Smoke and fire early warning device and method based on Internet of Things

Abstract

The invention relates to the technical field of fire early warning, in particular to a smoke and fire early warning device based on the Internet of Things, which comprises a historical fire data calling module used for historical fire data of a target area; the real-time sensor probability calculation module is used for calculating the real-time sensor probability of fire occurrence; the probability comparison module is used for comparing the real-time temperature probability with a real-time temperature threshold value, comparing the real-time smoke probability with a real-time smoke threshold value and comparing the real-time CO probability with a real-time CO threshold value; the probability weight data set calling module is used for calling probability weight data sets corresponding to the real-time temperature probability, the real-time smoke probability and the real-time CO probability; the target preset combination determining module is used for determining a corresponding target preset combination in the probability weight data set; and the alarm triggering module is used for triggering an alarm. The fire early warning accuracy of the smoke and fire early warning device in the large-scale commercial complex can be improved.

Description

Technical Field

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

[0002] With the acceleration of urbanization, the scale and complexity of large commercial complexes (such as shopping malls) are increasing, and the requirements for fire safety are becoming higher and higher. Traditional fire warning systems usually rely on data from a single sensor for judgment and are easily affected by environmental interference and false alarms.

[0003] At present, the existing fire and smoke warning systems have the following problems: due to the influence of environmental factors (such as smoke generated by cooking, high temperature weather, etc.), traditional systems are prone to false alarms; traditional systems lack the ability to dynamically adjust and cannot optimize the weights of sensors according to environmental changes.

[0004] In order to solve the above problems, the present invention proposes a fireworks warning method based on the Internet of Things, which significantly improves the accuracy, reliability and adaptability of the fireworks warning system through technical means such as multi-sensor fusion, combination of historical data and real-time data, dynamic weight adjustment, anomaly detection and fault diagnosis, and spatial correlation analysis.

[0005] From the above, we can see that how to improve the accuracy of fire warnings of smoke and fire warning devices in large commercial complexes remains to be solved. Summary of the invention

[0006] In order to improve the accuracy of fire warning by fireworks warning devices in large commercial complexes, the present application provides a fireworks warning device and method based on the Internet of Things.

[0007] In the first aspect, the present application provides a fire and smoke warning device based on the Internet of Things, which adopts the following technical solution: A fire and smoke warning device based on the Internet of Things, comprising: The historical fire data retrieval module is used to retrieve the historical fire data corresponding to the target area in the complex, and obtain the fire occurrence probability corresponding to the target area based on the historical fire data ; The real-time sensor probability calculation module obtains the real-time sensor data corresponding to the sensors in the target area and calculates the real-time sensor probability of fire in the target area based on the real-time sensor data reading based on the Bayesian formula. , where the sensors include temperature sensors, smoke sensors and CO sensors, the historical data include temperature historical data, smoke historical data and CO historical data, and the real-time sensor probability There are multiple,real-time sensor probabilities including real-time temperature probability,,real-time smoke probability, and real-time CO probability; The probability comparison module calls a preset sensor probability threshold set, the sensor probability threshold set includes a real-time temperature threshold, a real-time smoke threshold and a real-time CO threshold, and is used to compare the real-time temperature probability with the real-time temperature threshold, the real-time smoke probability with the real-time smoke threshold, and the real-time CO probability with the real-time CO threshold to obtain corresponding comparison results; The probability weight data set retrieval module is used to retrieve the probability weight data set corresponding to the real-time temperature probability, the real-time smoke probability and the real-time CO probability if one or more of the following situations exist in the comparison result: the real-time temperature probability is greater than the real-time temperature threshold, the real-time smoke probability is greater than the real-time smoke threshold, and the real-time CO probability is greater than the real-time CO threshold, wherein the probability weight data set includes multiple preset combinations of the three weight data of the real-time temperature weight, the real-time smoke weight and the real-time CO weight; a target preset combination determination module, which determines a corresponding target parameter based on the comparison result, and is used to determine a corresponding target preset combination in a probability weight data set based on the target parameter, wherein the target parameter includes one or more of a real-time temperature probability, a real-time smoke probability, and a real-time CO probability; The alarm triggering module determines the corresponding target fire probability based on the target preset combination and the real-time temperature probability, the real-time smoke probability and the real-time CO probability, compares the target fire probability with a preset fire warning threshold, and triggers an alarm if the target fire probability is greater than the fire warning threshold.

[0008] In the second aspect, the present application provides a fire and smoke warning method based on the Internet of Things, which adopts the following technical solutions: A fire and smoke early warning method based on the Internet of Things, comprising: retrieving historical fire data corresponding to a target area in a complex, and obtaining a fire occurrence probability corresponding to the target area based on the historical fire data ; Calculate the real-time sensor probability of fire in the target area based on the real-time sensor data readings based on the Bayesian formula , where the sensors include temperature sensors, smoke sensors and CO sensors, the historical data include temperature historical data, smoke historical data and CO historical data, and the real-time sensor probability There are multiple, the real-time sensor probability includes the real-time temperature probability, the real-time smoke probability and the real-time CO probability; call the preset sensor probability threshold set, the sensor probability threshold set includes the real-time temperature threshold, the real-time smoke threshold and the real-time CO threshold, compare the real-time temperature probability with the real-time temperature threshold, compare the real-time smoke probability with the real-time smoke threshold, and compare the real-time CO probability with the real-time CO threshold to obtain the corresponding comparison result; if the comparison result has one or more of the following situations: the real-time temperature probability is greater than the real-time temperature threshold, the real-time smoke probability is greater than the real-time smoke threshold, and the real-time CO probability is greater than the real-time CO threshold, call the real-time temperature probability, the real-time smoke probability and the real-time CO threshold a probability weight data set corresponding to the probability, wherein the probability weight data set includes multiple preset combinations of three weight data of real-time temperature weight, real-time smoke weight and real-time CO weight; determining the corresponding target parameters based on the comparison result, and determining the corresponding target preset combination in the probability weight data set based on the target parameters, wherein the target parameters include one or more of real-time temperature probability, real-time smoke probability and real-time CO probability; determining the corresponding target fire probability based on the target preset combination and the real-time temperature probability, real-time smoke probability and real-time CO probability, comparing the target fire probability with a preset fire warning threshold, and triggering an alarm if the target fire probability is greater than the fire warning threshold.

