A Fire Smoke Particle Detection Method and System for Smart Fire Protection

Abstract

The present application relates to the field of smoke particle detection technology, and specifically to a fire smoke particle detection method and system for smart firefighting. The method comprises: obtaining the fire smoke particle concentration and temperature at each location within each time window, obtaining the fire smoke particle abnormal fluctuation coefficient and temperature rise significance value at each location within each time window, and then obtaining the smoke particle diffusion significance at each location within each time window; predicting the smoke particle diffusion significance of adjacent locations at each location at the current moment to obtain the smoke particle state influence coefficient at each location at the current moment; and predicting the fire smoke concentration data at each location at the current moment. During the prediction process, the smoothing coefficient of the prediction algorithm is optimized according to the smoke particle state influence coefficient at each location at the current moment, so as to obtain the fire smoke particle concentration through the predicted value. The present application can improve the accuracy of fire smoke particle concentration distribution detection.

Description

Technical Field

[0001] The present application relates to the field of smoke particle detection technology, and specifically to a fire smoke particle detection method and system for smart firefighting. Background Art

[0002] Smoke is a significant fire product and a fundamental parameter for fire detection. Fire smoke primarily originates from incomplete and insufficient combustion and is composed of a high-temperature aerosol system consisting of solid particles, carbon monoxide, carbon dioxide, and decomposition products. Optical smoke sensors are widely used in fire smoke detection due to their simplicity and cost-effectiveness. Photoelectric smoke sensors emit a pulsed light signal. When this light signal encounters airborne particles, it scatters, and smoke concentration is detected by detecting the scattered light signal.

[0003] After a fire breaks out, real-time monitoring of smoke concentrations in different areas can quickly identify early signs of a fire. Smoke concentration data can also be used to determine the specific location and scope of the fire. By analyzing the changing trends of smoke concentrations in different areas, the direction and speed of fire spread can be predicted, facilitating the rational deployment of firefighting resources. During a building fire, smoke diffusion is easily affected by the building structure and complex airflow, resulting in a high degree of uncertainty in the speed and range of smoke diffusion, leading to inaccurate detection of smoke particle concentrations. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of this application is to provide a fire smoke particle detection method and system for smart firefighting. The technical solutions adopted are as follows:

[0005] The present application provides a method for detecting fire smoke particles for smart firefighting, including the following steps:

[0006] Obtain the fire smoke particle concentration and temperature data at each location at each collection moment within each time window;

[0007] Perform time series decomposition on the fire smoke particle concentration at each location in each time window. Based on the data variation degree of the trend item after time series decomposition and the data fluctuation of the residual item, obtain the abnormal fluctuation coefficient of the fire smoke particle at each location in each time window. Based on the temperature variation at each location in each time window, obtain the significant temperature rise value at each location in each time window. Combine the abnormal fluctuation coefficient of the fire smoke particle and the significant temperature rise value to obtain the smoke particle diffusion significance at each location in each time window.

[0008] For each location, a prediction algorithm is used to predict the smoke particle diffusion significance of its neighboring locations at the current moment based on the smoke particle diffusion significance of its neighboring locations in multiple time windows before the current moment, so as to obtain the smoke particle state influence coefficient of each location at the current moment;

[0009] Based on the average level of fire smoke particle concentration at each location in each time window before the current moment, the fire smoke concentration data of each location at the current moment is predicted. During the prediction process, the smoothing coefficient of the prediction algorithm is optimized according to the smoke particle state influence coefficient of each location at the current moment to obtain the fire smoke particle concentration through the predicted value.

[0010] Preferably, the time window is a set preset duration.

[0011] Preferably, the abnormal fluctuation coefficient of fire smoke particles at each location in each time window is:

[0012] For the trend term and residual term of the fire smoke particle concentration at each location in each time window, the trend term is fitted to obtain the corresponding curve, the cumulative sum of the slopes of all data points in the curve is calculated, and the product of the standard deviation and the mean of the residual term is obtained. The ratio of the cumulative sum to the product is used as the abnormal fluctuation coefficient of the fire smoke particles at each location in each time window.

