A smoke alarm testing method and device

Abstract

The present application relates to the technical field of smoke alarm testing, and discloses a smoke alarm testing method and device. The method comprises: setting parameters on a multifunctional testing platform to obtain a testing environment simulating various fire scenarios; performing multi-parameter collaborative testing on the smoke alarm in the testing environment to obtain multi-dimensional raw data; performing multi-sensor fusion processing on the multi-dimensional raw data to obtain dimension-reduced feature data; performing multi-factor comprehensive analysis on the dimension-reduced feature data to obtain the performance evaluation result of the smoke alarm; performing smoke alarm simulation and configuration genetic optimization based on the performance evaluation result to obtain an optimized configuration scheme; performing continuous sampling and Kalman filtering processing on the optimized configuration scheme to generate an intelligent smoke tracking strategy. The present application improves the test accuracy of the smoke alarm.

Description

Technical Field

[0001] The present application relates to the technical field of smoke alarm testing, and in particular to a method and device for testing a smoke alarm. Background Art

[0002] Traditional smoke alarm testing methods are often limited to static testing under a single environmental condition, making it difficult to fully evaluate the performance of the alarm in complex and changing actual fire scenarios. At the same time, existing testing methods lack coordinated consideration of multiple environmental factors and cannot accurately simulate the smoke characteristics and diffusion laws of different types of fires, resulting in a large deviation between the test results and actual application scenarios.

[0003] In addition, the performance evaluation of smoke alarms usually relies on simple response time and sensitivity indicators, ignoring the impact of environmental factors on the long-term stability and reliability of the alarm, making it difficult to fully reflect the actual working performance of the alarm. Existing alarm layout and configuration schemes often use empirical methods, lack scientific optimization methods, and are difficult to adapt to different building structures and fire risk characteristics, affecting the overall effectiveness of the alarm system. Summary of the invention

[0004] The present application provides a smoke alarm testing method and device, which are used to improve the testing accuracy of the smoke alarm.

[0005] In a first aspect, the present application provides a method for testing a smoke alarm, the method comprising:

[0006] Parameters of the multifunctional test platform are set to obtain a test environment that simulates various fire scenarios;

[0007] Performing a multi-parameter collaborative test on the smoke alarm in the test environment to obtain multi-dimensional raw data;

[0008] Performing multi-sensor fusion processing on the multi-dimensional raw data to obtain dimension-reduced feature data;

[0009] Performing a multi-factor comprehensive analysis on the dimension-reduced feature data to obtain a performance evaluation result of the smoke alarm;

[0010] Based on the performance evaluation results, smoke alarm simulation and configuration genetic optimization are performed to obtain an optimized configuration solution;

[0011] The optimized configuration scheme is continuously sampled and processed by Kalman filtering to generate an intelligent smoke tracking strategy.

[0012] In a second aspect, the present application provides a test device for a smoke alarm, the test device for a smoke alarm comprising:

[0013] A setting module is used to set parameters of the multifunctional test platform to obtain a test environment simulating various fire scenarios;

[0014] A test module, used for performing a multi-parameter collaborative test on the smoke alarm in the test environment to obtain multi-dimensional raw data;

[0015] A processing module, used for performing multi-sensor fusion processing on the multi-dimensional raw data to obtain dimension-reduced feature data;

[0016] An analysis module, used for performing a multi-factor comprehensive analysis on the dimension reduction feature data to obtain a performance evaluation result of the smoke alarm;

[0017] An optimization module, used for performing smoke alarm simulation and configuration genetic optimization based on the performance evaluation results to obtain an optimized configuration solution;

[0018] The generation module is used to perform continuous sampling and Kalman filtering on the optimization configuration scheme to generate an intelligent smoke tracking strategy.

[0019] A third aspect of the present application provides a smoke alarm testing device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the smoke alarm testing device executes the above-mentioned smoke alarm testing method.

[0020] A fourth aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the above-mentioned smoke alarm testing method.

[0021] In the technical solution provided by the present application, the parameter setting of the multifunctional test platform realizes the accurate simulation of various fire scenes, improves the authenticity and comprehensiveness of the test environment, and provides test conditions closer to actual application for the performance evaluation of smoke alarms. The multi-parameter collaborative test method is adopted, and the influence of various environmental factors such as temperature, humidity, and airflow is considered, and more comprehensive multi-dimensional raw data is obtained, which effectively improves the accuracy and reliability of the test results. The multi-sensor fusion processing technology is introduced to realize the dimensionality reduction and feature extraction of the raw data, improve the efficiency and accuracy of data analysis, and provide more refined and effective data support for subsequent performance evaluation. Through multi-factor comprehensive analysis, a multi-dimensional evaluation index system is constructed to comprehensively evaluate the sensitivity, detection range, stability and other performance of smoke alarms. The configuration optimization is carried out by using genetic algorithms to realize the automatic optimization of the smoke alarm layout scheme, improve the overall coverage efficiency and response speed of the alarm system, and reduce the influence of human experience factors. The Kalman filter technology is used to realize the real-time estimation and prediction of the smoke diffusion state, generate an intelligent smoke tracking strategy, improve the dynamic response capability of the alarm system to the development of fire, and enhance the timeliness and accuracy of early warning. Through the hierarchical alarm strategy and dynamic adjustment mechanism, the intelligent management of the alarm system is realized, the probability of false alarms and missed alarms is reduced, and the reliability and practicality of the system are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0023] Figure 1 A schematic diagram of an embodiment of a method for testing a smoke alarm in an embodiment of the present application;

[0024] Figure 2 Schematic diagram of an embodiment of a testing device for a smoke alarm in an embodiment of the present application. DETAILED DESCRIPTION

[0025] The embodiments of the present application provide a method and device for testing a smoke alarm. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0026] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 , an embodiment of the test method of the smoke alarm in the embodiment of the present application includes:

[0027] Step S101, setting parameters of the multifunctional test platform to obtain a test environment simulating various fire scenarios;

[0028] It is understandable that the execution subject of the present application may be a test device for a smoke alarm, or a terminal or a server, which is not limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0029] Specifically, the type of smoke generator is selected to determine different smoke types, and according to these types, the smoke concentration control unit is gradient set to obtain a changing smoke concentration sequence. At the same time, the temperature range of the temperature control device is set to generate a temperature change sequence. According to the temperature change sequence, the humidity control device is adjusted accordingly to obtain a humidity change sequence corresponding to the temperature change. The speed and direction of the airflow generating device are set to form a parameter combination that can simulate different airflow conditions. According to the selected smoke type, smoke concentration sequence, temperature change sequence and humidity change sequence, the central control unit is programmed with a test matrix to obtain a series of test condition combinations. These combinations determine the simulation conditions under different test scenarios. In the test preparation process, the positions of multiple sensors in the test box are optimized to ensure that the most accurate data collection point distribution can be obtained under different conditions. According to these distribution conditions, the sampling frequency of different sensors is set to form a specific sampling strategy for different sensors. The smoke concentration measuring instrument is calibrated and set to ensure the accuracy of the measured smoke concentration data, and a reference data source for real-time monitoring of smoke concentration is obtained, which is used for real-time monitoring and correction of smoke concentration. According to the test condition combination, sampling strategy and benchmark data source, the sampling strategy and benchmark data source are integrated and verified to ensure that all links and equipment can work together to form a test environment that can simulate multiple fire scenarios. This environment is used to comprehensively evaluate the response performance and reliability of smoke alarms in different fire scenarios.

