A method and device for detecting and alarming the concentration of combustible gas

Abstract

The present invention discloses a method and device for detecting and alarming the concentration of combustible gas, and relates to the technical field of gas detection and alarm. The method comprises: collecting combustible gas concentration data in the environment through multiple types of sensors, generating real-time concentration signals, performing noise suppression processing, and obtaining pre-processed concentration data; performing leakage prediction, generating gas leakage prediction results, and performing similarity calculations in combination with historical sensing scene data to determine the gas leakage risk level; triggering a multi-level alarm strategy, and executing the multi-level alarm strategy to link intelligent devices for safety control. The method solves the technical problems existing in the prior art that a variety of combustible gases cannot be fully detected, the detection data is easily affected by the environment, there is a lack of forward-looking prediction of the gas leakage risk level, and data privacy protection is insufficient, which leads to insufficient accuracy of combustible gas concentration detection and insufficient timely and accurate alarms. The linkage of smart homes is realized, and the technical effects of improving the accuracy of combustible gas detection and reducing the false alarm rate are achieved.

Description

Technical Field

[0001] The present application relates to the technical field related to combustible gas detection, and in particular to a combustible gas concentration detection alarm method and device. Background Art

[0002] If combustible gases such as natural gas and liquefied petroleum gas leak and mix with air to a certain concentration, they may cause explosions, fires and other serious accidents when encountering excitation sources such as electric sparks generated by turning on electrical appliances. Traditional household combustible gas detection equipment has many shortcomings. Most of them use a single sensor and can only detect specific types of combustible gases. They cannot cope with the coexistence of multiple combustible gases in complex home environments such as kitchens. Moreover, they are easily affected by environmental interference such as the noise of kitchen appliances and Wi-Fi signals during operation, resulting in inaccurate gas concentration data and difficulty in accurately determining whether a gas leak has occurred, which seriously misleads the accurate judgment of combustible gas leaks. More importantly, traditional detection methods lack early warning capabilities. They usually alarm only when the gas leak reaches a high concentration and danger is about to occur, and cannot provide early warning at the early stage of the leak. In addition, there is no comprehensive judgment based on home scenarios (such as whether someone is at home, whether there is an open flame source, etc.), which makes the risk level classification inaccurate and difficult to link smart home devices for effective safety processing, such as being unable to close gas valves and start ventilation equipment in time.

[0003] Therefore, the current relevant technologies have the problems of being unable to comprehensively detect multiple combustible gases, detection data being easily affected by environmental interference, lacking forward-looking predictions of gas leakage risk levels, and insufficient data privacy protection, which in turn lead to technical problems such as insufficient accuracy in combustible gas concentration detection and insufficiently timely and accurate alarms. Summary of the invention

[0004] The present application provides a combustible gas concentration detection alarm method and device, which solves the technical problems existing in the prior art, such as the inability to comprehensively detect multiple combustible gases, the susceptibility of detection data to environmental interference, the lack of forward-looking prediction of gas leakage risk levels, and insufficient data privacy protection, which in turn lead to insufficient combustible gas concentration detection accuracy and untimely and inaccurate alarms. The application realizes smart home linkage and achieves the technical effect of improving the accuracy of combustible gas detection and reducing the false alarm rate.

[0005] The present application provides a combustible gas concentration detection and alarm method, which includes: collecting combustible gas concentration data in an environment through multiple types of sensors to generate a real-time concentration signal, performing noise suppression processing on the real-time concentration signal to obtain pre-processed concentration data; performing leakage prediction based on the pre-processed concentration data to generate a gas leakage prediction result, performing similarity calculation based on the gas leakage prediction result combined with historical sensor scene data to determine the gas leakage risk level; triggering a multi-level alarm strategy according to the gas leakage risk level, and executing the multi-level alarm strategy to link smart devices for safety control.

[0006] In a possible implementation, the combustible gas concentration detection and alarm method also performs the following processing: obtaining a sensor data set, dynamically calibrating the sensor data set according to ambient temperature parameters and ambient humidity parameters, and determining sensor calibration data; performing gas concentration analysis based on the sensor calibration data, and generating the real-time concentration signal according to the concentration value; traversing the real-time concentration signal for pulse processing, and extracting abnormal pulse signals; using a wavelet transform algorithm to filter the abnormal pulse signal to generate a smooth concentration curve; and correcting the concentration value according to the smooth concentration curve to obtain the preprocessed concentration data.

[0007] In a possible implementation, the combustible gas concentration detection and alarm method also performs the following processing: constructing a time series concentration distribution diagram based on the smooth concentration curve, performing feature analysis according to the time series concentration distribution diagram, and determining dynamic leakage characteristics; retrieving historical sensing scene data to perform scene feature analysis and determine scene feature vectors; performing multi-dimensional similarity matching on the dynamic leakage characteristics and the scene feature vectors to generate a scene matching degree set; performing weighted clustering analysis on the scene matching degree set to obtain multiple scene feature clusters, traversing the multiple scene feature clusters for association screening, and determining an associated historical scene group, wherein the associated historical scene group includes multiple leakage event probabilities; constructing a leakage prediction model based on the multiple leakage event probabilities and the dynamic leakage characteristics, performing leakage prediction through the leakage prediction model, and generating a gas leakage prediction result; and performing risk decision analysis in combination with the gas leakage prediction result to determine the gas leakage risk level.

[0008] In a possible implementation, the method for detecting and alarming the concentration of combustible gas further performs the following processing: obtaining a plurality of historical gas usage scenario data, where the plurality of historical gas usage scenario data includes gas equipment type data, usage time period data, and environmental parameter data; performing correlation analysis on the environmental parameter data and the gas equipment type data to determine leakage event data; calculating a similarity matrix between real-time sensing scenario data and the historical sensing scenario data according to the usage time period data; performing linear regression analysis on the leakage event data based on the similarity matrix to construct the leakage prediction model.

[0009] In a possible implementation, the method for detecting and alarming the concentration of combustible gas further performs the following processing: establishing a sliding time window according to the smoothed concentration curve, calculating the concentration according to the sliding time window to obtain a concentration change rate; detecting the concentration duration of the sliding time window based on the concentration change rate to extract the peak duration; counting the number of times the concentration value in the sliding time window exceeds a preset warning line to determine the fluctuation frequency; combining the concentration change rate, the peak duration, and the fluctuation frequency as a three-dimensional feature vector to obtain the dynamic leakage feature.

[0010] In a possible implementation, the method for detecting and alarming the concentration of combustible gas further performs the following processing: performing multi-dimensional sensing acquisition based on a historical scenario to obtain multi-dimensional sensing data; performing time alignment processing on the multi-dimensional sensing data to generate a scenario data stream; using an autoencoder to perform feature dimensionality reduction on the scenario data stream to generate a scenario feature vector; associatively storing a leakage event label and the scenario feature vector to generate the historical sensing scenario data.

[0011] In a possible implementation, the method for detecting and alarming the concentration of combustible gas further performs the following processing: matching the gas leakage risk level with a multi-level alarm strategy library to generate an alarm control instruction set; parsing based on the alarm control instruction set to generate an alarm priority sequence, triggering a local alarm module to execute an alarm according to the alarm priority sequence to generate a linkage instruction; distributing the linkage instruction to a target intelligent device to control the target intelligent device to perform a safety operation.

