appropriately increased to prioritize the rapid and consistent judgment of sensor data trends. In contrast, in scenarios where fire location accuracy is crucial, the weight β of the reasonable time-distance value can be increased. This comprehensive and integrated evaluation method provides a more scientific and reliable basis for subsequent accurate judgment of sensor status and fire occurrence, significantly improving the performance and adaptability of the entire fire detection system.

[0016] Optionally, the step of calculating first sensing fusion data of the first sensing target data and the second sensing target data further includes the following sub-steps: Calculating the first sensing target data as a first normalized value using a maximum-minimum normalization algorithm; Calculating the second sensing target data as a second normalized value using a maximum-minimum normalization algorithm; The first sensing fusion data = w1×first normalized value + w2×second normalized value; the weight of the first sensing target data is w1, the weight of the second sensing target data is w2, w1+w2=1.

[0017] By adopting this technical solution, since the first and second sensor target data come from different sensor types and have significantly different dimensions and value ranges, the normalization algorithm maps them uniformly to the [0,1] interval. This prevents dimensional differences from causing certain data to dominate the fusion results, thereby improving the fairness and accuracy of data fusion. Raw sensor target data is susceptible to large fluctuations due to various factors. Normalization smoothes the data and makes it more stable, helping the subsequent fire detection model learn data patterns and characteristics, improving model generalization and reducing the risk of misjudgment. By setting weights, the importance of the two sensor data in the fusion results can be flexibly adjusted according to the actual scenario, enhancing the adaptability and practicality of the fire detection system. Furthermore, the maximum and minimum normalization algorithm is computationally simple, reducing data processing complexity and resource consumption, and improving computational efficiency. The weighted summation fusion calculation is easy to understand and implement, facilitating its application and widespread adoption in practical fire detection systems.

[0018] Optionally, the step of calculating the first process matching value of the first timestamp difference and the first position data difference further includes the following sub-steps: Obtaining a calculation amount of a first process matching value for calculating a difference between the first timestamp and the first position data; If the calculation amount is less than or equal to the set calculation amount threshold, or the difference between the first position data is less than or equal to the set position threshold, the first sensor data and the second sensor data are sent to the background system for calculation; If the calculation amount is greater than the set calculation amount threshold, or the difference of the first position data is greater than the set position threshold, the first sensor data and the second sensor data are sent to the corresponding relay station for calculation, and then sent by the relay station to the background system for calculation; Based on multiple relay stations, the master and slave are defined based on the fire detector where the first sensor is located. The master is the relay station connected to the fire detector where the first sensor is currently located, and the remaining relay stations are slaves. The master and slaves are associated; Obtain host status feedback information at a set frequency; If the status feedback information indicates an abnormal state, a connection is established with the nearest slave.

[0019] By employing the above technical solution, when calculating the process matching value between the timestamp difference and the location data difference, computational tasks are dynamically allocated based on the computational effort and the comparison result between the location data difference and a set threshold. When the computational effort is low or the two sensors are located in the same fire detector, the first and second sensor data are sent to the backend system for computation. When the computational effort is high or the two sensors are located in different fire detectors, the corresponding relay station performs computation first and then sends the data to the backend system. This solution, by rationally allocating computational tasks, effectively improves computational efficiency, reduces data transmission latency, enhances system reliability and fault tolerance, optimizes resource utilization, reduces energy consumption, and improves the system's adaptability and scalability for diverse application scenarios. In the step of sending the first and second sensor data to the corresponding relay station for computation and then transferring it to the backend system for computation, a master (i.e., the relay station connected to the fire detector where the first sensor is located) and slaves (the remaining relay stations) are defined and associated based on multiple relay stations. Master status feedback is obtained at a set frequency. If the master status is abnormal, a connection is established with the nearest slave. This solution improves data processing reliability, ensures the stable operation of the fire detection system, enhances the system's fault tolerance, and can cope with relay station failures in complex environments. It also optimizes resource utilization and efficiency, rationally allocates resources, reduces transmission delays and resource waste, and achieves a balance between resource utilization and system efficiency.

[0020] Optionally, the power supply system of the composite fire detector includes a power supply and a backup battery, and the method further includes the following steps: Obtaining a power input voltage, and if the power input voltage is within an abnormal power range, disconnecting the power supply, switching to a backup battery, and issuing a warning prompt; If the power input voltage is within the normal power range, disconnect the backup battery and switch to the power supply to charge the backup battery; The entire power supply is electrically connected to the anode of the first Schottky diode, the positive electrode of the backup battery is connected to the anode of the second Schottky diode, the cathodes of the first Schottky diode and the second Schottky diode are connected together and then connected to the positive electrode of the load.

[0021] By employing this technical solution, the power supply input voltage is monitored in real time. When the voltage falls outside the abnormal range, the power supply is immediately disconnected, the backup battery is switched to, and an alert is issued. When the voltage returns to normal, the battery is disconnected, and the power supply is switched back to charging. Furthermore, two Schottky diodes are used to isolate the power supply from the battery. Under normal conditions, the battery does not supply power to the load to prevent overdischarge, while seamlessly taking over power in the event of a fault. This solution effectively ensures stable system operation, optimizes battery management to extend its life, enhances circuit safety to prevent current backflow, and reduces complexity and cost due to its simple circuit structure.

[0022] In a second aspect, the present application provides a multivariable composite fire detection system, which adopts the following technical solutions: A multivariable composite fire detection system comprises a processor, wherein the processor executes the steps of any one of the multivariable composite fire detection methods described above.

[0023] In a third aspect, the present application provides a storage medium that adopts the following technical solution: A storage medium stores a program, which, when executed by a processor, implements the steps of any one of the multi-variable composite fire detection methods described above.

[0024] In a fourth aspect, the present application provides a multivariable composite fire detection device, which adopts the following technical solutions: A multivariable composite fire detection device, comprising: The basic functional module includes a basic module, an outer shell and an expansion plug. The basic module includes a control / storage module, a communication / transmission module, a signal / button module and a backup battery. The outer shell and the expansion plug are detachably connected. Multi-sensor fusion module, including two or more sensors among smoke, temperature, infrared, ultraviolet, image, and gas, for multi-dimensional, multi-sensor centralized monitoring of the surrounding environment; The matching interface is electrically connected to the basic functional module and is used to link with the fire sound and light alarm, fire emergency cut-off and fire extinguisher equipment and facilities; A data storage module, including a storage element for storing data; Transmission module, including Ethernet TCP interface, Internet of Things interface, fire alarm controller dedicated interface, RS485 interface; Power supply module, including power supply and backup battery; Positioning module, including Beidou / GPS positioning unit, used to obtain real-time positioning information; The alarm unit is electrically connected to the basic function module and is used to send out an alarm signal.

[0025] By adopting the above technical solution, this device adopts an integrated packaged basic functional module, integrating point-type smoke, temperature and infrared composite fire detectors, and combines with a split-type special quick-function connector to achieve modular expansion; the basic functional module consists of a module body, an outer shell and an expansion plug. The outer shell and the plug are connected through a detachable and interchangeable connection to form a protective structure. The module integrates communication, control and upgrade interfaces; the special connector has a built-in elastic guide structure to support the rapid replacement of sensors, communication interfaces, etc., and has impact resistance and high and low temperature resistance; the matching interface uses an elastic probe to ensure reliable contact, and the module's external guide structure and the sealing structure are separated to ensure accurate positioning. This design significantly improves fire detection accuracy, maintenance convenience and adaptability to complex environments, supports flexible functional expansion, and reduces operation and maintenance costs.

[0026] In summary, the present application includes at least one of the following beneficial technical effects: using the first and second sensors to respectively obtain sensor data, extract information such as sensing target data, location and timestamp, and detect a variety of state parameters, breaking through the limitations of a single sensor, comprehensively sensing fire risks, and improving detection accuracy and reliability. By calculating the timestamp, location data difference and process matching value, it is possible to determine sensor failure, avoid false alarms and missed alarms, and judge fires based on matching values ​​and output corresponding prompts. In each sub-step, such as processing data from different sensors, it is possible to accurately extract fire characteristics, determine time and location, construct a curve to quantify data change consistency, normalize and flexibly set weights when fusing data, and reasonably allocate computing tasks to the background or relay station to ensure stable system operation, optimize resource utilization, and enhance adaptability and scalability. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a step diagram of a multivariable composite fire detection method.

