A temperature sensor-based fire extinguishing system response time testing method

Abstract

The application discloses a kind of based on temperature sensor's fire extinguishing system response time test method, comprising the following steps: constructing fire extinguishing system response time test environment;Set the excitation parameter of fire source excitation device, temperature data sampling frequency and test starting condition;Form temperature time series data;Pretreatment is obtained to standardized temperature time series data;Temperature evolution state trajectory is constructed using delay embedding method, and state characteristic parameter is extracted;Classification determination is carried out using improved ThermoONet model, and the classification result of fire source development state and fire extinguishing intervention state is obtained, and state transition moment is determined;Fire extinguishing system response time test value is obtained, test result is formed, realizes the dynamic identification and accurate determination of fire extinguishing system response process, and the accuracy and stability of response time test are improved.

Description

Technical Field

[0001] This invention relates to the field of fire safety detection and intelligent testing, and in particular to a method for testing the response time of a fire extinguishing system based on a temperature sensor. Background Technology

[0002] Fire suppression systems, as crucial safety devices in various buildings, industrial facilities, and special environments, directly impact the effectiveness of initial fire control and the safety of personnel and property. In practical applications, the response time of a fire suppression system—the time interval between the appearance of a fire source or the achievement of triggering conditions and the activation of the system to provide effective intervention—is typically considered a key performance indicator. Therefore, accurately, stably, and repeatably testing the response time of fire suppression systems has become a critical technical issue in the field of fire protection engineering testing.

[0003] Existing technologies for testing the response time of fire extinguishing systems mainly include manual observation, simple temperature threshold triggering, and time recording based on single-point sensors. Manual observation typically relies on testers to visually assess changes in the fire source and the activation time of the fire extinguishing system through visual inspection or video playback. This method is highly subjective, easily influenced by human experience and environmental conditions, and struggles to achieve high precision and standardized testing. Temperature threshold triggering uses single or a small number of temperature sensors, recording the time point when the detected temperature reaches a preset threshold as a reference. However, this method often ignores the dynamic temperature changes during fire development and is easily affected by local temperature fluctuations or environmental disturbances, leading to deviations in response time determination. Furthermore, testing methods based on single points or a small number of measuring points cannot reflect the spatial distribution characteristics of the temperature field within the test area, making it difficult to accurately characterize the overall state changes during fire source development and the fire extinguishing process. Summary of the Invention

[0004] One objective of this invention is to propose a method for testing the response time of a fire extinguishing system based on a temperature sensor. This invention fully utilizes multi-point temperature sensor data acquisition and preprocessing technology, introduces a delayed embedding method to model the state of the temperature time series, and uses an improved ThermoONet model to determine the temperature evolution state. Furthermore, based on state space and attractor region identification, it realizes the state transition identification during the fire extinguishing system response process, thereby completing the accurate calculation of the response time. This invention can continuously characterize the dynamic process of temperature evolution and has the advantages of high testing accuracy, strong anti-interference ability, and high result stability.

[0005] A method for testing the response time of a fire extinguishing system based on a temperature sensor, according to an embodiment of the present invention, includes the following steps:

[0006] Construct a test environment for the response time of a fire extinguishing system. The test environment for the response time of a fire extinguishing system includes the fire extinguishing system under test, the test area, the fire source excitation device, the temperature sensor, the data acquisition unit, and the data processing unit.

[0007] Each temperature sensor was calibrated and time-synchronized. The excitation parameters of the fire source excitation device, the temperature data sampling frequency, and the test start conditions were set.

[0008] The fire source excitation device is activated and controlled thermal excitation is applied to the test area according to the excitation parameters. When the temperature reaches the preset temperature threshold, the fire extinguishing system under test is activated. The data acquisition unit synchronously collects the real-time temperature data of each temperature sensor according to the temperature data sampling frequency to form temperature time series data.

[0009] The temperature time series data is preprocessed by the data processing unit to obtain standardized temperature time series data;

[0010] Based on standardized temperature time series data, a delayed embedding method is used to construct the temperature evolution state trajectory and extract state feature parameters that characterize the dynamic evolution behavior of temperature.

[0011] The state feature parameters are input into the improved ThermoONet model to classify and determine the temperature evolution state, obtain the classification results of the fire source development state and the fire extinguishing intervention state, and determine the state transition time when the temperature evolution state changes based on the classification results.

[0012] The start time corresponding to the test start condition is taken as the test start time, and the state transition time is taken as the fire extinguishing system response time. The corresponding time difference is calculated to obtain the test value of the fire extinguishing system response time, thus forming the test result.

[0013] Optionally, the fire extinguishing system under test is installed in a closed or semi-closed test area, and the spatial range of the test area is determined according to the spray range and effective area of ​​the fire extinguishing system under test. Temperature sensors are arranged in the test area according to the preset spatial positions, and the temperature sensors are electrically connected to the data acquisition unit. The data acquisition unit and the data processing unit establish a data communication connection. A fire source excitation device is set in the test area. The fire extinguishing system under test, the fire source excitation device, the temperature sensors, the data acquisition unit and the data processing unit are configured as a whole to form a test environment for the response time of the fire extinguishing system.

[0014] Optionally, the setting of the excitation parameters, temperature data sampling frequency, and test start conditions of the fire source excitation device specifically includes:

[0015] Each temperature sensor is calibrated using a standard temperature source. The calibration coefficient and offset of each temperature sensor are obtained. The output value of the temperature sensor is corrected based on the calibration coefficient and offset to obtain the calibrated temperature data of each temperature sensor at each time.

[0016] Based on the calibrated temperature data, time synchronization processing is performed on each temperature sensor, a unified time reference is set, and the time offset of each temperature sensor is calculated. The acquisition time of each temperature sensor is then corrected to obtain time-aligned temperature data.

[0017] Based on time-aligned temperature data, a temperature data sampling frequency is set, and a sampling time interval is determined according to the temperature data sampling frequency to unify the data acquisition time interval of each temperature sensor.

[0018] Under a unified data acquisition time interval, the excitation parameters of the fire source excitation device are set;

[0019] Based on the excitation parameters and temperature data sampling frequency, the test start conditions are set.

[0020] Optionally, the formation of the temperature time series data specifically includes:

[0021] The fire source excitation device is activated at the start time corresponding to the test start conditions, and controlled thermal excitation is continuously applied to the test area according to the excitation parameters;

[0022] During the continuous operation of controlled thermal excitation, each temperature sensor continuously monitors the temperature field within the test area and outputs corresponding real-time temperature data.