[0013] Preferably, obtaining the significant temperature rise value at each location in each time window further includes:

[0014] The fire smoke concentration data of each location in each time window are fitted, and the collection time of the data point corresponding to the maximum curvature value in the fitting curve is obtained. The range of the temperature data at the collection time of the data point corresponding to the maximum curvature value and the N times before and after it is calculated. The ratio of the range to N is used as the temperature rise significance value of each location in each time window.

[0015] Preferably, the smoke particle diffusion significance at each location in each time window is the product of the abnormal fluctuation coefficient of the fire smoke particles at each location in each time window and the temperature rise significance value.

[0016] Preferably, the acquisition of the smoke particle state influence coefficient at each position at the current moment includes: for each position, setting its adjacent positions, taking the average of the predicted values of the smoke particle diffusion significance of the adjacent positions at the current moment as the smoke particle state influence coefficient at each position at the current moment.

[0017] Preferably, the method for obtaining the adjacent positions of each position is: taking each position as the center, taking the positions within a circular area with a preset radius as the adjacent positions of each position.

[0018] Preferably, the predicting of the fire smoke concentration data at each location at the current moment further includes:

[0019] The mean fire smoke concentration data of each location in each time window of a preset number of time windows before the current moment is obtained and used as the input of the exponential smoothing algorithm to predict the fire smoke concentration of each location at the current moment.

[0020] Preferably, the optimizing the smoothing coefficient of the prediction algorithm further comprises:

[0021] The smoke particle state influence coefficient at each position at the current moment is normalized, and the obtained normalized result is used as the smoothing coefficient of the exponential smoothing algorithm.

[0022] An embodiment of the present application also provides a fire smoke particle detection system for smart firefighting, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of any one of the above-mentioned fire smoke particle detection methods for smart firefighting are implemented.

[0023] As can be seen from the above, the fire smoke particle detection method and system for smart fire protection provided by this application have at least the following beneficial effects:

[0024] This application calculates the sustained diffusion coefficient of smoke particle concentration by deeply analyzing the changing characteristics of fire smoke particle concentration data collected during the pyrolysis and smoldering processes in the early stages of a fire, as well as the correlation between fire smoke concentration and temperature changes. Furthermore, the application combines the mutual influence of smoke particle concentrations between adjacent locations within the fire building under rapid diffusion to calculate the sustained diffusion coefficient of smoke particle concentration. Compared to methods that rely solely on smoke sensor detection, this method considers the diffusion characteristics of smoke particle concentration generated in the early stages of combustible combustion, enabling prediction of smoke particle concentration variation, thereby improving the accuracy of fire smoke particle concentration distribution detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0026] Figure 1 This application provides a flowchart of the steps of a fire smoke particle detection method for smart firefighting. DETAILED DESCRIPTION

[0027] In order to further illustrate the technical means and effects adopted by this application to achieve the predetermined invention objectives, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features and effects of a fire smoke particle detection method and system for smart firefighting proposed in this application. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics of one or more embodiments may be combined in any suitable form.

[0028] Unless otherwise specified and limited, terms such as "comprises", "includes" or any other variants thereof are intended to cover non-exclusive inclusion, so that a circuit structure, article or device comprising a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the article or device comprising the element. In addition, the term "and\or" used herein includes any and all combinations of one or more related listed items. All technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs.

[0029] The following describes in detail a fire smoke particle detection method and system for smart firefighting provided by the present application with reference to the accompanying drawings.

[0030] See also Figure 1 , which shows a flowchart of a fire smoke particle detection method for smart firefighting provided by an embodiment of the present application, including the following steps:

[0031] Step 1: Obtain the fire smoke particle concentration and temperature data at each location at each collection moment within each time window.