[0009] By adopting the above technical solution, by using historical data to predict the probability of fire, a baseline is provided to help more accurately assess the current situation; by using the Bayesian formula for probability calculation, the possibility of fire can be dynamically adjusted according to prior knowledge and real-time data to improve the accuracy of prediction; in addition, by setting reasonable thresholds, it is possible to effectively distinguish between normal status and potential fire risks, reduce false alarms and missed alarms, and use different weight combinations according to the contribution of different sensors to ensure that the most reasonable judgment can be made in different situations. Based on the above process, the accuracy of fire warning can be improved.

[0010] Optionally, it includes: retrieving historical data corresponding to the sensor, and calculating the mean and standard deviation corresponding to each sensor based on the historical data; obtaining the latest sensor data corresponding to the sensor, and calculating the corresponding abnormal value based on the mean and standard deviation of the latest sensor data, wherein is the mean, is the standard deviation, and x is the latest sensor data; retrieving a pre-set abnormal threshold, comparing the absolute value of the abnormal value with the abnormal threshold, if the absolute value of the abnormal value is greater than the abnormal threshold, determining the corresponding target abnormal sensor data, and generating a corresponding abnormal report for visual display.

[0011] By adopting the above technical solutions, through real-time monitoring and analysis of sensor data, abnormal conditions can be detected before a fire occurs, thereby achieving early warning; by using the statistical parameters (mean and standard deviation) of historical data, it is possible to more accurately determine whether the current data is within the normal range, thereby reducing false positives and false negatives. Once abnormal data is detected, the system can immediately generate an abnormal report and remind relevant personnel through a visual interface to speed up the response.

[0012] Optionally, after each fire occurs, sensor probability thresholds corresponding to all sensors in the target area are collected; the cumulative number of times for each sensor that the real-time temperature probability is greater than the real-time temperature threshold, the real-time smoke probability is greater than the real-time smoke threshold, and the real-time CO probability is greater than the real-time CO threshold is calculated respectively, and the cumulative number of times includes the temperature number, the smoke number, and the CO number; based on the temperature number, the smoke number, and the CO number, the sensor adjustment priority in the target area is determined, wherein the sensor adjustment priority includes the priority between different types of sensors and the priority between sensors of the same type at different installation positions; and parameters of different types of sensors in the target area are adjusted based on the sensor adjustment priority.

[0013] By adopting the above technical solution and analyzing the data after the fire, it is possible to identify which sensors performed poorly or frequently triggered alarms when a fire occurred, so that targeted adjustments can be made to improve the performance and accuracy of the sensors; by adjusting the priority of the sensors, key signals can be detected more quickly when a fire occurs, speeding up the response and improving the overall system's reaction efficiency; in addition, dynamically adjusting sensor thresholds and priorities can help reduce false alarms and missed alarms caused by improper sensor settings, ensuring that the system can respond accurately in actual fires.

[0014] Optionally, the method includes: obtaining environmental change data in a target area through a built-in or external sensor in the target area, wherein the environmental change data includes one or more of temperature change data, humidity change data, and light change data; calling a preset environmental model database, and determining a corresponding target typical environmental characteristic model based on the three-dimensional data of the target area, wherein the three-dimensional data is 3D model data of the target area, and the environmental model database stores typical environmental characteristic models of different installation environments and different types of buildings; inputting the environmental change data into the target typical environmental characteristic model to obtain working parameters corresponding to a fireworks warning device, and adjusting all sensors in the target area based on the working parameters, wherein the working parameters include at least a trigger threshold, a sampling frequency, and an alarm delay time of the sensor.

[0015] By adopting the above technical solutions, through real-time monitoring and analysis of environmental change data, the system can automatically adjust the working parameters of the sensor to better adapt to the current environmental conditions and improve the accuracy and reliability of detection. Dynamic adjustment of the sensor's trigger threshold, sampling frequency and alarm delay time helps reduce false alarms and missed alarms caused by environmental changes, ensuring that the system can respond accurately under different conditions.

[0016] Optionally, it includes: acquiring historical fire data and real-time environmental data, preprocessing the historical fire data and the real-time environmental data, the historical fire data including the time, location, cause, environmental conditions, building type and structural information of the fire in the target area; configuring a machine learning algorithm to train the historical fire data to generate a fire risk assessment model; inputting the real-time environmental data into the fire risk assessment model, calculating the fire risk probability in the future, and providing risk warnings based on the fire risk probability and a preset risk level classification.

[0017] By adopting the above technical solutions, the system can provide early warning before the fire actually occurs by predicting the probability of fire risk in the future; the system can dynamically adjust the fire risk assessment according to real-time environmental data, adapt to changes in different time and environmental conditions, and improve the flexibility and response speed of the system; according to different risk levels, the system can help optimize the allocation of fire protection resources and ensure that more resources and attention are invested in high-risk areas.