[0028] Figure 2 This is a step diagram of a multivariable composite fire detection method based on a third sensor.

[0029] Figure 3 This is a modular structure diagram of the multivariable composite fire detection system of this application.

[0030] Figure 4 The present invention is a step diagram of obtaining first sensing data based on a first sensor, and extracting first sensing target data, a first sensing position and a first timestamp from the first sensing data.

[0031] Figure 5 The present invention is a step diagram of obtaining second sensing data based on a second sensor, and extracting second sensing target data, a second sensing position, and a second timestamp from the second sensing data.

[0032] Figure 6 It is a schematic diagram of the overall structure of a multi-variable composite fire detection device.

[0033] Figure 7 It is a structural diagram of a multi-variable composite fire detection device, mainly used to show the schematic diagram of the bottom "PIR" position.

[0034] Figure 8 This is a diagram showing the location of the ports on the side of the device.

[0035] Figure 9 It is a schematic diagram used to show the location of the expansion plug.

[0036] Figure 10 It is a first-person exploded view of a multivariable composite fire detection device.

[0037] Figure 11 It is a second-perspective exploded view of a multi-variable composite fire detection device.

[0038] Description of the accompanying drawings: 1. Module body; 2. Outer shell; 3. Expansion plug. DETAILED DESCRIPTION

[0039] Embodiments of the present application are described in detail below, examples of which are illustrated in the accompanying drawings.

[0040] Throughout this specification, reference to the terms "certain embodiments," "one embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiments or examples are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0041] The present application discloses a multivariable composite fire detection method, referring to Figure 1 , including the following steps: First sensor data is acquired based on the first sensor, and first sensing target data, a first sensing location, and a first timestamp are extracted from the first sensor data. The first sensing target data is physical quantity information directly related to fire, such as smoke concentration and flame intensity. The first sensing location specifies the specific location of the fire-related phenomenon sensed by the sensor in space, such as the precise location on a floor or in a building. The first timestamp indicates the exact moment the first sensor data was acquired, providing a benchmark for subsequent temporal analysis.

[0042] Second sensor data is acquired based on the second sensor, and second sensing target data, a second sensing location, and a second timestamp are extracted from the second sensor data. The second sensing target data is physical quantity information directly related to the fire, such as smoke concentration and flame intensity. The second sensing location specifies the specific location of the sensor, such as the floor or room number in a building. The second timestamp indicates the exact moment the second sensor data was acquired, providing a benchmark for subsequent temporal analysis.

[0043] Calculate the first timestamp difference between the first timestamp and the second timestamp. The first timestamp difference can reflect the time difference when the two sensors acquire data. During the occurrence and development of a fire, the time when sensors at different locations perceive fire-related phenomena may have a certain order. This difference can be used to analyze the time characteristics of the fire spread.

[0044] Calculate the first position data difference between the first sensing location and the second sensing location. The first position data difference clarifies the spatial distance between the two sensor sensing locations, which helps determine the impact range of the fire and the direction in which the fire may spread.

[0045] The first process matching value of the first timestamp difference and the first position data difference is calculated. The first process matching value comprehensively considers the differences in the two dimensions of time and space, and is an important basis for judging fire-related conditions.

[0046] When the calculated first process match value is less than the preset reference match value, a first target difference between the first and second sensing target data is calculated to further determine the system status. If this first target difference is greater than the preset reference target value, the system will be able to keenly detect the possibility of a sensor failure and immediately issue a sensor failure warning. Simultaneously, the first and second sensing target data are output, facilitating further troubleshooting of the cause of the failure and ensuring the normal operation of all aspects of the fire detection system. This avoids erroneous fire judgments caused by sensor failures and effectively prevents false alarms or missed alarms.

[0047] If the first process matching value is greater than the preset reference matching value, this means that the data from the two sensors show consistent characteristics related to the occurrence of fire in the time and space dimensions. At this time, the first sensing fusion data of the first sensing target data and the second sensing target data is calculated, and the data of the two sensors are organically integrated through a specific fusion algorithm to give full play to the complementary advantages of different sensor data. Then, the first sensing fusion data is compared with the preset fusion reference value. If the first sensing fusion data is greater than the preset fusion reference value, the system will immediately output a fire alarm prompt to warn relevant personnel that a fire has occurred and has reached a relatively serious level, and emergency measures such as fire extinguishing and evacuation must be taken quickly; on the contrary, if the first sensing fusion data does not exceed the preset fusion reference value, a fire warning prompt will be output to remind relevant personnel that the fire is in the embryonic or early stage and needs to be closely monitored and hidden dangers checked in time.

[0048] By acquiring sensor data from different sensors (a first sensor and a second sensor) and extracting information such as the target data, location, and timestamp, this innovative multivariable composite fire detection method overcomes the traditional limitations of each fire detector, which is limited to detecting a single state parameter. This method enables a single fire detector to comprehensively perceive multiple potential fire risks from multiple dimensions, significantly improving the accuracy and reliability of fire detection. Compared to existing technologies that only detect a single state parameter under specific environmental conditions, this method calculates the difference in the first timestamp, first position data, and first process matching value between the two sensors and compares them with preset values, enabling a more comprehensive and in-depth analysis of the fire situation. When the first process matching value is less than the preset reference matching value, a timely sensor failure warning is issued to ensure stable operation of the fire detection system. When the first process matching value is greater than the preset reference matching value, the occurrence and severity of the fire are accurately determined, and a fire alarm or warning is issued promptly. Furthermore, because a single fire detector can detect multiple state parameters, there is no need to deploy additional sensors and perform cumbersome secondary wiring to detect different parameters, as is required in existing technologies. This not only avoids problems caused by secondary wiring, such as unsightly building appearance and the need for secondary disassembly and assembly during renovation, but also does not cause delays in construction schedules. It effectively improves construction efficiency and greatly reduces construction costs and complexity. It has significant advantages and wide promotion value in practical applications.

[0049] Reference Figure 2 On the basis of the existing fire detection based on the first sensor and the second sensor, a third sensor is further introduced to enhance the comprehensiveness and accuracy of fire detection. The method includes the following steps: Acquire third sensing data based on a third sensor, and extract third sensing target data, a third sensing location, and a third timestamp from the third sensing data. The third sensor may be a smoke sensor, a flame sensor, or a temperature sensor installed at a different location within the building. Extract the third sensing target data, the third sensing location, and a third timestamp from the third sensing data.

[0050] For example, if the third sensor is a smoke sensor, the third sensing target data is the detected smoke concentration; if it is a flame sensor, it is the detected flame intensity; if it is a temperature sensor, it is the detected ambient temperature, etc. For example, if a third sensor is installed in the warehouse area of ​​a large shopping mall, when smoke begins to form in the warehouse, the smoke concentration value detected by the sensor is the third sensing target data.

[0051] The specific location of the fire-related phenomenon detected by the third sensor in space is clearly defined. In a building, for example, this might be a specific room number on a floor, a sign for a specific area, etc. For example, in an office building, if the third sensor is installed in the conference room on the 10th floor, then "10th Floor Conference Room" would be the third sensing location.

[0052] Third timestamp: This records the exact moment when the third sensor data was acquired, providing a benchmark for subsequent time-based analysis. For example, if the third sensor data was acquired at xx:xx:xx on xx / xx / xx, this time is the third timestamp.

[0053] Calculate the difference between the third and second timestamps. This difference reflects the difference in the timing of data acquisition between the third and second sensors. During a fire, sensors at different locations may perceive fire-related phenomena in a certain order. For example, if the second sensor is installed in a shopping mall lobby and the third sensor is installed in a warehouse, when a fire occurs in the warehouse, smoke and flames may not spread to the lobby until later due to the warehouse's relative isolation. Therefore, the third sensor may acquire data earlier than the second sensor. By calculating the difference in the second timestamps, the time characteristics of the fire's spread to different locations can be analyzed.