[0032] Combustion can be divided into three stages: pyrolysis, smoldering, and open flame. The early stages of a fire, during the pyrolysis and smoldering of the substance, are characterized by the release of gases and aerosols. Aerosols, also known as smoke, are small particles suspended in the air. Aerosols produced by different substances during the pyrolysis and smoldering stages of combustion are concentrated within a certain particle diameter range. The particle diameters of water vapor, dust, and oil smoke particles are larger than those produced by combustion. Photoelectric smoke sensors essentially determine smoke concentration by measuring the intensity of light scattered by smoke particles. The intensity of the scattered light received within the sensor increases as the concentration of particulate matter increases. Conventional photoelectric smoke sensors only detect the presence of a certain concentration of particulate matter, but non-fire smoke, such as water vapor, dust, and oil smoke particles, also produce similar scattered light.

[0033] The dual-wavelength photoelectric smoke alarm emits two light sources of different wavelengths. As the size of the particles changes, the scattering intensity ratio of the different wavelength light sources will change, thereby identifying different types of smoke. Therefore, in order to obtain the concentration of fire smoke particles generated when a fire occurs, this embodiment uses a dual-wavelength photoelectric smoke alarm to collect fire smoke concentration data at each location in the room and the corridor. The temperature during the spread of the fire has a certain degree of influence on the detection of smoke particles, and temperature data is collected at each smoke alarm location through a temperature sensor. In this embodiment, taking any electric smoke alarm in a building fire as an example, starting from the collection of fire smoke concentration data, every 5 minutes is set as a time window, and in this embodiment, the time interval for collecting fire smoke particle data and temperature data is set to 0.1 seconds.

[0034] Step 2: Perform time series decomposition on the fire smoke particle concentration at each location in each time window. Based on the data change degree of the trend item after time series decomposition and the data fluctuation of the residual item, obtain the abnormal fluctuation coefficient of the fire smoke particles at each location in each time window. According to the temperature change at each location in each time window, obtain the significant temperature rise value at each location in each time window. Combined with the abnormal fluctuation coefficient of the fire smoke particles and the significant temperature rise value in each time window, obtain the smoke particle diffusion significance at each location in each time window.

[0035] When a fire breaks out in a building, combustible materials first produce flammable gases, followed by large amounts of smoke. Then, with sufficient combustion-supporting gases, the combustibles begin to burn, generating significant heat, which raises the ambient temperature and further spreads the fire. As can be seen, smoke is generated in the initial stages of a fire. During this period, the ambient temperature has not yet reached its peak, and small aerosol particles produced by the burning materials are released into the air. Dual-wavelength photoelectric smoke alarms installed indoors typically detect small aerosol particles during this phase and sound a fire alarm when the concentration of small aerosol particles exceeds the threshold. Before the fire is fully ignited, the rate of increase in smoke concentration gradually increases, and this increase can fluctuate within a small range. The higher the rate of increase in smoke concentration, the faster the smoke spreads during a fire. Once the fire reaches the open flame stage, the fire spreads rapidly. The heat generated by the combustion causes the ambient temperature to continue to rise. During this temperature rise, the smoke concentration in the environment rises to a maximum and then remains relatively stable. When the temperature begins to increase at a high rate and the smoke concentration has just reached its maximum, it indicates that smoke particles from the fire are rapidly dispersing. Therefore, this embodiment deeply considers the changing characteristics of the smoke particle concentration generated in the early stage of a fire, as well as the correlation between the smoke concentration state and the temperature rise degree, so as to detect smoke particles.

[0036] In addition, as the smoke concentration increases, the smoke will spread to the surroundings at a certain speed, thereby affecting the detection of smoke particles in the surrounding environment. Based on this feature, it is helpful to more accurately identify and predict the concentration of fire smoke particles.