[0018] In a third aspect, the present application provides a fire and smoke warning device based on the Internet of Things, which adopts the following technical solution: A fire and smoke warning device based on the Internet of Things comprises a processor, wherein the processor runs a program of any one of the fire and smoke warning methods based on the Internet of Things described above.

[0019] In a fourth aspect, the present application provides a storage medium, which adopts the following technical solution: A storage medium stores a program of any one of the above-mentioned fireworks warning methods based on the Internet of Things.

[0020] In summary, the present application includes at least one of the following beneficial technical effects: Improve the accuracy of early warning: By integrating historical data and real-time data, the system can more accurately assess fire risks and improve the accuracy of early warnings; Early warning: Through real-time monitoring and anomaly detection, the system can provide early warnings before the fire actually occurs, buying more time for preventive measures; Reduce false alarms and missed alarms: By setting reasonable thresholds and using statistical parameters, the system can effectively distinguish between normal conditions and potential fire risks, reducing false alarms and missed alarms; Dynamic adjustment: The system can dynamically adjust the working parameters and priorities of sensors based on real-time environmental data and post-fire data analysis, thereby improving the flexibility and response speed of the system; Decision support: The risk warnings and abnormal reports provided can provide decision support for managers and help them take appropriate preventive measures; Optimize resource allocation: According to different risk levels, the system can help optimize the allocation of fire-fighting resources to ensure that more resources and attention are invested in high-risk areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 The figure is a flow chart of a fire and smoke warning method based on the Internet of Things according to an exemplary embodiment.

[0022] Figure 2 It is a structural block diagram of a smoke and fire warning device based on the Internet of Things according to an exemplary embodiment. DETAILED DESCRIPTION

[0023] Embodiments of the present application are described in detail below, examples of which are shown in the accompanying drawings.

[0024] In the description of this specification, the description with 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. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0025] The present application embodiment discloses a fire warning method based on the Internet of Things, referring to Figure 1 ,include: S100, retrieve the historical fire data corresponding to the target area in the complex, and obtain the fire occurrence probability corresponding to the target area based on the historical fire data .

[0026] Among them, historical fire data in the target area is extracted from the database or storage system. The historical fire data may include the time, location, cause, environmental conditions (such as temperature, humidity, wind speed, etc.), building type and structural information, etc. of the fire. The historical fire data can provide a basis for subsequent probability calculations. By analyzing this data, a baseline can be established to assess the fire risk in the current situation.

[0027] In addition, by analyzing historical fire data, statistical methods or machine learning algorithms are used to calculate the probability of fire in the target area. This can be achieved by calculating the frequency of fires over a certain period of time or using more complex models. Thus, historical data can be used to obtain a baseline for assessing the fire risk in the current situation, so as to more accurately predict the possibility of future fires and provide background information for the processing of real-time sensor data.

[0028] S110, obtaining real-time sensor data corresponding to sensors in the target area, and calculating the real-time sensor probability of fire in the target area based on the real-time sensor data readings based on the Bayesian formula .

[0029] The sensors include temperature sensors, smoke sensors and CO sensors, the historical data include temperature historical data, smoke historical data and CO historical data, and the real-time sensor probability There are multiple,real-time sensor probabilities including real-time temperature probability,,real-time smoke probability, and real-time CO probability.

[0030] In addition, real-time data is obtained from temperature sensors, smoke sensors, and CO sensors in the target area. The sensors continuously monitor the temperature, smoke concentration, and CO level in the environment; the actual conditions of the current environment can serve as a key basis for determining whether there is a fire risk, allowing the system to respond to potential fire threats in a timely manner.

[0031] For real-time sensor probability ,It should be pointed out here that the Bayesian formula is used to ,combine historical data and real-time sensor data to calculate the real-time probability of fire ,occurrence under the current reading. Specifically, for each sensor (temperature, smoke, CO), the probability of fire ,occurrence under the current reading is calculated.

[0032] The general form of the Bayesian formula is: ,in, is the probability of A occurring given B (posterior probability), is the probability (likelihood) of B occurring if A occurs, is the prior probability of A occurring, is the marginal probability of B occurring. The Bayesian formula can dynamically adjust the possibility of fire based on prior knowledge (historical data) and new evidence (real-time data) to improve the accuracy of prediction. The above scheme takes into account the combined impact of historical data and real-time data, making the calculation of fire probability more scientific and reliable.

[0033] S120, calling a preset sensor probability threshold set, the sensor probability threshold set includes a real-time temperature threshold, a real-time smoke threshold and a real-time CO threshold, comparing the real-time temperature probability with the real-time temperature threshold, comparing the real-time smoke probability with the real-time smoke threshold, and comparing the real-time CO probability with the real-time CO threshold to obtain corresponding comparison results.

[0034] Among them, the sensor probability threshold set includes real-time temperature threshold, real-time smoke threshold and real-time CO threshold. The above threshold data is set according to experience or data analysis during the system design phase, and can be used to distinguish between normal status and potential fire risks. By setting reasonable thresholds, the system can reduce false alarms and missed alarms, ensuring timely warnings when there is a real fire risk.

[0035] The real-time temperature probability is compared with the real-time temperature threshold, the real-time smoke probability is compared with the real-time smoke threshold, and the real-time CO probability is compared with the real-time CO threshold to obtain the comparison result of each sensor probability with the corresponding threshold. It should be pointed out here that the result of comparing the real-time smoke probability with the real-time smoke threshold may be that the real-time smoke probability is greater than the real-time smoke threshold, the real-time smoke probability is equal to the real-time smoke threshold, and the real-time smoke probability is less than the real-time smoke threshold. According to the same principle, there are three results for comparing the real-time smoke probability with the real-time smoke threshold and comparing the real-time CO probability with the real-time CO threshold.