[0054] Calculate the difference in second-position data between the third sensing location and the second sensing location. This difference in second-position data clarifies the spatial distance between the sensing locations of the third and second sensors, helping to determine the fire's impact range and potential direction of spread. For example, suppose the second sensor is in the first-floor lobby of a shopping mall, and the third sensor is in a second-floor warehouse. By calculating information such as the floor difference and horizontal distance between them, the difference in second-position data can be calculated. A large difference indicates that the fire may take a long time to spread from one location to the other; a small difference indicates that the fire may spread quickly.

[0055] Calculate the second-process matching value of the difference between the second timestamp and the second location data. The second-process matching value comprehensively considers differences in both time and space and is an important basis for determining fire-related conditions. For example, if the second timestamp difference is large but the second location data difference is small, it indicates that the fire took a long time to spread over a short distance, possibly due to a small fire or being blocked by some obstacles. Conversely, if the second timestamp difference is small but the second location data difference is large, it indicates that the fire spread quickly and is likely to be more severe.

[0056] If the second process matching value is less than the preset reference matching value, it means that the temporal and spatial differences between the two sensor data do not conform to the characteristics of normal fire spread, and there may be a sensor failure. In this case, it is necessary to calculate the second target difference between the third sensing target data and the second sensing target data.

[0057] For example, if the second sensor detects a smoke concentration of 10% and the third sensor detects a smoke concentration of 80%, the second target difference is 70%. A comprehensive target difference is then calculated based on the first and second target differences. Assuming the first target difference is 30% and the second target difference is 70%, the comprehensive target difference can be calculated using a simple weighted average method. If the comprehensive target difference exceeds the preset reference target value, the system will be able to detect a possible sensor failure and immediately issue a sensor failure warning. It will also output the first, second, and third target data to facilitate further troubleshooting. For example, based on the output data, staff can verify the correct installation position of each sensor and whether there is any external interference. This ensures the proper operation of all aspects of the fire detection system, avoids incorrect fire detections due to sensor failure, and effectively prevents false alarms or missed alarms.

[0058] If the second process match value is greater than the preset reference match value, this indicates that the data from the third and second sensors exhibit consistent characteristics related to fire occurrence in both temporal and spatial dimensions. At this point, the second sensing fusion data is calculated for the third and second sensing target data. Using a specific fusion algorithm, the data from the two sensors is organically integrated, leveraging the complementary advantages of the different sensor data. For example, a weighted average method can be used to fuse the third and second sensing target data, assigning different weights based on factors such as the reliability of the two sensors.

[0059] Then, comprehensive sensing fusion data is calculated based on the first and second sensing fusion data. Appropriate fusion algorithms, such as weighted averaging, can also be used. If the comprehensive sensing fusion data exceeds a preset fusion reference value, it indicates that a fire has occurred and has reached a relatively serious level. The system will immediately output a fire alarm, warning relevant personnel to quickly take emergency measures such as fire extinguishing and evacuation. For example, if the comprehensive sensing fusion data shows that indicators such as smoke concentration and flame intensity exceed the preset fusion reference value, it indicates that the fire has spread to a large area and requires evacuation and fire extinguishing as soon as possible. Conversely, if the comprehensive sensing fusion data does not exceed the preset fusion reference value, a fire warning is output, reminding relevant personnel that the fire is in its incipient or early stages and requires close attention and timely investigation for potential hazards. For example, if the comprehensive sensing fusion data shows that indicators such as smoke concentration and flame intensity have increased but have not yet reached a serious level, relevant personnel can be notified in advance to conduct inspections and prevent further development of the fire. Among them, the calculation methods of the second timestamp difference, the second position data difference, the second process matching value, the second induction fusion data and other values ​​are the same as the calculation methods of the corresponding first timestamp difference, the first position data difference, the first process matching value, and the first induction fusion data.

[0060] The first sensor, the second sensor, and the third sensor are each a smoke sensor, a temperature sensor, an infrared sensor, an ultraviolet sensor, an image sensor, or a gas sensor, and the first sensor, the second sensor, and the third sensor are different sensors. Preferably, the first sensor, the second sensor, and the third sensor are, respectively, a smoke sensor, a temperature sensor, and an infrared sensor. Remote on-site fire confirmation can be achieved through images, eliminating the need for on-site personnel to confirm. In other embodiments, more than three sensors may be included, each of which is a smoke sensor, a temperature sensor, an infrared sensor, an ultraviolet sensor, an image sensor, or a gas sensor.

[0061] Reference Figure 3 This solution adopts a modular structure and building block design. Basic functions are combined with various preset sensors to realize multi-sensor fusion technology. It is equipped with matching interfaces to connect with different equipment and facilities. It meets the fire detection and dynamic environment monitoring needs of different industries and scenarios.

[0062] Reference Figure 4 In other embodiments, the step of acquiring first sensing data based on the first sensor and extracting first sensing target data, first sensing position, and first timestamp from the first sensing data further includes the following sub-steps: The first sensor is an image sensor, and the first sensor data is the image data captured by the image sensor. The image sensor converts the actual scene of the fire into digital image data. This image data captures the fire scene in a moment, contains rich and intuitive fire-related information, and provides a foundation for subsequent data analysis and processing.

[0063] Using a preset first recognition algorithm, the image data is extracted to identify fire, temperature, and smoke signatures. The first sensing target data is presented as the image confidence values ​​for these signatures. This value provides a visual assessment of the severity of these signatures. A higher image confidence value indicates a larger fire, higher smoke density, and a more severe fire situation. Conversely, a lower image confidence value indicates a fire in its early stages or at a lower intensity.

[0064] The time at which the flame, temperature, and smoke characteristics were detected is extracted as the first timestamp. This first timestamp allows the time information of the image sensor data to be closely linked with data from other sensors. For example, when the smoke sensor detects a change in smoke concentration, the timestamp can be used to compare and analyze the flame, temperature, and smoke characteristics captured by the image sensor at that moment, thereby clearly illustrating the timeline of the fire's development. This allows for a clear visualization of the entire dynamic process, from the initial appearance of smoke to the ignition of the flames and the gradual spread of the fire. This provides strong temporal support for firefighters to promptly grasp the fire's development and formulate effective response measures. During the crucial fire rescue timeframe, every second counts. Accurate time information helps rescuers make swift decisions, improve rescue efficiency, and minimize fire damage.

[0065] The image location of the flame, temperature, and smoke signature targets in the image data is calculated, and a physical location is matched from a pre-set location database based on the image location as the first sensing location. By precisely calculating the pixel coordinates in the image, the image location of the flame, temperature, and smoke signature targets in the image data is determined. Although this image location is based on coordinate information in the image space, it lays the foundation for subsequently determining the actual physical location. Subsequently, based on a pre-set location database containing a large amount of actual physical space location information corresponding to the image space, a precise matching algorithm is used to accurately map the image location to the actual physical space, thereby matching the corresponding physical location as the first sensing location. Accurate location information is crucial in fire rescue scenarios. Upon receiving a fire alarm, firefighters must arrive at the scene quickly and accurately. The first sensing location provides them with a precise location of the fire. Whether in complex environments such as large commercial complexes, high-rise buildings, or industrial plants, it helps firefighters quickly locate the fire source, saving time and enabling effective rescue operations. Accurate positioning not only improves rescue efficiency but also ensures the personal safety of rescuers, avoiding the waste of precious time blindly searching for the fire source in complex environments.

[0066] Reference Figure 5 The step of acquiring second sensing data based on the second sensor and extracting second sensing target data, a second sensing position, and a second timestamp from the second sensing data further includes the following sub-steps: The second sensor is a non-image sensor, and the second sensor data is the electrical signal data collected by the non-image sensor. Non-image sensors can keenly perceive various changes in fire-related physical quantities in the surrounding environment and convert these changes into electrical signal data. For example, a smoke sensor can detect changes in the concentration of smoke particles in the air, a temperature sensor can sense increases in ambient temperature, and a gas sensor can identify specific gas components produced by a fire. These changes in physical quantities are ultimately recorded in the form of electrical signals, forming a rich set of second sensor data.