[0023] The data acquisition unit synchronously samples the real-time temperature data of each temperature sensor based on the temperature data sampling frequency to obtain discrete time-series temperature data.

[0024] The real-time temperature data from each temperature sensor at each sampling time are combined and arranged in chronological order of sampling time to form temperature time series data.

[0025] Optionally, obtaining the standardized temperature time series data specifically includes:

[0026] The data processing unit performs noise reduction processing on the temperature time series data to obtain the corresponding denoised temperature data.

[0027] Based on the denoised temperature data, the data processing unit performs outlier removal processing to obtain temperature time series data after removing outlier data.

[0028] Based on the temperature time series data after removing outlier data, the data processing unit performs data completion processing on the missing data positions generated after the removal. Linear interpolation is used to complete the missing data positions to obtain complete temperature time series data.

[0029] Based on complete temperature time series data, the data processing unit performs standardization processing to obtain standardized temperature time series data.

[0030] Optionally, the extraction of the state feature parameters specifically includes:

[0031] Based on standardized temperature time series data, a delayed embedding process is performed on each temperature sensor at each sampling time to construct the corresponding temperature evolution state vector;

[0032] At each sampling moment, the temperature evolution state vectors corresponding to each temperature sensor are combined to form the system temperature evolution state vector.

[0033] According to the sampling time sequence, the system temperature evolution state vectors corresponding to each sampling time are connected sequentially to construct a continuously changing temperature evolution state trajectory.

[0034] Based on the temperature evolution state trajectory, the data processing unit compares the system temperature evolution state vectors corresponding to adjacent sampling times to obtain the state change between each sampling time.

[0035] Based on the temperature evolution trajectory and state change, the data processing unit extracts state characteristic parameters that characterize the dynamic evolution behavior of temperature.

[0036] Optionally, determining the state transition time specifically includes:

[0037] The improved ThermoONet model incorporates state feature parameters as input. This improved model includes a temperature evolution state embedding module, an operator mapping backbone module, a spatiotemporal coupling enhancement module, and a state determination output module. The improvements are as follows: Traditional ThermoONet models rely on function-to-function mappings and operator learning on the input sequence, depending on instantaneous feature mapping results during state determination. The improved ThermoONet model introduces a temperature evolution state embedding module, mapping state feature parameters to a unified state space to form a state sequence with temporal evolution continuity. The operator mapping backbone module uses a Branch-Trunk dual-pathway structure to couple the state function with the time basis function. The spatiotemporal coupling enhancement module performs structured adjustment of the time lag component and sensor channel component based on state changes. The state determination output module constructs a temperature evolution state phase space and divides the state convergence behavior based on attractor regions, resulting in a state determination result oriented towards the dynamic process of temperature evolution.

[0038] In the temperature evolution state embedding module, the state feature parameters corresponding to each sampling time are embedded into a unified state space in the order of time evolution to form an embedded state feature sequence.

[0039] In the operator mapping backbone module, the embedded state feature sequence is input into the Branch-Trunk dual-path operator learning structure. The Branch-Trunk dual-path operator learning structure includes a Branch path and a Trunk path. The Branch path performs feature mapping on the embedded state feature sequence, and the Trunk path performs basis function mapping on the temporal position information corresponding to each sampling time. The mapping results of the Branch path and the Trunk path are coupled to obtain the implicit representation of the temperature evolution state corresponding to each sampling time.

[0040] In the spatiotemporal coupling enhancement module, cross-delay gating enhancement and cross-sensor channel enhancement are applied to the implicit representation of temperature evolution state corresponding to each sampling time to obtain the enhanced implicit representation of temperature evolution state.

[0041] In the state determination output module, a temperature evolution state phase space is constructed based on the enhanced temperature evolution state implicit representation. The attractor region of the temperature evolution state is identified in the temperature evolution state phase space. The enhanced temperature evolution state implicit representation is divided according to the attractor region to which it converges. The enhanced temperature evolution state implicit representation that converges to the attractor region corresponding to continuous temperature rise is determined as the fire source development state. The enhanced temperature evolution state implicit representation that converges to the attractor region corresponding to temperature drop or stabilization is determined as the fire extinguishing intervention state. The classification results of the fire source development state and the fire extinguishing intervention state corresponding to each sampling time are obtained.

[0042] The classification results at each sampling time are continuously compared according to the sampling time sequence to determine the sampling time when the classification result changes from the fire source development state to the fire extinguishing intervention state, and the sampling time is determined as the state transition time when the temperature evolution state changes.

[0043] Optionally, the formation of the test results specifically includes:

[0044] Based on the start time corresponding to the test start conditions, the start time is determined as the test start time, and the state transition time is determined as the fire extinguishing system response time;

[0045] The data processing unit calculates the time difference between the test start time and the fire extinguishing system response time to obtain the test value of the fire extinguishing system response time.

[0046] The data processing unit associates and stores the test value of the fire extinguishing system response time with the corresponding test start time, the fire extinguishing system response time, and the classification results corresponding to each sampling time to form the test results.

[0047] The beneficial effects of this invention are:

[0048] This invention constructs a standardized testing environment that includes a fire source excitation device, multi-point temperature sensors, and a data processing unit. This enables multi-dimensional continuous acquisition of the temperature field in the testing area. By combining temperature sensor calibration with time synchronization processing, the consistency of temperature time series data in time and space is ensured. This effectively overcomes the testing error problems caused by data deviation and asynchronous acquisition in the prior art, and improves the reliability and comparability of the original data.

[0049] This invention systematically preprocesses temperature time-series data, including denoising, outlier removal, and standardization, ensuring the data has a uniform scale and a high signal-to-noise ratio before entering subsequent modeling stages. This significantly reduces the impact of environmental disturbances and occasional anomalies on test results. Simultaneously, by introducing a delayed embedding method to construct a temperature evolution trajectory, the original temperature data is expanded from a single-point-in-time value to a state representation containing historical information. This reflects the dynamic continuity and evolutionary patterns of temperature changes, overcoming the shortcomings of existing technologies that rely solely on instantaneous features for judgment.