[0012] In a possible implementation, the combustible gas concentration detection and alarm method also performs the following processing: when the gas leakage risk level is level one, the multi-level alarm strategy library is searched to determine the first alarm control instruction, the local sound and light alarm is triggered and an early warning notification is pushed through the first alarm control instruction, and level one leakage record data is generated; when the gas leakage risk level is level two, the multi-level alarm strategy library is searched to determine the second alarm control instruction, the smart gas valve is closed and the ventilation equipment is started through the second alarm control instruction, and level two leakage record data is generated; when the gas leakage risk level is level three, the multi-level alarm strategy library is searched to determine the third alarm control instruction, an emergency broadcast signal is sent to the security system through the third alarm control instruction, and level three leakage record data is generated; the level one leakage record data, the level two leakage record data, and the level three leakage record data are alarm integrated to determine a leakage alarm record log, and the alarm control instruction set is extracted according to the leakage alarm record log.

[0013] In a possible implementation, the combustible gas concentration detection alarm method also performs the following processing: mapping the alarm control instruction set with the device control permission table of the target smart device to determine the alarm priority sequence; traversing the target smart device to perform status verification, and when the verification passes, triggering the local alarm module according to the alarm priority sequence to generate device operation instructions; encapsulating the device operation instructions into a standardized Internet of Things data packet and sending it to the device gateway to execute the alarm, and determining the linkage instruction.

[0014] The present application also provides a combustible gas concentration detection and alarm device, which includes: a pre-processed concentration data acquisition module, which is used to collect combustible gas concentration data in the environment through multiple types of sensors, generate real-time concentration signals, perform noise suppression processing on the real-time concentration signals, and obtain pre-processed concentration data; a gas leakage risk level determination module, which is used to predict leakage based on the pre-processed concentration data, generate gas leakage prediction results, perform similarity calculations based on the gas leakage prediction results combined with historical sensor scene data, and determine the gas leakage risk level; a multi-level alarm strategy execution module, which is used to trigger a multi-level alarm strategy according to the gas leakage risk level, and execute the multi-level alarm strategy to link smart devices for safety control.

[0015] A method and device for detecting and alarming the concentration of combustible gas proposed in this application collect the concentration data of combustible gas in the environment through multiple types of sensors, generate real-time concentration signals, perform noise suppression processing to obtain preprocessed concentration data; perform leakage prediction, generate gas leakage prediction results, perform similarity calculation in combination with historical sensing scenario data to determine the gas leakage risk level; trigger a multi-level alarm strategy, and execute the multi-level alarm strategy to link intelligent devices for safety control. It solves the technical problems existing in the prior art, such as the inability to comprehensively detect multiple combustible gases, the detection data being easily interfered by the environment, the lack of forward-looking prediction of the gas leakage risk level, and insufficient data privacy protection, which in turn lead to insufficient accuracy of combustible gas concentration detection and untimely and inaccurate alarming. It realizes intelligent home linkage and achieves the technical effects of improving the accuracy of combustible gas detection and reducing the false alarm rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments of the present disclosure will be briefly introduced below. Flowcharts are used in this application to illustrate the operations performed by the devices according to the embodiments of the present application. It should be understood that the operations described above or below do not necessarily need to be executed precisely in sequence. On the contrary, according to the need, they can be executed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 It is a schematic flowchart of a method for detecting and alarming the concentration of combustible gas provided by an embodiment of the present application.

[0018] Figure 2 It is a schematic flowchart of a method for detecting and alarming the concentration of combustible gas provided by an embodiment of the present application to determine the gas leakage risk level.

[0019] Figure 3 It is a schematic structural diagram of a device for detecting and alarming the concentration of combustible gas provided by an embodiment of the present application.

[0020] Description of reference numerals: Preprocessed concentration data acquisition module 10, gas leakage risk level determination module 20, multi-level alarm strategy execution module 30. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] The above description is only an overview of the technical solutions of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following specifically illustrates the specific embodiments of the present application.

[0022] To make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be construed as limitations on the present application. All other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present application.

[0023] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments. However, it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments and can be combined with each other without conflict. The terms "first / second" involved are only used to distinguish similar objects and do not represent a specific order for the objects. The terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, device, product or server that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application.

[0024] An embodiment of the present application provides a method for detecting and alarming the concentration of combustible gas, as Figure 1 shown. The method includes:

[0025] Step S100: Collect the concentration data of combustible gas in the environment through multi-type sensors, generate a real-time concentration signal, and perform noise suppression processing on the real-time concentration signal to obtain preprocessed concentration data.

[0026] Preferably, the concentration data of combustible gas in the environment is collected through multi-type sensors (infrared sensors, electrochemical sensors, and semiconductor sensors). Specifically, the absorption characteristics of combustible gas for infrared light with a specific wavelength are used to detect the gas concentration. Different combustible gases have different absorption degrees for infrared light. By measuring the change in the intensity of the absorbed infrared light, the concentration of the combustible gas can be calculated. Infrared sensors have the advantages of high accuracy, good stability, and strong anti-interference ability, and are particularly suitable for detecting some combustible gases with obvious absorption characteristics in the infrared band, such as methane, ethane, etc.; an oxidation-reduction reaction occurs on the electrode of the combustible gas to generate a current, and the magnitude of the current is proportional to the concentration of the combustible gas. Electrochemical sensors have high selectivity and sensitivity to specific combustible gases and can detect low-concentration combustible gases, and are commonly used to detect combustible gases such as carbon monoxide and hydrogen.

[0027] Preferably, when the semiconductor material comes into contact with a combustible gas, its resistance value will change. When the combustible gas adsorbs on the semiconductor surface, it will cause a change in the carrier concentration inside the semiconductor, thereby resulting in a resistance change. By measuring the resistance change, the concentration of the combustible gas can be determined. The semiconductor sensor has the advantages of fast response speed, low cost, small size, etc., but its selectivity is relatively poor and it is easily interfered by other gases. By comprehensively using these sensors, the concentration information of combustible gases in the environment is collected in real time and converted into corresponding electrical signals, that is, real-time concentration signals, to achieve a comprehensive detection of multiple combustible gases and improve the accuracy and reliability of detection. During the process of data collection by the sensor, the real-time concentration signal obtained by the sensor may be interfered by various noises, such as electromagnetic interference in the environment, noises of electrical equipment, etc. The noise will cause fluctuations or deviations in the collected gas concentration data, affecting the accurate judgment of the combustible gas concentration. Therefore, noise suppression processing is performed on the real-time concentration signal. For example, noise suppression is carried out through digital filtering, wavelet analysis, etc. to remove the noise components in the concentration signal and retain the useful concentration information, thereby obtaining preprocessed concentration data, which can more accurately reflect the true concentration of combustible gases in the environment.

[0028] Further, step S100 further includes step S110 of obtaining a sensing data set, dynamically calibrating the sensing data set according to the environmental temperature parameter and the environmental humidity parameter to determine sensing calibration data; step S120 of performing gas concentration analysis based on the sensing calibration data and generating the real-time concentration signal according to the concentration value; step S130 of traversing the real-time concentration signal for pulse processing to extract abnormal pulse signals; step S140 of filtering the abnormal pulse signals by using a wavelet transform algorithm to generate a smooth concentration curve; step S150 of correcting the concentration value according to the smooth concentration curve to obtain the preprocessed concentration data.