[0067] Using a pre-set second recognition algorithm, the electrical signal data is extracted for flame, temperature, and smoke characteristics. The second sensing target data is the values ​​of the electrical signal data for these characteristics. This method accurately extracts features related to flame, temperature, and smoke from complex electrical signal data. For example, when smoke concentration increases, the electrical signal exhibits a specific pattern of changes. The algorithm can identify this pattern and interpret it as part of the flame, temperature, and smoke characteristics. The second sensing target data is the values ​​of the electrical signal data corresponding to these flame, temperature, and smoke characteristics. This method captures fire-related information from a dimension distinct from that of images. Image sensors provide a visual representation of the fire scene, allowing for intuitive visualization of the flame and smoke patterns. Non-image sensors, on the other hand, detect information such as smoke concentration, temperature, and gas composition from the perspective of changes in physical quantities. These two types of information complement each other, enabling a comprehensive understanding of the fire scene. This results in more accurate fire assessments and significantly reduces the potential for misjudgment based on a single sensor. For example, relying solely on image sensors may lead to misjudgment due to factors such as light and obstruction; however, combined with electrical signal data from non-image sensors, it is possible to verify from multiple aspects whether a fire has actually occurred, thereby improving the comprehensiveness and accuracy of fire detection.

[0068] The times of the flame, temperature, and smoke characteristics are extracted as the second timestamp. Different types of sensors capture fire-related information at different times. By comparing and correlating these two timestamps, we can more accurately analyze the temporal correlation between different physical phenomena during the fire process. For example, an image sensor may first capture the image of an open flame, while a non-image sensor later detects rising smoke concentrations. By comparing these timestamps, we can clearly understand the time interval between the ignition of the fire and the generation of smoke, further untangling the temporal context of the fire's development. This is crucial for early warning and rapid response to fires. Prompt detection and action in the early stages of a fire can effectively control its spread and minimize losses. The second timestamp provides a powerful temporal clue, enabling optimal response.

[0069] The physical location of the non-image sensor is matched from a preset location database as the second sensing location. This method offers flexibility. Preset location database matching is suitable for fixed scenarios, providing reliable positioning of the fire location. Combined with the location information determined by the first sensor, this provides dual confirmation of the fire location and more accurate positioning, providing critical location information support for firefighters to quickly arrive at the fire scene and carry out effective rescue operations, thereby improving the efficiency and success rate of fire rescue operations.

[0070] The step of calculating the process matching value of the first timestamp difference and the first position data difference further includes the following sub-steps: A first characteristic change curve for the first sensor is constructed, with time as the horizontal axis and first sensing target data as the vertical axis. The first sensor is an image sensor, and the first sensing target data represents the magnitude of fire, temperature, and smoke features in the image, such as their area, brightness, and other fire-related characteristics. By connecting the first sensing target data corresponding to different time points, the resulting curve can intuitively demonstrate the changes in fire-related features in the image over time.

[0071] With time as the horizontal axis and the second sensing target data as the vertical axis, a second characteristic change curve for the second sensor is constructed. Similarly, the second sensor is typically a smoke sensor, and the second sensing target data is related data such as smoke concentration. This curve clearly shows the changing trend of smoke concentration over time.

[0072] Based on the first characteristic change curve and the second characteristic change curve, the changing trends of the first characteristic change curve and the second characteristic change curve are identified, and the consistency of the changing trends is calculated as the curve matching value. For example, when the image sensor curve shows a rapid increase in the smoke characteristic value, it means that the smoke is rapidly increasing. Then, within the corresponding time range, the smoke sensor curve should also show a similar upward trend in the smoke concentration data value detected. If the two curves show a high degree of consistency in these key features, the algorithm will give a higher matching score. Using advanced data analysis algorithms and pattern recognition technology, in-depth analysis is performed on the rise, fall, fluctuation and other characteristics of the curve. Compared with relying solely on the data of a single sensor for judgment, this method of correlating and analyzing different sensor data through curve matching values ​​greatly improves the accuracy and comprehensiveness of fire feature identification and effectively reduces the possibility of misjudgment.

[0073] A reasonable time difference and a reasonable position data difference are calculated according to the first timestamp difference and the first position data difference.

[0074] A reasonable time difference = first position data difference / first sensing target data; a reasonable position data difference = first timestamp difference × first sensing target data. In real-world fire scenarios, the time and distance between smoke spreading from the fire location to the smoke sensor location have a reasonable correspondence. The first position data difference is linked to the time change on the two constructed data change curves. Assuming an ideal situation, once a fire occurs at the location identified by the image sensor, smoke will spread at a certain speed. If the smoke sensor is within its range, then over time, the relationship between the first timestamp difference and the first position data difference should conform to a reasonable time-distance relationship for smoke diffusion. For example, if the smoke diffusion speed is v and the position data difference is d, then the timestamp difference Δt from the fire onset (first timestamp) to the smoke sensor detecting smoke (second timestamp) should theoretically satisfy Δt = d / v (ignoring some complex environmental factors).

[0075] The difference between the reasonable time difference and the first timestamp difference is calculated as the comprehensive time difference; the difference between the reasonable position data difference and the first position data difference is calculated as the comprehensive position difference. The two comprehensive differences further quantify the deviation between the actual detection data and the theoretical reasonable value.

[0076] Calculate the reasonable time-distance value based on the comprehensive time difference and comprehensive position difference; Reasonable value of time-distance = k / (comprehensive position difference 2 +Comprehensive time difference 2 ), where k is a tuning parameter. The tuning parameter k can be adjusted based on different application scenarios and actual needs to ensure that the reasonable time-distance value more accurately reflects the rationality of the temporal and spatial relationships in a fire scene. The reasonable time-distance value calculated using this formula effectively verifies the rationality of sensor detection data in both temporal and spatial dimensions. If the actual timestamp difference and location data difference are close to the calculated reasonable values, it indicates that the sensor detection results conform to the normal laws of fire smoke diffusion, further enhancing the reliability of fire diagnosis.

[0077] According to the curve matching value and the reasonable value of time-distance, the process matching value is calculated: The first-pass matching value = α × curve matching value + β × (time-distance reasonable value), where α + β = 1. By setting the weights α and β, the importance of the curve matching value and the time-distance reasonable value in the calculation of the first-pass matching value can be flexibly adjusted according to the needs of different application scenarios. For example, in scenarios where timely fire response is crucial, such as large shopping malls and hospitals with large crowds, the weight α of the curve matching value can be appropriately increased to ensure rapid fire detection and timely action. Prioritize the rapid consistency of sensor data trends. Conversely, in scenarios where fire location accuracy is critical, such as industrial plants and warehouses, the weight β of the time-distance reasonable value can be increased to accurately determine the fire location so that firefighters can more accurately extinguish and rescue.

[0078] This comprehensive evaluation method fully considers the consistency of sensor data feature changes and their temporal and spatial rationality, providing a more scientific and reliable basis for subsequent accurate judgment of sensor status and fire status. This significantly improves the performance and adaptability of the entire fire detection system, enabling it to better cope with various complex real-world scenarios and play a greater role in protecting people's lives and property. Similarly, the calculation method for the second-stage matching value is the same as that for the first-stage matching value.

[0079] The step of calculating the sensing fusion data of the first sensing target data and the second sensing target data further includes the following sub-steps: The first-normalized value of the first-sensing target data is calculated using a maximum-minimum normalization algorithm. This algorithm is a simple and effective data preprocessing method. Its core principle is to find the maximum and minimum values ​​in the first-sensing target data and map each data point to the interval [0, 1] according to a specific formula. Specifically, for any value in the first-sensing target data, the minimum value of the data group is first subtracted, and then divided by the difference between the maximum and minimum values ​​to obtain the corresponding first-normalized value. This method aims to unify the first-sensing target data into a standardized range, eliminating the effects of differences in the original data dimensions and value range.