[0036] Through the above comparison, it is possible to determine which sensors' data are outside the normal range, indicating that there may be a fire risk. This comparison helps to quickly identify sensor data that requires further attention, thereby improving the system's response speed and accuracy.

[0037] S130, if the comparison result shows that the real-time temperature probability is greater than the real-time temperature threshold, the real-time smoke probability is greater than the real-time smoke threshold, and the real-time CO probability is greater than the real-time CO threshold, one or more of the following situations is found, the probability weight data set corresponding to the real-time temperature probability, the real-time smoke probability, and the real-time CO probability is retrieved.

[0038] Among them, firstly, by comparing the real-time temperature probability with the real-time temperature threshold, comparing the real-time smoke probability with the real-time smoke threshold, comparing the real-time CO probability with the real-time CO threshold, and obtaining the results of these comparisons, it is determined which sensors' probabilities exceed their corresponding thresholds. If there is at least one sensor whose probability is greater than its corresponding threshold in the comparison result (for example, the real-time temperature probability is greater than the real-time temperature threshold, the real-time smoke probability is greater than the real-time smoke threshold, and the real-time CO probability is greater than the real-time CO threshold), the pre-set probability weight data set is retrieved. It should be pointed out here that the probability weight data set includes multiple preset combinations of three weight data: real-time temperature weight, real-time smoke weight, and real-time CO weight, and each combination includes real-time temperature weight, real-time smoke weight, and real-time CO weight. By retrieving the appropriate weight data set, the system can dynamically adjust the fire risk assessment method according to the current environmental conditions and sensor data, which helps to improve the flexibility and accuracy of the system.

[0039] S140, determining corresponding target parameters based on the comparison result, and determining corresponding target preset combinations in the probability weight data set based on the target parameters.

[0040] Among them, the target parameters include one or more of the real-time temperature probability, the real-time smoke probability and the real-time CO probability; according to the comparison result, the sensors exceeding the threshold and their corresponding probabilities are determined. For example, if the real-time temperature probability is greater than the real-time temperature threshold, and the real-time smoke probability is greater than the real-time smoke threshold, then the target parameters include the real-time temperature probability and the real-time smoke probability.

[0041] In addition, based on the determined target parameters, the most appropriate preset combination is selected from the probability weight data set. For example, if only the real-time temperature probability exceeds the threshold, a weight combination focusing on the temperature effect is selected; if both the real-time temperature probability and the real-time smoke probability exceed the threshold, a weight combination that comprehensively considers the temperature and smoke effects is selected; if the probabilities of all sensors exceed the threshold, a weight combination that comprehensively considers the effects of all factors is selected.

[0042] By comparing the results, we can determine which sensors’ data are beyond the normal range, so as to identify the sensors that need special attention. Different sensors have different impacts on fire risk. By selecting the appropriate weight combination, we can more reasonably assess the fire risk and improve the flexibility and accuracy of the system.

[0043] S150, determine the corresponding target fire probability based on the target preset combination and the real-time temperature probability, real-time smoke probability and real-time CO probability, compare the target fire probability with the preset fire warning threshold, and trigger an alarm if the target fire probability is greater than the fire warning threshold.

[0044] The final target fire probability is calculated by using the weights in the selected target preset combination, combining the real-time temperature probability, the real-time smoke probability, and the real-time CO probability. The specific calculation method may be weighted average or other appropriate algorithms. For example: ,in, , and are the selected weights, , and is the corresponding real-time sensor probability. The calculated target fire probability is compared with the preset fire warning threshold. The fire warning threshold is a threshold set based on experience or data analysis during the system design phase to determine whether to trigger an alarm; if the target fire probability is greater than the fire warning threshold, an alarm is triggered. The alarm can be issued by sound, light or other means to remind relevant personnel to take emergency measures.

[0045] By comprehensively considering the data from multiple sensors and their weights, a more comprehensive and accurate fire risk probability is calculated. By setting a reasonable fire warning threshold, the system can decide whether to trigger an alarm, ensuring that a warning is issued in a timely manner when there is a real fire risk.

[0046] In the embodiment of the present application, for the real-time data of each sensor, data anomaly detection is also performed, and the specific process includes: S111, retrieve historical data corresponding to the sensor, and calculate the mean and standard deviation corresponding to each sensor based on the historical data.

[0047] The historical data of each sensor (such as temperature sensor, smoke sensor, CO sensor) in the target area is extracted from the database or storage system. These data usually include reading records over a period of time.

[0048] Calculate the mean: For each sensor's historical data, calculate its mean. The mean is the average of all data points, reflecting the central trend of the data; the formula is: , where N is the number of data points, is the ith data point.

[0049] Calculate standard deviation: Calculate the standard deviation of historical data for each sensor. The standard deviation measures the dispersion of the data, that is, the degree of deviation between the data points and the mean; the formula is: .

[0050] S112, obtaining the latest sensor data corresponding to the sensor, based on the average of the latest sensor data and standard deviation Calculate the corresponding abnormal value ,in, is the mean, is the standard deviation, and x is the latest sensor data.

[0051] Among them, the latest reading data is obtained from each sensor (such as temperature sensor, smoke sensor, CO sensor) in the target area, and the abnormal value represents the degree of deviation between the latest reading and the historical mean, in units of standard deviation.

[0052] By normalizing the latest sensor data into anomaly values, the system can more easily compare data from different sensors and determine the extent to which they deviate from the normal range. Anomaly values ​​can help the system quickly identify which sensor data may be outside the normal range for further analysis.

