Intelligent data management system for building fire safety monitoring

Abstract

The present application relates to the field of data processing, and specifically to an intelligent data management system for building fire safety monitoring, the system comprising: a data acquisition module for collecting temperature data of each temperature sensor at all times; a spatial feature analysis module for determining the positional spatial feature value of temperature data by analyzing the change trend of the data collected by the temperature sensors at adjacent positions, and the change characteristics of the temperature data with the spatial distance; a monitoring data processing module for completing the intelligent management of temperature data based on the positional spatial feature value and temperature data of each temperature sensor at each data collection moment. The present application completes data classification by analyzing the temperature data fluctuation characteristics caused by the smoldering phenomenon of combustibles in the building and the influence of the positional spatial distance on the temperature data difference, thereby improving the processing accuracy of the monitoring data and realizing the intelligent processing of the monitoring data.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to an intelligent data management system for building fire safety monitoring. Background Art

[0002] Building fire safety monitoring is to ensure the fire safety of buildings by dynamically monitoring and regularly testing multiple links in the building fire protection system. Among them, early fire detection is one of the important detection tasks to check for fire hazards in buildings, discover potential fire safety hazards, and improve the fire safety level of buildings.

[0003] The very early stage of a fire is the smoldering stage. The most important feature of this stage is the change in temperature inside the building. At present, a variety of intelligent monitoring equipment, such as dust detectors, smoke alarms, temperature and humidity sensors, are usually deployed inside buildings to achieve all-weather monitoring of fires. Temperature sensors are used to control the area around combustibles in the building, collect data through temperature sensors, and then analyze the data to achieve early monitoring of building fires. However, all-weather monitoring causes sensors to generate a large amount of data, which requires effective data management to filter out valid data.

[0004] For fire monitoring, abnormal data in the monitoring process is valid data that may have safety hazards and is of high importance. Therefore, it is necessary to accurately segment abnormal data and normal data in the monitoring process. However, at present, the segmentation of sensor monitoring data is often done through the global threshold segmentation method, which ignores the impact of the sensor installation environment on the data value, resulting in low efficiency of data management in the monitoring process and failure to provide effective data support for fire management. Summary of the invention

[0005] The present application provides an intelligent data management system for building fire safety monitoring to solve the existing problems.

[0006] This application proposes an intelligent data management system for building fire safety monitoring, the system comprising:

[0007] A data acquisition module, used to collect temperature data of each temperature sensor at all times;

[0008] The spatial feature analysis module is used to determine the spatial feature value of the temperature data by analyzing the change trend of the data collected by the temperature sensors at adjacent locations and the change characteristics of the temperature data with the spatial distance, specifically:

[0009] (1) Obtain the abnormal points of the temperature data, divide the temperature data into multiple local data sequences with each abnormal point as a segmentation point, and determine the feature array of the abnormal point based on all temperature values ​​in the local data sequence, the numerical value of the abnormal point, and the number of temperature data; obtain the fluctuation coefficient of each temperature data based on the distance between the feature arrays of the abnormal points of the same order before each temperature data of the temperature sensor and the distance difference between the temperature sensors; determine the fluctuation credibility of each temperature data based on the difference between the numerical value of the abnormal point before the temperature data and the numerical value of the temperature data and the fluctuation coefficient of the temperature data;

[0010] (2) With a temperature sensor as the center, set several radii, sort the differences between the temperature of the temperature sensor and the temperatures of all temperature sensors within its radius in order from small to large, and then perform first-order differences to obtain a correction sequence; determine the correction coefficient of each temperature sensor at each data collection moment based on the difference between the number of negative numbers in the correction sequence and the total number;

[0011] (3) At each acquisition moment, the fluctuation credibility of each temperature data is corrected by forward fusion of the fluctuation credibility according to the difference between the correction coefficient and the maximum value of the correction coefficient, and the position space characteristic value at each data acquisition moment is obtained;

[0012] The monitoring data processing module obtains the associated characteristic values ​​based on the mean, variance and minimum time interval of adjacent abnormal points of all abnormal point temperatures before each collection moment and the position space characteristic values ​​at the collection moment; all associated characteristic values ​​are classified into classes with different danger levels to complete the intelligent management of temperature data.

[0013] Preferably, determining the fluctuation credibility of each temperature data based on the difference of temperature data between temperature sensors includes:

[0014] Determine the fluctuation coefficient of each temperature data collected by each temperature sensor based on the difference of the characteristic arrays between the abnormal points in the temperature data collected by different temperature sensors;

[0015] The fluctuation credibility of each temperature data is determined based on the overall fluctuation degree of the temperature data of each temperature sensor before each data collection moment and the fluctuation coefficient.

[0016] Preferably, the method for obtaining the fluctuation coefficient of each temperature data includes:

[0017] The RRCF algorithm is used to detect abnormal points in the temperature data collected by each temperature sensor, and the feature array of each abnormal point is obtained;

[0018] Calculate the ratio of the DTW distance between each temperature sensor and any other temperature sensor's feature array of any abnormal point of the same order before each moment to the position distance of the two temperature sensors;

[0019] The cumulative result of the ratio at all abnormal points of all temperature sensors in the building before each moment is taken as the fluctuation coefficient of each temperature data collected by each sensor.