[0030] Step S102, performing a multi-parameter collaborative test on the smoke alarm in the test environment to obtain multi-dimensional raw data;

[0031] Specifically, by optimizing the installation position of the smoke alarm, the best test point is determined so that accurate and effective data can be obtained in the test. Based on this best test point, the smoke concentration is gradually increased, and the response time at each concentration is recorded to obtain the response time series. The response time series is statistically analyzed to draw the sensitivity curve of the smoke alarm and show its response ability at different concentrations. According to the sensitivity curve, the temperature change sequence is comprehensively traversed to obtain the temperature influence factor. In order to ensure the accuracy and comparability of the data, the temperature influence factor is normalized to obtain the temperature correction coefficient. The humidity change sequence is traversed using the temperature correction coefficient, and the response characteristics of the smoke alarm under different humidity conditions are recorded to obtain the humidity influence factor. The humidity influence factor is polynomially fitted to generate a humidity correction function. According to the temperature correction coefficient and the humidity correction function, the airflow condition is matrix tested to form an airflow influence matrix to evaluate the influence of the airflow speed and direction on the performance of the smoke alarm. The airflow influence matrix is ​​subjected to principal component analysis to extract the main influencing factors of the airflow, which are the key parameters affecting the performance of the smoke alarm. Combining the temperature correction coefficient, humidity correction function and main influencing factors of airflow, all test data are comprehensively corrected and correlated, and the comprehensive impact of various environmental factors on the response characteristics of smoke alarms is evaluated to obtain comprehensive multi-dimensional raw data.

[0032] Step S103, performing multi-sensor fusion processing on the multi-dimensional raw data to obtain dimension-reduced feature data;

[0033] Specifically, the multi-dimensional raw data is processed by sliding window averaging to eliminate the noise and fluctuation in the data and obtain a smooth pre-processed data set. The pre-processed data set is normalized and the data is converted into a standardized data matrix to ensure that sensor data of different dimensions can be compared and analyzed under the same benchmark. After the standardized data matrix, confidence analysis is performed to evaluate the reliability of each sensor data and obtain the confidence distribution. Based on the obtained confidence distribution, the sensor data is weighted fused, and different weights are assigned according to the confidence of each data point to generate preliminary fusion results, highlighting the importance of reliable data while reducing the influence of low-confidence data. The preliminary fusion results are analyzed to extract the feature vector after dimensionality reduction. The feature vector reflects the key information in the multi-dimensional data. According to the feature vector after dimensionality reduction, the feature model of normal state and alarm state is constructed to form a dual-state feature space to distinguish the performance of the smoke alarm in different states. The dual-state feature space is processed by the support vector machine method and a state classifier is trained. The classifier can classify the newly collected data in real time and determine whether the current state of the smoke alarm is normal or alarm. Perform time series analysis on the state judgment results, calculate the transition probability between each state, and form a state transition probability matrix. Generate dimensionality reduction feature data based on the state transition probability matrix and the dual-state feature space.

[0034] Step S104, performing a multi-factor comprehensive analysis on the dimension-reduced feature data to obtain a performance evaluation result of the smoke alarm;

[0035] Specifically, cluster analysis is performed on the reduced dimension feature data to identify the response feature clusters under different smoke concentrations. These feature clusters represent the different response behaviors of the smoke alarm in various concentration environments. According to the response feature clusters, the response time of the smoke alarm is statistically analyzed to obtain the response time distribution function, which can describe the response speed of the alarm under different smoke concentrations. The response time distribution function is parameter-fitted to generate a sensitivity model, which helps determine the sensitivity characteristics of the alarm. Using the sensitivity model, the detection limit of the smoke alarm is numerically solved to obtain the minimum detectable concentration, which indicates the minimum detection capability of the alarm. At the same time, a multivariate regression analysis is performed on the environmental factors in the reduced dimension feature data to generate an environmental impact coefficient matrix, which reflects the degree of influence of various environmental conditions on the detection performance of the alarm. Based on this matrix, the detection range of the smoke alarm is simulated and analyzed to determine the maximum detection range distribution, which indicates the area that the alarm can effectively cover under different conditions. The probability density of the maximum detection range distribution is estimated to obtain the effective coverage area function of the smoke alarm. This function indicates the detection coverage area of ​​the alarm under specific conditions. Based on this, the false alarm rate and missed alarm rate of the alarm were analyzed by ROC curve to determine the optimal working point, which represents the best detection sensitivity and balanced false alarm and missed alarm rate. A sensitivity analysis was performed on the optimal working point to evaluate the stability of the working point under different environmental conditions and obtain a stability score. Combining the sensitivity model, environmental impact coefficient matrix, effective coverage area function and stability score, a multidimensional evaluation index system was constructed to comprehensively evaluate the detection performance, coverage and reliability of smoke alarms in different environments, and generate comprehensive performance evaluation results.

[0036] Step S105: Perform smoke alarm simulation and configuration genetic optimization based on the performance evaluation results to obtain an optimized configuration solution;

[0037] Specifically, a set of key performance indicators is extracted from the performance evaluation results, including the response time, detection sensitivity and false alarm rate of the smoke alarm. Using the key performance indicators, a numerical model of smoke diffusion is constructed, and a virtual indoor environment is generated to simulate the diffusion behavior of smoke under different conditions. The virtual indoor environment is gridded to form a discretized computational domain for further numerical calculation and simulation analysis. Based on the discretized computational domain, Monte Carlo simulation technology is applied to generate a variety of smoke diffusion scenarios, which represent the diffusion path and concentration changes of smoke under different conditions. The configuration genetic optimization of a variety of smoke diffusion scenarios is performed to obtain the sensor layout scheme of the initial population, and the fitness function of each scheme is calculated, and these schemes are scored according to the fitness function. According to the scoring results, the sensor layout scheme is gradually optimized through the selection, crossover and mutation operations of the genetic algorithm, and the detection efficiency of the sensor layout is continuously improved. After multiple iterative analyses, this process continues until the preset termination condition is reached, and the optimal sensor layout scheme is finally determined. The sensitivity parameters of the optimal sensor layout scheme are optimized to ensure the response speed and accuracy of the sensor under different concentrations and environmental conditions. This process includes dual-level sensitivity configuration, which adjusts the sensor response at different detection levels to improve the detection efficiency and reliability of the overall system. Based on the optimal sensor layout and dual-level sensitivity configuration, an optimized configuration scheme for smoke alarms is comprehensively constructed. This scheme not only improves the detection capability and sensitivity of smoke alarms in various complex environments, but also optimizes the layout and configuration of sensors, so that they can respond to smoke alarm signals more accurately and efficiently in actual use.