[0029] Preferably, an infrared sensor, an electrochemical sensor and a semiconductor sensor are used to monitor the combustible gases in the environment, continuously collect relevant data such as the concentration of combustible gases, environmental temperature parameters, environmental humidity parameters, etc., to form a sensing data set. Among them, environmental temperature and humidity will affect the performance of the sensor, resulting in measurement errors. For example, high temperature may change the physical properties of the sensor element, affecting its response to gas concentration; high humidity may form a water vapor film on the sensor surface, interfering with the normal reaction between the gas and the sensor. The sensing data set is dynamically calibrated according to the environmental temperature parameter and the environmental humidity parameter. Specifically, through data analysis, the relationship between the gas concentration data output by the sensor and the environmental temperature and humidity is studied. According to the analyzed relationship, a polynomial model is selected to describe the influence of environmental temperature and humidity on gas concentration measurement, a calibration model is constructed, and then the sensing data set is input into the calibration model for dynamic adjustment to finally obtain more accurate sensing calibration data.

[0030] Preferably, gas concentration analysis is performed on the sensing calibration data, that is, features related to gas concentration are extracted from the sensing calibration data. For example, an electrochemical sensor outputs a current signal proportional to the gas concentration, and an infrared sensor outputs a change in light absorption intensity at a specific wavelength. The gas concentration is calculated based on the working principle and mathematical model of the sensor. For example, the signals output by some sensors have a linear relationship with the gas concentration, and the concentration can be calculated through a simple linear equation C = kS + b, where C is the gas concentration, S is the sensor output signal, k is the slope, and b is the intercept. The values of k and b can be obtained by measuring and fitting gases with known concentrations; then the calculated gas concentration value is converted into a signal form suitable for transmission and processing, representing the values of the combustible gas concentration in the environment at different times, such as analog signals (such as voltage signals, current signals) and digital signals. For example, the concentration value is converted into a 4-20 mA current signal, or it is converted into a digital signal through an analog-to-digital converter (ADC).

[0031] Preferably, the real-time concentration signal is traversed for pulse processing to extract abnormal pulse signals. Specifically, each data point in the real-time concentration signal is checked one by one, and the corresponding gas concentration value at each moment is viewed in turn. Pulse signals are identified by setting appropriate thresholds and time windows. Assuming that the normal fluctuation range of the gas concentration is within ±5%, when the change amplitude of a certain data point exceeds 10% and this change only lasts for 1-2 sampling periods, it may be a pulse signal; after identifying the pulse signal, it is classified into different types according to the characteristics of the pulse, such as amplitude, duration, shape, etc. For example, some pulses may be caused by momentary interference of the sensor, with a small amplitude and a short duration, while some pulses may be caused by sudden gas leakage, with a large amplitude and a long duration; then an abnormal determination rule is formulated, and pulse signals with a large amplitude, a long duration, or a high occurrence frequency are judged as abnormal pulses, which may indicate sudden leakage of combustible gas, sensor failure, etc.

[0032] Preferably, the wavelet transform algorithm is used to filter the abnormal pulse signal, that is, to remove the noise and interference in the abnormal pulse signal and generate a smooth concentration curve to more accurately reflect the true gas concentration change trend. Among them, wavelet transform can decompose the signal into wavelet coefficients of different scales and positions. Wavelets are like a set of "basis functions" with different frequencies and shapes. By performing a convolution operation with the original signal, the signal can be analyzed on different time-frequency scales. For signals with mutation characteristics such as abnormal pulse signals, wavelet transform can effectively decompose them at different scales, thereby distinguishing the useful components and noise components in the signal. Specifically, the abnormal pulse signal is input into the wavelet transform algorithm, and the algorithm will decompose the signal into wavelet coefficients of different scales according to the selected wavelet basis function. At different scales, the characteristics of the signal will be presented in different ways. For example, at a larger scale, it mainly reflects the overall trend and low-frequency components of the signal; at a smaller scale, it more reflects the details and high-frequency components of the signal. The noise and interference in the abnormal pulse signal usually appear as high-frequency components, while the true concentration change signal is mainly contained in the low-frequency components. Then, a threshold is set according to the characteristics of the signal and the noise level. Those less than the threshold are attenuated, and those greater than the threshold are retained or appropriately enhanced to effectively suppress the noise. The wavelet coefficients after threshold processing are then reconstructed through inverse wavelet transform to obtain the filtered signal, which accurately reflects the true change of the gas concentration. Finally, these filtered data points are connected in chronological order to generate a smooth concentration curve, which more clearly shows the change trend of the gas concentration over time.

[0033] Preferably, the concentration value is corrected according to the smooth concentration curve, that is, the original concentration value corresponding to each time point is replaced with the concentration value corrected according to the smooth concentration curve, so as to obtain the corrected concentration data, which is called preprocessed concentration data, and to a certain extent, it eliminates the influence of environmental factors, sensor errors, and noise and other interferences, and can more accurately reflect the gas concentration in the actual environment.

[0034] Step S200, based on the preprocessed concentration data, perform leakage prediction to generate a gas leakage prediction result, and perform similarity calculation according to the gas leakage prediction result combined with historical sensing scenario data to determine the gas leakage risk level.

[0035] Preferably, a gas leakage prediction model is established by using preprocessed concentration data through methods such as data analysis and machine learning. Specifically, a leakage prediction model is constructed based on the collaborative filtering algorithm, that is, by analyzing the similarity between the concentration data at different time points or different sensor positions in the preprocessed concentration data, potential patterns and associations in the data are discovered. Using the historical preprocessed concentration data and known gas leakage events as training data, and taking the characteristics of the concentration data (such as the concentration change rate, the duration of high concentration periods, etc.) as inputs, a leakage prediction model based on the collaborative filtering algorithm is constructed; then the current preprocessed concentration data is input into the trained prediction model to predict whether a gas leakage is likely to occur and the likelihood of leakage, and a prediction result is output to obtain the gas leakage prediction result. The prediction result may be a probability value indicating the likelihood of a gas leakage occurring within a future time period; it may also be a simple "yes" or "no" judgment indicating whether a gas leakage is predicted.

[0036] Preferably, historical sensing scenario data is obtained. The historical sensing scenario data includes gas concentration data, environmental parameters (such as temperature, humidity, etc.), and information on whether a gas leakage has actually occurred under different past times and environmental conditions. A similarity measurement method is selected to compare the similarity between the current gas leakage prediction result and the historical sensing scenario data, such as Euclidean distance, cosine similarity, etc. For a dataset containing multiple features (such as gas concentration, temperature, humidity, etc.), the Euclidean distance between the current prediction result and the historical data points in the feature space can be calculated. The smaller the distance, the more similar it is; then the similarity between the gas leakage prediction result and each data point in the historical sensing scenario data is calculated to find the historical scenario most similar to the current situation. Suppose the current gas leakage prediction result is a probability value of 0.6 and includes information such as the current gas concentration, temperature, and humidity. Through similarity calculation, it is found that the gas concentration change trend, temperature, humidity, and the probability of gas leakage at that time in a certain historical scenario are most similar to the current situation.

[0037] Preferably, according to the results of similarity calculation and combined with the actual gas leakage situations in historical scenarios, the current gas leakage risk is classified into levels. For example, the risk levels are divided into three levels: low, medium, and high. If there is no gas leakage in the historical scenario that is most similar to the current situation and the similarity is high, then the current risk level is determined to be low; if a certain proportion of situations in the similar historical scenarios have gas leakage, then it is determined to be a medium risk; if most of the similar historical scenarios have gas leakage, then the current risk level is high. Thus, the risk level is initially determined and then adjusted in combination with environmental conditions, such as whether the current environmental conditions are more severe than the historical scenarios (such as factors like high temperature and high humidity that may increase the gas leakage risk), or whether there are some potential fault hazards in the current equipment status, etc., so as to more accurately reflect the actual gas leakage risk situation.