[0053] S113, calling a preset abnormal threshold, comparing the absolute value of the abnormal value with the abnormal threshold, if the absolute value of the abnormal value is greater than the abnormal threshold, determining the corresponding target abnormal sensor data, generating a corresponding abnormal report for visual display.

[0054] Among them, the preset abnormal threshold is called. The abnormal threshold is a value set according to experience or data analysis during the system design phase to distinguish normal data from abnormal data. The absolute value of the calculated abnormal value is compared with the abnormal threshold. If the absolute value of the abnormal value is greater than the abnormal threshold, the sensor data is considered abnormal.

[0055] By setting reasonable anomaly thresholds, the system can effectively distinguish between normal data and abnormal data and reduce false positives and negatives; once abnormal data is detected, the system can immediately take corresponding measures to improve response speed and accuracy.

[0056] In addition, if the absolute value of the abnormal value is greater than the abnormal threshold, the system generates an abnormal report, which contains the data of the abnormal sensor, timestamp, abnormal value and other information. The abnormal report is displayed to the user or system administrator through a visual interface. The visual interface can include charts, alarm lights, sound alarms, etc., so that users can intuitively see the abnormal situation.

[0057] At the same time, in the process of monitoring the environmental data of the target area, it is necessary to determine whether the environmental data has changed. The specific process includes: S114, obtaining environmental change data in the target area through a built-in or external sensor in the target area.

[0058] Among them, built-in or external sensors are installed in the target area. These sensors may include temperature sensors, humidity sensors, and light sensors. The sensors continuously monitor and collect environmental change data in the target area, including one or more of temperature change data, humidity change data, and light change data. By collecting multiple types of environmental data (such as temperature, humidity, and light), it is helpful to have a more comprehensive understanding of the environmental status and improve the accuracy of early warning.

[0059] S115, calling a preset environment model database, and determining a corresponding target typical environment feature model based on the three-dimensional data of the target area, Among them, the pre-set environmental model database is retrieved, which stores typical environmental feature models of different installation environments (such as indoors and outdoors) and different types of buildings (such as residences, factories, shopping malls, etc.). Based on the three-dimensional data (3D model data) of the target area, the typical environmental feature model that best matches the target area is selected from the environmental model database. The three-dimensional data provides the spatial structure information of the target area, which helps to accurately select the appropriate model.

[0060] By using a preset environmental model database, the system can select the most appropriate operating parameters according to different environments and building types, which can improve the adaptability and accuracy of the system. And based on historical data and typical environmental characteristic models, the system can better predict and adjust operating parameters to adapt to current environmental conditions.

[0061] S116, inputting the environmental change data into the target typical environmental characteristic model to obtain the working parameters corresponding to the fireworks warning device, and adjusting all sensors in the target area based on the working parameters.

[0062] Among them, the environmental change data (such as temperature, humidity, and light) obtained from the sensor is input into the selected target typical environmental characteristic model. The target typical environmental characteristic model calculates the working parameters of the fireworks warning device suitable for the current environmental conditions based on the input environmental change data. These working parameters include at least the sensor's trigger threshold, sampling frequency, and alarm delay time.

[0063] According to the calculated trigger threshold, adjust the trigger threshold of each sensor (such as temperature sensor, smoke sensor, CO sensor). The trigger threshold determines under what conditions the sensor will sound an alarm.

[0064] Adjust the sampling frequency of each sensor based on the calculated sampling frequency. The sampling frequency determines how often the sensor reads data, and a higher sampling frequency can provide more timely data updates.

[0065] According to the calculated alarm delay time, adjust the alarm delay time of the sensor after detecting an abnormal situation. Reasonable alarm delay time can reduce false alarms and ensure that the alarm is issued in time when there is a real risk of fire.

[0066] By executing steps S114 to S116, the system can dynamically adjust the working parameters of the fireworks warning device according to the real-time environmental change data of the target area, thereby improving the adaptability and accuracy of the system.

[0067] For the monitoring of commercial complexes, real-time monitoring is required, and the system needs certain predictive performance to take timely remedial measures, so the following steps are also performed: S117, acquiring historical fire data and real-time environmental data, and preprocessing the historical fire data and the real-time environmental data.

[0068] Among them, historical fire data in the target area is extracted from the database or storage system. These data include the time, location, cause, environmental conditions (such as temperature, humidity, wind speed, etc.), building type and structural information, etc. of the fire. Real-time environmental data in the target area is collected through built-in or external sensors (such as temperature sensors, humidity sensors, and light sensors), including temperature change data, humidity change data, and light change data.

[0069] Preprocessing includes: data cleaning: removing invalid or erroneous data points and filling missing values; standardization / normalization: converting data of different dimensions to the same scale for subsequent processing; feature extraction: extracting useful features from the raw data, such as temperature change rate, smoke concentration peak, CO level fluctuation, etc.; data segmentation: dividing historical fire data into training sets and test sets for model training and verification.

[0070] Historical fire data provides important information about the frequency, causes, and environmental conditions of fires in the area, providing a basis for subsequent probability calculations; while real-time environmental data reflects the actual conditions of the current environment and is the key basis for determining whether there is a fire risk.

[0071] S118, configuring a machine learning algorithm to train historical fire data and generate a fire risk assessment model.

[0072] The selected machine learning algorithm is trained using historical fire data in the training set, and the trained model is verified using the test set to evaluate its performance indicators, such as accuracy, recall, F1 score, etc. If the model performance is poor, the algorithm parameters can be adjusted or other algorithms can be tried until satisfactory performance is achieved. After training and verification, the final fire risk assessment model is generated. The model can predict the probability of fire risk in the future based on the input historical fire data and real-time environmental data.