[0038] Step S106: Continuously sample and perform Kalman filtering on the optimized configuration scheme to generate an intelligent smoke tracking strategy.

[0039] Specifically, the sensor arrangement in the optimization configuration scheme is initialized to form a sensor network topology. Based on the sensor network topology, a two-stage weighted average method model is established to integrate sensor data and improve the accuracy of data fusion. The two-stage weighted average method model is simulated and analyzed to obtain the positioning error distribution and show the error of each sensor in actual monitoring. According to the positioning error distribution, a smoke diffusion state estimation model is created, which is used to estimate the diffusion state of smoke under different environmental conditions. The parameters of the smoke diffusion state estimation model are optimized to obtain the optimal filter parameter set, which is used to enhance the accuracy and reliability of data processing. Using the optimal filter parameter set, the sensor data is continuously sampled and filtered to generate a real-time smoke concentration distribution estimation to reflect the concentration distribution of smoke in the current environment. Based on the real-time smoke concentration distribution estimation, a smoke diffusion trend prediction model is constructed to predict the diffusion path and speed of smoke. According to the smoke diffusion trend prediction model, a smoke diffusion simulation is performed to obtain a potential danger area warning map, which is used to identify areas that may be affected by smoke. The potential danger area warning map is threshold segmented, and a hierarchical alarm strategy is formulated. The alarm threshold is set according to different concentration levels so that the corresponding alarm mechanism can be triggered under different degrees of danger. Combining the hierarchical alarm strategy and real-time smoke concentration distribution estimation, the sampling frequency and sensitivity of the sensor are dynamically adjusted to ensure that the sensor can respond in time and provide accurate data when the smoke concentration changes. The dynamic adjustment process ultimately forms an intelligent smoke tracking strategy, which aims to improve the overall efficiency and responsiveness of the smoke detection system and ensure accurate and timely monitoring and early warning of smoke diffusion in actual applications.

[0040] In the embodiment of the present application, through the parameter setting of the multifunctional test platform, accurate simulation of various fire scenes is achieved, the authenticity and comprehensiveness of the test environment are improved, and test conditions closer to actual application are provided for the performance evaluation of smoke alarms. The multi-parameter collaborative test method is adopted, and the influence of various environmental factors such as temperature, humidity, and airflow is considered, and more comprehensive multi-dimensional raw data is obtained, which effectively improves the accuracy and reliability of the test results. The multi-sensor fusion processing technology is introduced to achieve dimensionality reduction and feature extraction of raw data, improve the efficiency and accuracy of data analysis, and provide more refined and effective data support for subsequent performance evaluation. Through multi-factor comprehensive analysis, a multi-dimensional evaluation index system is constructed to comprehensively evaluate the sensitivity, detection range, stability and other performance of smoke alarms. The configuration optimization is carried out using genetic algorithms to realize the automatic optimization of the smoke alarm layout scheme, improve the overall coverage efficiency and response speed of the alarm system, and reduce the influence of human experience factors. The Kalman filter technology is used to realize the real-time estimation and prediction of the smoke diffusion state, generate an intelligent smoke tracking strategy, improve the dynamic response capability of the alarm system to the development of fire, and enhance the timeliness and accuracy of early warning. Through the hierarchical alarm strategy and dynamic adjustment mechanism, the intelligent management of the alarm system is realized, the probability of false alarms and missed alarms is reduced, and the reliability and practicality of the system are improved.

[0041] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0042] (1) Select the type of smoke generator to obtain the smoke type, and according to the smoke type, set the gradient of the smoke concentration control unit to obtain the smoke concentration sequence;

[0043] (2) Setting the temperature range of the temperature control device to obtain a temperature change sequence, and adjusting the relative humidity of the humidity control device according to the temperature change sequence to obtain a humidity change sequence;

[0044] (3) Setting the speed and direction of the airflow generating device to obtain parameter combinations that simulate different airflow conditions;

[0045] (4) According to the smoke type, smoke concentration sequence, temperature change sequence and humidity change sequence, the central control unit is programmed with a test matrix to obtain a combination of test conditions;

[0046] (5) Optimize the position arrangement of multiple sensors in the test box to obtain the best data collection point distribution, and set the data sampling frequency according to the best data collection point distribution to obtain the sampling strategy for different sensors;

[0047] (6) The smoke density measuring instrument is calibrated and set up to obtain a benchmark data source for real-time monitoring of smoke density. The instrument is then integrated and verified based on the test condition combination, sampling strategy, and benchmark data source to obtain a test environment that simulates a variety of fire scenarios.

[0048] Specifically, determine the type of smoke required. Choose according to the different characteristics of the fire scene, including dust smoke, oil smoke, plastic burning smoke, etc. For example, when simulating a fire in an industrial environment, it is more appropriate to choose an oil smoke type smoke generator, because oily substances often burn in industrial scenes. By adjusting the output power of the smoke generator, the smoke concentration control unit is set to a gradient to simulate different concentrations of smoke in the test. Assume that a smoke generator that can generate a maximum concentration of The smoke generator can range from zero to It is divided into several levels at equal intervals, each level represents a specific concentration. The formula for this process is:

[0049] ;

[0050] in, represents the smoke density of the i-th level, is the total number of levels, is the index of the current level. Through the gradient setting, a series of smoke concentration sequences with different concentrations are obtained, which are used in subsequent test scenarios. The temperature range is set for the temperature control device to simulate the fire scene under different temperature conditions. The process of setting the temperature range is similar to the concentration gradient setting. The required temperature range (for example, 0°C to 100°C) is divided into several temperature levels, each corresponding to a specific temperature value. A temperature change sequence is generated to test the response of the smoke alarm under different temperature conditions. According to the set temperature change sequence, the parameters of the humidity control device are adjusted to obtain the corresponding humidity change sequence. This sequence reflects the relative humidity changes that may occur under different temperature conditions. For example, in a high temperature environment, the humidity may be low, while in a low temperature environment, the humidity may be high. In order to simulate the fire scene more realistically, the speed and direction of the airflow generating device are set to form different airflow conditions. Different airflow speeds and directions will affect the diffusion path and concentration distribution of the smoke. For example, when simulating an outdoor environment with a high wind speed, a larger airflow speed and a specific direction need to be set. These airflow conditions can be defined by a combination of parameters, such as using wind speed and wind direction To indicate that is the air flow velocity, is the airflow direction. The central control unit is programmed with a test matrix to generate a set of test condition combinations based on smoke type, smoke concentration sequence, temperature change sequence and humidity change sequence. Each row of the test matrix represents a specific test scenario, including specific parameters such as smoke concentration, temperature and humidity. The matrix ensures that all possible condition combinations are tested to fully evaluate the performance of the smoke alarm. The multiple sensors in the test chamber are optimally arranged to ensure that the best data can be obtained under various test conditions. The location optimization of the sensors usually takes into account factors such as the diffusion path of the smoke, temperature gradient and airflow influence. By optimizing the arrangement, the best data collection point distribution is obtained to ensure that the sensors can collect data at the appropriate location. Based on the optimized collection point distribution, the sampling frequency of each sensor is determined to ensure the accuracy and timeliness of the data. The sampling strategy of different sensors may be different, depending on the parameters they detect (such as concentration, temperature or humidity). The smoke concentration meter is calibrated and set up to ensure that the measuring instrument can accurately reflect the actual smoke concentration during the entire test. A reference data source is obtained by calibrating the measuring instrument with standard smoke of known concentration. The entire system is integrated and verified based on all the set test condition combinations, sampling strategies and benchmark data sources. The integration process includes checking the operating status of each device, ensuring the normal data collection and transmission, and verifying the correctness and consistency of all data to obtain a test environment that can simulate multiple fire scenarios.