[0050] This invention employs an improved ThermoONet model to analyze temperature evolution. It utilizes an operator learning structure to map and model the dynamic temperature process, and combines a spatiotemporal coupling enhancement mechanism to unify the processing of time-lag information and multi-sensor spatial information. This allows the model to simultaneously perceive the temporal evolution characteristics and spatial distribution features of temperature changes. Furthermore, by constructing a phase space of the temperature evolution state and identifying attractor regions, the problem of temperature change during fire extinguishing is transformed from a simple numerical change problem into a problem of identifying state convergence behavior. This achieves a stable characterization of the transition process between the fire source development state and the fire extinguishing intervention state, effectively avoiding misjudgments caused by short-term fluctuations in traditional methods.

[0051] This invention accurately determines the state transition moment when the temperature evolution state changes by continuously comparing the state classification results, and calculates the corresponding value of the fire extinguishing system response time by comparing it with the test start time. This method not only improves the accuracy and stability of response time determination, but also enhances the repeatability and consistency of test results under different test conditions, thus more realistically reflecting the actual response performance of the fire extinguishing system and possessing high engineering application value. Attached Figure Description

[0052] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0053] Figure 1 This is an overall flowchart of a fire extinguishing system response time testing method based on a temperature sensor proposed in this invention;

[0054] Figure 2 This is a schematic diagram illustrating the construction of temperature time series data for a fire extinguishing system response time testing method based on a temperature sensor proposed in this invention.

[0055] Figure 3 This is a schematic diagram of the structure of the improved ThermoONet model for a fire extinguishing system response time testing method based on a temperature sensor proposed in this invention. Detailed Implementation

[0056] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0057] refer to Figures 1-3 A method for testing the response time of a fire extinguishing system based on a temperature sensor, comprising the following steps:

[0058] Construct a test environment for the response time of a fire extinguishing system. The test environment for the response time of a fire extinguishing system includes the fire extinguishing system under test, the test area, the fire source excitation device, the temperature sensor, the data acquisition unit, and the data processing unit.

[0059] Each temperature sensor was calibrated and time-synchronized. The excitation parameters of the fire source excitation device, the temperature data sampling frequency, and the test start conditions were set.

[0060] The fire source excitation device is activated and controlled thermal excitation is applied to the test area according to the excitation parameters. When the temperature reaches the preset temperature threshold, the fire extinguishing system under test is activated. The data acquisition unit synchronously collects the real-time temperature data of each temperature sensor according to the temperature data sampling frequency to form temperature time series data.

[0061] The temperature time series data is preprocessed by the data processing unit to obtain standardized temperature time series data;

[0062] Based on standardized temperature time series data, a delayed embedding method is used to construct the temperature evolution state trajectory and extract state feature parameters that characterize the dynamic evolution behavior of temperature.

[0063] The state feature parameters are input into the improved ThermoONet model to classify and determine the temperature evolution state, obtain the classification results of the fire source development state and the fire extinguishing intervention state, and determine the state transition time when the temperature evolution state changes based on the classification results.

[0064] The start time corresponding to the test start condition is taken as the test start time, and the state transition time is taken as the fire extinguishing system response time. The corresponding time difference is calculated to obtain the test value of the fire extinguishing system response time, thus forming the test result.

[0065] In this embodiment, the fire extinguishing system under test is installed in a closed or semi-closed test area. The spatial range of the test area is determined according to the spray range and effective area of ​​the fire extinguishing system under test. Temperature sensors are arranged in the test area according to preset spatial locations, including locations near the fire source activation device, locations far from the fire source activation device, and locations corresponding to the spray path of the fire extinguishing system under test. The temperature sensors are electrically connected to the data acquisition unit, and the data acquisition unit establishes a data communication connection with the data processing unit. A fire source activation device is set in the test area. The fire extinguishing system under test, the fire source activation device, the temperature sensors, the data acquisition unit, and the data processing unit are configured as a whole to form a test environment for the response time of the fire extinguishing system.

[0066] In this embodiment, the setting of the excitation parameters of the fire source excitation device, the temperature data sampling frequency, and the test start conditions specifically includes:

[0067] Each temperature sensor is calibrated using a standard temperature source. The calibration coefficient and offset of each temperature sensor are obtained. The output value of the temperature sensor is corrected based on the calibration coefficient and offset to obtain the calibrated temperature data of each temperature sensor at each time.

[0068] The specific process for obtaining calibrated temperature data is as follows: Each temperature sensor is placed in a stable temperature environment formed by a standard temperature source in sequence. The original temperature data corresponding to each temperature sensor is collected under multiple known temperature points. The original temperature data is then correlated with the actual temperature value of the standard temperature source. Based on the correspondence, the calibration coefficient and offset of each temperature sensor are calculated. After determining the calibration coefficient and offset, the original temperature data output by each temperature sensor at each moment is collected in real time during the actual test. The original temperature data is multiplied by the corresponding calibration coefficient and then the offset is added to obtain the calibrated temperature data of each temperature sensor at each moment.

[0069] Based on the calibrated temperature data, time synchronization processing is performed on each temperature sensor, a unified time reference is set, and the time offset of each temperature sensor is calculated. The acquisition time of each temperature sensor is then corrected to obtain time-aligned temperature data.

[0070] The process of obtaining time-aligned temperature data is as follows: using the system clock of the data acquisition unit as a unified time reference, the calibrated temperature data output by each temperature sensor within the same sampling period are time-stamped; the sampling time of the reference temperature sensor is selected as the reference time series, and the calibrated temperature data collected by other temperature sensors within the corresponding time period are matched with the reference time series; by comparing the characteristic change times of each temperature sensor in the same temperature change process, the time offset of each temperature sensor relative to the reference time series is determined; the sampling time of each temperature sensor is uniformly corrected according to the time offset to obtain time-aligned temperature data.

[0071] Based on time-aligned temperature data, a temperature data sampling frequency is set, and a sampling time interval is determined according to the temperature data sampling frequency to unify the data acquisition time interval of each temperature sensor.

[0072] The process of unifying the data acquisition time interval of each temperature sensor is as follows: Based on time-aligned temperature data analysis, the rate and magnitude of temperature change during the temperature change process are determined. The characteristic time scale of temperature change is determined according to the change of temperature difference between adjacent moments during the temperature change process, and the temperature data sampling frequency is set in combination with the characteristic time scale. After determining the temperature data sampling frequency, the number of samples per unit time is converted into the time interval between two adjacent samples. The time interval corresponding to a single sample is obtained by dividing the unit time length by the temperature data sampling frequency. The time interval is written as the sampling period parameter of the data acquisition unit into the timing control module of the data acquisition unit. The timing control module periodically triggers each temperature sensor to acquire data according to the sampling period parameter, thus completing the unification of the data acquisition time interval.