[0073] S119, inputs real-time environmental data into the fire risk assessment model, calculates the probability of fire risk in the future, and provides risk warnings based on the fire risk probability and the preset risk level classification.

[0074] Among them, real-time environmental data (such as temperature, humidity, light, etc.) are input into the generated fire risk assessment model, and the fire risk probability in the future period is calculated based on the input real-time environmental data.

[0075] Different risk levels are pre-set, such as low risk, medium risk, and high risk. Each risk level corresponds to a range of fire risk probability. For example: low risk: fire risk probability < 0.2; medium risk: 0.2 ≤ fire risk probability < 0.5; high risk: fire risk probability ≥ 0.5.

[0076] According to the calculated fire risk probability, the risk level is determined and corresponding risk warnings are provided. Risk warnings can be communicated to relevant personnel through a visual interface, sound alarm or other means. For example, if the calculated fire risk probability is 0.6, it belongs to the high risk level, the system will issue a high risk alarm and recommend emergency measures.

[0077] The notification executes steps S117 to S119. By predicting the probability of fire risk in the future, the system can provide early warning before the fire actually occurs, buying more time for preventive measures. According to different risk levels, the system can help optimize the allocation of firefighting resources and ensure that more resources and attention are invested in high-risk areas.

[0078] Finally, in the embodiment of the present application, the following steps are also performed: S121, after each fire occurs, sensor probability thresholds corresponding to all sensors in the target area are collected.

[0079] Among them, after each fire occurs, the real-time data of all sensors (such as temperature sensors, smoke sensors, and CO sensors) in the target area are collected, and the real-time readings of each sensor during the fire and the corresponding probability thresholds are recorded. These thresholds are pre-set during the system design phase or based on historical data. By collecting sensor data and thresholds during the fire, the system can obtain the actual performance in the fire situation and provide basic data for subsequent analysis.

[0080] S122, respectively calculating the cumulative number of times for each sensor that the real-time temperature probability is greater than the real-time temperature threshold, the real-time smoke probability is greater than the real-time smoke threshold, and the real-time CO probability is greater than the real-time CO threshold.

[0081] For each sensor type (temperature, smoke, CO), compare its real-time probability during a fire with the corresponding probability threshold. Calculate the number of times the real-time temperature probability is greater than the real-time temperature threshold (temperature times); calculate the number of times the real-time smoke probability is greater than the real-time smoke threshold (smoke times); calculate the number of times the real-time CO probability is greater than the real-time CO threshold (CO times).

[0082] By counting the number of times the threshold is exceeded, the system can quantify the performance of each sensor during a fire and determine which sensors trigger alarms more frequently; a high frequency of triggering may indicate that the sensor is too sensitive or improperly set up and needs further adjustment.

[0083] S123, determining the sensor adjustment priority in the target area based on the temperature count, the smoke count, and the CO count.

[0084] Among them, the adjustment priority between different types of sensors is determined based on the temperature times, smoke times and CO times. For example, if the temperature sensor is triggered significantly more times than other sensors, the temperature sensor has a higher adjustment priority. At the same time, for sensors of the same type (such as multiple temperature sensors), the adjustment priority between them is determined based on the trigger times at different installation locations.

[0085] By determining the adjustment priority, the system can allocate resources reasonably and give priority to adjusting those sensors that perform poorly or frequently trigger alarms. At the same time, reasonable adjustment priorities help improve the overall accuracy and reliability of the system and reduce false positives and negatives.

[0086] S124, adjusting parameters of various sensors in the target area based on the sensor adjustment priority.

[0087] According to the determined adjustment priority, adjust the parameters of various sensors in the target area. These parameters include but are not limited to: trigger threshold: adjust the trigger threshold of the sensor to make it more suitable for the current environmental conditions and reduce false alarms; sampling frequency: adjust the sampling frequency of the sensor to increase the response speed or reduce power consumption; alarm delay time: adjust the alarm delay time of the sensor after detecting an abnormal situation to ensure that the alarm is issued in time when there is a real risk of fire. By adjusting the sensor parameters, the system can better adapt to the current environmental conditions and improve the accuracy and reliability of detection.

[0088] The present application embodiment discloses a fire warning device based on the Internet of Things, referring to Figure 2 , including but not limited to: The historical fire data retrieval module 200 is used to retrieve the historical fire data corresponding to the target area in the complex, and obtain the fire occurrence probability corresponding to the target area based on the historical fire data; A real-time sensor probability calculation module 210 obtains real-time sensor data corresponding to sensors in a target area, and calculates the real-time sensor probability of a fire occurring in the target area under the real-time sensor data reading based on the Bayesian formula, wherein the sensor includes a temperature sensor, a smoke sensor, and a CO sensor, the historical data includes temperature historical data, smoke historical data, and CO historical data, and there are multiple real-time sensor probabilities, and the real-time sensor probabilities include real-time temperature probability, real-time smoke probability, and real-time CO probability; The probability comparison module 220 retrieves a preset sensor probability threshold set, the sensor probability threshold set includes a real-time temperature threshold, a real-time smoke threshold, and a real-time CO threshold, and is used to compare the real-time temperature probability with the real-time temperature threshold, compare the real-time smoke probability with the real-time smoke threshold, and compare the real-time CO probability with the real-time CO threshold to obtain corresponding comparison results; The probability weight data set retrieval module 230 is used to retrieve the probability weight data set corresponding to the real-time temperature probability, the real-time smoke probability and the real-time CO probability if one or more of the following situations exist in the comparison result: the real-time temperature probability is greater than the real-time temperature threshold, the real-time smoke probability is greater than the real-time smoke threshold, and the real-time CO probability is greater than the real-time CO threshold, wherein the probability weight data set includes multiple preset combinations of the three weight data of the real-time temperature weight, the real-time smoke weight and the real-time CO weight; A target preset combination determination module 240 determines a corresponding target parameter based on the comparison result, and is used to determine a corresponding target preset combination in a probability weight data set based on the target parameter, wherein the target parameter includes one or more of a real-time temperature probability, a real-time smoke probability, and a real-time CO probability; The alarm trigger module 250 determines the corresponding target fire probability based on the target preset combination and the real-time temperature probability, the real-time smoke probability and the real-time CO probability, compares the target fire probability with the preset fire warning threshold, and triggers an alarm if the target fire probability is greater than the fire warning threshold.