[0037] First, the relevant variation characteristics of the smoke concentration data during its rise are obtained. For each location, taking the u-th location as an example in this embodiment, the fire smoke particle concentration data collected in each time window exhibits a certain degree of high and low fluctuations due to the non-smooth rise process of smoke concentration. The rate of rise and the degree of non-smooth fluctuation can reflect the corresponding fire development status. Therefore, to obtain the overall variation trend characteristics of the smoke particle data, this embodiment uses the time series decomposition algorithm STL (Seasonal Trend decomposition using Loess) to decompose the fire smoke particle data in each time window at the u-th location, obtaining the corresponding trend term and residual term.

[0038] For the fire smoke particle concentration data at the u-th location in the i-th time window, where the resulting trend term is discrete data, the least squares method is used to fit the trend term to obtain a corresponding curve to facilitate analysis of the rate of change at each data point in the trend term. The slope of the curve at each data point is then determined. When smoke concentration increases rapidly, it is usually in the early stages of smoldering combustion. The rate of increase in smoke concentration can reflect the severity of the fire, and the higher the rate of increase in fire smoke concentration, the faster the smoke spread at that location. The slope of each data point in the curve is then obtained, and the cumulative sum of all these slopes is used as the smoke concentration trend significance at the u-th location in the i-th time window. The smoke concentration trend significance reflects the rate of change of the detected fire smoke particles.

[0039] Furthermore, rising smoke can be affected by airflow changes or insufficient concentration, resulting in uneven fluctuations. The greater the unevenness, the less concentrated the distribution of smoke particles at that location. Conversely, when smoke concentration steadily rises, it indicates that smoke particles in the air are densely distributed, and the fire at that location is relatively severe. The residual term can reflect the uneven fluctuation characteristics of the data. This embodiment analyzes the data fluctuations of the residual term by calculating its standard deviation and mean. For ease of understanding and presentation, this embodiment uses the product of the standard deviation and mean of the residual term at the u-th location in the i-th time window as the random fluctuation coefficient of smoke concentration at the u-th location in the i-th time window. The random fluctuation coefficient of smoke concentration reflects the changing characteristics of smoke particles during the rise of a fire.

[0040] Furthermore, the ratio of the smoke concentration trend significance at each location in each time window to the random fluctuation coefficient of smoke concentration is used as the abnormal fluctuation coefficient of fire smoke particles at each location in each time window, and the abnormal fluctuation coefficient of fire smoke particles at the u-th location in the i-th time window is recorded as The larger the abnormal fluctuation coefficient of fire smoke particles is, the more it indicates that the smoke particle concentration at the corresponding position in the time window is in a stage of rapid abnormal growth.

[0041] Furthermore, as the fire spreads, the fire smoke concentration at a certain position gradually reaches a maximum value, and within the local range where the maximum value is reached, the corresponding temperature data may be in a rapid growth stage, indicating that a significant open flame may soon appear at the corresponding position. Therefore, in this embodiment, the fire smoke concentration data in each time window of the u-th position is fitted, and the acquisition time of the data point corresponding to the maximum curvature value in the fire smoke concentration data fitting curve is obtained, and then the range of the temperature data at the acquisition time of the data point corresponding to the maximum curvature value and the N moments before and after it is calculated, where N is 10 and can be set by the implementer. Further, the ratio of the range to N is used as the significant temperature rise value of each position in each time window, and the significant temperature rise value of the u-th position in the i-th time window is recorded as , which reflects the degree of change in ambient temperature when fire smoke concentration reaches its maximum value. A larger abnormal fluctuation coefficient of fire smoke particles and a higher significant temperature rise value within the time window indicate a more severe smoke particle diffusion state at that location within the corresponding time window.

[0042] Furthermore, this embodiment combines the abnormal fluctuation coefficient of fire smoke particles and the significant value of temperature rise in each time window to obtain the smoke particle diffusion significance at each location in each time window. The smoke particle diffusion significance at each location in each time window is the product of the abnormal fluctuation coefficient of fire smoke particles and the significant value of temperature rise at each location in each time window. For ease of understanding, the specific calculation formula in this embodiment is: , where is the smoke particle diffusion significance at the u-th position in the i-th time window. The smoke particle diffusion significance reflects the change in smoke concentration during the early pyrolysis and smoldering process of a fire, as well as the correlation characteristics between the smoke concentration state and temperature changes. The greater the smoke particle diffusion significance, the more likely the smoke particle concentration at this position is to rise further.