[0038] Furthermore, as Figure 2 shown, step S200 further includes step S210, constructing a time-series concentration distribution map based on the smoothed concentration curve, performing feature analysis according to the time-series concentration distribution map, and determining dynamic leakage characteristics; step S220, retrieving historical sensing scenario data for scenario feature analysis and determining a scenario feature vector; step S230, performing multi-dimensional similarity matching between the dynamic leakage characteristics and the scenario feature vector to generate a set of scenario matching degrees; step S240, performing weighted clustering analysis on the set of scenario matching degrees to obtain multiple scenario feature clusters, traversing the multiple scenario feature clusters for associated screening, and determining an associated historical scenario group, where the associated historical scenario group contains multiple leakage event probabilities; step S250, constructing a leakage prediction model according to the multiple leakage event probabilities and the dynamic leakage characteristics, performing leakage prediction through the leakage prediction model, and generating a gas leakage prediction result; step S260, performing risk decision analysis in combination with the gas leakage prediction result to determine the gas leakage risk level.

[0039] Preferably, according to the smoothed concentration curve, arrange the concentration values at different times in chronological order to construct a time-series concentration distribution map, which intuitively shows the change of gas concentration over time. Then analyze the time-series concentration distribution map and extract the features that can reflect the dynamic change of gas leakage, such as the change rate of concentration (rapid increase, slow increase, decrease, etc.), concentration peak, duration of concentration remaining higher than a certain threshold, and fluctuation frequency, etc., to judge the possibility and severity of gas leakage; retrieve historical sensing scenario data, that is, obtain the sensing data collected in different scenarios in the past from the database, including gas concentration, environmental parameters (such as temperature, humidity, pressure, etc.), sensor location information, and relevant data such as whether leakage occurred. Analyze these historical sensing scenario data and extract the key features of each scenario, such as specific combinations of environmental parameters, correlation patterns of sensor data, etc. Then quantify and encode the extracted scenario features to obtain a scenario feature vector, which contains various feature information of the scenario.

[0040] Preferably, compare the current dynamic leakage features with the scenario feature vectors of each historical scenario in multiple dimensions, calculate the similarity between them. For example, compare in multiple aspects such as concentration change features and environmental parameter features, and use Euclidean distance, cosine similarity, etc. to measure the similarity. Then calculate the similarity value between each historical scenario and the current dynamic leakage features and form a set of scenario matching degrees to reflect the similarity between the current situation and each historical scenario; then judge the importance of gas leakage risk according to different features, assign different weights to the similarity of each dimension, and then perform clustering analysis on the set of scenario matching degrees, group the historical scenarios with higher similarity into one category to form multiple scenario feature clusters; then traverse each scenario feature cluster, analyze the characteristics of the historical scenarios in it, and screen out the historical scenarios with a higher correlation with the current situation, that is, screen out the associated historical scenario group with a matching degree higher than the preset threshold. Among them, the associated historical scenario group contains the probability records of gas leakage occurring in it, that is, multiple leakage event records, which are used for subsequent leakage prediction and risk assessment.

[0041] Preferably, by using the probabilities of multiple leakage events in the associated historical scenario group and the current dynamic leakage characteristics, a leakage prediction model is constructed through a collaborative filtering algorithm and trained with historical data to establish the relationship between the leakage event probability and the dynamic leakage characteristics. Then, the current dynamic leakage characteristics are input into the constructed leakage prediction model. Based on the knowledge and relationships it has learned, the model predicts whether a gas leakage will occur and the likelihood of leakage, and outputs the gas leakage prediction result. Finally, by comprehensively considering the gas leakage prediction result, the possible consequences of the leakage (such as threats to personnel safety, property losses, etc.) and other relevant factors (such as the sensitivity of the current environment, etc.), a risk decision-making analysis is carried out. According to the results of the risk decision-making analysis, the gas leakage risk is divided into different levels, such as high, medium, and low risks. Different risk levels correspond to different response measures and management strategies to effectively respond to possible gas leakage events in a timely manner.

[0042] Further, step S200 further includes step S270 of performing multi-dimensional sensing acquisition based on historical scenarios to obtain multi-dimensional sensing data; step S280 of performing time alignment processing on the multi-dimensional sensing data to generate a scenario data stream; step S290 of using an autoencoder to perform feature dimensionality reduction on the scenario data stream to generate a scenario feature vector; and step S2100 of associatively storing the leakage event label with the scenario feature vector to generate the historical sensing scenario data.

[0043] Preferably, based on various historical scenarios related to combustible gases (such as different indoor environments, different equipment usage conditions, etc.), multiple types of sensors are used for data acquisition, including the concentration of combustible gases, which is a key data directly reflecting the presence and degree of gas leakage; environmental temperature and humidity, which affect the diffusion, aggregation of combustible gases, and the performance of sensors; the operating state of equipment, such as the on / off state and operating parameters of gas equipment, and abnormal operation of the equipment may be related to gas leakage; user behavior logs, which record the user's operation behaviors on gas equipment (such as the time of turning on / off the equipment, adjusting the settings of the equipment, etc.), and improper user operations may also trigger gas leakage risks. Since the time intervals and time origins of data acquisition by different types of sensors may be different, time alignment processing is performed on the multi-dimensional sensing data, that is, the data of each dimension are arranged in chronological order, and an accurate timestamp is added to each data point, so that the data of different dimensions are consistent in time, generating a scenario data stream, which is an ordered sequence containing multi-dimensional data and each data with a timestamp. For example, at a specific time point, the scenario data stream will simultaneously include relevant information such as the concentration of combustible gases, environmental temperature and humidity, the operating state of equipment, and user behavior logs at that moment.

[0044] Preferably, an autoencoder is used to perform feature dimensionality reduction on the scene data stream. An autoencoder is an unsupervised learning neural network model composed of an encoder and a decoder. Specifically, feature dimensionality reduction is to compress high-dimensional raw data (i.e., sensing data containing multiple dimensions) into low-dimensional feature representations. By learning the internal patterns and structures in the data, the autoencoder can extract the key features that best represent the raw data, remove redundant information, obtain the scene feature vector, retain the information related to combustible gas in the raw data, and at the same time reduce the computational amount and storage space requirements. Then, the leakage event label is associated and stored with the scene feature vector. Specifically, the leakage event label is used to mark whether a combustible gas leakage event has occurred in a certain scene (which can be represented by "yes" or "no", or by the numbers 0 and 1). The leakage event label corresponding to each scene is associated with the scene feature vector obtained through feature dimensionality reduction, that is, the information of the leakage event is bound to the feature information of the scene, and then stored to form historical sensing scene data for the training of the leakage prediction model. The relationship between different scene features and leakage events can be learned according to the annotation information, so as to improve the accuracy and reliability of combustible gas leakage analysis and prediction.

[0045] Further, step S210 further includes step S211, establishing a sliding time window according to the smoothed concentration curve, calculating the concentration according to the sliding time window to obtain the concentration change rate; step S212, detecting the concentration duration of the sliding time window based on the concentration change rate to extract the peak duration; step S213, counting the number of times the concentration value in the sliding time window exceeds the preset warning line to determine the fluctuation frequency; step S214, combining the concentration change rate, the peak duration and the fluctuation frequency as a three-dimensional feature vector to obtain the dynamic leakage feature.

