Intelligent fire-fighting spraying linkage control method and system based on smoke diffusion trend

Abstract

The invention discloses an intelligent fire-fighting spraying linkage control method and system based on a smoke diffusion trend, and relates to the technical field of fire-fighting control. The intelligent fire-fighting spraying linkage control method and system based on the smoke diffusion trend comprises the following steps: S1, acquiring smoke sensing data, and preprocessing the smoke sensing data to obtain preprocessed smoke sensing data; s2, extracting a disturbance index of each sensing node, evaluating the credibility of smoke sensing data in the current sampling period, and distinguishing normal nodes from abnormal nodes; s3, the diffusion credibility of each sensing node is evaluated, and the sensing nodes in the rising trend are marked based on the evaluation result; s4, constructing a spraying control area, analyzing the spraying response level of the control area, generating a corresponding spraying strategy, and issuing an execution instruction; the spraying period is continuously monitored, and spraying is restarted when abnormity is monitored in the window keeping period. The problem of diffusion trend misjudgment caused by reading jitter superposition errors in a low-concentration stage in the multi-node smoke sensing network is solved.

Description

Technical Field

[0001] The present invention relates to the field of fire control technology, and in particular to an intelligent fire sprinkler linkage control method and system based on smoke diffusion trends. Background Art

[0002] With the increasing density of urban buildings and the continued expansion of complex industrial scenarios, fire early warning and rapid initial response have become key development directions for fire protection systems driven by intelligent monitoring. Traditional fire alarm systems often rely on point-type smoke sensors, which trigger an alarm signal when smoke concentration exceeds a set threshold. Some systems can also activate sprinklers for initial fire suppression.

[0003] However, such traditional systems have significant shortcomings in practical applications: on the one hand, the fixed threshold trigger mechanism cannot effectively respond to the dynamic changes in the weak smoke stage at the beginning of a fire, and it is difficult to meet the accurate identification of abnormal smoke behavior based on the data evolution trend of intelligent monitoring. It is easy to cause false alarms or missed alarms due to factors such as sensor self-heating drift, environmental disturbances or quantization errors; on the other hand, the sprinkler linkage strategy usually lacks a comprehensive analysis of the smoke diffusion trend, the spatial distribution of sensor nodes and environmental disturbance factors, making it difficult to achieve accurate perception of the development of the fire and intelligent judgment of the response level.

[0004] At the same time, the current system generally ignores the problems of data credibility screening and automatic identification of abnormal sensor nodes in multi-node collaboration, resulting in some non-real diffusion trends being misjudged as fire development, triggering unnecessary spraying actions, causing water resource waste and reduced system stability.

[0005] Therefore, in response to the above problems, there is an urgent need for an intelligent fire sprinkler linkage control method and system based on the smoke diffusion trend. Summary of the Invention

[0006] Technical problems solved

[0007] In response to the shortcomings of the existing technology, the present invention provides an intelligent fire sprinkler linkage control method and system based on smoke diffusion trend, which solves the problem of misjudgment of diffusion trend caused by jitter superposition error of readings in the low concentration stage in a multi-node smoke sensor network.

[0008] Technical Solution

[0009] To achieve the above objectives, the present invention is implemented through the following technical solutions: an intelligent fire sprinkler linkage control method and system based on smoke diffusion trend, including: S1, obtaining smoke perception data, and performing temperature and pressure correction, trend filtering and normalization processing on the smoke perception data to obtain pre-processed smoke perception data; S2, extracting the disturbance index of each sensor node based on the pre-processed smoke perception data, evaluating the credibility of the smoke perception data in the current sampling period, distinguishing normal nodes from abnormal nodes, and dynamically maintaining the node credibility status; S3, retrieving the smoke perception data of all normal nodes, analyzing the smoke concentration fluctuation amplitude and change rate, evaluating the diffusion credibility of each sensor node, and marking the sensor nodes in an upward trend based on the evaluation results; S4, constructing a spraying control area based on the spatial position of the upward trend sensor node, analyzing the spraying response level of the control area, generating a corresponding spraying strategy and issuing an execution instruction; continuous monitoring during spraying, stopping spraying if a continuous low level occurs, and restarting spraying if an abnormality is detected within the window period.

[0010] Furthermore, smoke sensing data is obtained, and temperature and pressure correction, trend filtering and normalization processing are performed on the smoke sensing data. The specific steps to obtain the pre-processed smoke sensing data are as follows: the concentration evolution characteristics, temperature and humidity fluctuation state, local airflow disturbance form, sensor self-heating change behavior, shading rate jump characteristics, atmospheric pressure stress response, voltage signal output offset, current power supply stability and particulate matter transmission environment in the early stage of fire smoke diffusion evolution process are collected synchronously in multiple dimensions throughout the whole process to obtain smoke sensing data. The smoke sensing data includes smoke concentration, sensor temperature, ambient temperature, wind speed, relative humidity, voltage signal The smoke sensing data is corrected for thermal drift using the temperature-pressure-concentration correlation correction algorithm to compensate for the measurement deviation caused by sensor self-heating and environmental fluctuations. The wind speed and shading rate sequences in the smoke sensing data are subjected to boundary denoising using the polynomial residual smoothing algorithm to suppress abnormal jumps caused by short-term disturbances. The smoke sensing data is subjected to anomaly elimination and local backfilling using the spatial trend consistency discrimination algorithm to avoid the influence of single-point errors on the diffusion trend judgment. The smoke sensing data is scaled uniformly using the zero-mean standard deviation normalization algorithm to achieve normalized processing of the smoke sensing data.

[0011] Furthermore, based on the pre-processed smoke sensing data, the disturbance index of each sensor node is extracted, and the specific steps for evaluating the credibility of the smoke sensing data in the current sampling period are as follows: set a fixed-length sampling period to collect smoke sensing data, and extract the disturbance index of each sensor node when updating the pre-processed smoke sensing data in each sampling period: calculate the standard deviation of the voltage signal in the current sampling period to obtain the electrical fluctuation term; calculate the absolute value of the difference between the sensor temperature and the ambient temperature to obtain the temperature difference disturbance term; calculate the absolute value of the difference between the smoke concentration and the smoke concentration in the previous sampling period to obtain the concentration mutation term; calculate the absolute value of the difference between the smoke shading rate and the average value of the smoke shading rate, and divide it by The average value of the smoke shading rate is used to obtain the shading anomaly item; if any two of the smoke concentration, sensor temperature, and voltage signals of the sensor node are missing, the key field missing item is marked as 1, otherwise it is 0; based on the various disturbance indicators within the sampling period, the credibility of the smoke perception data of the current sampling period is evaluated: the sum of the electrical fluctuation item, the temperature difference disturbance item, and the concentration mutation item is calculated, and divided by the atmospheric pressure plus 1 to obtain the basic disturbance factor; the shading anomaly item is calculated plus 1 to obtain the shading fluctuation amplification factor; the product of the key field missing item and the missing penalty coefficient is calculated to obtain the missing penalty item; the basic disturbance factor is multiplied by the shading fluctuation amplification factor, and the missing penalty item is added to obtain the perception credibility assessment value.