[0051] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0052] (1) Optimize the installation position of the smoke alarm to obtain the best test point, and increase the smoke concentration gradient according to the best test point to obtain the response time series;

[0053] (2) Perform statistical analysis on the response time series to obtain a sensitivity curve, and based on the sensitivity curve, perform traversal tests on the temperature change series to obtain the temperature influence factor;

[0054] (3) Normalize the temperature influence factor to obtain the temperature correction coefficient, and then perform a traversal test on the humidity change sequence based on the temperature correction coefficient to obtain the humidity influence factor;

[0055] (4) Perform polynomial fitting on the humidity influencing factors to obtain the humidity correction function, and perform matrix testing on the airflow conditions based on the temperature correction coefficient and the humidity correction function to obtain the airflow influence matrix;

[0056] (5) Perform principal component analysis on the airflow influence matrix to obtain the main influencing factors of the airflow. Perform comprehensive correction and correlation analysis on the test data based on the temperature correction coefficient, humidity correction function and the main influencing factors of the airflow to obtain multidimensional original data.

[0057] Specifically, the characteristics of smoke diffusion and the structure of building space are analyzed. Smoke diffusion is affected by environmental factors such as airflow, temperature and humidity, and may show different concentration distributions at different locations in the building. Computational fluid dynamics simulation is used to analyze the concentration changes of smoke at different locations. By simulating smoke diffusion, the concentration distribution function at a certain location is determined. ,in is the spatial coordinate. Select the area with the highest concentration as the best test point for installing the alarm to ensure that the alarm can quickly detect smoke. Increase the smoke concentration gradient according to the best test point and measure the response time of the alarm. Assume that the output of the smoke generator can steadily increase the concentration, for example, at a rate (mg / m3 / s) increases, then at time The smoke concentration at can be expressed as:

[0058] ;

[0059] in, is the initial smoke concentration. Record the response time of the alarm at different concentrations , we can get the response time series The response time series is statistically analyzed to obtain the sensitivity curve. The sensitivity curve shows the response of the alarm at different concentrations. Its sensitivity is usually defined as the rate at which the response time changes with the concentration, that is:

[0060] ;

[0061] The sensitivity curve is analyzed to determine the performance of the alarm. In order to evaluate the impact of environmental factors, temperature and humidity changes are tested. In the temperature change test, the ambient temperature is set from arrive Changes, record the response time and sensitivity changes at each temperature. The temperature influence factor can be defined as:

[0062] ;

[0063] in, is the sensitivity at the reference temperature. Then the temperature effect factor is normalized to obtain the temperature correction factor. This correction factor helps to standardize the sensitivity performance under different temperature conditions. The humidity change is tested, and the response time and sensitivity change are recorded in a similar way to obtain the humidity effect factor. The polynomial fitting technique is used to fit the humidity effect factor data to obtain the humidity correction function. Assume that the fitting function of the humidity effect factor is:

[0064] ;

[0065] in, Represents humidity, is the parameter obtained by fitting. The polynomial fitting function can be used to correct the effect of humidity on the alarm response. The airflow conditions are tested in matrix form. The speed and direction of the airflow will significantly affect the diffusion path and concentration distribution of the smoke. By changing the airflow speed and direction, the sensitivity changes under each condition are recorded to form an airflow influence matrix. Each element in the matrix represents the sensitivity change value under a specific airflow condition. The airflow influence matrix is ​​subjected to principal component analysis to extract the main airflow influencing factors. The main factors help to reduce the complexity of the data and highlight the key environmental variables. Combined with the temperature correction coefficient, humidity correction function and the main influencing factors of the airflow, comprehensive correction and correlation analysis are performed to obtain multidimensional raw data. These data reflect the response characteristics of the smoke alarm under multiple environmental factors such as smoke concentration, temperature, humidity and airflow.

[0066] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0067] (1) Perform sliding window averaging on the multi-dimensional raw data to obtain a preprocessed data set, and normalize the preprocessed data set to obtain a standardized data matrix;

[0068] (2) Perform confidence analysis on the standardized data matrix to obtain the confidence distribution, and perform weighted fusion of sensor data based on the confidence distribution to obtain a preliminary fusion result;

[0069] (3) Analyze the preliminary fusion results to obtain the feature vector after dimensionality reduction, and construct the feature model of normal state and alarm state based on the feature vector after dimensionality reduction to obtain the dual-state feature space;

[0070] (4) Perform support vector machine processing on the dual-state feature space to obtain a state classifier, and perform real-time classification based on the state classifier to obtain the smoke state judgment result;

[0071] (5) Perform time series analysis on the smoke state judgment results to obtain the state transition probability matrix, and generate reduced-dimensional feature data based on the state transition probability matrix and the dual-state feature space.