[0080] The second-sensed target data is calculated using a maximum-minimum normalization algorithm as the second normalized value. Similarly, the second-sensed target data is also calculated using the same maximum-minimum normalization algorithm. This method maps the first and second-sensed target data from different sensor types to the same [0,1] interval, creating an excellent foundation for subsequent data fusion.

[0081] The sensor fusion data is calculated as w1 × the first normalized value + w2 × the second normalized value. The weight of the first sensor target data is w1, and the weight of the second sensor target data is w2, so w1 + w2 = 1. By properly setting the values ​​of w1 and w2, the relative importance of the two sensor data in the fusion result can be flexibly adjusted.

[0082] Because the primary and secondary sensing target data come from different types of sensors, their dimensions and value ranges often differ significantly. For example, the primary sensing target data might be the area of ​​smoke from an open fire in an image captured by an image sensor, with values ​​ranging from a few square centimeters to hundreds of square meters. The secondary sensing target data might be the smoke concentration detected by a smoke sensor, with values ​​ranging from very low ppm (parts per million) to high percentages. Without normalization, the significant differences in dimensions and value ranges can cause certain data to dominate the fusion process, obscuring important information contained in other data and affecting the fairness and accuracy of data fusion. The normalization algorithm, however, uniformly maps these data to the interval [0, 1], ensuring that each data point is treated equally during the fusion process. This avoids irrational fusion results caused by dimensional differences and significantly improves the quality of data fusion.

[0083] Raw sensing target data is susceptible to various factors, such as environmental interference and sensor errors, leading to significant data fluctuations. This fluctuating data hinders the subsequent fire identification model from learning the data's patterns and features. When processing highly volatile data, the model may mistake random fluctuations for signs of a fire, increasing the risk of misjudgment. Normalization, on the other hand, smooths the data and makes it more stable. By mapping the data to a fixed interval, the data's fluctuations are reduced, making the data's changing trends clearer. This allows subsequent fire identification models to more easily learn truly fire-related patterns and features from stable data, improving the model's generalization ability and reducing the likelihood of misjudgment.

[0084] By setting weights w1 and w2, the importance of the two sensor data in the fusion result can be flexibly adjusted based on the specific needs of the scenario. Different sensor types may have different reliability and importance in different application scenarios. For example, in a large, well-ventilated warehouse, a smoke sensor may be more susceptible to airflow, resulting in inaccurate detection results. In this case, the weight w2 of the second sensor target data (smoke concentration) can be appropriately reduced, while the weight w1 of the first sensor target data (the area of ​​open flame smoke in the image) can be increased, thereby relying more on the image sensor data for fire detection. In a relatively closed indoor space, the smoke sensor's detection results may be more reliable, so the weight of the second sensor target data can be increased. This flexible weight adjustment mechanism enables the fire detection system to better adapt to a variety of real-world scenarios, enhancing the system's adaptability and practicality.

[0085] The maximum-minimum normalization algorithm is inherently computationally simple. It requires only simple addition, subtraction, multiplication, and division operations, eliminating the need for complex mathematical models and extensive computing resources. This allows for rapid data normalization in actual data processing, reducing processing complexity and resource consumption, and improving computational efficiency. Furthermore, the weighted summation fusion calculation method is intuitive and easy to understand, making its principles and implementation easily accessible to both professional technicians and casual users. This straightforward calculation method facilitates its application and widespread adoption in practical fire detection systems, enabling fire detection technology to serve society more broadly and play a greater role in protecting people's lives and property.

[0086] The step of calculating the first process matching value of the first timestamp difference and the first position data difference further includes the following sub-steps: Obtain the computational effort required to calculate the first process matching value of the first timestamp difference and the first location data difference. For example, calculating the timestamp difference may involve subtraction of time data, while calculating the location data difference may require coordinate distance calculations, which may include more complex mathematical operations such as squaring and square root extraction. Furthermore, calculating the process matching value may also involve the calculation and integration of multiple intermediate results, such as curve matching values ​​and reasonable time-distance values. The number, complexity, and required computing resources of these operations together constitute the final computational effort. To accurately obtain this computational effort, the system needs to utilize advanced resource monitoring and evaluation technologies to track, in real time, the CPU, memory, and other resources consumed during the computational process, thereby accurately measuring the severity of the computational task. The calculation of the second process matching value is similar.

[0087] Computational tasks are dynamically allocated based on the acquired computational effort and the comparison of the position data difference with a set threshold. Two important thresholds are set here: the computational effort threshold and the positional effort threshold. The computational effort threshold is a pre-set standard based on the system's computing power and performance requirements, representing the upper limit of the computational effort the system can easily handle. The positional effort threshold is set based on the distribution of sensors and the characteristics of data transmission in actual application scenarios. It is used to determine whether the positional relationship between two sensors will have a significant impact on data transmission and computation.

[0088] If the computational effort is less than or equal to the set computational effort threshold, or if the difference in the first position data is less than or equal to the set position threshold, this means the current computational task is relatively easy, or the two sensors are relatively close together, resulting in low data transmission latency and complexity. In this case, sending the first and second sensor data to the backend system for computation is a more appropriate option. Backend systems typically possess powerful computing capabilities and abundant computing resources, enabling them to efficiently complete these relatively simple computational tasks and centrally manage and analyze the data, facilitating subsequent processing and decision-making.

[0089] If the amount of computation is greater than the set computation threshold, or the difference in the first position data is greater than the set position threshold, it indicates that the current computational task is too heavy, or the distance between the two sensors is too far. Directly transmitting the data to the backend system may result in significant delays and waste of resources. In this case, the first and second sensor data are sent to the corresponding relay stations for computation, and the relay stations then transmit the computational results to the backend system. Relay stations are typically located near the sensors and have a certain level of computing and data processing capabilities. By assigning some computational tasks to relay stations, the burden on the backend system can be reduced, while also reducing the distance and delay of data transmission and improving computational efficiency.

[0090] From the perspective of computational efficiency, rational task allocation optimizes the utilization of computing resources. When the computational load is small, the backend system can quickly process data, avoiding the overhead associated with assigning simple tasks to relay stations. When the computational load is large, the relay stations offload some of the computational tasks, enabling the entire computation process to proceed in parallel, significantly shortening computation time and improving system response speed. Regarding data transmission, performing partial computation at relay stations near the sensors reduces the amount and distance of data transmission, effectively minimizing data transmission latency. This is crucial for fire detection systems with high real-time requirements, ensuring timely and accurate access to fire-related information, providing strong support for early warning and rapid response. From the perspective of system reliability and fault tolerance, this solution enhances system stability. In the event of a relay station failure, the system can continue to operate because some computational tasks can be performed in the backend system, preventing the entire system from collapsing due to a relay station failure. Furthermore, the presence of multiple relay stations provides a degree of redundancy, further enhancing the system's fault tolerance. In terms of resource utilization and energy consumption, this dynamic allocation of computational tasks avoids excessive resource waste. The relay station and the background system reasonably share the tasks according to the actual computing needs, so that the computing resources are fully utilized, unnecessary energy consumption is reduced, and the requirements of energy conservation and environmental protection are met.

[0091] This solution also improves the system's adaptability and scalability to diverse application scenarios. Sensor distribution and computing requirements may vary across different scenarios. By adjusting the computation and location thresholds, we can flexibly adapt to these changes, ensuring the system operates efficiently in diverse environments. Furthermore, as the system continues to grow and expand, new sensors and relay stations can be easily added. Simply adjusting the task allocation strategy based on actual conditions allows for smooth system upgrades and expansions.

[0092] The step of sending the first sensor data and the second sensor data to the corresponding relay station for calculation, and then sending the data from the relay station to the background system for calculation, further includes the following sub-steps: Based on the existence of multiple relay stations, the master and slave devices are precisely defined based on the fire detector where the primary sensor is located. The master is specifically the relay station connected to the fire detector where the primary sensor is located. The remaining relay stations are defined as slaves. A close relationship is established between the master and slave devices, not just a simple connection, but a collaborative partnership. Together, they form a distributed data processing network, designed to more efficiently perform data calculation and transmission, providing a solid data foundation for subsequent fire assessment and early warning.