[0012] Furthermore, the specific steps for distinguishing normal nodes from abnormal nodes and dynamically maintaining the node trust status are as follows: real-time comparison of the perception trust evaluation value and the perception trust threshold; if the perception trust evaluation value is less than or equal to the perception trust threshold, the smoke perception data of the current sampling period is retained, and the corresponding sensor node is marked as a normal node; if the perception trust evaluation value is greater than the perception trust threshold, the smoke perception data of the current sampling period is discarded, and the corresponding sensor node is marked as an abnormal node, and participation in subsequent diffusion trend judgment and spray response evaluation is suspended; if the perception trust evaluation value of the abnormal node is less than or equal to the perception trust threshold in three consecutive sampling periods, its abnormal mark is removed and the data processing authority of the sensor node is restored.

[0013] Furthermore, the smoke sensing data of all normal nodes are retrieved, the smoke concentration fluctuation amplitude and change rate are analyzed, and the specific steps for evaluating the diffusion credibility of each sensor node are as follows: at the end of each sampling period, the smoke sensing data of all normal nodes are retrieved, and the difference between the maximum and minimum smoke concentrations of each sensor node within the sampling period is calculated to obtain the smoke concentration fluctuation amplitude; and the concentration change rate of each sensor node per unit time is calculated to obtain the smoke concentration change rate; the diffusion credibility of each sensor node is evaluated: the smoke concentration change rate is calculated and multiplied by the wind speed to obtain the diffusion driving term; the absolute value of the difference between the sensor temperature and the ambient temperature is calculated to obtain the thermal disturbance term; the smoke concentration fluctuation amplitude is calculated and divided by the relative humidity value to obtain the fluctuation disturbance term; the inverse of the atmospheric pressure is calculated to obtain the pressure disturbance term; the thermal disturbance term, the fluctuation disturbance term and the pressure disturbance term are added together to obtain the environmental disturbance term; the diffusion driving term is divided by the environmental disturbance term to obtain the diffusion credibility evaluation value.

[0014] Furthermore, the specific steps for marking sensor nodes in an upward trend based on the evaluation results are as follows: real-time comparison of the diffusion credible evaluation value and the diffusion credible threshold, elimination of the smoke concentration data corresponding to the sensor nodes whose diffusion credible evaluation value is less than the diffusion credible threshold, and differential analysis of the smoke concentration change rate sequence of the retained nodes: if the direction of the concentration change rate alternates in three consecutive sampling periods, and the smoke concentration fluctuation amplitude in each sampling period is less than the amplitude threshold, then the smoke concentration data corresponding to the sensor node is determined to be a pseudo-diffusion trend and eliminated; for the remaining sensor nodes after pseudo-diffusion elimination, determine whether the smoke concentration change rate in the current sampling period is greater than the change rate threshold, and whether the diffusion credible evaluation value is greater than the diffusion credible threshold. If both conditions are met at the same time, the sensor node is marked as an upward trend; a delayed release window is set for the sensor node marked as a credible upward trend. If the smoke concentration change rate is negative in three consecutive sampling periods and the diffusion credible evaluation value is lower than the diffusion credible threshold, the upward trend mark is removed.

[0015] Furthermore, the spraying control area is constructed based on the spatial position of the upward trend sensor node, and the specific steps for analyzing the spraying response level of the control area are as follows: according to the physical space coordinate information of the sensor node, the monitoring area is divided into several fixed unit areas by using the grid division method, and all the sensor nodes in the upward trend state are mapped to the corresponding grid units to construct the spraying control area; the average value of the smoke perception data and the average value of the diffusion credible assessment value in each spraying control area are calculated; based on the average smoke perception data in each control area, the spraying response level of the control area is analyzed: the ratio of the average smoke concentration to the average relative humidity is calculated to obtain the humidity modulation concentration term; the humidity modulation concentration term is added to the average smoke shading rate to obtain the smoke intensity term; the product of the average smoke concentration change rate and the average wind speed is calculated to obtain the diffusion coupling term, and 1 is added to the diffusion coupling term to obtain the diffusion amplification factor; the smoke intensity term, the diffusion amplification factor and the average diffusion credible assessment value are multiplied to obtain the spraying response assessment value.

[0016] Furthermore, the specific steps of generating the corresponding spraying strategy and issuing the execution instruction are as follows: compare the spraying response evaluation value with the spraying response threshold, and generate the spraying strategy according to the grading rule: if the spraying response evaluation value is less than or equal to the first-level spraying response threshold, a low-level response is triggered, and the single-point sprinkler and low-pressure short-time spraying are started; if the spraying response evaluation value is greater than the second-level spraying response threshold and less than the first-level spraying response threshold, a medium-level response is triggered, and the regional sprinkler and medium-pressure medium-time spraying are started; if the spraying response evaluation value is greater than or equal to the second-level spraying response threshold, a high-level response is triggered, and the full-coverage sprinkler group and high-pressure continuous spraying are started; generate a spraying control instruction according to the spraying strategy, the instruction content includes the target sprinkler number, spraying pressure, spraying duration and start time, and send it to the execution end for spraying operation.

[0017] Furthermore, continuous monitoring is performed during the spraying period, and spraying is stopped when continuous low levels appear. The specific steps for restarting the spraying are as follows: During the spraying execution, the smoke perception data in the corresponding control area is continuously monitored, and the spraying response evaluation value is dynamically updated; if the spraying response evaluation values ​​of all rising trend sensor nodes in the area are continuously at a low level for three consecutive sampling periods, a nozzle closing command is issued to terminate the current spraying behavior; after the spraying is completed, the control area enters the spraying maintenance window period, retaining the node high-frequency sampling task and trend priority analysis algorithm; if the rising trend sensor node is detected again during the window period, and its corresponding spraying response evaluation value is greater than the start threshold, the restart process of the spraying linkage control strategy is immediately triggered, the spraying intensity level of the control area is reallocated, and a new spraying control command is generated.

[0018] The second aspect of the present invention provides an intelligent fire sprinkler linkage control system based on smoke diffusion trend, including: a smoke perception data acquisition and preprocessing module, an abnormal state identification and node control module, a diffusion trend evaluation and marking judgment module and a spray response classification and linkage control module, wherein: the smoke perception data acquisition and preprocessing module is used to obtain smoke perception data, and perform temperature and pressure correction, trend filtering and normalization processing on the smoke perception data to obtain preprocessed smoke perception data; the abnormal state identification and node control module is used to extract the disturbance index of each sensor node based on the preprocessed smoke perception data, and evaluate the credibility of the smoke perception data in the current sampling period. The system uses a diffusion trend assessment and marking discrimination module to retrieve smoke perception data from all normal nodes, analyze the fluctuation amplitude and change rate of smoke concentration, evaluate the diffusion credibility of each sensor node, and mark the sensor nodes with an upward trend based on the assessment results. The spraying response classification and linkage control module is used to construct a spraying control area based on the spatial position of the upward trend sensor node, analyze the spraying response level of the control area, generate the corresponding spraying strategy and issue execution instructions. The system continuously monitors during the spraying period and stops spraying if a continuous low level occurs. If an abnormality is detected within the window period, the spraying is restarted.