[0072] Specifically, define the size of the sliding window, that is, the number of time steps or data points contained in the window. Sliding window averaging is a method of smoothing data by calculating the average of data points in each window to reduce noise in the data and reveal potential trends. Assume that the window size is , the original data sequence is , then The average value of the windows can be expressed as:

[0073] ;

[0074] in, For the The average value of windows. The obtained preprocessed data set can effectively smooth short-term fluctuations and extract long-term trends. The preprocessed data set is normalized to eliminate the influence of different dimensions and make the data easier to compare. Normalization usually scales the data to a specific range (such as [0, 1] or [-1, 1]). Through normalization, a standardized data matrix is ​​obtained. Each element in the matrix is ​​in the same scale range, which is convenient for subsequent analysis and processing. Confidence analysis is performed on the standardized data matrix to evaluate the credibility of each data point or sensor data. Confidence analysis can be implemented through statistical methods or machine learning methods, such as calculating the error distribution of each sensor or using a model based on historical data to estimate the credibility of current data to obtain a confidence distribution, which represents the credibility level of each data point. According to the confidence distribution, the sensor data is weighted fused. Different weights are assigned to data from different data sources according to the confidence. The higher the confidence, the greater the weight of the data. Assume that the first The data confidence of each sensor is , the original data is , then the weighted data can be expressed as:

[0075] ;

[0076] in, The result after weighted fusion is a combination of data from multiple sensors, taking into account the credibility of the data. The preliminary fusion results are analyzed to extract the feature vector after dimensionality reduction. Dimensionality reduction can be achieved through principal component analysis or other dimensionality reduction techniques, mapping high-dimensional data to low-dimensional space while retaining the main information of the data. The feature vector after dimensionality reduction represents the key features of the data, simplifies the structure of the data, and makes subsequent processing more efficient. Based on the feature vector after dimensionality reduction, the feature models of normal state and alarm state are constructed. These two models describe the typical characteristics of the system under normal and alarm conditions. The dual-state feature space is a multidimensional space composed of these two models, which is used to classify and identify the system state. The dual-state feature space is processed by support vector machine. Support vector machine is a commonly used classification algorithm that divides data into different categories by finding the best hyperplane. Support vector machine is used to distinguish between normal state and alarm state, and a state classifier is trained. The state classifier is used to classify new data in real time to obtain the smoke state judgment result, which indicates whether the current state of the system is normal or alarm. The state judgment result is analyzed in time series to obtain the state transition probability matrix. This matrix describes the probability of the system transitioning between different states and provides dynamic information about the system state changes. Reduced-dimensional feature data is generated based on the state transition probability matrix and the dual-state feature space. Reduced-dimensional feature data is a simplified data set that retains the key features and dynamic information of the data.

[0077] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0078] (1) Perform cluster analysis on the reduced dimension feature data to obtain the response feature clusters under different smoke concentrations. Based on the response feature clusters, perform statistical analysis on the response time of the smoke alarm to obtain the response time distribution function.

[0079] (2) Perform parameter fitting on the response time distribution function to obtain a sensitivity model, and based on the sensitivity model, perform numerical solution on the detection limit of the smoke alarm to obtain the minimum detectable concentration;

[0080] (3) Perform multiple regression analysis on the environmental factors in the reduced-dimensional feature data to obtain the environmental impact coefficient matrix. Based on the environmental impact coefficient matrix, simulate and analyze the detection range of the smoke alarm to obtain the maximum detection range distribution.

[0081] (4) Estimating the probability density of the maximum detection range distribution to obtain the effective coverage area function of the smoke alarm. Based on the effective coverage area function, perform ROC curve analysis on the false alarm rate and missed alarm rate of the smoke alarm to obtain the optimal working point.

[0082] (5) Conduct sensitivity analysis on the optimal operating point and obtain a stability score;

[0083] (6) Based on the sensitivity model, environmental impact coefficient matrix, effective coverage area function and stability score, a multidimensional evaluation index system is constructed to generate the comprehensive performance evaluation results of the smoke alarm.

[0084] Specifically, the K-means clustering algorithm is used to perform cluster analysis on the reduced-dimensional feature data. Assume that the data point is (i=1, 2, ..., n), the cluster center is (j=1, 2, ..., k), the goal of the K-means algorithm is to minimize the following objective function:

[0085] ;

[0086] in, is an indicator variable, when the data point When it belongs to the jth cluster, ,otherwise . represents the Euclidean distance. By repeatedly adjusting the cluster center and assign data points To the nearest cluster, the algorithm eventually converges to the local optimal solution and forms a response feature cluster. The response time of each cluster is statistically analyzed to generate a response time distribution function. Assuming the response time is , the concentration is , fit a model to describe the response time distribution. A common choice is to use the lognormal distribution, whose probability density function is:

[0087] ;

[0088] in, and is the parameter of the lognormal distribution, which can be obtained by maximum likelihood estimation. Based on this distribution, the sensitivity model is calculated, usually using the mean of the lognormal distribution and standard deviation To express:

[0089] ;

[0090] Through the sensitivity model, the minimum detectable concentration of the smoke alarm is solved. Assume that the sensitivity threshold of the alarm is , find the minimum concentration that makes the sensitivity reach the threshold

[0091] ;

[0092] Perform multiple regression analysis on the environmental factors in the reduced dimension feature data to determine the environmental factors (such as temperature ,humidity , air flow velocity The multiple regression model can be expressed as:

[0093] ;

[0094] in, is the response variable, such as response time or detection sensitivity, is the intercept, is the regression coefficient of each factor, is the error term. The regression coefficient matrix It can be solved by the least squares method:

[0095] ;

[0096] in, is the matrix containing all the independent variables, is the vector of response variables. When estimating the probability density of the maximum detection range distribution, the kernel density estimation method is usually used to obtain the estimated density function

[0097] ;

[0098] in, is a kernel function, such as a Gaussian kernel, is the bandwidth parameter. The kernel density estimation method provides a probability density estimate of the maximum detection range, which can be used to calculate the effective coverage area function of the alarm. It is expressed as:

[0099] ;

[0100] in, is the maximum detection radius, The radius The density function at . According to the effective coverage area function, the false alarm rate and false negative rate of the smoke alarm are analyzed by ROC curve. The ROC curve is a tool for evaluating the performance of a classifier, which shows the true positive rate (TPR) and false positive rate (FPR) at different thresholds. Through the ROC curve, the optimal working point is found, which represents the best detection threshold and can strike a balance between the false alarm rate and the false negative rate. The formula is expressed as:

[0101] ;

[0102] in, is the number of true positives, is the number of false negatives, is the number of false positives, is the number of true negatives. The optimal working point is usually obtained by maximizing the Youden index ( ). By changing the model parameters and analyzing their impact on performance indicators (such as AUC), a sensitivity analysis of the optimal operating point is performed. The sensitivity score can be expressed as:

[0103] ;

[0104] in, Indicates the parameters Combining all these analysis results, a multidimensional evaluation index system is constructed, including sensitivity model, environmental impact coefficient matrix, effective coverage area function and stability score.

[0105] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0106] (1) Parameters are extracted from the performance evaluation results to obtain the key performance indicator set of the smoke alarm. Based on the key performance indicator set, a numerical model of smoke diffusion is constructed and a virtual indoor environment is generated.

[0107] (2) The virtual indoor environment is meshed to obtain a discretized computational domain, and Monte Carlo simulation is performed based on the discretized computational domain to obtain a variety of smoke diffusion scenarios;

[0108] (3) Perform genetic optimization on various smoke diffusion scenarios to obtain the sensor layout scheme of the initial population. Based on the sensor layout scheme of the initial population, calculate the fitness function of each sensor layout scheme and obtain the scheme score.