[0089] Furthermore, the device includes but is not limited to: A historical data retrieval module is used to retrieve the historical data corresponding to the sensor and calculate the mean and standard deviation corresponding to each sensor based on the historical data; The latest sensor data acquisition module is used to obtain the latest sensor data corresponding to the sensor, based on the latest sensor data mean and standard deviation Calculate the corresponding abnormal value ,in, is the mean, is the standard deviation, x is the latest sensor data; The abnormal threshold retrieval module is used to retrieve the pre-set abnormal threshold and compare the absolute value of the abnormal value with the abnormal threshold. If the absolute value of the abnormal value is greater than the abnormal threshold, the corresponding target abnormal sensor data is determined and the corresponding abnormal report is generated for visual display.

[0090] Furthermore, the device includes but is not limited to: A sensor probability threshold collection module is used to collect sensor probability thresholds corresponding to all sensors in the target area after each fire occurs; The cumulative times calculation module is used to calculate the cumulative times of each sensor in which the real-time temperature probability is greater than the real-time temperature threshold, the real-time smoke probability is greater than the real-time smoke threshold, and the real-time CO probability is greater than the real-time CO threshold. The cumulative times include temperature times, smoke times, and CO times. A sensor adjustment priority determination module, which is used to determine the sensor adjustment priority in the target area based on the temperature count, the smoke count and the CO count, wherein the sensor adjustment priority includes the priority between different types of sensors and the priority between sensors of the same type at different installation positions; The adjustment module is used to adjust parameters of various sensors in the target area based on the sensor adjustment priority.

[0091] Furthermore, the device includes but is not limited to: An environmental change data acquisition module is used to acquire environmental change data in a target area through a built-in or external sensor in the target area, and the environmental change data includes one or more of temperature change data, humidity change data, and light change data; An environment model database retrieval module is used to retrieve a preset environment model database, and determine a corresponding target typical environment feature model based on the three-dimensional data of the target area, wherein the three-dimensional data is the 3D model data of the target area, and the environment model database stores typical environment feature models of different installation environments and different types of buildings; The working parameter acquisition module inputs the environmental change data into the target typical environmental characteristic model to obtain the corresponding working parameters of the fireworks warning device, and adjusts all sensors in the target area based on the working parameters, wherein the working parameters at least include the sensor's trigger threshold, sampling frequency and alarm delay time.

[0092] Furthermore, the device includes but is not limited to: A preprocessing module acquires historical fire data and real-time environmental data and is used to preprocess the historical fire data and real-time environmental data. The historical fire data includes the time, location, cause, environmental conditions, building type and structural information of the fire in the target area; The training module is used to configure the machine learning algorithm to train historical fire data and generate a fire risk assessment model; The fire risk probability calculation module inputs real-time environmental data into the fire risk assessment model to calculate the fire risk probability in the future and provide risk warnings based on the fire risk probability and preset risk level classification.

[0093] The embodiment of the present application further discloses a fire and smoke warning method based on the Internet of Things, including a processor, wherein the program of any one of the fire and smoke warning methods based on the Internet of Things described above is run in the processor.

[0094] The embodiment of the present application further discloses a storage medium storing a program of any one of the above-mentioned fireworks warning methods based on the Internet of Things.

[0095] 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 fire and smoke warning device based on the Internet of Things, characterized in that: include: The historical fire data retrieval module is used to retrieve the historical fire data corresponding to the target area in the complex, and obtain the fire occurrence probability corresponding to the target area based on the historical fire data ; The real-time sensor probability calculation module obtains the real-time sensor data corresponding to the sensors in the target area and calculates the real-time sensor probability of fire in the target area based on the real-time sensor data reading based on the Bayesian formula. , where the sensors include temperature sensors, smoke sensors and CO sensors, the historical data include temperature historical data, smoke historical data and CO historical data, and the real-time sensor probability There are multiple,real-time sensor probabilities including real-time temperature probability,,real-time smoke probability, and real-time CO probability; The probability comparison module calls a preset sensor probability threshold set, the sensor probability threshold set includes a real-time temperature threshold, a real-time smoke threshold and a real-time CO threshold, and is used to compare the real-time temperature probability with the real-time temperature threshold, the real-time smoke probability with the real-time smoke threshold, and the real-time CO probability with the real-time CO threshold to obtain corresponding comparison results; The probability weight data set retrieval module is used to retrieve the probability weight data set corresponding to the real-time temperature probability, the real-time smoke probability and the real-time CO probability if one or more of the following situations exist in the comparison result: the real-time temperature probability is greater than the real-time temperature threshold, the real-time smoke probability is greater than the real-time smoke threshold, and the real-time CO probability is greater than the real-time CO threshold, wherein the probability weight data set includes multiple preset combinations of the three weight data of the real-time temperature weight, the real-time smoke weight and the real-time CO weight; a target preset combination determination module, which determines a corresponding target parameter based on the comparison result, and is used to determine a corresponding target preset combination in a probability weight data set based on the target parameter, wherein the target parameter includes one or more of a real-time temperature probability, a real-time smoke probability, and a real-time CO probability; The alarm triggering module determines the corresponding target fire probability based on the target preset combination and the real-time temperature probability, the real-time smoke probability and the real-time CO probability, compares the target fire probability with a preset fire warning threshold, and triggers an alarm if the target fire probability is greater than the fire warning threshold.