[0019] Beneficial effects

[0020] The present invention has the following beneficial effects:

[0021] (1) The intelligent fire sprinkler linkage control method and system based on smoke diffusion trend constructs a perception trust evaluation value in each sampling period and compares the perception trust threshold in real time, automatically marks and processes the sensor nodes with abnormal data, ensures the use of trusted data sources to participate in subsequent diffusion trend judgment and response evaluation, and thus improves the overall reliability of the data of each node in the multi-node smoke perception network.

[0022] (2) The intelligent fire sprinkler linkage control method and system based on smoke diffusion trend constructs the diffusion driving term by calculating the product of the smoke concentration change rate and the wind speed, superimposes the thermal disturbance term, the fluctuation disturbance term and the air pressure disturbance term to form the environmental disturbance term, and obtains the diffusion credible assessment value based on this. Then, combined with the differential sequence analysis of the concentration change rate, the node data with pseudo-diffusion characteristics is eliminated, and the real smoke diffusion trend is accurately identified and dynamically updated, thereby enhancing the accuracy of response judgment in the weak diffusion stage.

[0023] (3) The intelligent fire sprinkler linkage control method and system based on smoke diffusion trend constructs a spray response evaluation value. According to the hierarchical comparison results of the evaluation value and the response threshold, it intelligently generates three types of spray control strategies: low-level single-point low-pressure spraying, medium-level regional medium-pressure spraying, and high-level full-coverage high-pressure spraying, thereby improving the spraying efficiency and the rationality of resource utilization.

[0024] (4) The intelligent fire sprinkler linkage control method and system based on smoke diffusion trend continuously updates the spray response evaluation value of each rising trend sensor node in the spray control area during the spraying execution, identifies the continuous low-level response state, triggers the automatic stop spraying instruction, and sets the spray holding window for high-frequency sampling monitoring. If a valid rising trend is re-detected in the window and the spray response evaluation value meets the start-up conditions, the spraying process is immediately restarted to achieve adaptive linkage and closed-loop control of fire changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a flow chart of the intelligent fire sprinkler linkage control method based on smoke diffusion trend;

[0026] Figure 2 This is a flow chart of the intelligent fire sprinkler linkage control system based on smoke diffusion trends;

[0027] Figure 3 It is the distribution diagram of the perception trust evaluation status within the sampling period;

[0028] Figure 4 Distribution map of spray response levels for each control area. DETAILED DESCRIPTION

[0029] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0030] See also Figures 1-4, an embodiment of the present invention provides a technical solution: an intelligent fire sprinkler linkage control method and system based on smoke diffusion trend, including: S1, acquiring smoke perception data, and performing temperature and pressure correction, trend filtering and normalization processing on the smoke perception data to obtain pre-processed smoke perception data; S2, extracting the disturbance index of each sensor node based on the pre-processed smoke perception data, evaluating the credibility of the smoke perception data in the current sampling period, distinguishing normal nodes from abnormal nodes, and dynamically maintaining the node credibility status; S3, retrieving the smoke perception data of all normal nodes, analyzing the smoke concentration fluctuation amplitude and change rate, evaluating the diffusion credibility of each sensor node, and marking the sensor nodes in an upward trend based on the evaluation results; S4, constructing a spraying control area based on the spatial position of the upward trend sensor node, analyzing the spraying response level of the control area, generating a corresponding spraying strategy and issuing an execution instruction; continuous monitoring during spraying, stopping spraying if a continuous low level occurs, and restarting spraying if an abnormality is detected within the window period.

[0031] Specifically, smoke sensing data is acquired, and temperature and pressure correction, trend filtering and normalization processing are performed on the smoke sensing data. The specific steps to obtain the pre-processed smoke sensing data are as follows: the concentration evolution characteristics, temperature and humidity fluctuation state, local airflow disturbance form, sensor self-heating change behavior, shading rate jump characteristics, atmospheric pressure stress response, voltage signal output offset, current power supply stability and particle transmission environment in the early stage of fire smoke diffusion evolution are collected synchronously throughout the whole process to acquire smoke sensing data. The smoke sensing data includes smoke concentration, sensor temperature, ambient temperature, wind speed, relative humidity, voltage signal, atmospheric pressure, smoke shading rate and sensor node spatial coordinates; among them, smoke concentration is used to reflect the cumulative intensity of smoke in the local area, sensor temperature and ambient temperature are used to identify self-heating offset, wind speed is used to characterize the influence of airflow disturbance on particle distribution, relative humidity is used to correct particle hygroscopic characteristics, and voltage signal is used to identify sensor working State stability, atmospheric pressure is used to measure the interference of altitude changes on measurement results, smoke shading rate is used to indirectly determine the change in particle density, and the spatial coordinates of the sensor nodes are used for regional identification of subsequent diffusion trends; the smoke sensing data is corrected for thermal drift through the temperature-pressure-concentration correlation correction algorithm to compensate for the measurement deviation caused by sensor self-heating and environmental fluctuations, and improve the response accuracy in the low concentration range; the wind speed and shading rate sequences in the smoke sensing data are subjected to boundary denoising through the polynomial residual smoothing algorithm to suppress abnormal jumps caused by short-term disturbances and enhance the stability of trend identification; the smoke sensing data is subjected to anomaly elimination and local backfilling through the spatial trend consistency discrimination algorithm to avoid the influence of single-point errors on diffusion trend judgment and improve the reliability of multi-node fusion analysis; the smoke sensing data is scaled uniformly through the zero-mean standard deviation normalization algorithm to achieve balanced processing of smoke sensing data under multiple index inputs and provide a standardized input basis for subsequent model calculations.

[0032] In this implementation plan, an intelligent monitoring mechanism for early fire identification is constructed by performing multi-dimensional synchronous acquisition of smoke perception data throughout the entire process, combined with temperature-pressure-concentration correlation correction, residual smoothing, spatial consistency judgment, and zero-mean standard deviation normalization. This significantly improves the measurement accuracy and output stability of key perception variables such as smoke concentration, wind speed, shading rate, voltage signal, sensor temperature, ambient temperature, atmospheric pressure, relative humidity, and spatial coordinates of sensor nodes, enhances the intelligent recognition capability of subtle anomalies in the early stages of low-concentration fires, and ensures that the smoke perception data relied upon for diffusion trend judgment and spraying control processes have high consistency, strong continuity, and excellent credibility.