[0109] (4) Perform selection, crossover and mutation operations on the scheme scores to obtain the target sensor layout scheme, and perform iterative analysis based on the target sensor layout scheme until the termination condition is reached to obtain the optimal sensor layout scheme;

[0110] (5) The sensitivity parameters of the optimal sensor layout scheme are optimized to obtain a two-level sensitivity configuration. Based on the optimal sensor layout scheme and the two-level sensitivity configuration, an optimal configuration scheme for the smoke alarm is constructed.

[0111] Specifically, determine the set of key performance indicators. These indicators may include response time, detection sensitivity, false alarm rate and missed alarm rate. By analyzing these indicators, determine the performance of each indicator under different conditions, such as response time under different smoke concentrations, temperatures and humidity. The extraction of the key performance indicator set can be completed by statistical analysis and regression analysis of the data, from which factors that have a significant impact on the performance of the alarm are extracted. These factors are used as input parameters for the subsequent construction of numerical models. Based on the extracted set of key performance indicators, a numerical model of smoke diffusion is constructed. The model helps to understand the distribution of smoke under different conditions by simulating the diffusion process of smoke in indoor environments. Numerical models usually use fluid dynamics methods, taking into account the diffusion rate, concentration gradient and turbulence effects of smoke. It is assumed that the diffusion of smoke can be expressed by the diffusion equation:

[0112] ;

[0113] in, Indicates smoke density. is the diffusion coefficient, is the air velocity, is the Laplace operator. By solving this partial differential equation, the distribution of smoke concentration at different time and space positions is obtained. When generating a virtual indoor environment, it is meshed to obtain a discretized computational domain. The purpose of meshing is to discretize the continuous indoor space for numerical calculation. The fineness of the grid directly affects the calculation accuracy and time. The smaller the grid unit, the more accurate the calculation result, but the greater the amount of calculation. In order to balance the calculation accuracy and time, the appropriate grid size is selected according to the specific situation. Based on the discretized computational domain, Monte Carlo simulation is performed to estimate the behavior of complex systems through random sampling. In the smoke diffusion scenario, multiple smoke diffusion scenarios are generated by multiple random simulations of different initial conditions and boundary conditions. These scenarios can cover different smoke concentrations, temperatures and airflow conditions, providing comprehensive test data. Genetic optimization is performed on multiple smoke diffusion scenarios to determine the best sensor layout scheme. Genetic algorithm is a bionic optimization algorithm that gradually optimizes the solution by simulating the natural selection process. Generate the sensor layout scheme of the initial population, and each individual represents a sensor layout method. For each scheme, its fitness function is calculated, and the fitness function is usually expressed as:

[0114] ;

[0115] in, is the value of the fitness function, It is The weight of the performance indicator, It is The algorithm calculates the scores of the performance indicators. These scores can be calculated based on performance indicators such as detection coverage and response time. The schemes are scored according to the value of the fitness function, and the schemes with higher fitness are selected for the next step. In the genetic algorithm, the scheme scores are subjected to selection, crossover and mutation operations to generate new schemes. These operations simulate the natural selection, hybridization and mutation processes in biological evolution. The selection operation selects excellent individuals according to the fitness value; the crossover operation exchanges part of the genes of two individuals to generate new individuals; the mutation operation randomly changes the genes of individuals to increase the diversity of the population. Through these operations, the target sensor layout scheme is obtained. The iterative process of the genetic algorithm continues until the preset termination condition is reached, such as reaching the maximum number of iterations or the improvement of the fitness function is less than a certain threshold. During the iteration process, the sensor layout is continuously optimized to obtain the optimal sensor layout scheme. The sensitivity parameters of the optimal sensor layout scheme are optimized to obtain a dual-level sensitivity configuration. This process involves adjusting the sensitivity of the sensor at different detection levels to improve the overall detection efficiency. The goal of the dual-level sensitivity configuration is to ensure rapid response under low-concentration smoke while maintaining accuracy under high-concentration smoke. This configuration is achieved by adjusting the detection threshold or amplification factor of the sensor. Based on the optimal sensor layout and dual-level sensitivity configuration, an optimized configuration of the smoke alarm is constructed. For example, suppose in a simulation, the optimal sensor layout finds that the sensors should be placed in the four corners of the room, and the dual-level sensitivity configuration requires the sensor to trigger an early warning at a smoke concentration below 0.1mg / m³ and an alarm at a concentration above 1mg / m³. This configuration not only ensures early warning, but also provides accurate alarm information under dense smoke conditions, thereby effectively improving the overall performance of the smoke alarm system.

[0116] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0117] (1) Initialize the sensor layout in the optimal configuration scheme to obtain the sensor network topology, and construct a two-stage weighted average method model based on the sensor network topology;

[0118] (2) The two-stage weighted average method model is simulated to obtain the positioning error distribution, and a smoke diffusion state estimation model is created based on the positioning error distribution;

[0119] (3) Optimize the parameters of the smoke diffusion state estimation model to obtain the optimal filter parameter set, and continuously sample and filter the sensor data based on the optimal filter parameter set to obtain a real-time smoke concentration distribution estimate;

[0120] (4) Generate a smoke diffusion trend prediction model based on the real-time smoke concentration distribution estimate, and perform smoke diffusion simulation based on the smoke diffusion trend prediction model to obtain a warning map of potential danger areas;

[0121] (5) Threshold segmentation is performed on the warning map of potential danger areas to obtain a graded alarm strategy. Based on the graded alarm strategy and real-time smoke concentration distribution estimation, the sensor sampling frequency and sensitivity are dynamically adjusted to obtain an intelligent smoke tracking strategy.

[0122] Specifically, the physical location and connection mode of the sensors are determined to form a complete sensor network topology. The sensor network topology determines the path of data collection and transmission and is the basis of the smoke detection system. The topology can be star, ring or grid, etc., depending on the environment and monitoring requirements. For example, in a large open space, a grid structure may be more suitable to provide more uniform coverage and higher redundancy. According to the sensor network topology, a two-stage weighted average method model is constructed to integrate data from different sensors to obtain a more accurate estimate of smoke concentration. The first stage is the data preprocessing stage, which improves data quality through denoising and standardization. The second stage is the weighted average stage, which gives different data different weights according to the confidence or historical performance of the sensor and calculates the weighted average. The two-stage weighted average method model is simulated to obtain the positioning error distribution. The positioning error refers to the difference between the estimated smoke concentration and the actual concentration. The accuracy and reliability of different sensor layout schemes are evaluated through simulation. The simulation results are usually displayed as a probability distribution of errors, which is used to judge the accuracy level of the system. According to the positioning error distribution, a smoke diffusion state estimation model is created. The model combines sensor data and positioning errors to estimate the current state of smoke diffusion. The model can use a Kalman filter to handle dynamic changes and uncertainties. The state equation of the Kalman filter can be expressed as:

[0123] ;

[0124] in, is the state vector (e.g., concentration distribution), is the state transition matrix, is the control matrix, is the process noise, is the observation vector, is the observation matrix, is the observation noise. The smoke diffusion state estimation model is optimized to obtain the optimal filter parameter set. These parameters include the state transfer matrix , control matrix , process noise covariance matrix and the observation noise covariance matrix Etc. The goal of optimization is to minimize the estimation error so that the model can more accurately predict the diffusion state of smoke. Through the optimal set of filtering parameters, the sensor data can be continuously sampled and filtered to obtain a real-time estimate of the smoke concentration distribution. Based on the real-time smoke concentration distribution estimate, a smoke diffusion trend prediction model is generated. The model is used to predict future smoke diffusion and help identify potential dangerous areas. The trend prediction model can predict future concentration changes through time series analysis methods such as autoregressive models or moving average models. Using this model, smoke diffusion simulation is performed to obtain a warning map of potential dangerous areas. The warning map of potential dangerous areas shows the smoke concentration distribution and possible high-risk areas at a certain point in the future. The warning map is threshold segmented, and different concentration thresholds are set to divide the risk level to obtain a hierarchical alarm strategy. For example, areas with concentrations exceeding a certain threshold are marked as high-risk areas, indicating that immediate action is needed, while areas below the threshold but still requiring attention are marked as medium-risk areas. Based on the hierarchical alarm strategy and real-time smoke concentration distribution estimates, the sampling frequency and sensitivity of the sensor are dynamically adjusted. The adjustment of the sampling frequency can be determined based on the rate of concentration change, such as increasing the sampling frequency in rapidly changing areas and reducing the frequency in relatively stable areas. The sensitivity is adjusted based on the absolute value and change trend of the concentration, ensuring that smoke can be detected even when the concentration is low, while avoiding false alarms when the concentration is high. Through adjustment and optimization, an intelligent smoke tracking strategy is formed.

[0125] The above describes the test method of the smoke alarm in the embodiment of the present application. The following describes the test device of the smoke alarm in the embodiment of the present application. Figure 2 In one embodiment of the present application, a test device for a smoke alarm includes:

[0126] The setting module 201 is used to set parameters of the multifunctional test platform to obtain a test environment simulating various fire scenarios;

[0127] The test module 202 is used to perform a multi-parameter collaborative test on the smoke alarm in the test environment to obtain multi-dimensional raw data;

[0128] The processing module 203 is used to perform multi-sensor fusion processing on the multi-dimensional raw data to obtain dimension-reduced feature data;

[0129] An analysis module 204 is used to perform a multi-factor comprehensive analysis on the dimension-reduced feature data to obtain a performance evaluation result of the smoke alarm;

[0130] An optimization module 205 is used to perform smoke alarm simulation and configuration genetic optimization based on the performance evaluation results to obtain an optimized configuration solution;

[0131] The generation module 206 is used to perform continuous sampling and Kalman filtering on the optimized configuration scheme to generate an intelligent smoke tracking strategy.

[0132] Through the cooperation of the above components and the parameter setting of the multifunctional test platform, the accurate simulation of various fire scenes is achieved, the authenticity and comprehensiveness of the test environment are improved, and the test conditions closer to actual application are provided for the performance evaluation of smoke alarms. The multi-parameter collaborative test method is adopted, and the influence of various environmental factors such as temperature, humidity, and airflow is considered, and more comprehensive multi-dimensional raw data is obtained, which effectively improves the accuracy and reliability of the test results. The introduction of multi-sensor fusion processing technology realizes the dimension reduction and feature extraction of raw data, improves the efficiency and accuracy of data analysis, and provides more refined and effective data support for subsequent performance evaluation. Through multi-factor comprehensive analysis, a multi-dimensional evaluation index system is constructed to comprehensively evaluate the sensitivity, detection range, stability and other performance of smoke alarms. The configuration optimization is carried out by using genetic algorithms to realize the automatic optimization of the smoke alarm layout plan, improve the overall coverage efficiency and response speed of the alarm system, and reduce the influence of human experience factors. The Kalman filter technology is used to realize the real-time estimation and prediction of the smoke diffusion state, generate an intelligent smoke tracking strategy, improve the dynamic response capability of the alarm system to the development of fire, and enhance the timeliness and accuracy of early warning. Through the hierarchical alarm strategy and dynamic adjustment mechanism, the intelligent management of the alarm system is realized, the probability of false alarms and missed alarms is reduced, and the reliability and practicality of the system are improved.

[0133] The present application also provides a smoke alarm testing device, the smoke alarm testing device comprising a memory and a processor, the memory storing computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes the steps of the smoke alarm testing method in the above-mentioned embodiments.

[0134] The present application also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the smoke alarm testing method.

[0135] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0136] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0137] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for testing a smoke alarm, characterized in that: The method comprises: Parameters of the multifunctional test platform are set to obtain a test environment that simulates various fire scenarios; Performing a multi-parameter collaborative test on the smoke alarm in the test environment to obtain multi-dimensional raw data; Performing multi-sensor fusion processing on the multi-dimensional raw data to obtain dimension-reduced feature data; Performing a multi-factor comprehensive analysis on the dimension-reduced feature data to obtain a performance evaluation result of the smoke alarm; Based on the performance evaluation results, smoke alarm simulation and configuration genetic optimization are performed to obtain an optimized configuration solution; The optimized configuration scheme is continuously sampled and processed by Kalman filtering to generate an intelligent smoke tracking strategy.

2. The smoke alarm testing method according to claim 1, characterized in that: The multifunctional test platform is parameterized to obtain a test environment simulating various fire scenarios, including: Selecting a type of smoke generator to obtain a smoke type, and setting a gradient of a smoke concentration control unit according to the smoke type to obtain a smoke concentration sequence; Setting a temperature range for the temperature control device to obtain a temperature change sequence, and adjusting the relative humidity for the humidity control device according to the temperature change sequence to obtain a humidity change sequence; Setting the speed and direction of the airflow generating device to obtain parameter combinations simulating different airflow conditions; According to the smoke type, the smoke concentration sequence, the temperature change sequence and the humidity change sequence, a central control unit is programmed with a test matrix to obtain a test condition combination; Optimizing the positions of multiple sensors in the test box to obtain an optimal data collection point distribution, and setting the data sampling frequency according to the optimal data collection point distribution to obtain sampling strategies for different sensors; The smoke density measuring instrument is calibrated and set to obtain a reference data source for real-time monitoring of smoke density, and is integrated and verified according to the test condition combination, the sampling strategy and the reference data source to obtain a test environment simulating various fire scenarios.