[0073] Under a unified data acquisition time interval, the excitation parameters of the fire source excitation device are set;

[0074] The specific process for setting the excitation parameters is as follows: Under a unified data acquisition time interval, the thermal power and duration of the fire source excitation device during the test are determined based on the spatial range of the test area, the effective range of the fire extinguishing system under test, and the expected simulated fire source intensity level; by selecting historical test data or preset test conditions, the reference heat release level under the corresponding fire source intensity level is obtained, and the reference heat release level is used as the target thermal power of the fire source excitation device; the duration of the fire source excitation device is set according to the activation characteristics of the fire extinguishing system under test and the time range required for the response process.

[0075] Based on the excitation parameters and temperature data sampling frequency, test start conditions are set, including the start time of the fire source excitation device and the initial ambient temperature.

[0076] The specific process for setting the test start conditions is as follows: the test area is monitored in its initial state. Under the condition that the fire source excitation device is not activated, the temperature sensor continuously collects the temperature data in the test area according to the temperature data sampling frequency. When the temperature data collected by each temperature sensor remains stable within the continuous sampling period and the fluctuation range is within the preset range, the temperature value under the stable state is determined as the initial ambient temperature. After determining the initial ambient temperature, the start time of the fire source excitation device is set, and the start time is used as the time reference in the test start conditions. The start time is associated with the sampling control of the data acquisition unit to complete the setting of the test start conditions.

[0077] In this embodiment, the formation of temperature time series data specifically includes:

[0078] The fire source excitation device is activated at the start time corresponding to the test start conditions, and controlled thermal excitation is continuously applied to the test area according to the excitation parameters;

[0079] The process of continuously applying controlled thermal excitation is as follows: at the start time corresponding to the test start conditions, the control unit of the fire source excitation device outputs thermal excitation to the test area according to the preset excitation parameters. The control unit adjusts the power of the heating element of the fire source excitation device according to the set thermal power, so that the fire source excitation device maintains a stable thermal output throughout the excitation process. During the continuous excitation process, the control unit monitors the output status of the fire source excitation device in real time and makes dynamic adjustments according to the set thermal power to compensate for the impact of environmental heat dissipation or device fluctuations on the thermal output, so that the fire source excitation device continuously outputs controlled thermal excitation to the test area within the set duration until the duration is reached and the thermal excitation output stops.

[0080] During the continuous operation of controlled thermal excitation, each temperature sensor continuously monitors the temperature field within the test area and outputs corresponding real-time temperature data.

[0081] The real-time temperature data output process is as follows: During the continuous operation of controlled thermal excitation, each temperature sensor is deployed within the test area to continuously sense the temperature field and convert the sensed temperature changes into corresponding electrical signals; the signal conditioning circuit inside the temperature sensor amplifies and filters the electrical signals; the conditioned electrical signals are input to the data acquisition unit, which performs analog-to-digital conversion on the electrical signals to obtain the corresponding digital temperature data; the data acquisition unit periodically reads the output of each temperature sensor according to the temperature data sampling frequency and outputs the digital temperature data collected at each moment as the real-time temperature data at the corresponding moment;

[0082] The data acquisition unit synchronously samples the real-time temperature data of each temperature sensor based on the temperature data sampling frequency. Taking the start time corresponding to the test start condition as the starting time point, the real-time temperature data of each temperature sensor at each sampling time is collected according to a uniform sampling time interval to obtain discrete time-series temperature data.

[0083] The real-time temperature data from each temperature sensor at each sampling time are combined and arranged in chronological order of sampling time to form temperature time series data.

[0084] In this embodiment, obtaining standardized temperature time series data specifically includes:

[0085] The data processing unit performs noise reduction on the temperature time series data. The real-time temperature data of each temperature sensor at each sampling time is processed by the sliding window averaging method. At each sampling time, the real-time temperature data of that time and the previous consecutive sampling times are summed and averaged to obtain the corresponding noise-reduced temperature data.

[0086] Based on the denoised temperature data, the data processing unit performs abnormal data removal processing. It calculates the corresponding average value and standard deviation of the denoised temperature data of each temperature sensor. When the absolute value of the difference between the denoised temperature data at the sampling time and the average value is greater than the product of the preset threshold coefficient and the standard deviation, the denoised temperature data at the sampling time is judged as abnormal data and removed, thus obtaining the temperature time series data after removing abnormal data.

[0087] Based on the temperature time series data after removing outlier data, the data processing unit performs data completion processing on the missing data positions generated after the removal. Linear interpolation is used to complete the missing data positions to obtain complete temperature time series data.

[0088] The process of obtaining complete temperature time series data is as follows: The data processing unit traverses the temperature time series data after removing outliers to identify missing data locations caused by the removal of outliers; for each missing data location, the temperature data corresponding to the two adjacent valid sampling times that were not removed in the time series are selected; based on the time interval between the two valid sampling times and the corresponding temperature difference, linear interpolation is used to estimate the temperature data at the missing data location. The linear interpolation method is to linearly distribute the temperature data of the two valid sampling times according to the time axis; after interpolation and completion of all missing data locations, the completed temperature data is integrated with the original non-missing temperature data according to the sampling time order to obtain complete temperature time series data;

[0089] Based on complete temperature time series data, the data processing unit performs standardization processing, which removes the mean from the temperature data of each temperature sensor at each sampling time and scales it according to the standard deviation to obtain standardized temperature time series data.

[0090] In this embodiment, the extraction of state feature parameters specifically includes:

[0091] Based on standardized temperature time series data, a delayed embedding process is performed on each temperature sensor at each sampling time to construct the corresponding temperature evolution state vector;

[0092] The specific process of constructing the temperature evolution state vector is as follows: Based on standardized temperature time series data, each temperature sensor is subjected to delayed embedding processing at each sampling time. At any sampling time, the standardized temperature data corresponding to the sampling time is taken as the current state component, and the standardized temperature data of the historical sampling times are selected as the delayed state components in sequence. The current state component and each delayed state component are arranged and combined in chronological order to construct a temperature evolution state vector containing information of the current time and information of the historical time.