[0033] Specifically, based on the preprocessed smoke sensing data, the disturbance index of each sensor node is extracted, and the specific steps for evaluating the credibility of the smoke sensing data in the current sampling period are as follows: a sampling period of fixed length is set to collect smoke sensing data, and when the preprocessed smoke sensing data is updated in each sampling period, the disturbance index corresponding to the smoke concentration evolution, voltage signal output, current power supply form, sensor temperature response, ambient temperature background, shading rate change characteristics and spatial coordinate position of each sensor node is extracted in turn; the standard deviation of the voltage signal in the current sampling period is calculated to obtain the electrical fluctuation term, which reflects the short-term fluctuation degree of the electrical energy supply path; the absolute value of the difference between the sensor temperature and the ambient temperature is calculated to obtain the temperature difference disturbance term, which measures the intensity of the impact of local thermal disturbance on the sensing performance; the absolute value of the difference between the smoke concentration and the smoke concentration in the previous sampling period is calculated to obtain the concentration mutation term, which is used to identify the mutation behavior of the smoke particle aggregation speed in a short period; the absolute value of the difference between the smoke shading rate and the average value of the smoke shading rate is calculated, and divided by the average value of the smoke shading rate to obtain the shading anomaly term, which characterizes the relative variability of the optical obstruction degree; If any two of the sensor node's smoke concentration, sensor temperature, or voltage signals are missing, the missing key field item is marked as 1; otherwise, it is marked as 0, quantifying the data integrity of the key variables within the current sampling period. The credibility of the smoke perception data within the current sampling period is assessed based on various disturbance indicators within the sampling period. The basic disturbance factor is calculated by summing the electrical fluctuation term, the temperature difference disturbance term, and the concentration mutation term, and dividing it by the atmospheric pressure plus 1. This is used to normalize the degree of disturbance caused by pressure differences. The shading anomaly term is added by 1 to obtain the shading fluctuation amplification factor, which acts as an optical interference enhancement term to participate in disturbance amplification. The missing key field item is multiplied by the missing penalty coefficient to obtain the missing penalty term, which is used to construct a constraint factor for incomplete data. The missing penalty coefficient is obtained by using the missing key field ratio of each sensor node within the historical sampling period and an empirical weighted adjustment algorithm. The missing penalty coefficient ranges from 0.05 to 0.30. The basic disturbance factor is multiplied by the shading fluctuation amplification factor and added to the missing penalty term to obtain the perception credibility assessment value, which is used for subsequent trusted node screening and dynamic status management.

[0034] Among them, the specific calculation formula of the perception credibility evaluation value is:

[0035]

[0036] Where F represents the perceived credibility evaluation value, σ V represents the electrical fluctuation term, ΔT represents the temperature difference disturbance term, ΔC represents the concentration mutation term, P represents the atmospheric pressure, ΔS represents the shading anomaly term, A represents the key field missing term, and λ represents the missing penalty coefficient.

[0037] In this embodiment, Table 1 is a perception credibility evaluation value data table, which lists the disturbance index of each sensor node in 5 sampling periods and its corresponding perception credibility evaluation value, which is used to comprehensively evaluate the credibility of the perception data in the current period. Among them: in sampling period T1, the electrical fluctuation term is 0.18, the temperature difference disturbance term is 0.35, the concentration mutation term is 0.22, the atmospheric pressure is 1.0, the shading abnormality term is 0.10, the key field missing term is 0, the missing penalty coefficient is 0.15, and the calculated perception credibility evaluation value is 0.41; in T2, the electrical fluctuation term is 0.16, the temperature difference disturbance term is 0.30, the concentration mutation term is 0.16, the atmospheric pressure is 1.1, the shading abnormality term is 0.05, the key field missing term is 0, the missing penalty coefficient is 0.15, and the perception credibility evaluation value is 0.31; in T3, the electrical fluctuation term is 0.17, the temperature difference disturbance term is 0.28, the concentration mutation term is 0.19, the atmospheric pressure is 1.0, the shading abnormality term is 0.10, the key field missing term is 0, the missing penalty coefficient is 0.15, and the perception credibility evaluation value is 0.31. The force is 1.05, the shading anomaly item is 0.08, the key field missing item is 0, the missing penalty coefficient is 0.15, and the perception credibility evaluation value is 0.34; in T4, the electrical fluctuation item is 0.24, the temperature difference disturbance item is 0.40, the concentration mutation item is 0.26, the atmospheric pressure is 1.0, the shading anomaly item is 0.15, the key field missing item is 0, the missing penalty coefficient is 0.15, and the perception credibility evaluation value is 0.67; in T5, the electrical fluctuation item is 0.19, the temperature difference disturbance item is 0.32, the concentration mutation item is 0.21, the atmospheric pressure is 1.0, the shading anomaly item is 0.06, the key field missing item is 0, the missing penalty coefficient is 0.15, and the perception credibility evaluation value is 0.38.

[0038] Table 1 Perception credibility evaluation value data table

[0039] Sampling period <![CDATA[σ V ]]> ΔT ΔC P ΔS A λ F T1 0.18 0.35 0.22 1.0 0.10 0 0.15 0.41 T2 0.16 0.30 0.16 1.1 0.05 0 0.15 0.31 T3 0.17 0.28 0.19 1.05 0.08 0 0.15 0.34 T4 0.24 0.40 0.26 1.0 0.15 1 0.15 0.67 T5 0.19 0.32 0.21 1.0 0.06 0 0.15 0.38

[0040] like Figure 3 The figure below shows the distribution of sensor node trustworthiness assessment status within a sampling period, illustrating the distribution of sensor node trustworthiness assessment values ​​and node status over five sampling periods. The horizontal axis represents the sampling period, and the vertical axis represents the trustworthiness assessment value. The blue dashed line represents the trustworthiness threshold. Green bars represent normal nodes, and red bars represent abnormal nodes. The assessment value during sampling period T4 was 0.67, exceeding the trustworthiness threshold, and the node was classified as abnormal. All other periods show normal nodes. Figure 3 It intuitively reflects the trustworthy status of sensor nodes in each sampling period, which is helpful for the implementation of anomaly detection and node screening strategies.

[0041] In this implementation plan, by synchronously extracting electrical fluctuation items, temperature difference disturbance items, concentration mutation items, shading anomaly items and key field missing items, an intelligent monitoring mechanism for the stability of sensor nodes is constructed to realize disturbance monitoring of five key perception variables: smoke concentration, voltage signal, sensor temperature, ambient temperature, and smoke shading rate. In addition, a perception credibility assessment value is constructed in combination with atmospheric pressure and missing penalty coefficient to effectively identify sensor nodes with drastic data fluctuations, high field missing rates or perception instability behaviors in the low-concentration and high-disturbance stage, thereby improving the overall reliability of smoke perception data and the precision control capability of subsequent diffusion trend judgment.

[0042] Specifically, the steps for distinguishing normal nodes from abnormal nodes and dynamically maintaining the node trust status are as follows: The perception trust evaluation value is compared with the perception trust threshold in real time. If the perception trust evaluation value is less than or equal to the perception trust threshold, the smoke perception data for the current sampling period, including smoke concentration, sensor temperature, ambient temperature, wind speed, relative humidity, voltage signal, atmospheric pressure, smoke shading rate, and the spatial coordinates of the sensor node, are retained. The corresponding sensor node is marked as a normal node and allowed to continue participating in the subsequent diffusion trust evaluation value calculation and spray response evaluation value generation process. If the perception trust evaluation value is greater than the perception trust threshold, the smoke perception data for the current sampling period, along with its original records of smoke concentration, voltage signal, and sensor temperature, are immediately discarded. The corresponding sensor node is marked as an abnormal node and its participation in the diffusion trend determination process and spray control area construction process is suspended. If the perception trust evaluation value of an abnormal node is less than or equal to the perception trust threshold for three consecutive sampling periods, and all fields of its smoke perception data are complete and the disturbance value is stable, its abnormal mark is automatically removed, the normal data processing authority of the sensor node is restored, and it is re-integrated into the trend analysis and control strategy generation process.