[0093] The host's status feedback information is obtained at a set frequency; this set frequency has been carefully considered and verified through experiments to ensure that any abnormal conditions that may occur in the host can be discovered in a timely manner, while not obtaining the status too frequently to avoid unnecessary waste of resources. The host's status feedback information includes key content from multiple aspects, such as the hardware operating status of the relay station, whether there are problems such as overheating and insufficient memory; the network connection status, whether it is stable, whether there is packet loss, etc.; and the execution status of the computing task, whether it is completed on time, whether there are any computing errors, etc. By continuously monitoring this status information, the system can grasp the working status of the host in real time and provide an accurate basis for subsequent decision-making.

[0094] If the status feedback indicates an abnormal state, the system quickly activates its emergency response mechanism and establishes a connection with the nearest slave. This process requires the system to possess rapid and accurate judgment capabilities and efficient communication capabilities. Using built-in positioning algorithms and network topology, the system quickly determines the location of the slave closest to the master and establishes a connection with it using a stable and reliable communication protocol. Once the connection is established, computing tasks originally undertaken by the master are quickly transferred to the slave, ensuring that data calculation and transmission are not interrupted by master anomalies.

[0095] In terms of data processing reliability, this solution provides dual guarantees for data processing in the fire detection system. If the master node experiences an anomaly, the slave node can quickly take over, ensuring continuous data computation and transmission. This prevents data loss or processing interruptions caused by a single relay station failure, thereby ensuring stable operation of the fire detection system. In terms of system fault tolerance, this solution significantly enhances the system's ability to cope with complex environments. In real-world fire detection scenarios, relay stations may face a variety of complex conditions, such as severe weather conditions and electromagnetic interference, which can cause relay station failures. By implementing a master-slave switching mechanism, the system can quickly respond to relay station failures and continue processing data, ensuring normal system operation. In terms of resource utilization and efficiency, this solution ensures optimal resource allocation. The collaborative operation between the master and slave nodes allows for flexible allocation of computing tasks based on the actual needs of each relay station, preventing overloading of some relay stations while leaving others idle, thereby improving resource utilization efficiency. Furthermore, because the slaves are closer to the master, switching to the slaves for calculations when the master experiences an anomaly reduces data transmission distance and time, lowering transmission latency and further improving overall system efficiency. This solution strikes a perfect balance between resource utilization and system efficiency, ensuring efficient system operation while avoiding resource waste, providing strong support for the long-term stable operation of the fire detection system.

[0096] That is, this solution forms a self-healing network, automatically adjusting the role and connection method of the device under different network conditions. When the device cannot connect to the LAN, it connects to the platform or backend through a 4G / 5G module or ring network to ensure that the device is always online. Task events are triggered through API functions, software timers or system message passing mechanisms to achieve the network's self-starting, self-organizing and self-healing functions. When multiple detectors form a LAN through interfaces, a certain device will automatically become the host and establish contact with the platform or backend, and the others are all slaves; when a device becomes the host and cannot get in touch with the platform or backend, it will replace it with another device to suggest contact; when a device cannot connect to the LAN, the device will automatically start the network discovery process and try to join other available networks; when the device reconnects to the LAN, the device will automatically detect the existence of the network and become part of the LAN by rejoining the network; when the entire LAN is offline, the coordinator will detect the network failure and start the self-healing mechanism.

[0097] The self-healing function is implemented through a coordinator, routers, and terminal nodes. Terminal nodes are authorized and grouped through the platform or backend and integrated into the network system. The coordinator is responsible for establishing and managing the entire network, allocating network addresses, and maintaining the network topology. The router is responsible for data forwarding, maintaining neighbor node information, and supporting multi-hop communication. The terminal nodes are responsible for data collection and upload, and can be sensors or other devices.

[0098] After power-on, the coordinator, router, and node devices each initialize the network, and the router establishes the network; the preset or authorized node device searches for the network, obtains network information, and determines whether to join the network; the coordinator can select another device (such as a pre-configured backup coordinator) to take over the network, or reconnect to the platform or backend through other channels (such as 4G / 5G modules or ring networks).

[0099] This solution incorporates on-board storage components to maintain operational data for a specific period. This allows for data to be reconnected after a short period of disconnection. It also communicates with connected devices, enabling dual data backup on both local and remote platforms.

[0100] This solution features an Ethernet TCP interface, an IoT interface, a dedicated interface for the fire alarm controller, and an RS485 interface on the detector, enabling both local wired and remote wireless access, allowing simultaneous sharing of local status with multiple parties. It supports both wired and wireless transmission modes, automatically switching based on actual conditions, ensuring stable remote upgrades and communication / transmission in various network environments. RS232 communication is optional for communication / transmission.

[0101] The power supply system of the composite fire detector includes a power supply and a backup battery. The method further includes the following steps: The system continuously acquires the power input voltage in real time, which is achieved through the voltage monitoring module. This module accurately measures the power input voltage and promptly feeds the measurement results back to the system control center. The control center then quickly and accurately determines the acquired voltage value based on the pre-set normal power supply range.

[0102] If the power input voltage falls within the abnormal power range, the system will respond quickly. This abnormal range includes voltage that is too low, such as below the lower limit of the normal operating voltage. This may be caused by unstable mains power supply, increased resistance due to aging lines, and other reasons; it also includes voltage that is too high, above the upper limit of the normal operating voltage. This may be caused by sudden situations such as power supply system failures and lightning strikes. Once an abnormality is detected, the system will immediately trigger the corresponding circuit switching mechanism, disconnecting the power supply from the load and seamlessly switching to the backup battery to power the load. At the same time, the system will issue an early warning prompt and convey information about the power anomaly to relevant management personnel through various means such as audio and visual alarm devices and remote communication modules, so that timely investigation and maintenance can be carried out.

[0103] Conversely, if the power input voltage is within the normal power range, the system will also perform the corresponding operation. At this time, the system will disconnect the backup battery from the load, switch back to the power source to power the load, and start charging the backup battery. The charging process uses intelligent charge management technology, which can automatically adjust the charging current and voltage based on factors such as the current charge level of the backup battery and battery type to ensure safe and efficient charging of the battery, thereby extending the battery life.

[0104] The power supply system utilizes two Schottky diodes to effectively isolate the power supply from the battery. The entire power supply is electrically connected to the anode of the first Schottky diode, meaning the power supply's output current can only flow in the direction of the Schottky diode's conduction. When the power supply is operating normally, its output voltage exceeds the backup battery's voltage. The first Schottky diode conducts, allowing current to flow smoothly through the diode to the load, providing stable power. Simultaneously, due to the unidirectional conductivity of the Schottky diode, the second Schottky diode connected to the backup battery's positive terminal is in a cutoff state. This effectively prevents the battery from supplying power to the load, preventing unnecessary discharge during normal power conditions and thus extending the battery's lifespan. The backup battery's positive terminal is connected to the anode of the second Schottky diode. In the event of a power failure, when the input voltage drops or disappears, the backup battery's voltage becomes higher than the power supply's output voltage. At this point, the second Schottky diode conducts, allowing the battery current to flow through the diode to the load. This ensures seamless power takeover during a power failure, ensuring continuous power supply to the load and maintaining the proper operation of the fire detector. The cathodes of the first Schottky diode and the second Schottky diode are connected together and then connected to the positive electrode of the load. This connection method builds a stable and reliable power supply circuit, ensuring the reasonable flow of current under different power supply states.

[0105] This solution utilizes a dual power supply system (power supply + backup battery). This ensures uninterrupted device operation by rapidly switching to the backup battery in the event of a primary power failure. In the event of an input power outage or power anomaly (such as undervoltage or overvoltage), the backup battery allows continued operation for a limited time. When power is restored, the system automatically switches to the original power supply and charges the backup battery. When the battery reaches a certain capacity, the system extends its operating life by reducing the frequency of reporting and sampling.