[0093] At each sampling moment, the temperature evolution state vectors corresponding to each temperature sensor are combined to form the system temperature evolution state vector.

[0094] The specific process of forming the system temperature evolution state vector is as follows: at each sampling moment, the data processing unit obtains the temperature evolution state vector corresponding to each temperature sensor, and takes the sampling moment as a unified time node, arranges and combines the temperature evolution state vectors of each temperature sensor in a preset order, and connects the temperature evolution state vectors of different spatial locations in sequence to form an overall vector, thereby constructing a system temperature evolution state vector that reflects the overall temperature change of the test area at that sampling moment.

[0095] According to the sampling time sequence, the system temperature evolution state vectors corresponding to each sampling time are connected sequentially to construct a continuously changing temperature evolution state trajectory.

[0096] Based on the temperature evolution state trajectory, the data processing unit compares the system temperature evolution state vectors corresponding to adjacent sampling times to obtain the state change between each sampling time.

[0097] The extraction process of state change is as follows: Based on the temperature evolution state trajectory, the data processing unit reads the system temperature evolution state vector corresponding to each sampling time in the order of sampling time. At any sampling time, the system temperature evolution state vector corresponding to the sampling time is calculated element by element with the system temperature evolution state vector corresponding to the previous sampling time. The elements at the same position in the two vectors are subtracted to obtain the difference at the corresponding position. The differences at all positions are arranged in the original order to form the state change corresponding to the sampling time.

[0098] Based on the temperature evolution trajectory and state change, the data processing unit extracts state feature parameters that characterize the dynamic evolution behavior of temperature.

[0099] The extraction process of state feature parameters is as follows: Based on the temperature evolution state trajectory and state change, the data processing unit processes the system temperature evolution state vector and corresponding state change at each sampling time according to the sampling time sequence. At each sampling time, the values ​​at each position in the system temperature evolution state vector are statistically calculated and combined with the state change at the corresponding position for joint analysis to obtain the average change amplitude, maximum change amplitude, and change trend information at that sampling time. The state change at consecutive sampling times is accumulated to obtain the overall change intensity during the temperature evolution process, and the direction of temperature change is determined based on the increase or decrease of state change between adjacent sampling times. The average change amplitude, maximum change amplitude, overall change intensity, and change direction information obtained at each sampling time are combined according to the sampling time sequence to form state feature parameters characterizing the dynamic evolution behavior of temperature.

[0100] In this embodiment, determining the state transition time specifically includes:

[0101] The improved ThermoONet model incorporates state feature parameters as input. This improved model includes a temperature evolution state embedding module, an operator mapping backbone module, a spatiotemporal coupling enhancement module, and a state determination output module. The improvements are as follows: Traditional ThermoONet models rely on function-to-function mappings and operator learning on the input sequence, depending on instantaneous feature mapping results during state determination. The improved ThermoONet model introduces a temperature evolution state embedding module, mapping state feature parameters to a unified state space to form a state sequence with temporal evolution continuity. The operator mapping backbone module uses a Branch-Trunk dual-pathway structure to couple the state function with the time basis function. The spatiotemporal coupling enhancement module performs structured adjustment of the time lag component and sensor channel component based on state changes. The state determination output module constructs a temperature evolution state phase space and divides the state convergence behavior based on attractor regions, resulting in a state determination result oriented towards the dynamic process of temperature evolution.

[0102] The improved ThermoONet model includes a temperature evolution state embedding module, an operator mapping backbone module, a spatiotemporal coupling enhancement module, and a state determination output module. These modules are connected in series according to the data flow order. The state feature parameters are input to the temperature evolution state embedding module as a multidimensional vector sequence arranged according to sampling time, and are mapped to a sequence of state points in a unified state space. This sequence of state points is then input to the operator mapping backbone module as a multidimensional state data sequence. Feature dimension mapping is performed in the Branch path, and the sampling time position is encoded as a basis function vector in the Trunk path. The two are then multiplied element-wise at the corresponding sampling times and the result is calculated. The output is a sequence of implicit representations of temperature evolution states arranged according to sampling time. This sequence is then input into the spatiotemporal coupling enhancement module, where it is divided into components based on delay positions in the time dimension and gating adjustments are made in conjunction with state changes. In the spatial dimension, it is divided into components based on sensor channels and channel reconstruction is performed. The output is an enhanced sequence of implicit representations of temperature evolution states with the same length as the original sequence. Finally, this enhanced sequence of implicit representations of temperature evolution states is input into the state determination output module in the form of multi-dimensional state points. Distribution mapping and attractor region identification are performed in a unified state space, and the discrete state category results corresponding to each sampling time are output.

[0103] In the training process of the improved ThermoONet model, the training data comes from temperature time series data obtained under different fire source excitation parameters and test conditions. After preprocessing and delayed embedding, state feature parameters are formed. The state corresponding to each sampling time is labeled according to the known start time of the fire extinguishing system. The samples before the fire extinguishing system is started are labeled as the fire source development state, and the samples after the start are labeled as the fire extinguishing intervention state. During the training process, the difference between the model output state judgment result and the corresponding label result is used as the optimization objective. A loss function based on classification error is constructed, and the loss function is jointly optimized by combining the constraint term on the continuity of state changes between adjacent sampling times. The training parameters include learning rate, batch size, and number of iterations. The model parameters are updated by iterating through the training data for multiple rounds. When the change of the loss function is less than the preset threshold in multiple consecutive iterations and the classification consistency of the state judgment result on the validation data remains stable, the model is considered to have reached the convergence condition, thus completing the model training.

[0104] In the temperature evolution state embedding module, the state feature parameters corresponding to each sampling time are embedded into a unified state space in the order of time evolution to form an embedded state feature sequence.

[0105] The specific process of forming the embedded state feature sequence is as follows: taking the sampling time order as the main order, the state feature parameters of each sampling time are mapped time by time, and the state feature parameters of each sampling time are mapped to a state point in a unified state space. During the mapping process, the relative positional relationship of the corresponding features of each temperature sensor in the state space is maintained, and the state points corresponding to each sampling time are connected in sequence according to the sampling time order to form the embedded state feature sequence that changes continuously in the temperature evolution process.