[0043] In this implementation plan, through dynamic comparison based on perception credibility assessment value and perception credibility threshold, sensor nodes with insufficient perception credibility are accurately identified and isolated, ensuring that when there is fluctuation interference or field missing in key perception indicators such as smoke concentration, sensor temperature, ambient temperature, wind speed, relative humidity, voltage signal, atmospheric pressure, smoke shading rate and sensor node spatial coordinates, the system only retrieves node data with stable perception credibility for diffusion trend judgment and spray response strategy generation, thereby improving the robustness of the overall smoke perception data processing process and the accuracy of response control.

[0044] Specifically, the smoke perception data of all normal nodes are retrieved, the smoke concentration fluctuation amplitude and change rate are analyzed, and the specific steps for evaluating the diffusion credibility of each sensor node are as follows: at the end of each sampling period, the smoke perception data of all nodes marked as normal are retrieved, covering dimensional information such as smoke concentration, sensor temperature, ambient temperature, wind speed, relative humidity, atmospheric pressure, smoke shading rate and sensor node spatial coordinates. The difference between the maximum and minimum smoke concentrations of each sensor node in the sampling period is calculated based on a complete and continuous numerical sequence to obtain the smoke concentration fluctuation amplitude; and based on the ratio of the concentration difference at continuous sampling moments to the cycle duration, the concentration change rate of each sensor node per unit time is calculated to obtain the smoke concentration change rate; and then the smoke concentration change rate of each sensor node is evaluated. The diffusion credibility of each sensor node is calculated as follows: the smoke concentration change rate is multiplied by the wind speed to obtain the diffusion driving term, which is used to characterize the diffusion potential energy in the current period; the absolute value of the difference between the sensor temperature and the ambient temperature is calculated to obtain the thermal disturbance term, which reflects the intensity of heat source interference; the smoke concentration fluctuation amplitude is calculated and divided by the relative humidity value to obtain the fluctuation disturbance term, which measures the concentration instability under the influence of humidity; the inverse of the atmospheric pressure is calculated to obtain the air pressure disturbance term, which quantifies the impact of air pressure changes on the diffusion pattern; the thermal disturbance term, the fluctuation disturbance term and the air pressure disturbance term are added together to construct the environmental disturbance composite term; finally, the diffusion driving term is divided by the environmental disturbance composite term to form a diffusion credibility assessment value, which is used to characterize the credibility level of the diffusion trend of the current node in the sampling period.

[0045] The specific calculation formula of the diffusion credibility assessment value is:

[0046]

[0047] In the formula, K represents the diffusion credible assessment value, C represents the smoke concentration, represents the rate of change of smoke concentration, v represents wind speed, T s Indicates the sensor temperature, T a Indicates the ambient temperature, C b Indicates the fluctuation amplitude of smoke concentration, RH indicates relative humidity, and P indicates atmospheric pressure.

[0048] In this implementation plan, by retrieving the smoke sensing data of all normal nodes, the smoke concentration fluctuation amplitude and the smoke concentration change rate are comprehensively calculated based on the smoke concentration, sensor temperature, ambient temperature, wind speed, relative humidity, atmospheric pressure and smoke shading rate, and the diffusion driving term, thermal disturbance term, fluctuation disturbance term and air pressure disturbance term are constructed to form a diffusion credible assessment value, which can accurately reflect the evolution characteristics of the early fire diffusion process under various environmental disturbance conditions, effectively improve the accuracy of identifying weak diffusion trends, and significantly enhance the system's ability to discriminate diffusion trends in complex dynamic environments.

[0049] Specifically, the specific steps for marking sensor nodes that are in an upward trend based on the evaluation results are as follows: compare the diffusion credible evaluation value and the diffusion credible threshold in real time, and retrieve the smoke concentration, smoke concentration change rate, wind speed, sensor temperature, ambient temperature, relative humidity, atmospheric pressure and smoke concentration fluctuation amplitude in the current sampling period for each sensor node; eliminate the smoke concentration data corresponding to the sensor nodes whose diffusion credible evaluation value is less than the diffusion credible threshold, and perform differential analysis on the smoke concentration change rate sequence of the retained nodes: if the direction of the smoke concentration change rate changes alternately in three consecutive sampling periods, and the smoke concentration fluctuation amplitude in each sampling period is less than the amplitude threshold, then the smoke concentration data corresponding to the sensor node is judged to be pseudo-diffusion trend and eliminate them; for the remaining sensor nodes after pseudo-diffusion elimination, determine whether the smoke concentration change rate in the current sampling period is greater than the change rate threshold, and whether the diffusion credibility assessment value is greater than the diffusion credibility threshold. If both the smoke concentration change rate and the diffusion credibility assessment value conditions are met, the sensor node is marked as an upward trend based on its concentration evolution trend and environmental response state; for the sensor nodes marked as a credible upward trend, a delayed release window with a fixed length is set to continuously monitor the smoke concentration change rate and diffusion credibility assessment value in the subsequent sampling period. If the smoke concentration change rate is always negative for three consecutive sampling periods, and the diffusion credibility assessment value is lower than the diffusion credibility threshold, the upward trend mark of the sensor node is removed.

[0050] In this implementation scheme, by comprehensively analyzing the correspondence between the smoke concentration, smoke concentration change rate, smoke concentration fluctuation amplitude, diffusion credibility assessment value and diffusion credibility threshold of the sensor node, based on the multi-period differential trend and continuous fluctuation state, the upward trend sensor nodes in the real diffusion process are accurately identified, and the pseudo-diffusion trend interference caused by local disturbances, random noise or perception errors is significantly reduced, thereby improving the reliability and stability of the diffusion trend judgment, and enhancing the accuracy of the smoke sensing system's response to early fire dynamics and the robustness of node status judgment.

[0051] Specifically, the spraying control area is constructed based on the spatial position of the rising trend sensor node, and the specific steps for analyzing the spraying response level of the control area are as follows: according to the physical space coordinate information of the sensor node, the monitoring area is divided into several fixed unit areas using the grid division method, and all the sensor nodes in the rising trend state are mapped to the corresponding grid units to construct the spraying control area; for each rising trend sensor node in the spraying control area, the smoke concentration, sensor temperature, ambient temperature, wind speed, relative humidity, voltage signal, atmospheric pressure, and smoke shading rate are extracted respectively, and the weighted mean of each field in the control area is calculated to obtain the average smoke concentration, average relative humidity, average smoke shading rate, and smoke concentration change. The average smoke perception data of the rate average, wind speed average and diffusion credible assessment value average are used; based on the average smoke perception data in each control area, the spraying response level of the control area is analyzed: the ratio of the average smoke concentration to the average relative humidity is calculated to obtain the humidity modulation concentration term; the humidity modulation concentration term is added to the average smoke shading rate to obtain the smoke intensity term; the product of the average smoke concentration change rate and the average wind speed is calculated to obtain the diffusion coupling term, and 1 is added to the diffusion coupling term to obtain the diffusion amplification factor; the smoke intensity term, the diffusion amplification factor and the average diffusion credible assessment value are multiplied to obtain the spraying response assessment value, so as to assign corresponding response levels and execution strategies to different control areas.