[0043] Step 3: For each location, based on the smoke particle diffusion significance of its neighboring locations in multiple time windows before the current moment, a prediction algorithm is used to predict the smoke particle diffusion significance of the neighboring locations at the current moment, so as to obtain the smoke particle state influence coefficient of each location at the current moment.

[0044] Because smoke particles are inherently hotter and less dense, they tend to diffuse upward due to thermal buoyancy. Smoke diffusion paths are complex, influenced by building structure and natural ventilation conditions. Within a burning building, smoke typically diffuses at a constant rate, preferentially upward and toward locations with greater temperature differences. Therefore, adjacent smoke alarms are more likely to affect each other.

[0045] Taking a smoke sensor at any location as an example, considering that air can circulate between adjacent locations, this embodiment will analyze the fire smoke particle data from adjacent locations. First, the adjacent locations of each location are set. In actual application scenarios, the implementer can set this freely, and this embodiment does not impose any special restrictions on this. Preferably, in this embodiment, taking the u-th location as an example, the u-th location is used as the center location, and the locations within a circular area with a radius of R are used as the adjacent locations of the u-th location. The value of R can be set by the implementer and is set to 1 meter in this embodiment.

[0046] Furthermore, for the u-th position, the smoke particle diffusion significance of each of its neighboring positions in M consecutive time windows before the current moment is arranged in ascending chronological order to obtain a significance value sequence corresponding to each neighboring position. In this embodiment, the value of M is 10. The higher the smoke particle diffusion significance of the neighboring position, the greater the impact on the smoke particle state at the central position. Thus, the smoke particle diffusion significance of each neighboring position at the current moment is predicted.

[0047] Specifically, for the u-th position, this embodiment uses a linear exponential smoothing algorithm to predict the saliency value sequence corresponding to each adjacent position, outputting a predicted smoke particle diffusion saliency value for each adjacent position at the current moment. The average of the predicted smoke particle diffusion saliency values for all adjacent positions at the current moment is taken as the smoke particle state influence coefficient for the central position at the current moment. The resulting smoke particle state influence coefficient reflects the degree to which smoke detection at the central position is influenced by the surrounding smoke state.

[0048] At this point, the above process of this embodiment is repeated to obtain the smoke particle state influence coefficient at each position at the current moment.

[0049] Step 4: Based on the average level of fire smoke particle concentration at each location in each time window before the current moment, the fire smoke concentration data at each location at the current moment is predicted. During the prediction process, the smoothing coefficient of the prediction algorithm is optimized according to the smoke particle state influence coefficient of each location at the current moment to obtain the fire smoke particle concentration through the predicted value.

[0050] Furthermore, fires can cause complex indoor airflow fluctuations, making smoke concentration data collected by smoke particle sensors susceptible to interference from airflow fluctuations, resulting in inaccurate sensor data. However, smoke diffusion characteristics can, to a certain extent, reflect the impact of airflow fluctuations on smoke concentration. Therefore, by incorporating the localized smoke diffusion characteristics, we can achieve accurate prediction of smoke particle data.

[0051] Specifically, the mean fire smoke concentration data for each location within each of the M time windows prior to the current moment is first calculated. A single exponential smoothing algorithm is then used to predict and analyze the smoke concentration data at each location at the current moment. During the prediction process, the smoothing coefficient of the exponential smoothing algorithm is optimized. Considering that a larger smoke particle state influence coefficient indicates a greater degree of short-term variability in the smoke concentration data at that location, a larger smoothing coefficient is required; a smaller smoke particle state influence coefficient indicates a smaller short-term variability in the smoke concentration data, and a smaller smoothing coefficient can be set. Therefore, in this embodiment, a sigmoid function is used to normalize the smoke particle state influence coefficient at each location at the current moment. The resulting normalization result is used as the smoothing coefficient. The mean fire smoke concentration data for each location within each of the M time windows prior to the current moment is used as input to the single exponential smoothing algorithm. The fire smoke concentration data at each location at the current moment is predicted, and the resulting predicted value is used as the corrected result of the fire smoke particle detection.