3. The smoke alarm testing method according to claim 2, characterized in that: The multi-parameter collaborative test is performed on the smoke alarm in the test environment to obtain multi-dimensional raw data, including: Optimizing the installation position of the smoke alarm to obtain an optimal test point, and increasing the smoke concentration gradient according to the optimal test point to obtain a response time series; Performing statistical analysis on the response time series to obtain a sensitivity curve, and performing traversal testing on the temperature change series according to the sensitivity curve to obtain a temperature influence factor; Normalizing the temperature influence factor to obtain a temperature correction coefficient, and performing a traversal test on the humidity change sequence according to the temperature correction coefficient to obtain a humidity influence factor; Performing polynomial fitting on the humidity influencing factor to obtain a humidity correction function, and performing a matrix test on the airflow condition according to the temperature correction coefficient and the humidity correction function to obtain an airflow influence matrix; The principal component analysis is performed on the airflow influence matrix to obtain the main influencing factors of the airflow, and comprehensive correction and correlation analysis of the test data are performed based on the temperature correction coefficient, the humidity correction function and the main influencing factors of the airflow to obtain multidimensional original data.

4. The method for testing a smoke alarm according to claim 3, characterized in that: The performing multi-sensor fusion processing on the multi-dimensional raw data to obtain dimension-reduced feature data includes: Performing sliding window averaging processing on the multidimensional raw data to obtain a preprocessed data set, and performing normalization processing on the preprocessed data set to obtain a standardized data matrix; Performing confidence analysis on the standardized data matrix to obtain a confidence distribution, and performing weighted fusion of sensor data according to the confidence distribution to obtain a preliminary fusion result; Analyzing the preliminary fusion result to obtain a feature vector after dimension reduction, and constructing a feature model of a normal state and an alarm state according to the feature vector after dimension reduction to obtain a dual-state feature space; Performing support vector machine processing on the dual-state feature space to obtain a state classifier, and performing real-time classification according to the state classifier to obtain a smoke state judgment result; A time series analysis is performed on the smoke state judgment result to obtain a state transition probability matrix, and dimension reduction feature data is generated according to the state transition probability matrix and the dual-state feature space.

5. The method for testing a smoke alarm according to claim 4, characterized in that: The multi-factor comprehensive analysis of the dimension reduction feature data to obtain the performance evaluation result of the smoke alarm includes: Performing cluster analysis on the dimension reduction feature data to obtain response feature clusters under different smoke concentrations, and performing statistical analysis on the response time of the smoke alarm based on the response feature clusters to obtain a response time distribution function; Performing parameter fitting on the response time distribution function to obtain a sensitivity model, and numerically solving the detection limit of the smoke alarm based on the sensitivity model to obtain a minimum detectable concentration; Performing a multivariate regression analysis on the environmental factors in the dimension reduction feature data to obtain an environmental impact coefficient matrix, and performing a simulation analysis on the detection range of the smoke alarm based on the environmental impact coefficient matrix to obtain a maximum detection range distribution; Performing probability density estimation on the maximum detection range distribution to obtain an effective coverage area function of the smoke alarm, and performing ROC curve analysis on the false alarm rate and missed alarm rate of the smoke alarm based on the effective coverage area function to obtain an optimal working point; Performing sensitivity analysis on the optimal working point to obtain a stability score; According to the sensitivity model, the environmental impact coefficient matrix, the effective coverage area function and the stability score, a multi-dimensional evaluation index system is constructed and a comprehensive performance evaluation result of the smoke alarm is generated.

6. The method for testing a smoke alarm according to claim 5, characterized in that: The smoke alarm simulation and configuration genetic optimization are performed based on the performance evaluation result to obtain an optimized configuration scheme, including: Extracting parameters from the performance evaluation results to obtain a key performance indicator set of the smoke alarm, and constructing a numerical model of smoke diffusion and generating a virtual indoor environment based on the key performance indicator set; Meshing the virtual indoor environment to obtain a discretized computational domain, and performing Monte Carlo simulation based on the discretized computational domain to obtain a variety of smoke diffusion scenarios; Performing genetic optimization on the configuration of the multiple smoke diffusion scenarios to obtain sensor layout schemes of an initial population, and calculating the fitness function of each sensor layout scheme based on the sensor layout scheme of the initial population to obtain a scheme score; Performing selection, crossover and mutation operations on the scheme scores to obtain a target sensor layout scheme, and performing iterative analysis according to the target sensor layout scheme until a termination condition is reached to obtain an optimal sensor layout scheme; The optimal sensor arrangement scheme is optimized for sensitivity parameters to obtain a dual-stage sensitivity configuration, and an optimized configuration scheme for the smoke alarm is constructed based on the optimal sensor arrangement scheme and the dual-stage sensitivity configuration.

7. The method for testing a smoke alarm according to claim 6, characterized in that: The step of continuously sampling and Kalman filtering the optimized configuration scheme to generate an intelligent smoke tracking strategy includes: Initializing the sensor arrangement in the optimization configuration scheme to obtain a sensor network topology structure, and constructing a two-stage weighted average method model according to the sensor network topology structure; Simulating the two-stage weighted average method model to obtain a positioning error distribution, and creating a smoke diffusion state estimation model based on the positioning error distribution; Optimizing the parameters of the smoke diffusion state estimation model to obtain an optimal filter parameter set, and continuously sampling and filtering the sensor data according to the optimal filter parameter set to obtain a real-time smoke concentration distribution estimate; Generate a smoke diffusion trend prediction model based on the real-time smoke concentration distribution estimation, and perform smoke diffusion simulation based on the smoke diffusion trend prediction model to obtain a potential danger area warning map; The potential danger area warning map is threshold segmented to obtain a graded alarm strategy, and the sensor sampling frequency and sensitivity are dynamically adjusted according to the graded alarm strategy and the real-time smoke concentration distribution estimation to obtain an intelligent smoke tracking strategy.

8. A smoke alarm testing device, characterized in that: The device for performing the smoke alarm testing method according to any one of claims 1 to 7 comprises: A setting module is used to set parameters of the multifunctional test platform to obtain a test environment simulating various fire scenarios; A test module, used for performing a multi-parameter collaborative test on the smoke alarm in the test environment to obtain multi-dimensional raw data; A processing module, used for performing multi-sensor fusion processing on the multi-dimensional raw data to obtain dimension-reduced feature data; An analysis module, used for performing a multi-factor comprehensive analysis on the dimension reduction feature data to obtain a performance evaluation result of the smoke alarm; An optimization module, used for performing smoke alarm simulation and configuration genetic optimization based on the performance evaluation results to obtain an optimized configuration solution; The generation module is used to perform continuous sampling and Kalman filtering on the optimization configuration scheme to generate an intelligent smoke tracking strategy.

9. A smoke alarm testing device, characterized in that: The smoke alarm test device comprises: a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instructions in the memory to enable the smoke alarm testing device to execute the smoke alarm testing method according to any one of claims 1 to 7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by a processor, a method for testing a smoke alarm according to any one of claims 1 to 7 is implemented.

Images