[0106] Dual power supply of power supply + backup battery, not limited to integrated IC, diode isolation, MOSFET switch and other methods.

[0107] Integrated IC method: Using integrated IC can achieve more complex power management functions. Integrated IC can manage both the mains and battery power at the same time, automatically switching the power supply path according to the power status.

[0108] The diode isolation method uses two Schottky diodes to isolate the power supply and battery, ensuring that when the power supply is normal, the battery will not supply power to the load; when the power supply fails, the battery can seamlessly take over the power supply.

[0109] In the MOSFET switching method, when the power supply is normal, the gate voltage of the MOSFET is higher than the source voltage, the MOSFET is in the off state, and the connection between the battery and the load is cut off. When the power supply fails, the gate voltage of the MOSFET is zero, the MOSFET is turned on, and the battery supplies power to the load.

[0110] Regardless of whether the power supply is normal or abnormal, the system automatically and quickly switches to a faulty state, ensuring continuous and stable operation of the combined fire detector. In fire detection scenarios, any brief power outage can result in missed or false fire alarms. This solution effectively prevents this, ensuring the fire detector is always operational and providing reliable power for fire warnings. It prevents battery overdischarge and utilizes intelligent charging technology to significantly extend the life of the backup battery. This not only reduces the frequency of battery replacements and maintenance costs, but also improves overall system reliability, ensuring the backup battery is fully operational during critical moments. The unidirectional conductivity of the Schottky diode effectively prevents current backflow. In complex circuit environments, current backflow can cause irreversible damage to the power supply, battery, and load. Through ingenious circuit design, this solution eliminates this safety hazard and enhances the safety of the entire circuit system.

[0111] Furthermore, the solution's relatively simple circuit structure reduces the number of electronic components and complicated wiring required compared to some complex power supply circuits. This not only reduces circuit complexity, making maintenance and troubleshooting easier, but also lowers hardware costs, improving the overall cost-effectiveness of the power supply system and providing strong support for large-scale applications.

[0112] This solution uses two or more composite fire detectors including smoke, temperature, infrared, ultraviolet, image, and gas. Through the combination and optimization of multiple sensors, it realizes multi-dimensional, multi-sensor centralized monitoring of the surrounding environment. It solves the problem of single-point, multi-functional, and three-dimensional exploration with a one-time construction. It can detect fire hazards at an extremely early stage, judge the fire by itself, and automatically upload it to the fire alarm controller or remote platform.

[0113] An embodiment of the present application further discloses a multi-variable composite fire detection system, comprising a processor, wherein the processor executes the steps of any one of the multi-variable composite fire detection methods described above.

[0114] An embodiment of the present application further discloses a storage medium, wherein the storage medium stores a program, and when the program is executed by a processor, the steps of any one of the multi-variable composite fire detection methods described above are implemented.

[0115] Reference Figures 6-11, the embodiment of the present application also discloses a multi-variable composite fire detection device, including an integrated packaged basic functional module, which integrates a point-type smoke detector, a temperature detector and an infrared composite fire detector to realize multi-dimensional fire signal acquisition. The basic functional module further includes a control / storage module, a communication / transmission module (supporting RS485, Ethernet TCP and Internet of Things protocols), a signal / button module and a backup battery module to form a core functional unit. The basic functional module and the dedicated quick functional connector adopt a split packaging design, and the external functional modules (such as sensor interface, communication interface, matching interface) can be directly replaced through the connector. The dedicated quick functional connector is provided with an elastic guide structure inside to maintain a fixed installation direction when tightened, and has anti-impact, anti-vibration, high and low temperature resistance and high humidity protection performance. It also includes an alarm unit, which is electrically connected to the basic functional module for issuing an alarm signal.

[0116] The basic functional module consists of the module body, the outer shell and the expansion plug: The module body includes integrated buttons / indicators, transmission / communication interfaces (including RS485, Ethernet, and dry contacts), control / transmission interfaces, and upgrade / IO input and output interfaces.

[0117] The outer shell adopts a detachable and interchangeable connection method combined with an expansion plug to form a closed protective structure, providing installation and fixing space for the module body.

[0118] The matching interface ensures reliable electrical connection with external equipment (such as fire sound and light alarms, fire emergency cut-off devices, and fire extinguishers), supports equipment linkage function, and does not require secondary control of the fire alarm controller.

[0119] This solution adopts a self-healing networking and communication mechanism. The device supports self-healing networking functions, including a three-level architecture of coordinator, router and terminal node: The coordinator is responsible for network establishment, address allocation, and topology maintenance. When the network is offline, it connects to the platform through a 4G / 5G module or a ring network.

[0120] Routers are used to implement multi-hop data forwarding and maintain neighbor node information.

[0121] Terminal nodes are used to collect and upload data, including detectors and connected devices.

[0122] The self-healing process is as follows: After the device is powered on, the coordinator initializes the network and the preset nodes automatically join; When the host device loses connection with the platform, it automatically switches to the backup coordinator; When a single device is offline, it starts the network discovery mechanism and tries to join other available networks; When the entire LAN goes offline, the coordinator triggers a self-healing mechanism and restores the connection through the wireless module.

[0123] This device features built-in storage components, supporting periodic data storage and resumable data transmission. The communication module integrates the MQTT protocol, temporarily storing data in a local cache queue during network outages and sending it sequentially to the MQTT broker upon network restoration. It supports dual-mode transmission: local wired (RS485, Ethernet, USB, RS232, etc.) and remote wireless (4G / 5G), automatically switching based on network status.

[0124] The power supply of this device adopts a dual power supply design of power supply and backup battery, which can achieve seamless switching through the following methods: Integrated IC method: manages the main power supply and battery power paths and automatically switches the power supply path.

[0125] Diode isolation method: Use Schottky diodes to isolate the main power supply from the battery, so that the battery can provide seamless power when the main power supply fails.

[0126] MOSFET switching method: The gate voltage is used to control the on / off connection between the battery and the load, and the battery is turned on to supply power when the main power supply is abnormal.

[0127] When the battery capacity is insufficient, the reporting frequency and sampling times are automatically reduced to extend the battery life. After the main power is restored, it switches back to the original state and charges the battery.

[0128] This solution's positioning module includes an integrated Beidou / GPS positioning chip and antenna interface, enabling real-time device location information. The remote upgrade feature supports both FOTA (Free-Over-The-Air) wireless upgrades and local wired upgrades (UART, SPI, RS485, Ethernet, etc.). By distributing upgrade packages via a remote platform, the system automatically updates firmware, adjusts parameters (such as alarm thresholds and sensitivity), and transmits self-test results.

[0129] The power supply circuit includes: Non-polarity input circuit: full-bridge rectifier electronic components are used to achieve automatic correction of input voltage polarity.

[0130] Wide voltage design: The input range covers 100% to 200% of the nominal voltage, and overvoltage / undervoltage protection is achieved through a wide voltage integrated IC.

[0131] Anti-interference design: The external guide structure of the module body is separated from the blocking structure to ensure that the blocking operation does not affect the positioning of the internal module.

[0132] This device is suitable for complex indoor environments and long-distance and large-span scenarios. It supports extremely early fire warning, multi-device linkage and remote operation and maintenance management. Through multi-sensor fusion, modular expansion and self-healing network, it significantly improves system reliability and reduces operation and maintenance costs.

[0133] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A multivariable composite fire detection device, characterized in that: include: Basic functional modules include basic modules, outer shells and expansion plugs. Basic modules include control / storage modules, communication / transmission modules, signal / button modules and backup batteries. The outer shell and the expansion plug are detachably connected; Multi-sensor fusion module, including two or more sensors among smoke, temperature, infrared, ultraviolet, image, and gas, for multi-dimensional, multi-sensor centralized monitoring of the surrounding environment; The matching interface is electrically connected to the basic functional module and is used to link with the fire sound and light alarm, fire emergency cut-off and fire extinguisher equipment and facilities; A data storage module, including a storage element for storing data; Transmission module, including Ethernet TCP interface, Internet of Things interface, fire alarm controller dedicated interface, RS485 interface; Power supply module, including power supply and backup battery; Positioning module, including Beidou / GPS positioning unit, used to obtain real-time positioning information; The alarm unit is electrically connected to the basic function module and is used to send out an alarm signal.