[0052] The specific calculation formula for the spray response evaluation value is:

[0053]

[0054] Where X represents the spray response evaluation value, The average smoke concentration, Represents the average relative humidity, represents the average smoke shading rate, Indicates the average value of the smoke concentration change rate, represents the average wind speed, Represents the average value of the diffusion credibility assessment.

[0055] In this embodiment, Table 2 is a table of spray response evaluation values, which lists the average smoke perception data of the five control areas and their corresponding spray response evaluation values, which are used to comprehensively determine the fire response level and spray control strategy of different areas. Among them: in control area K1, the average smoke concentration is 12.5, the average relative humidity is 50, the average smoke shading rate is 0.35, the average smoke concentration change rate is 0.6, the average wind speed is 0.8, the average diffusion credibility evaluation value is 0.98, and the calculated spray response evaluation value is 0.87; in control area K2, the average smoke concentration is 35.0, the average relative humidity is 65, the average smoke shading rate is 0.60, the average smoke concentration change rate is 1.2, the average wind speed is 1.5, the average diffusion credibility evaluation value is 0.23, and the spray response evaluation value is 0.76; in control area K3, the average smoke concentration is 22.0, the average relative humidity is 55, and the average smoke shading rate is 0.45. The average value of the smoke concentration change rate was 0.8, the average value of the wind speed was 1.0, the average value of the diffusion credible assessment value was 0.67, and the spraying response assessment value was 1.03; in the control area K4, the average value of the smoke concentration was 45.0, the average value of the relative humidity was 60, the average value of the smoke shading rate was 0.70, the average value of the smoke concentration change rate was 1.5, the average value of the wind speed was 2.0, the average value of the diffusion credible assessment value was 0.16, and the spraying response assessment value was 0.95; in the control area K5, the average value of the smoke concentration was 18.0, the average value of the relative humidity was 58, the average value of the smoke shading rate was 0.40, the average value of the smoke concentration change rate was 0.7, the average value of the wind speed was 1.2, the average value of the diffusion credible assessment value was 0.67, and the spraying response assessment value was 0.88.

[0056] Table 2 Spray response evaluation value data table

[0057]

[0058] like Figure 4 The figure shows the distribution of spray response levels for each control area, displaying the spray response assessment values ​​and their corresponding spray response levels for the five control areas: Control areas K1, K4, and K5 have spray response assessment values ​​of 0.87, 0.95, and 0.88, respectively, falling between the first and second level spray response thresholds and being classified as medium-level responses; control area K2 has an assessment value of 0.76, indicating a low-level response; and control area K3 has an assessment value of 1.03, indicating a high-level response. The figure, with green and red dashed lines representing the first and second level spray response thresholds, respectively, visually illustrates the spray response status and classification of each control area in the current cycle, providing a reliable basis for coordinated decision-making regarding spray control strategies.

[0059] In this implementation plan, the spatial coordinate information of the rising trend sensor node and the average smoke perception data are integrated to accurately construct the spraying control area, and the average smoke concentration, relative humidity, smoke shading rate, smoke concentration change rate, wind speed and diffusion credibility assessment value are integrated to calculate the spraying response assessment value, thereby achieving quantitative judgment of the spraying response level and effectively enhancing the regional adaptability and response accuracy of the linkage spraying control strategy in multi-source perception scenarios.

[0060] Specifically, the specific steps of generating the corresponding spraying strategy and issuing the execution instruction are as follows: compare the spraying response evaluation value with the spraying response threshold, and generate the spraying strategy according to the spraying response level classification standard and the classification rule: if the spraying response evaluation value is less than or equal to the first-level spraying response threshold, trigger the low-level response, call the corresponding nozzle number of all rising trend sensor nodes in the current control area, start the single-point nozzle, and set the low-pressure short-time spraying parameters; if the spraying response evaluation value is greater than the second-level spraying response threshold and less than the first-level spraying response threshold, trigger the medium-level response, based on the control The regional grid boundary identifies the associated regional sprinklers, starts the regional sprinklers, and sets the medium-pressure medium-time spraying parameters; if the spray response evaluation value is greater than or equal to the secondary spray response threshold, a high-level response is triggered, all controllable sprinkler groups in the control area are loaded, the full-coverage sprinkler group is started, and the high-pressure continuous spraying parameters are set; based on the final determined spray response level and the control area spraying strategy, a spray control instruction is generated, which includes the target sprinkler number, spraying pressure, spraying duration and start time, and is sent to the execution end through the control interface to drive the corresponding sprinkler into the spraying state.

[0061] In this implementation plan, by performing a graded comparison between the spray response evaluation value and the spray response threshold, a spray strategy generation mechanism is constructed, which can accurately determine the spray response level, dynamically match the nozzle number, spray pressure, spray duration and start time, and effectively improve the response accuracy and execution efficiency of the spray linkage control system under the conditions of multi-source heterogeneous smoke perception data, and realize a hierarchical response, fixed-point linkage control, and parameter-adjustable intelligent spraying operation mode.

[0062] Specifically, continuous monitoring is performed during the spraying period, and spraying is stopped when continuous low levels appear. The specific steps for restarting the spraying are as follows: During the spraying execution, the smoke perception data in the corresponding control area is continuously monitored. The smoke perception data includes smoke concentration, sensor temperature, ambient temperature, wind speed, relative humidity, voltage signal, atmospheric pressure, smoke shading rate and sensor node spatial coordinates, and the spraying response evaluation value is dynamically updated based on the updated average smoke concentration, average relative humidity, average smoke shading rate, average smoke concentration change rate, average wind speed and average diffusion credible evaluation value; if the spraying response evaluation values ​​of all sensor nodes in the area that are in an upward trend state are lower than the first-level spraying response threshold within three consecutive sampling periods, a nozzle closing command is issued to terminate the current The system detects the previous spraying behavior and records the spraying end time; after the spraying is completed, the control area enters the spraying maintenance window period, and the high-frequency sampling tasks of the sensor nodes and the trend priority analysis algorithm based on the diffusion credibility evaluation value and the concentration change rate are maintained during the spraying maintenance window period; if the upward trend sensor node is detected again during the spraying maintenance window period, and the corresponding smoke concentration average value, relative humidity average value, smoke shading rate average value, smoke concentration change rate average value, wind speed average value and diffusion credibility evaluation value average value are recalculated to obtain a spraying response evaluation value greater than the spraying strategy startup threshold, the restart process of the spraying linkage control strategy is immediately triggered, the spraying intensity level is reallocated to the control area, and a new spraying control instruction is generated based on the current smoke perception data, so as to realize the precise linkage response to the dynamic changes of the fire spread.

[0063] In this implementation plan, by continuously monitoring the smoke perception data in the control area, dynamically updating the spraying response evaluation value, and combining the continuous state changes of the rising trend sensor node, it is ensured that the spraying is terminated in time when the spraying response evaluation value is continuously in a low-level state to avoid waste of water resources; at the same time, a spraying maintenance window period is set after the spraying is completed, and the high-frequency sampling task and trend priority analysis mechanism are retained. When the rising trend sensor node is detected again and the spraying response evaluation value exceeds the spraying strategy start-up threshold again, the spraying linkage control strategy is accurately restarted, which enhances the system's adaptability and response sensitivity to repeated fluctuations in the fire situation, and improves the intelligent scheduling level and safety and reliability of the spraying control.