[0106] In the operator mapping backbone module, the embedded state feature sequence is input into the Branch-Trunk dual-path operator learning structure. The Branch-Trunk dual-path operator learning structure includes a Branch path and a Trunk path. The Branch path performs feature mapping on the embedded state feature sequence, and the Trunk path performs basis function mapping on the temporal position information corresponding to each sampling time. The mapping results of the Branch path and the Trunk path are coupled to obtain the implicit representation of the temperature evolution state corresponding to each sampling time.

[0107] The specific process for obtaining the implicit representation of the temperature evolution state at each sampling time is as follows: The embedded state feature sequence is input into the Branch-Trunk dual-path operator learning structure. The embedded state feature sequence is a multidimensional state data sequence arranged in the order of sampling time. The Branch path is a mapping structure that performs nonlinear feature transformation on the input embedded state feature sequence, extracts high-dimensional feature information from the state feature sequence, and outputs the corresponding feature representation. The Trunk path is a mapping structure that expands the basis function on the temporal position information corresponding to each sampling time. The temporal position information is the position identifier of each sampling time in the time series. By mapping the position identifier to a set of basis function values ​​with different variation rules, the corresponding time representation is formed. After completing the mapping between the Branch path and the Trunk path, the feature representation output by the Branch path and the time representation output by the Trunk path are multiplied element-wise at the corresponding sampling time, and the product result is summed to obtain the fusion representation corresponding to the sampling time. The fusion representation is used as the implicit representation of the temperature evolution state at the corresponding sampling time.

[0108] In the spatiotemporal coupling enhancement module, cross-delay gating enhancement and cross-sensor channel enhancement are applied to the implicit representation of temperature evolution state corresponding to each sampling time to obtain the enhanced implicit representation of temperature evolution state.

[0109] The process of obtaining the enhanced temperature evolution state latent representation is as follows: The temperature evolution state latent representation corresponding to each sampling time is processed. In cross-delay gated enhancement, the different time lag positions formed by delay embedding in the temperature evolution state latent representation are gated according to the state change at each sampling time. For time lag positions with large state changes, the corresponding components of the temperature evolution state latent representation are retained; for time lag positions with small state changes, the corresponding components of the temperature evolution state latent representation are suppressed, resulting in a time-enhanced representation reflecting the significant part of the temperature dynamic change. In cross-sensor channel enhancement, based on the consistency of the temperature evolution state latent representation changes between adjacent sampling times, the state components corresponding to different temperature sensors are reconstructed. State components with similar change trends are combined to obtain a spatial enhancement representation. After completing cross-delay gated enhancement and cross-sensor channel enhancement, the time-enhanced representation and the spatial enhancement representation are integrated to obtain the enhanced temperature evolution state latent representation.

[0110] In the state determination output module, a temperature evolution state phase space is constructed based on the enhanced temperature evolution state implicit representation. The attractor region of the temperature evolution state is identified in the temperature evolution state phase space. The enhanced temperature evolution state implicit representation is divided according to the attractor region to which it converges. The enhanced temperature evolution state implicit representation that converges to the attractor region corresponding to continuous temperature rise is determined as the fire source development state. The enhanced temperature evolution state implicit representation that converges to the attractor region corresponding to temperature drop or stabilization is determined as the fire extinguishing intervention state. The classification results of the fire source development state and the fire extinguishing intervention state corresponding to each sampling time are obtained.

[0111] The construction process of the temperature evolution state phase space is as follows: the enhanced temperature evolution state implicit representation at each sampling time is taken as a state point in a unified state space, and the enhanced temperature evolution state implicit representation is a multi-dimensional data representation composed of multiple state components; using the sampling time order as an index, the state points corresponding to each sampling time are sequentially mapped to the same coordinate system, so that each state point forms a spatial distribution in the coordinate system according to the numerical values ​​of its state components, thus constituting the temperature evolution state phase space.

[0112] An attractor region is a stable clustered region formed by multiple enhanced implicit representations of temperature evolution states in the phase space of temperature evolution states. State points located within the same attractor region exhibit similar trends of change between adjacent sampling times and gradually concentrate in that region during the evolution process of multiple consecutive sampling times.

[0113] The classification results at each sampling time are continuously compared according to the sampling time sequence to determine the sampling time when the classification result changes from the fire source development state to the fire extinguishing intervention state, and the sampling time is determined as the state transition time when the temperature evolution state changes.

[0114] The continuous comparison process is as follows: the classification results corresponding to each sampling time are read sequentially according to the sampling time order, and the classification results of each sampling time are compared with the classification results of the previous sampling time. During the comparison process, when the classification result of the sampling time changes from the fire source development state to the fire extinguishing intervention state, the sampling time is marked as the candidate state transition time. The consistency of the classification results of several consecutive sampling times after the candidate state transition time is checked. When the classification results of consecutive sampling times all remain in the fire extinguishing intervention state, the candidate state transition time is confirmed as the state transition time when the temperature evolution state changes.

[0115] In this embodiment, the formation of the test results specifically includes:

[0116] Based on the start time corresponding to the test start conditions, the start time is determined as the test start time, and the state transition time is determined as the fire extinguishing system response time;

[0117] The data processing unit calculates the time difference between the test start time and the fire extinguishing system response time, and subtracts the test start time from the fire extinguishing system response time to obtain the test value of the fire extinguishing system response time.

[0118] The data processing unit associates and stores the test value of the fire extinguishing system response time with the corresponding test start time, the fire extinguishing system response time, and the classification results corresponding to each sampling time to form the test results.

[0119] The specific process of forming the test results is as follows: After obtaining the test value of the fire extinguishing system response time, the data processing unit organizes the test value of the fire extinguishing system response time, the test start time, the fire extinguishing system response time, and the classification results corresponding to each sampling time in a unified manner. The above data are associated and identified according to the same test number, and the classification results of each sampling time are aligned with the corresponding timestamp according to the sampling time order. The test start time and the fire extinguishing system response time are embedded into the corresponding time series to form a structured data set containing time information and status information. The structured data set is encapsulated and stored to generate complete test results.