[0046] Preferably, a sliding time window of a fixed length is set on the smoothed concentration curve and slides on the curve over time for focusing on the concentration data within a period of time. Specifically, for the concentration data within each sliding time window, the first derivative of its concentration is calculated, representing the rate of change of the concentration over time. By calculating the first derivative of the concentration within the window as the concentration change rate, it is possible to understand how fast the gas concentration rises or falls during this period. For example, if the concentration change rate is positive and large, it indicates that the gas concentration is rising rapidly, and there may be a gas leak with a relatively fast leakage rate; if the change rate is negative, it means the concentration is decreasing, which may be due to the leakage being controlled or the environment diluting the gas, etc. Within the sliding time window, the length of the continuous period during which the detected concentration exceeds the safety threshold is measured. Here, the safety threshold is a pre-set standard value used to judge safety risks. By detecting the length of the period continuously exceeding the safety threshold, it is possible to understand the duration during which the gas concentration is in a dangerous state. During the process when the concentration exceeds the safety threshold, there may be a peak value of the concentration, that is, the highest value reached by the concentration. The peak duration refers to the length of time during which the concentration remains at a relatively high level near the peak value, which can reflect the severity and duration state of the gas leak. If the peak duration is long, it indicates that the gas leak situation is relatively serious and lasts for a long time, posing a greater threat to safety.

[0047] Preferably, the number of times the concentration value within the sliding time window exceeds the preset warning line is counted. The preset warning line is used to judge the fluctuation of the concentration. When the concentration frequently exceeds the warning line, it indicates that the gas concentration is in an unstable state. There may be intermittent leakage from the leakage source or a greater impact of environmental factors on the gas distribution, resulting in frequent concentration fluctuations. The higher the fluctuation frequency, the more complex the dynamic change of the gas leak. Finally, the calculated concentration change rate, the extracted peak duration, and the statistically obtained fluctuation frequency are combined into a three-dimensional feature vector, which can comprehensively describe the dynamic characteristics of the gas leak. The concentration change rate reflects the speed of concentration change, the peak duration reflects the severity and duration of the leak, and the fluctuation frequency indicates the stability of the concentration and the complexity of the change, thereby more accurately analyzing and judging the gas leak situation.

[0048] Furthermore, step S250 further includes step S251 of obtaining a plurality of historical gas usage scenario data, where the plurality of historical gas usage scenario data includes gas equipment type data, usage period data, and environmental parameter data; step S252 of performing correlation analysis on the environmental parameter data and the gas equipment type data to determine leakage event data; step S253 of calculating a similarity matrix between the real-time sensing scenario data and the historical sensing scenario data according to the usage period data; and step S254 of performing linear regression analysis on the leakage event data based on the similarity matrix to construct the leakage prediction model.

[0049] Preferably, multiple historical gas usage scenario data are obtained from various data sources (such as the databases of gas companies, records of relevant monitoring devices, etc.). The multiple historical gas usage scenario data include gas equipment type data, usage time period data, and environmental parameter data. Among them, the gas equipment type data records the types of gas equipment used, such as gas stoves, gas water heaters, etc.; the usage time period data records the usage time of the gas equipment in a day, accurate to specific time periods; the environmental parameter data covers the environmental factors at that time, such as temperature, humidity, air pressure, etc. The environmental parameter data and the gas equipment type data are comprehensively analyzed. Specifically, for different types of gas equipment under different environmental conditions, the possibility and characteristics of leakage may be different. For example, some gas equipment is more likely to have seal aging in high-temperature and high-humidity environments, resulting in gas leakage; through association analysis, the potential relationship between environmental parameters and gas equipment types is found, and then in which cases gas leakage events occur is determined, forming leakage event data, which specifically includes information such as the corresponding gas equipment type and environmental parameters when the leakage occurs.

[0050] Preferably, the real-time acquired sensing scenario data (i.e., gas-related data in the current environment, which may include real-time gas concentration, environmental parameters, etc.) is compared with the historical sensing scenario data. According to the usage time period data, the corresponding relationship in time between the real-time data and the historical data is analyzed. For example, the real-time data and the historical data in the same time period (such as 7 pm to 9 pm) are compared, and then the cosine similarity, Euclidean distance, etc. are used to calculate the similarity between the real-time sensing scenario data and each historical sensing scenario data, and the calculated similarity values are established as a similarity matrix to show the similarity degree between the real-time data and each historical data; then, using the obtained similarity matrix, combined with the leakage event data, linear regression analysis is carried out to establish the prediction relationship between some eigenvalue in the similarity matrix and related environmental parameters, gas equipment types and other factors, and whether leakage occurs or the probability of leakage. By performing linear regression analysis on the leakage event data, the influence weight of each independent variable on the dependent variable is determined, so as to construct a leakage prediction model, which can predict the possibility of gas leakage in the current situation according to the similarity between the real-time sensing scenario data and the historical data and other relevant factors, and ensure the accuracy rate of gas concentration prediction.

[0051] Step S300, trigger a multi-level alarm strategy according to the gas leakage risk level, and execute the multi-level alarm strategy to link intelligent devices for safety control.

[0052] Preferably, different levels of alarm mechanisms are formulated according to the determined gas leakage risk levels, and each level corresponds to different alarm methods and contents. Specifically, the low-risk level alarm only issues a prompt message, such as displaying a yellow warning icon on the relevant monitoring interface, informing that there is a certain gas leakage risk at present, but the situation is relatively mild; the medium-risk level alarm may issue a sound alarm at a specific frequency, the warning icon on the monitoring interface becomes orange and flashes, and at the same time, detailed alarm information, including the approximate location of the leakage, the possible affected range, etc., is sent to more relevant personnel; the high-risk level alarm triggers a strong alarm signal, such as a harsh alarm sound, a red flashing light, etc., and clearly indicates that the leakage situation is critical. At the same time, an emergency notice is sent to all relevant personnel, the emergency response plan is activated, and immediate emergency measures are required to be taken; after different levels of alarms are triggered, intelligent home devices are linked for safety control to reduce the risk brought by gas leakage. For example, the intelligent ventilation device is started to accelerate air circulation and reduce the concentration of combustible gas; non-critical gas equipment is automatically shut down to prevent dangers such as explosion caused by electric sparks generated by the operation of the equipment; the main gas supply valve is immediately cut off to prevent the leakage of combustible gas, and at the same time, the fire sprinkler system is started to cool down and reduce dust in the possibly affected area, reducing the explosion risk. Through the multi-level alarm strategy and the safety control of intelligent device linkage, corresponding measures are taken according to the severity of the gas leakage risk to ensure safety to the greatest extent.

[0053] Further, step S300 further includes step S310 of matching the gas leakage risk level with the multi-level alarm strategy library to generate an alarm control instruction set; step S320 of parsing based on the alarm control instruction set to generate an alarm priority sequence, triggering the local alarm module to execute an alarm according to the alarm priority sequence, and generating a linkage instruction; step S330 of distributing the linkage instruction to the target intelligent device to control the target intelligent device to perform a safety operation.