[0064] like Figure 2As shown, the second aspect of the present invention provides an intelligent fire sprinkler linkage control system based on smoke diffusion trend, including: a smoke perception data acquisition and preprocessing module, an abnormal state identification and node control module, a diffusion trend evaluation and marking judgment module and a spray response classification and linkage control module, wherein: the smoke perception data acquisition and preprocessing module is used to obtain smoke perception data, and perform temperature and pressure correction, trend filtering and normalization processing on the smoke perception data to obtain preprocessed smoke perception data; the abnormal state identification and node control module is used to extract the disturbance index of each sensor node based on the preprocessed smoke perception data, and evaluate the current sampling period of smoke perception data. Credibility, distinguishing normal nodes from abnormal nodes, and dynamically maintaining the node credibility; diffusion trend assessment and marking discrimination module, used to retrieve the smoke perception data of all normal nodes, analyze the fluctuation amplitude and change rate of smoke concentration, evaluate the diffusion credibility of each sensor node, and mark the sensor nodes in an upward trend based on the assessment results; spraying response classification and linkage control module, used to construct the spraying control area based on the spatial position of the upward trend sensor node, analyze the spraying response level of the control area, generate the corresponding spraying strategy and issue execution instructions; continuous monitoring during spraying, stop spraying if continuous low levels occur, and restart spraying if abnormalities are detected within the window period.

[0065] In this implementation plan, by integrating the smoke perception data acquisition and preprocessing module, the abnormal state identification and node management module, the diffusion trend assessment and marking discrimination module, and the spraying response classification and linkage control module, an intelligent monitoring system for the entire process of fire evolution is constructed, realizing dynamic management of the entire process of smoke perception data from acquisition, preprocessing, credibility assessment, trend identification to spraying response; improving data accuracy through temperature and pressure correction, trend filtering, and normalization processing, ensuring node data quality through disturbance index extraction and perception credibility assessment, accurately identifying the fire evolution process through diffusion credibility analysis and rising trend discrimination, and improving the system's rapid response capability to the early spread of fire and the refined scheduling level of spraying control through response level identification and spraying linkage control strategy, comprehensively enhancing the stability, reliability and timeliness of the intelligent fire protection system.

[0066] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0067] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. An intelligent fire sprinkler linkage control method based on smoke diffusion trend, characterized in that: The following steps are involved: S1, acquiring smoke perception data, and performing temperature and pressure correction, trend filtering, and normalization processing on the smoke perception data to obtain pre-processed smoke perception data; S2, based on the pre-processed smoke sensing data, extracts the disturbance index of each sensor node, evaluates the credibility of the smoke sensing data in the current sampling period, distinguishes normal nodes from abnormal nodes, and dynamically maintains the node credibility; S3, retrieves smoke sensing data from all normal nodes, analyzes smoke concentration fluctuation amplitude and change rate, evaluates the diffusion credibility of each sensor node, and marks sensor nodes with an upward trend based on the evaluation results; S4, based on the spatial position of the rising trend sensor node, constructs the spraying control area, analyzes the spraying response level of the control area, generates the corresponding spraying strategy and issues execution instructions; continuously monitors during spraying, stops spraying if continuous low levels occur, and restarts spraying if abnormalities are detected within the window period.

2. The intelligent fire sprinkler linkage control method based on smoke diffusion trend according to claim 1 is characterized by: The specific steps of acquiring smoke sensing data, performing temperature and pressure correction, trend filtering, and normalization processing on the smoke sensing data to obtain pre-processed smoke sensing data are as follows: The smoke diffusion process in the early stage of a fire is synchronously collected in multiple dimensions throughout the entire process, including smoke concentration evolution characteristics, temperature and humidity fluctuations, local airflow disturbance patterns, sensor self-heating behavior, shading rate jump characteristics, atmospheric pressure stress response, voltage signal output offset, current supply stability, and particle transmission environment. Smoke sensing data is obtained, including smoke concentration, sensor temperature, ambient temperature, wind speed, relative humidity, voltage signal, atmospheric pressure, smoke shading rate, and sensor node spatial coordinates. The smoke sensing data is corrected for thermal drift using a temperature-pressure-concentration correlation correction algorithm to compensate for measurement deviations caused by sensor self-heating and environmental fluctuations. The wind speed and shading rate sequences in the smoke sensing data are subjected to boundary denoising using a polynomial residual smoothing algorithm to suppress abnormal jumps caused by short-term disturbances. The smoke sensing data is anomaly eliminated and partially filled using a spatial trend consistency discrimination algorithm to prevent single-point errors from affecting diffusion trend judgment. The smoke perception data is scaled and normalized using the zero-mean standard deviation normalization algorithm to achieve normalization of the smoke perception data.

3. The intelligent fire sprinkler linkage control method based on smoke diffusion trend according to claim 1 is characterized in that: The specific steps of extracting the disturbance index of each sensor node based on the preprocessed smoke perception data and evaluating the credibility of the smoke perception data in the current sampling period are as follows: A fixed-length sampling period is set for smoke sensing data collection. When the pre-processed smoke sensing data is updated in each sampling period, the disturbance index of each sensor node is extracted: the standard deviation of the voltage signal in the current sampling period is calculated to obtain the electrical fluctuation term; the absolute value of the difference between the sensor temperature and the ambient temperature is calculated to obtain the temperature difference disturbance term; the absolute value of the difference between the smoke concentration and the smoke concentration in the previous sampling period is calculated to obtain the concentration mutation term; the absolute value of the difference between the smoke shading rate and the average smoke shading rate is calculated and divided by the average smoke shading rate to obtain the shading anomaly term; if any two of the smoke concentration, sensor temperature, and voltage signals of the sensor node are missing, the key field missing item is marked as 1, otherwise it is marked as 0; Based on the various disturbance indicators within the sampling period, the credibility of the smoke sensing data in the current sampling period is evaluated: the sum of the electrical fluctuation term, the temperature difference disturbance term, and the concentration mutation term is calculated and divided by the atmospheric pressure plus 1 to obtain the basic disturbance factor; Calculate the shading anomaly item and add 1 to obtain the shading fluctuation amplification factor; calculate the product of the key field missing item and the missing penalty coefficient to obtain the missing penalty item; multiply the basic disturbance factor by the shading fluctuation amplification factor and add the missing penalty item to obtain the perception credibility assessment value.

4. The intelligent fire sprinkler linkage control method based on smoke diffusion trend according to claim 1 is characterized in that: The specific steps for distinguishing normal nodes from abnormal nodes and dynamically maintaining the node trust status are as follows: Compare the perception credibility assessment value with the perception credibility threshold in real time. If the perception credibility assessment value is less than or equal to the perception credibility threshold, retain the smoke perception data of the current sampling period and mark the corresponding sensor node as a normal node. If the perception credibility evaluation value is greater than the perception credibility threshold, the smoke perception data of the current sampling period will be discarded, and the corresponding sensor node will be marked as an abnormal node, and will be suspended from participating in subsequent diffusion trend determination and spray response evaluation; If the perception credibility evaluation value of the abnormal node is less than or equal to the perception credibility threshold in three consecutive sampling periods, its abnormal mark is removed and the data processing authority of the sensor node is restored.