[0120] Example 1: During the testing of an automatic sprinkler fire suppression system in a large logistics warehouse, it was necessary to test the response time of the installed ceiling sprinkler system. The warehouse space was approximately 3000 square meters with a ceiling height of 12 meters, containing a large quantity of paper-packaged goods, exhibiting typical characteristics of concentrated flammable load and rapid fire development. Traditional testing methods primarily employed single-point temperature probes combined with manual timing. When the temperature at the probe point reached a set threshold, the time was recorded, and the sprinkler system's activation time was manually observed and compared for calculation. However, in actual testing, it was found that due to complex airflow within the warehouse, lag in localized temperature rise, and human error, the test results fluctuated significantly. Multiple tests under the same conditions showed obvious deviations, making it difficult to accurately reflect the fire suppression system's response performance.

[0121] In this embodiment, the fire extinguishing system response time test method based on temperature sensors of the present invention is used to test the warehouse fire extinguishing system. First, 16 temperature sensors are evenly distributed within the test area, covering the central area of ​​the fire source and the surrounding diffusion area. Real-time data is synchronously acquired through a data acquisition unit and a data processing unit. Before the test, all temperature sensors are uniformly calibrated, and a time synchronization mechanism ensures that the data from each sensor has a unified time reference. Simultaneously, based on the characteristics of the warehouse environment, the thermal power of the fire source excitation device is set to 12kW, the duration to 120 seconds, and the temperature data sampling frequency is set to 10Hz to ensure that the detailed process of temperature changes can be captured.

[0122] During the test, after the fire source excitation device was activated, controlled thermal excitation was applied to the test area. Temperature sensors collected temperature data in real time and formed temperature time series data. After denoising, anomaly removal, and standardization of the collected data by the data processing unit, a delayed embedding method was used to construct the temperature evolution trajectory and extract state feature parameters. Subsequently, the state feature parameters were input into an improved ThermoONet model, and the model was used to analyze the temperature evolution state to obtain the classification results of the fire source development state and the fire extinguishing intervention state. In the state determination process, by constructing a phase space of the temperature evolution state and identifying attractor regions, the convergence behavior of temperature changes was analyzed to determine the state transition time. Finally, the fire source excitation activation time was used as the test start time, and the difference between the start time and the state transition time was calculated to obtain the fire extinguishing system response time.

[0123] To verify the effectiveness of the method of this invention, a comparative test was conducted with the traditional single-point temperature threshold method. Under the same test conditions, multiple repeated tests were performed, and the response time and fluctuations were recorded.

[0124] Table 1 Comparative Analysis of Fire Extinguishing System Response Time Test Methods

[0125]

[0126] As shown in Table 1, the traditional single-point temperature threshold method, due to its reliance on local temperature changes, exhibits a high average response time with a large fluctuation range, a maximum deviation of 6.2 seconds, and a standard deviation of 3.8 seconds. This indicates poor stability of the test results and the occurrence of misjudgments in multiple tests. While the multi-point averaging method improves upon the limitations of single-point measurement to some extent, it still fails to model the temperature evolution process, resulting in significant fluctuations in the test results.

[0127] The method of this invention preprocesses temperature data systematically and models the temperature evolution process using delayed embedding and operator learning models. This makes the average response time closer to the actual system startup time, while reducing the maximum deviation to 1.3 seconds and the standard deviation to only 0.7 seconds, indicating that the test results have high consistency and stability. Furthermore, no misjudgments occurred during the entire test, demonstrating that this method can effectively identify the state transition between fire source development and fire extinguishing intervention, avoiding the misjudgment problems caused by temperature fluctuations in traditional methods.

[0128] This invention constructs a temperature evolution state phase space, transforming temperature change from a simple numerical change into an analytical problem of state distribution and convergence behavior. This allows the system to identify the overall trend of temperature change rather than local anomalies, thereby improving the accuracy of the judgment. Simultaneously, the fusion of multi-sensor data enables the test results to reflect the temperature field changes across the entire test area, rather than local information from a single location, significantly improving the representativeness of the test results.

[0129] In summary, this embodiment demonstrates that the method of the present invention can effectively solve the problems of large testing errors, poor stability, and high misjudgment rate in the prior art under complex environments, and achieves high-precision measurement of the response time of fire extinguishing systems, which has good prospects for engineering applications.

[0130] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for testing the response time of a fire extinguishing system based on a temperature sensor, characterized in that, Includes the following steps: Construct a test environment for the response time of a fire extinguishing system. The test environment for the response time of a fire extinguishing system includes the fire extinguishing system under test, the test area, the fire source excitation device, the temperature sensor, the data acquisition unit, and the data processing unit. Calibrate each temperature sensor and complete time synchronization; set the excitation parameters of the fire source excitation device, the temperature data sampling frequency, and the test start conditions. The fire source excitation device is activated and controlled thermal excitation is applied to the test area according to the excitation parameters. When the temperature reaches the preset temperature threshold, the fire extinguishing system under test is activated. The data acquisition unit synchronously collects the real-time temperature data of each temperature sensor according to the temperature data sampling frequency to form temperature time series data. The temperature time series data is preprocessed by the data processing unit to obtain standardized temperature time series data; Based on standardized temperature time series data, a delayed embedding method is used to construct the temperature evolution state trajectory and extract state feature parameters that characterize the dynamic evolution behavior of temperature. The state feature parameters are input into the improved ThermoONet model to classify and determine the temperature evolution state, obtain the classification results of the fire source development state and the fire extinguishing intervention state, and determine the state transition time when the temperature evolution state changes based on the classification results. The start time corresponding to the test start condition is taken as the test start time, and the state transition time is taken as the fire extinguishing system response time. The corresponding time difference is calculated to obtain the test value of the fire extinguishing system response time, thus forming the test result.

2. The method for testing the response time of a fire extinguishing system based on a temperature sensor according to claim 1, characterized in that, The fire extinguishing system under test is installed in a closed or semi-closed test area. The spatial range of the test area is determined according to the spray range and effective area of ​​the fire extinguishing system under test. Temperature sensors are arranged in the test area according to the preset spatial positions. The temperature sensors are electrically connected to the data acquisition unit, and the data acquisition unit establishes a data communication connection with the data processing unit. A fire source excitation device is set in the test area. The fire extinguishing system under test, the fire source excitation device, the temperature sensors, the data acquisition unit, and the data processing unit are configured as a whole to form a test environment for the response time of the fire extinguishing system.