[0054] Preferably, a multi-level alarm policy library pre-stores a variety of detailed alarm policies for different risk levels. Each policy specifies corresponding alarm methods, alarm objects, alarm contents, etc. Compare and match the determined gas leakage risk level with the contents in the multi-level alarm policy library to find the corresponding alarm policy. Then generate specific alarm control instructions according to the matched alarm policy to form an alarm control instruction set. For example, if the risk level is high risk, the matched alarm policy may include emitting a high-decibel alarm sound, sending an emergency notice to all relevant personnel, etc. The corresponding instructions will include instructions to control the alarm device to emit an alarm sound and information such as the specific content and recipient of the notice; then parse the generated alarm control instruction set, analyze the importance and urgency of each instruction, and sort the alarm control instructions according to the parsing results to form an alarm priority sequence. For example, in the case of high risk, the instruction to immediately cut off the gas supply may have the highest priority, while the instruction to send a notice has a relatively lower priority; then trigger the local alarm module to perform corresponding alarm operations in sequence according to the alarm priority sequence. Among them, the local alarm module may include an audible and visual alarm, etc., and can emit different forms of alarms according to the alarm control instructions; and during the execution of the alarm operation, generate linkage instructions according to the alarm control instructions and the actual execution situation to control the intelligent home devices to work together to cope with the combustible gas leakage situation. For example, when the local alarm module issues a high-risk alarm, the linkage instructions include closing the gas valve, starting the ventilation equipment, etc.

[0055] Preferably, send the generated linkage instructions to the corresponding target intelligent home devices, such as intelligent gas valves, intelligent ventilation systems, intelligent environmental monitoring devices, etc. These devices perform safety operations according to the requirements of the linkage instructions. For example, after receiving the closing instruction, the intelligent gas valve will automatically close the valve to cut off the gas supply; after receiving the start instruction, the intelligent ventilation system will start working to accelerate air circulation to reduce the concentration of combustible gas; and during the process of the intelligent home devices performing safety operations, real-time monitor the execution status of the devices (such as whether the valve is successfully closed, the operating power of the ventilation equipment, etc.) and the change of the environmental concentration (obtain the combustible gas concentration data in real time through the environmental monitoring device), and dynamically adjust the execution intensity of the alarm policy according to the monitoring information. For example, if the concentration of combustible gas drops slowly after the ventilation equipment runs, the operating power of the ventilation equipment may be increased to achieve intelligent monitoring and control of combustible gas.

[0056] Further, step S310 further includes step S311. When the gas leakage risk level is at the first level, retrieve the multi-level alarm policy library to determine the first alarm control instruction. Trigger the local audible and visual alarm through the first alarm control instruction and push a warning notice, and generate first-level leakage record data; step S312, when the gas leakage risk level is at the second level, retrieve the multi-level alarm policy library to determine the second alarm control instruction. Perform a closing operation on the intelligent gas valve through the second alarm control instruction and start the ventilation equipment, and generate second-level leakage record data; step S313, when the gas leakage risk level is at the third level, retrieve the multi-level alarm policy library to determine the third alarm control instruction. Send an emergency broadcast signal to the security system through the third alarm control instruction, and generate third-level leakage record data; step S314, integrate the first-level leakage record data, the second-level leakage record data, and the third-level leakage record data for alarm, determine the leakage alarm record log, and extract the alarm control instruction set according to the leakage alarm record log.

[0057] Preferably, when it is determined that the gas leakage risk level is at the first level (usually indicating a relatively low risk), retrieve in the multi-level alarm policy library to determine the first alarm control instruction applicable to the first-level risk. The main function of this instruction is to trigger the local audible and visual alarm, making it emit sound and light alarms to alert nearby personnel of the potential gas leakage situation. At the same time, push a warning notice, such as sending messages to the mobile phones, computers and other terminal devices of relevant staff to inform them of the current first-level gas leakage risk. At the same time, generate first-level leakage record data, recording relevant information about this first-level gas leakage incident, such as the time, location of the leakage, and triggered alarm actions, etc.; when the gas leakage risk level rises to the second level (indicating an increase in the risk level), also retrieve in the multi-level alarm policy library to determine the second alarm control instruction applicable to the second-level risk, which is mainly used to operate smart home devices, that is, perform a closing operation on the intelligent gas valve through this instruction to cut off the gas supply from the source and prevent the leakage from expanding further; at the same time, start the ventilation equipment to accelerate air circulation and reduce the concentration of combustible gas in the environment, and then generate second-level leakage record data, recording the details of the second-level gas leakage incident, including information such as the valve closing time and the ventilation equipment startup time.

[0058] Preferably, when the gas leakage risk level reaches level three (indicating a relatively high risk and an urgent situation), a search is still conducted in the multi-level alarm strategy library to determine the third alarm control instruction applicable to level-three risks, which is used to send an emergency broadcast signal to the security system, that is, to play an emergency notice to relevant areas (such as the entire building or a specific dangerous area), informing people that the situation is critical and they need to quickly take response measures such as evacuation. Furthermore, level-three leakage record data is generated to record relevant information about the level-three gas leakage event, such as the time and content of the broadcast. Then, the level-one leakage record data, level-two leakage record data, and level-three leakage record data are integrated for alarm, that is, the leakage record data of different risk levels are aggregated to form a leakage alarm record log, and all alarm control instructions are extracted according to the leakage alarm record log to form an alarm control instruction set, which is used to evaluate the execution effect of the alarm strategy or, when a similar gas leakage event occurs, as a reference to quickly determine the measures to be taken. Thus, an orderly response to and recording of gas leakage events are achieved to ensure the reduction of safety risks.

[0059] Further, step S320 further includes step S321 of mapping according to the alarm control instruction set and the device control authority table of the target intelligent device to determine the alarm priority sequence; step S322 of traversing the target intelligent device for status verification, and when the verification is passed, triggering the local alarm module according to the alarm priority sequence to generate a device operation instruction; step S323 of encapsulating the device operation instruction into a standardized Internet of Things data packet and sending it to the device gateway to execute the alarm to determine the linkage instruction.

[0060] Preferably, the device control authority table of the target intelligent device records the control authorities of each target intelligent device (such as intelligent gas valves, ventilation devices, security broadcast systems, etc.), for example, which instructions can operate the device, as well as information such as the importance level and priority of different operations. The instructions in the alarm control instruction set are correspondingly matched with the device control authority table. According to the priority regulations for different operations in the device control authority table, the priority of each instruction in the current situation is determined, and then an alarm priority sequence sorted by priority is formed, which determines the execution order of security operations to ensure that important operations can be executed first. Then, all target intelligent devices are checked one by one to verify their online status (whether the device is connected to the network and can communicate normally) and controllability (whether the device can operate according to the instructions). If it is found that a certain device is in an offline state, it is automatically switched to a backup control node (backup communication path or control method) to ensure that the device can be controlled. When all target intelligent devices pass the status verification, according to the alarm priority sequence, the local alarm module is triggered in turn to execute corresponding alarm operations (such as emitting sound and light alarms, etc.). At the same time, according to the alarm control instruction set and the current state of the device, specific operation instructions for each target intelligent device are generated, clarifying the actions that the device needs to execute, such as the closing instruction for the intelligent gas valve, the starting instruction for the ventilation device, etc.

[0061] Preferably, these instructions are encapsulated into standardized Internet of Things data packets to ensure that device operation instructions can be accurately and securely transmitted in the Internet of Things environment. In addition to the device operation instructions, the data packets will also be appended with a timestamp (recording the time when the instruction is sent, used for time synchronization and operation traceability) and a digital signature (used to verify the integrity of the data packet and the identity of the sender to prevent the data packet from being tampered with or forged). The encapsulated data packets are sent to the device gateway (an intermediate device connecting intelligent devices and the network), and then transferred to the corresponding target intelligent devices to execute alarm and control operations. In this process, linkage instructions are determined, that is, it is clarified how each intelligent device collaborates to respond to a gas leakage event, and then the linkage control between devices is realized, achieving the effective management and control of target intelligent devices in the gas leakage alarm system, thereby improving reliability and security.