5. The intelligent fire sprinkler linkage control method based on smoke diffusion trend according to claim 1 is characterized in that: The specific steps of retrieving smoke sensing data from all normal nodes, analyzing smoke concentration fluctuation amplitude and change rate, and evaluating the diffusion credibility of each sensor node are as follows: At the end of each sampling period, the smoke sensing data of all normal nodes are retrieved, and the difference between the maximum and minimum smoke concentrations of each sensor node within the sampling period is calculated to obtain the smoke concentration fluctuation amplitude; And calculate the concentration change rate of each sensor node per unit time to obtain the smoke concentration change rate; Evaluate the diffusion credibility of each sensor node: calculate the smoke concentration change rate multiplied by the wind speed to obtain the diffusion driving term; Calculate the absolute value of the difference between the sensor temperature and the ambient temperature to obtain the thermal disturbance term; Calculate the smoke concentration fluctuation amplitude and divide it by the relative humidity value to obtain the fluctuation disturbance term; calculate the inverse of the atmospheric pressure to obtain the pressure disturbance term; Add the thermal disturbance term, the fluctuation disturbance term and the pressure disturbance term to obtain the total environmental disturbance term. The diffusion driving term is divided by the environmental disturbance term to obtain the diffusion credibility assessment value.

6. The intelligent fire sprinkler linkage control method based on smoke diffusion trend according to claim 1 is characterized by: The specific steps of marking the sensor nodes in an upward trend based on the evaluation results are as follows: The diffusion credibility assessment value and the diffusion credibility threshold are compared in real time. The smoke concentration data corresponding to the sensor nodes whose diffusion credibility assessment value is less than the diffusion credibility threshold are eliminated. The smoke concentration change rate series of the retained nodes are differentially analyzed: if the direction of the concentration change rate changes alternately in three consecutive sampling periods, and the smoke concentration fluctuation amplitude in each sampling period is less than the amplitude threshold, the smoke concentration data corresponding to the sensor node is determined to be a false diffusion trend and is eliminated. For the remaining sensor nodes after pseudo-diffusion elimination, determine whether the smoke concentration change rate in the current sampling period is greater than the change rate threshold, and whether the diffusion credibility assessment value is greater than the diffusion credibility threshold. If both conditions are met at the same time, the sensor node is marked as an upward trend; a delayed release window is set for the sensor node marked as a credible upward trend. If the smoke concentration change rate is negative and the diffusion credibility assessment value is lower than the diffusion credibility threshold for three consecutive sampling periods, the upward trend mark is removed.

7. The intelligent fire sprinkler linkage control method based on smoke diffusion trend according to claim 1 is characterized in that: The specific steps of constructing the spraying control area based on the spatial position of the upward trend sensor node and analyzing the spraying response level of the control area are as follows: According to the physical space coordinate information of the sensor nodes, the monitoring area is divided into several fixed unit areas using the grid division method, and all the sensor nodes in the rising trend state are mapped to the corresponding grid cells to construct the spraying control area; Calculate the average smoke perception data and the average diffusion credibility assessment value in each spraying control area; Based on the average smoke perception data in each control area, the spraying response level of the control area is analyzed: the ratio of the average smoke concentration to the average relative humidity is calculated to obtain the humidity modulation concentration term; the humidity modulation concentration term is added to the average smoke shading rate to obtain the smoke intensity term; the product of the average smoke concentration change rate and the average wind speed is calculated to obtain the diffusion coupling term, and 1 is added to the diffusion coupling term to obtain the diffusion amplification factor; the smoke intensity term, the diffusion amplification factor and the average diffusion credible assessment value are multiplied to obtain the spraying response assessment value.

8. The intelligent fire sprinkler linkage control method based on smoke diffusion trend according to claim 1 is characterized in that: The specific steps of generating the corresponding spraying strategy and issuing the execution instruction are as follows: The spray response evaluation value is compared with the spray response threshold, and a spray strategy is generated according to the grading rules: if the spray response evaluation value is less than or equal to the first-level spray response threshold, a low-level response is triggered, and a single-point sprinkler and low-pressure short-time spraying are started; if the spray response evaluation value is greater than the second-level spray response threshold and less than the first-level spray response threshold, a medium-level response is triggered, and the regional sprinkler and medium-pressure medium-time spraying are started; if the spray response evaluation value is greater than or equal to the second-level spray response threshold, a high-level response is triggered, and a full-coverage sprinkler group and high-pressure continuous spraying are started; Generate spraying control instructions based on the spraying strategy. The instruction content includes the target nozzle number, spraying pressure, spraying duration and start time, and send it to the execution end for spraying operation.

9. The intelligent fire sprinkler linkage control method based on smoke diffusion trend according to claim 1 is characterized in that: The specific steps for continuously monitoring during the spraying period and stopping the spraying if a continuous low level occurs and restarting the spraying if an abnormality is detected within the window period are as follows: During the spraying process, the smoke sensing data in the corresponding control area is continuously monitored, and the spraying response evaluation value is dynamically updated. If the spraying response evaluation values ​​of all the upward trend sensor nodes in the area are continuously at a low level for three consecutive sampling periods, a nozzle shutdown command is issued to terminate the current spraying behavior. After the spraying is completed, the control area enters the spraying maintenance window period, retaining the node high-frequency sampling task and trend priority analysis algorithm; if the upward trend sensor node is detected again during the window period, and its corresponding spraying response evaluation value is greater than the start threshold, the restart process of the spraying linkage control strategy is immediately triggered, the spraying intensity level is reallocated to the control area, and a new spraying control instruction is generated.

10. Intelligent fire sprinkler linkage control system based on smoke diffusion trend, characterized by: include: Smoke perception data collection and preprocessing module, abnormal state identification and node control module, diffusion trend assessment and marking discrimination module, and spray response classification and linkage control module, including: The smoke sensing data acquisition and preprocessing module is used to obtain smoke sensing data, and perform temperature and pressure correction, trend filtering and normalization processing on the smoke sensing data to obtain preprocessed smoke sensing data; The abnormal state identification and node control module is used to extract the disturbance index of each sensor node based on the pre-processed smoke sensing data, evaluate the credibility of the smoke sensing data in the current sampling period, distinguish between normal nodes and abnormal nodes, and dynamically maintain the node credibility; The diffusion trend assessment and marking discrimination module is used to retrieve smoke sensing data from all normal nodes, analyze the smoke concentration fluctuation amplitude and change rate, evaluate the diffusion credibility of each sensor node, and mark the sensor nodes with an upward trend based on the assessment results; The spraying response classification and linkage control module is used to construct a spraying control area based on the spatial position of the rising trend sensor node, analyze the spraying response level of the control area, generate a corresponding spraying strategy and issue an execution instruction; continuously monitor during spraying, stop spraying when continuous low levels appear, and restart spraying if an abnormality is detected within the window period.

Images