[0052] This embodiment deeply analyzes the changing characteristics of fire smoke particle concentration data collected during the pyrolysis and smoldering processes in the early stages of a fire, as well as the correlation characteristics between the fire smoke concentration state and temperature changes. Combined with the mutual influence between adjacent smoke sensors in a building fire, this embodiment calculates the smoke particle state influence coefficient, which reflects the degree of influence of smoke particle diffusion on the detection of surrounding smoke particles. The smoothing coefficient is optimized based on the smoke particle state influence coefficient, and the smoke concentration data is corrected, which helps to improve the accuracy of particulate matter concentration detection.

[0053] Based on the same inventive concept as the above method, an embodiment of the present application also provides a fire smoke particle detection system for smart firefighting, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-mentioned fire smoke particle detection methods for smart firefighting.

[0054] It should be understood that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0055] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0056] The above content is only an implementation method of the present application and is not intended to limit the scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the scope of protection of the present application.

Claims

1. A fire smoke particle detection method for smart firefighting, characterized in that: The following steps are involved: Obtain the fire smoke particle concentration and temperature data at each location at each collection moment within each time window; Perform time series decomposition on the fire smoke particle concentration at each location in each time window, fit the trend term and residual term of the fire smoke particle concentration at each location in each time window to obtain a corresponding curve, calculate the cumulative sum of the slopes of all data points in the curve, and obtain the product of the standard deviation of the residual term and the mean, and use the ratio of the cumulative sum to the product as the abnormal fluctuation coefficient of the fire smoke particles at each location in each time window. Fit the fire smoke concentration data at each location in each time window, and obtain the collection time of the data point corresponding to the maximum curvature value in the fitting curve. Calculate the range of temperature data at the collection time of the data point corresponding to the maximum curvature value and N times before and after it, and use the ratio of the range to N as the temperature rise significance value at each location in each time window. Use the product of the abnormal fluctuation coefficient of the fire smoke particles at each location in each time window and the temperature rise significance value as the smoke particle diffusion significance at each location in each time window. For each location, its adjacent locations are set. Based on the smoke particle diffusion significance of its adjacent locations in multiple time windows before the current moment, a prediction algorithm is used to predict the smoke particle diffusion significance of the adjacent locations at the current moment. The predicted values of the smoke particle diffusion significance of the adjacent locations at the current moment are averaged as the smoke particle state influence coefficient of each location at the current moment. According to the average level of fire smoke particle concentration at each location in each time window before the current moment, the fire smoke concentration data of each location at the current moment is predicted. During the prediction process, the smoke particle state influence coefficient of each location at the current moment is normalized, and the obtained normalized result is used as the smoothing coefficient of the exponential smoothing algorithm to obtain the fire smoke particle concentration through the predicted value.

2. The fire smoke particle detection method for smart firefighting according to claim 1, characterized in that: The time window is a preset duration.

3. The fire smoke particle detection method for smart firefighting according to claim 1, characterized in that: The method for obtaining the adjacent positions of each position is: taking each position as the center, the positions within a circular area with a preset radius are used as the adjacent positions of each position.

4. The fire smoke particle detection method for smart firefighting according to claim 1, characterized in that: The prediction of the fire smoke density data at each location at the current moment further includes: The mean fire smoke concentration data of each location in each time window of a preset number of time windows before the current moment is obtained and used as the input of the exponential smoothing algorithm to predict the fire smoke concentration of each location at the current moment.

5. A fire smoke particle detection system for smart firefighting, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of a fire smoke particle detection method for smart firefighting as described in any one of claims 1 to 4 are implemented.

Images