[0062] In the above text, reference is made to Figure 1 A method for detecting and alarming the concentration of combustible gas according to an embodiment of the present invention is described in detail. Next, a device for detecting and alarming the concentration of combustible gas according to an embodiment of the present invention will be described with reference to Figure 3

[0063] A combustible gas concentration detection alarm device according to an embodiment of the present invention is used to solve the technical problems existing in the prior art, such as the inability to comprehensively detect multiple combustible gases, the detection data being susceptible to environmental interference, the lack of forward-looking prediction of gas leakage risk levels, and insufficient data privacy protection, which in turn lead to insufficient accuracy in combustible gas concentration detection and insufficient timely and accurate alarms. It realizes smart home linkage and achieves the technical effect of improving the accuracy of combustible gas detection and reducing the false alarm rate. Figure 3 As shown, a combustible gas concentration detection and alarm device includes: a pre-processing concentration data acquisition module 10, a gas leakage risk level determination module 20, and a multi-level alarm strategy execution module 30.

[0064] The pre-processed concentration data acquisition module 10 is used to collect the combustible gas concentration data in the environment through multiple types of sensors, generate a real-time concentration signal, perform noise suppression processing on the real-time concentration signal, and obtain pre-processed concentration data; the gas leakage risk level determination module 20 is used to predict leakage based on the pre-processed concentration data, generate a gas leakage prediction result, perform similarity calculation based on the gas leakage prediction result combined with historical sensor scene data, and determine the gas leakage risk level; the multi-level alarm strategy execution module 30 is used to trigger the multi-level alarm strategy according to the gas leakage risk level, and execute the multi-level alarm strategy to link smart devices for safety control.

[0065] The specific configuration of the pre-processing concentration data acquisition module 10 will be described in detail below. The pre-processing concentration data acquisition module 10 further includes: acquiring a sensor data set, dynamically calibrating the sensor data set according to the ambient temperature parameter and the ambient humidity parameter, and determining the sensor calibration data; performing gas concentration analysis based on the sensor calibration data, and generating the real-time concentration signal according to the concentration value; traversing the real-time concentration signal for pulse processing, and extracting abnormal pulse signals; filtering the abnormal pulse signal using a wavelet transform algorithm to generate a smoothed concentration curve; and correcting the concentration value according to the smoothed concentration curve to obtain the pre-processing concentration data.

[0066] Next, the specific configuration of the gas leakage risk level determination module 20 will be described in detail. The gas leakage risk level determination module 20 further includes: constructing a time series concentration distribution map based on the smoothed concentration curve, performing feature analysis according to the time series concentration distribution map to determine dynamic leakage features; retrieving historical sensing scenario data for scenario feature analysis to determine a scenario feature vector; performing multi-dimensional similarity matching on the dynamic leakage features and the scenario feature vector to generate a set of scenario matching degrees; performing weighted clustering analysis on the set of scenario matching degrees to obtain multiple scenario feature clusters, traversing the multiple scenario feature clusters for correlation screening to determine an associated historical scenario group, where the associated historical scenario group includes multiple leakage event probabilities; constructing a leakage prediction model based on the multiple leakage event probabilities and the dynamic leakage features, performing leakage prediction through the leakage prediction model to generate a gas leakage prediction result; and combining the gas leakage prediction result for risk decision analysis to determine the gas leakage risk level.

[0067] Next, the specific configuration of the gas leakage risk level determination module 20 will be further described in detail. The gas leakage risk level determination module 20 further includes: obtaining multiple historical gas usage scenario data, where the multiple historical gas usage scenario data includes gas equipment type data, usage time period data, and environmental parameter data; performing correlation analysis on the environmental parameter data and the gas equipment type data to determine leakage event data; calculating a similarity matrix between the real-time sensing scenario data and the historical sensing scenario data according to the usage time period data; and performing linear regression analysis on the leakage event data based on the similarity matrix to construct the leakage prediction model.

[0068] Next, the specific configuration of the gas leakage risk level determination module 20 will be further described in detail. The gas leakage risk level determination module 20 further includes: establishing a sliding time window according to the smoothed concentration curve, calculating the concentration based on the sliding time window to obtain a concentration change rate; performing concentration duration detection on the sliding time window based on the concentration change rate to extract the peak duration; counting the number of times the concentration value within the sliding time window exceeds a preset warning line to determine the fluctuation frequency; and combining the concentration change rate, the peak duration, and the fluctuation frequency as a three-dimensional feature vector to obtain the dynamic leakage features.

[0069] Next, the specific configuration of the gas leakage risk level determination module 20 will be further described in detail. The gas leakage risk level determination module 20 further includes: performing multi-dimensional sensing acquisition based on historical scenarios to obtain multi-dimensional sensing data; performing time alignment processing on the multi-dimensional sensing data to generate a scenario data stream; using an autoencoder to perform feature dimensionality reduction on the scenario data stream to generate a scenario feature vector; associating and storing the leakage event label with the scenario feature vector to generate the historical sensing scenario data.

[0070] Next, the specific configuration of the multi-level alarm policy execution module 30 will be described in detail. The multi-level alarm policy execution module 30 further includes: matching the gas leakage risk level with a multi-level alarm policy library to generate an alarm control instruction set; parsing based on the alarm control instruction set to generate an alarm priority sequence, and triggering the local alarm module to execute an alarm according to the alarm priority sequence to generate a linkage instruction; distributing the linkage instruction to a target intelligent device to control the target intelligent device to perform a safety operation.

[0071] Next, the specific configuration of the multi-level alarm policy execution module 30 will be further described in detail. The multi-level alarm policy execution module 30 further includes: when the gas leakage risk level is level one, retrieving the multi-level alarm policy library to determine a first alarm control instruction, triggering a local audible and visual alarm and pushing a warning notice through the first alarm control instruction to generate level one leakage record data; when the gas leakage risk level is level two, retrieving the multi-level alarm policy library to determine a second alarm control instruction, performing a closing operation on an intelligent gas valve and starting a ventilation device through the second alarm control instruction to generate level two leakage record data; when the gas leakage risk level is level three, retrieving the multi-level alarm policy library to determine a third alarm control instruction, sending an emergency broadcast signal to a security system through the third alarm control instruction to generate level three leakage record data; integrating the level one leakage record data, the level two leakage record data, and the level three leakage record data for alarm to determine a leakage alarm record log, and extracting the alarm control instruction set according to the leakage alarm record log.

[0072] Next, the specific configuration of the multi-level alarm policy execution module 30 will be further described in detail. The multi-level alarm policy execution module 30 further includes: mapping the alarm control instruction set with a device control authority table of a target intelligent device to determine the alarm priority sequence; traversing the target intelligent device for status verification, and when the verification is passed, triggering the local alarm module according to the alarm priority sequence to generate a device operation instruction; encapsulating the device operation instruction into a standardized Internet of Things data packet and sending it to a device gateway to execute an alarm to determine the linkage instruction.

[0073] A combustible gas concentration detection and alarm device provided by an embodiment of the present invention can execute a combustible gas concentration detection and alarm method provided by any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution of the method.