2. A multivariable composite fire detection method, based on the multivariable composite fire detection device according to claim 1, characterized in that: The steps include: acquiring first sensing data based on a first sensor, and extracting first sensing target data, a first sensing position, and a first timestamp from the first sensing data; acquiring second sensing data based on the second sensor, and extracting second sensing target data, a second sensing position, and a second timestamp from the second sensing data; Calculating a first timestamp difference between the first timestamp and the second timestamp; calculating a first position data difference between the first sensing position and the second sensing position; Calculating a first process matching value between the first timestamp difference and the first position data difference; If the first process matching value is less than a preset reference matching value, a first target difference between the first sensing target data and the second sensing target data is calculated; if the first target difference is greater than or equal to a preset reference target value, a sensor failure warning prompt is issued, and the first sensing target data and the second sensing target data are output; If the first process matching value is greater than or equal to a preset reference matching value, calculating first sensing fusion data of the first sensing target data and the second sensing target data; If the first sensing fusion data is greater than or equal to the preset fusion reference value, a fire alarm prompt is output; otherwise, a fire warning prompt is output.

3. The multivariable composite fire detection method according to claim 2, characterized in that: The method further comprises the steps of: acquiring third sensing data based on a third sensor, and extracting third sensing target data, a third sensing position, and a third timestamp from the third sensing data; Calculating a second timestamp difference between the third timestamp and the second timestamp; calculating a second position data difference between the third sensing position and the second sensing position; Calculating a second process matching value of the second timestamp difference and the second position data difference; If the second process matching value is less than a preset reference matching value, a second target difference between the third sensing target data and the second sensing target data is calculated, and a comprehensive target difference is calculated based on the first target difference and the second target difference. If the comprehensive target difference is greater than or equal to a preset reference target value, a sensor failure warning prompt is issued, and the first sensing target data, the second sensing target data, and the third sensing target data are output; If the second process matching value is greater than or equal to the preset reference matching value, the second sensing fusion data of the third sensing target data and the second sensing target data is calculated; the comprehensive sensing fusion data is calculated based on the first sensing fusion data and the second sensing fusion data. If the comprehensive sensing fusion data is greater than the preset fusion reference value, a fire alarm prompt is output, otherwise a fire warning prompt is output.

4. The multivariable composite fire detection method according to claim 3, characterized in that: The first sensor, the second sensor and the third sensor are each a smoke sensor, a temperature sensor, an infrared sensor, an ultraviolet sensor, an image sensor and a gas sensor, and the first sensor, the second sensor and the third sensor are different sensors.

5. The multivariable composite fire detection method according to claim 2, characterized in that: The step of acquiring first sensing data based on the first sensor and extracting first sensing target data, a first sensing position, and a first timestamp from the first sensing data further includes the following sub-steps: The first sensor is an image sensor, and the first sensor data is picture data captured by the image sensor; Extracting open flame, temperature, and smoke targets from the image data according to a preset first recognition algorithm, wherein the first sensing target data is image confidence values ​​of the open flame, temperature, and smoke targets; Extracting the time of the open flame, temperature and smoke targets as a first timestamp; The image positions of the open flame, temperature, and smoke targets in the image data are calculated, and a physical position is matched from a preset position database as the first sensing position according to the image position.

6. The multivariable composite fire detection method according to claim 5, characterized in that: The step of acquiring second sensing data based on the second sensor and extracting second sensing target data, second sensing position and second timestamp from the second sensing data further includes the following sub-steps: The second sensor is a non-image sensor, and the second sensor data is electrical signal data collected by the non-image sensor; extracting open flame, temperature, and smoke characteristics from the electrical signal data according to a preset second recognition algorithm, wherein the second sensing target data is the value of the electrical signal data of the open flame, temperature, and smoke characteristics; extracting the time of the open flame, temperature and smoke characteristics as a second timestamp; A physical position of the non-image sensor is acquired as the second sensing position.

7. The multivariable composite fire detection method according to claim 2 or 6, characterized in that: The method further comprises the steps of: With time as the horizontal axis and the first sensing target data as the vertical axis, a first characteristic change curve of the first sensor is constructed; With time as the horizontal axis and the second sensing target data as the vertical axis, construct a second characteristic change curve of the second sensor; identifying, based on the first characteristic change curve and the second characteristic change curve, change trends of the first characteristic change curve and the second characteristic change curve, and calculating consistency of the change trends as a curve matching value; Calculating a reasonable time difference and a reasonable position data difference respectively according to the first timestamp difference and the first position data difference; Reasonable time difference = first position data difference / first sensing target data; Reasonable position data difference = first timestamp difference × first sensing target data; Calculate the difference between the reasonable time difference and the first timestamp difference as a comprehensive time difference; Calculate the difference between the reasonable position data difference and the first position data difference as the comprehensive position difference; Calculating a reasonable time-distance value based on the integrated time difference and the integrated position difference; Reasonable value of time-distance = k / (comprehensive position difference 2 +Comprehensive time difference 2 ), where k is a tuning parameter; According to the curve matching value and the reasonable time-distance value, the first process matching value is calculated: The first process matching value = α × curve matching value + β × (time-distance reasonable value), where α + β = 1.

8. The multivariable composite fire detection method according to claim 2, characterized in that: The step of calculating first sensing fusion data of the first sensing target data and the second sensing target data further includes the following sub-steps: Calculating the first sensing target data as a first normalized value using a maximum-minimum normalization algorithm; Calculating the second sensing target data as a second normalized value using a maximum-minimum normalization algorithm; The first sensing fusion data = w1×first normalized value + w2×second normalized value; the weight of the first sensing target data is w1, the weight of the second sensing target data is w2, w1+w2=1.

9. The multivariable composite fire detection method according to claim 2, characterized in that: The step of calculating the first process matching value of the first timestamp difference and the first position data difference further includes the following sub-steps: Obtaining a calculation amount of a first process matching value for calculating a difference between the first timestamp and the first position data; If the calculation amount is less than or equal to the set calculation amount threshold, or the difference between the first position data is less than or equal to the set position threshold, the first sensor data and the second sensor data are sent to the background system for calculation; If the calculation amount is greater than the set calculation amount threshold, or the difference of the first position data is greater than the set position threshold, the first sensor data and the second sensor data are sent to the corresponding relay station for calculation, and then sent by the relay station to the background system for calculation; Based on multiple relay stations, the master and slave are defined based on the fire detector where the first sensor is located. The master is the relay station connected to the fire detector where the first sensor is currently located, and the remaining relay stations are slaves. The master and slaves are associated; Obtain host status feedback information at a set frequency; If the status feedback information indicates an abnormal state, a connection is established with the nearest slave.

10. The multivariable composite fire detection method according to claim 2, characterized in that: The power supply system of the composite fire detector includes a power supply and a backup battery. The method further includes the following steps: Obtaining a power input voltage, and if the power input voltage is within an abnormal power range, disconnecting the power supply, switching to a backup battery, and issuing a warning prompt; If the power input voltage is within the normal power range, disconnect the backup battery and switch to the power supply to charge the backup battery; The entire power supply is electrically connected to the anode of the first Schottky diode, the positive electrode of the backup battery is connected to the anode of the second Schottky diode, the cathodes of the first Schottky diode and the second Schottky diode are connected together and then connected to the positive electrode of the load.

11. A multivariable composite fire detection system, characterized in that: The method comprises a processor, wherein the processor executes the steps of the multivariable composite fire detection method according to any one of claims 2 to 10.

12. A storage medium, characterized in that: The medium stores a program, and when the program is executed by a processor, the steps of the multi-variable composite fire detection method according to any one of claims 2 to 10 are implemented.

Images