2. A fire and smoke early warning method based on the Internet of Things, characterized in that: include: Retrieve historical fire data corresponding to the target area in the complex, and obtain the fire occurrence probability corresponding to the target area based on the historical fire data ; Obtain the real-time sensor data corresponding to the sensors in the target area, and calculate the real-time sensor probability of fire in the target area based on the real-time sensor data readings based on the Bayesian formula , where the sensors include temperature sensors, smoke sensors and CO sensors, the historical data include temperature historical data, smoke historical data and CO historical data, and the real-time sensor probability There are multiple,real-time sensor probabilities including real-time temperature probability,,real-time smoke probability, and real-time CO probability; Retrieve a preset sensor probability threshold set, the sensor probability threshold set includes a real-time temperature threshold, a real-time smoke threshold, and a real-time CO threshold, compare the real-time temperature probability with the real-time temperature threshold, compare the real-time smoke probability with the real-time smoke threshold, and compare the real-time CO probability with the real-time CO threshold to obtain corresponding comparison results; If the comparison result shows one or more of the following situations: the real-time temperature probability is greater than the real-time temperature threshold, the real-time smoke probability is greater than the real-time smoke threshold, and the real-time CO probability is greater than the real-time CO threshold, retrieve the probability weight data set corresponding to the real-time temperature probability, the real-time smoke probability, and the real-time CO probability, wherein the probability weight data set includes multiple preset combinations of the three weight data of the real-time temperature weight, the real-time smoke weight, and the real-time CO weight; Determine a corresponding target parameter based on the comparison result, and determine a corresponding target preset combination in the probability weight data set based on the target parameter, wherein the target parameter includes one or more of a real-time temperature probability, a real-time smoke probability, and a real-time CO probability; The corresponding target fire probability is determined based on the target preset combination and the real-time temperature probability, real-time smoke probability and real-time CO probability, and the target fire probability is compared with the preset fire warning threshold. If the target fire probability is greater than the fire warning threshold, an alarm is triggered.

3. The method for early warning of fireworks based on the Internet of Things according to claim 2, characterized in that: include: Retrieving historical data corresponding to the sensor, and calculating the mean and standard deviation corresponding to each sensor based on the historical data; Get the latest sensor data corresponding to the sensor, based on the average of the latest sensor data and standard deviation Calculate the corresponding abnormal value ,in, is the mean, is the standard deviation, x is the latest sensor data; The preset abnormality threshold is retrieved, and the absolute value of the abnormal value is compared with the abnormality threshold. If the absolute value of the abnormal value is greater than the abnormality threshold, the corresponding target abnormal sensor data is determined, and the corresponding abnormality report is generated for visual display.

4. The method for early warning of fireworks based on the Internet of Things according to claim 2, characterized in that: include: After each fire occurs, the sensor probability thresholds corresponding to all sensors in the target area are collected; Calculate the cumulative number of each sensor in which the real-time temperature probability is greater than the real-time temperature threshold, the real-time smoke probability is greater than the real-time smoke threshold, and the real-time CO probability is greater than the real-time CO threshold. The cumulative number includes the temperature number, the smoke number, and the CO number. Determining the sensor adjustment priority in the target area based on the temperature times, the smoke times, and the CO times, wherein the sensor adjustment priority includes the priority between different types of sensors and the priority between sensors of the same type at different installation positions; Parameters of various sensors in the target area are adjusted based on the sensor adjustment priority.

5. The method for early warning of fireworks based on the Internet of Things according to claim 3, characterized in that: include: Acquiring environmental change data in the target area through a built-in or external sensor in the target area, wherein the environmental change data includes one or more of temperature change data, humidity change data, and light change data; Retrieving a preset environmental model database, and determining a corresponding target typical environmental feature model based on the three-dimensional data of the target area, wherein the three-dimensional data is 3D model data of the target area, and the environmental model database stores typical environmental feature models of different installation environments and different types of buildings; The environmental change data is input into the target typical environmental characteristic model to obtain the working parameters corresponding to the fire and smoke warning device, and all sensors in the target area are adjusted based on the working parameters, wherein the working parameters at least include the trigger threshold, sampling frequency and alarm delay time of the sensor.

6. The method for early warning of fireworks based on the Internet of Things according to claim 2, characterized in that: include: Acquire historical fire data and real-time environmental data, and pre-process the historical fire data and the real-time environmental data, wherein the historical fire data includes the time, location, cause, environmental conditions, building type and structural information of the fire in the target area; Configure machine learning algorithms to train historical fire data and generate fire risk assessment models; The real-time environmental data is input into the fire risk assessment model to calculate the fire risk probability in the future, and provide risk warnings based on the fire risk probability and the preset risk level classification.

7. A fire and smoke warning device based on the Internet of Things, characterized in that: The method comprises a processor, wherein the program of the fireworks warning method based on the Internet of Things as claimed in any one of claims 2 to 6 is run in the processor.

8. A storage medium, characterized in that: A program of a fireworks warning method based on the Internet of Things as described in any one of claims 2 to 6 is stored.

Images