[0074] Although the present application makes various references to certain modules in the device according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or server. The various units and modules included are only divided according to functional logic, but are not limited to the above division as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of mutual distinction and do not limit the protection scope of the present invention.

[0075] The above specific implementation manners do not constitute a limitation to the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present application shall be included within the protection scope of the present application.

Claims

1. A method for detecting and alarming the concentration of combustible gas, characterized in that, The method includes: Collecting combustible gas concentration data in the environment through multi-type sensors, generating a real-time concentration signal, and performing noise suppression processing on the real-time concentration signal to obtain preprocessed concentration data; Based on the preprocessed concentration data, performing leakage prediction, generating a gas leakage prediction result, and performing similarity calculation according to the gas leakage prediction result in combination with historical sensing scenario data to determine the gas leakage risk level; Triggering a multi-level alarm strategy according to the gas leakage risk level, and executing the multi-level alarm strategy to link intelligent devices for safety control; The multi-type sensors include an infrared sensor, an electrochemical sensor, and a semiconductor sensor, and the noise suppression processing includes: Obtaining a sensing data set, dynamically calibrating the sensing data set according to environmental temperature parameters and environmental humidity parameters to determine sensing calibration data; Based on the sensing calibration data, performing gas concentration analysis, and generating the real-time concentration signal according to the concentration value; Traversing the real-time concentration signal for pulse processing to extract abnormal pulse signals; Using a wavelet transform algorithm to filter the abnormal pulse signals to generate a smooth concentration curve; Correcting the concentration value according to the smooth concentration curve to obtain the preprocessed concentration data; The method for performing leakage prediction based on the preprocessed concentration data, generating a gas leakage prediction result, and performing similarity calculation according to the gas leakage prediction result in combination with historical sensing scenario data to determine the gas leakage risk level includes: Based on the smooth concentration curve, constructing a time-series concentration distribution map, and performing feature analysis according to the time-series concentration distribution map to determine dynamic leakage characteristics; Retrieving historical sensing scenario data for scenario feature analysis to determine scenario feature vectors; Performing multi-dimensional similarity matching between the dynamic leakage characteristics and the scenario feature vectors to generate a set of scenario matching degrees; Performing weighted clustering analysis on the set of scenario matching degrees to obtain multiple scenario feature clusters, traversing the multiple scenario feature clusters for correlation screening to determine an associated historical scenario group, and the associated historical scenario group includes multiple leakage event probabilities; Constructing a leakage prediction model according to the multiple leakage event probabilities and the dynamic leakage characteristics, and performing leakage prediction through the leakage prediction model to generate a gas leakage prediction result; Combining the gas leakage prediction result for risk decision analysis to determine the gas leakage risk level.

2. The combustible gas concentration detection and alarm method according to claim 1, characterized in that The method for constructing a leakage prediction model according to the multiple leakage event probabilities and the dynamic leakage characteristics includes: Obtaining multiple historical gas usage scenario data, and the multiple historical gas usage scenario data includes gas equipment type data, usage period data, and environmental parameter data; Performing correlation analysis on the environmental parameter data and the gas equipment type data to determine leakage event data; Calculating a similarity matrix between the real-time sensing scenario data and the historical sensing scenario data according to the usage period data; Based on the similarity matrix, performing linear regression analysis on the leakage event data to construct the leakage prediction model.

3. The method for detecting and alarming the concentration of combustible gas according to claim 1, characterized in that, Construct a time - series concentration distribution map based on the smoothed concentration curve, perform feature analysis according to the time - series concentration distribution map, and determine the dynamic leakage characteristics. The method includes: Establish a sliding time window according to the smoothed concentration curve, calculate the concentration according to the sliding time window, and obtain the concentration change rate; Detect the concentration duration of the sliding time window based on the concentration change rate, and extract the peak duration; Count the number of times the concentration value in the sliding time window exceeds the preset warning line to determine the fluctuation frequency; Combine the concentration change rate, the peak duration, and the fluctuation frequency as a three - dimensional feature vector to obtain the dynamic leakage characteristics.

4. The combustible gas concentration detection and alarm method according to claim 1, wherein The construction process of the historical sensing scenario data. The method includes: Perform multi - dimensional sensing acquisition based on the historical scenario to obtain multi - dimensional sensing data; Perform time alignment processing on the multi - dimensional sensing data to generate a scenario data stream; Use an auto - encoder to perform feature dimensionality reduction on the scenario data stream to generate a scenario feature vector; Associate and store the leakage event label with the scenario feature vector to generate the historical sensing scenario data.

5. The method for detecting and alarming the concentration of combustible gas according to claim 1, characterized in that, Trigger a multi - level alarm strategy according to the gas leakage risk level, and execute the multi - level alarm strategy to link intelligent devices for safety control. The method includes: Match the gas leakage risk level with the multi - level alarm strategy library to generate an alarm control instruction set; Parse based on the alarm control instruction set to generate an alarm priority sequence, and trigger the local alarm module to execute an alarm according to the alarm priority sequence to generate a linkage instruction; Distribute the linkage instruction to the target intelligent device to control the target intelligent device to perform safety operations.

6. The combustible gas concentration detection and alarm method according to claim 5, characterized in that, Match the gas leakage risk level with the multi - level alarm strategy library to generate an alarm control instruction set. The method includes: When the gas leakage risk level is level one, retrieve the multi - level alarm strategy library to determine the first alarm control instruction, trigger the local audible and visual alarm through the first alarm control instruction and push a warning notice, and generate level - one leakage record data; When the gas leakage risk level is level two, retrieve the multi - level alarm strategy library to determine the second alarm control instruction, perform a closing operation on the intelligent gas valve and start the ventilation equipment through the second alarm control instruction, and generate level - two leakage record data; When the gas leakage risk level is level three, retrieve the multi - level alarm strategy library to determine the third alarm control instruction, send an emergency broadcast signal to the security system through the third alarm control instruction, and generate level - three leakage record data; Integrate the level - one leakage record data, the level - two leakage record data, and the level - three leakage record data for alarm integration, determine the leakage alarm record log, and extract the alarm control instruction set according to the leakage alarm record log.

7. The combustible gas concentration detection and alarm method according to claim 5, characterized in that, Parse based on the alarm control instruction set to generate an alarm priority sequence, and trigger the local alarm module to execute an alarm according to the alarm priority sequence to generate a linkage instruction. The method includes: Map the alarm control instruction set with the device control permission table of the target intelligent device to determine the alarm priority sequence; Traversing the target intelligent device to perform status verification, and when the verification passes, triggering the local alarm module according to the alarm priority sequence to generate device operation instructions; The device operation instruction is encapsulated into a standardized Internet of Things data packet and sent to the device gateway to execute the alarm and determine the linkage instruction.

8. A combustible gas concentration detection and alarm device, characterized in that, The device is used to implement a combustible gas concentration detection and alarm method according to any one of claims 1 to 7, and the device comprises: A pre-processing concentration data acquisition module is used to collect combustible gas concentration data in the environment through multiple types of sensors, generate real-time concentration signals, perform noise suppression processing on the real-time concentration signals, and obtain pre-processing concentration data; A gas leakage risk level determination module, used to perform leakage prediction based on the pre-processed concentration data, generate a gas leakage prediction result, perform similarity calculation based on the gas leakage prediction result combined with historical sensing scene data, and determine the gas leakage risk level; The multi-level alarm strategy execution module is used to trigger the multi-level alarm strategy according to the gas leakage risk level, and execute the multi-level alarm strategy to link smart devices for safety control.

Images