3. The method for testing the response time of a fire extinguishing system based on a temperature sensor according to claim 1, characterized in that, The specific settings for the excitation parameters, temperature data sampling frequency, and test start conditions of the fire source excitation device include: Each temperature sensor is calibrated using a standard temperature source to obtain the calibration coefficient and offset of each temperature sensor. The output value of the temperature sensor is then corrected based on the calibration coefficient and offset to obtain the calibrated temperature data of each temperature sensor at each time point. Based on the calibrated temperature data, time synchronization processing is performed on each temperature sensor, a unified time reference is set, and the time offset of each temperature sensor is calculated. The acquisition time of each temperature sensor is then corrected to obtain time-aligned temperature data. Based on time-aligned temperature data, a temperature data sampling frequency is set, and a sampling time interval is determined according to the temperature data sampling frequency to unify the data acquisition time interval of each temperature sensor. Under a unified data acquisition time interval, the excitation parameters of the fire source excitation device are set; Based on the excitation parameters and temperature data sampling frequency, the test start conditions are set.

4. The method for testing the response time of a fire extinguishing system based on a temperature sensor according to claim 1, characterized in that, The formation of the temperature time series data specifically includes: The fire source excitation device is activated at the start time corresponding to the test start conditions, and controlled thermal excitation is continuously applied to the test area according to the excitation parameters; During the continuous operation of controlled thermal excitation, each temperature sensor continuously monitors the temperature field within the test area and outputs corresponding real-time temperature data. The data acquisition unit synchronously samples the real-time temperature data of each temperature sensor based on the temperature data sampling frequency to obtain discrete time-series temperature data. The real-time temperature data from each temperature sensor at each sampling time are combined and arranged in chronological order of sampling time to form temperature time series data.

5. The method for testing the response time of a fire extinguishing system based on a temperature sensor according to claim 1, characterized in that, The standardization of temperature time series data is obtained specifically through: The data processing unit performs noise reduction processing on the temperature time series data to obtain the corresponding denoised temperature data. Based on the denoised temperature data, the data processing unit performs outlier removal processing to obtain temperature time series data after removing outlier data. Based on the temperature time series data after removing outlier data, the data processing unit performs data completion processing on the missing data positions generated after the removal. Linear interpolation is used to complete the missing data positions to obtain complete temperature time series data. Based on complete temperature time series data, the data processing unit performs standardization processing to obtain standardized temperature time series data.

6. The method for testing the response time of a fire extinguishing system based on a temperature sensor according to claim 1, characterized in that, The extraction of the state feature parameters specifically includes: Based on standardized temperature time series data, a delayed embedding process is performed on each temperature sensor at each sampling time to construct the corresponding temperature evolution state vector; At each sampling moment, the temperature evolution state vectors corresponding to each temperature sensor are combined to form the system temperature evolution state vector. According to the sampling time sequence, the system temperature evolution state vectors corresponding to each sampling time are connected sequentially to construct a continuously changing temperature evolution state trajectory. Based on the temperature evolution state trajectory, the data processing unit compares the system temperature evolution state vectors corresponding to adjacent sampling times to obtain the state change between each sampling time. Based on the temperature evolution trajectory and state change, the data processing unit extracts state characteristic parameters that characterize the dynamic evolution behavior of temperature.

7. The method for testing the response time of a fire extinguishing system based on a temperature sensor according to claim 1, characterized in that, The determination of the state transition time specifically includes: The improved ThermoONet model incorporates state feature parameters as input. This improved model includes a temperature evolution state embedding module, an operator mapping backbone module, a spatiotemporal coupling enhancement module, and a state determination output module. The improvements are as follows: Traditional ThermoONet models rely on function-to-function mappings and operator learning on the input sequence, depending on instantaneous feature mapping results during state determination. The improved ThermoONet model introduces a temperature evolution state embedding module, mapping state feature parameters to a unified state space to form a state sequence with temporal evolution continuity. The operator mapping backbone module uses a Branch-Trunk dual-pathway structure to couple the state function with the time basis function. The spatiotemporal coupling enhancement module performs structured adjustment of the time lag component and sensor channel component based on state changes. The state determination output module constructs a temperature evolution state phase space and divides the state convergence behavior based on attractor regions, resulting in a state determination result oriented towards the dynamic process of temperature evolution. In the temperature evolution state embedding module, the state feature parameters corresponding to each sampling time are embedded into a unified state space in the order of time evolution to form an embedded state feature sequence. In the operator mapping backbone module, the embedded state feature sequence is input into the Branch-Trunk dual-path operator learning structure. The Branch-Trunk dual-path operator learning structure includes a Branch path and a Trunk path. The Branch path performs feature mapping on the embedded state feature sequence, and the Trunk path performs basis function mapping on the temporal position information corresponding to each sampling time. The mapping results of the Branch path and the Trunk path are coupled to obtain the implicit representation of the temperature evolution state corresponding to each sampling time. In the spatiotemporal coupling enhancement module, cross-delay gating enhancement and cross-sensor channel enhancement are applied to the implicit representation of temperature evolution state corresponding to each sampling time to obtain the enhanced implicit representation of temperature evolution state. In the state determination output module, a temperature evolution state phase space is constructed based on the enhanced temperature evolution state implicit representation. The attractor region of the temperature evolution state is identified in the temperature evolution state phase space. The enhanced temperature evolution state implicit representation is divided according to the attractor region to which it converges. The enhanced temperature evolution state implicit representation that converges to the attractor region corresponding to continuous temperature rise is determined as the fire source development state. The enhanced temperature evolution state implicit representation that converges to the attractor region corresponding to temperature drop or stabilization is determined as the fire extinguishing intervention state. The classification results of the fire source development state and the fire extinguishing intervention state corresponding to each sampling time are obtained. The classification results at each sampling time are continuously compared according to the sampling time sequence to determine the sampling time when the classification result changes from the fire source development state to the fire extinguishing intervention state, and the sampling time is determined as the state transition time when the temperature evolution state changes.

8. The method for testing the response time of a fire extinguishing system based on a temperature sensor according to claim 1, characterized in that, The formation of the test results specifically includes: Based on the start time corresponding to the test start conditions, the start time is determined as the test start time, and the state transition time is determined as the fire extinguishing system response time; The data processing unit calculates the time difference between the test start time and the fire extinguishing system response time to obtain the test value of the fire extinguishing system response time. The data processing unit associates and stores the test value of the fire extinguishing system response time with the corresponding test start time, the fire extinguishing system response time, and the classification results corresponding to each sampling time to form the test results.

Images