[0020] Preferably, the method for obtaining the fluctuation credibility of each temperature data includes:

[0021] Calculate the difference between the mean of all abnormal points before the corresponding collection time of each temperature data collected by each temperature sensor and the mean of all temperature data before the corresponding collection time;

[0022] The normalized result of the product of the difference and the fluctuation coefficient of each temperature data is used as the fluctuation credibility of each temperature data.

[0023] Preferably, the method for obtaining the feature array includes:

[0024] The RRCF algorithm is used to detect abnormal points in the temperature data collected by each temperature sensor;

[0025] Each abnormal point is used as a segmentation point to divide all temperature data into multiple local data sequences. The variance of all temperature data in a local data sequence adjacent to each abnormal point is calculated as the fluctuation radius of each abnormal point. The array consisting of the value of each abnormal point, the number of temperature data before each abnormal point and the fluctuation radius of each abnormal point is used as the feature array of each abnormal point.

[0026] Preferably, the method for obtaining the correction coefficient of each temperature sensor at each data collection moment includes:

[0027] The ratio of the number of elements less than 0 in the correction sequence of each temperature sensor at each data collection moment to the number of elements in the correction sequence is used as the correction coefficient of each temperature sensor at each data collection moment.

[0028] Preferably, the method for obtaining the correction sequence includes:

[0029] For any data collection moment, taking the position of each temperature sensor as the center of the sphere, obtaining the temperature data of all temperature sensors on the spherical surface at each data collection moment according to the radius r, and calculating the mean of the temperature data of all temperature sensors on the spherical surface at each data collection moment; taking the absolute value of the difference between the temperature data of each temperature sensor at each data collection moment and the mean as the single distance difference of each temperature sensor at each data collection moment;

[0030] The single distance difference of each temperature sensor at each data collection time under different radii is obtained in turn, all the single distance differences are sorted in order from small to large radius, and the first-order difference sequence of the sorting result is used as the correction sequence of each temperature sensor at each data collection time.

[0031] Preferably, the method for acquiring the position space characteristic value at each data collection moment includes:

[0032] Calculate the ratio of the correction coefficient of each temperature sensor at each data collection moment to the maximum value of the correction coefficient of each temperature sensor among all data collection moments;

[0033] The product of the ratio and the fluctuation credibility of the temperature data collected by each temperature sensor at each data collection time is used as the position space characteristic value at each data collection time.

[0034] Preferably, the method for completing intelligent management of temperature data based on the positional spatial characteristic value and temperature data of each temperature sensor at each data collection moment includes:

[0035] Determine the correlation characteristic value between the temperature data collected by each temperature sensor at each data collection moment and the super-early signs of fire based on the abnormal point of each temperature sensor before each data collection moment and the position spatial characteristic value of each temperature sensor at each data collection moment;

[0036] Obtaining correlation feature values ​​between the temperature data collected at all data collection moments of each temperature sensor and the early signs of fire, and dividing all the correlation feature values ​​into multiple clusters using a data clustering algorithm;

[0037] Calculate the mean of all associated feature values ​​in each cluster as the label value of each cluster, arrange the label values ​​of all clusters in order from large to small, and mark the clusters as the first monitoring danger level, the second monitoring danger level, and up to the Kth monitoring danger level based on the arrangement results; where K is the number of clusters.

[0038] Preferably, the method for obtaining the associated characteristic value includes:

[0039]

[0040] In the formula, is the correlation characteristic value between the temperature data of the cth temperature sensor at the tth moment and the early signs of fire, , are the mean and variance of the temperature data between all adjacent abnormal points of the cth temperature sensor before the tth time, is the minimum time interval between all adjacent abnormal points of the cth temperature sensor before the tth time. is the position space eigenvalue of the cth temperature sensor at the tth moment, is the minimum time interval between the tth moment and the corresponding moment of the abnormal point, is the tuning factor.

[0041] In the above scheme, the present application first determines the fluctuation coefficient of the temperature data of each temperature sensor at each data collection moment based on the characteristics that the temperature increases when the combustibles in the building are smoldering, and the heat gradually diffuses from the area close to the smoldering phenomenon to the area far away from the smoldering phenomenon; secondly, the fluctuation credibility of each temperature data is determined in combination with the data fluctuation characteristics when the smoldering phenomenon occurs. The beneficial effect is that the fluctuation characteristics of the temperature data of temperature sensors at different positions at each data collection moment can be determined; then, the stability of the change of temperature difference in the building under the premise of distance from small to large is analyzed to determine the correction coefficient of the fluctuation credibility of each temperature data, which is used to eliminate the temperature sensor in the all-weather monitoring process. The impact of data fluctuations caused by internal interference on data fluctuations caused by smoldering phenomena; then, based on the characteristic that smoldering phenomena will cause continuous high-temperature data between adjacent abnormal points and the spatial characteristic values ​​of the position of each temperature sensor at each data collection moment, the associated characteristic values ​​of the temperature data collected by each temperature sensor at each data collection moment and the ultra-early signs of fire are obtained, realizing a close correlation analysis between the monitoring data and the characteristics of ultra-early fire phenomena; finally, a data clustering algorithm is used to divide the temperature data collected by each temperature sensor at all data collection moments into multiple clusters based on the associated characteristic values ​​of the temperature data and the ultra-early signs of fire, and mark them, completing the intelligent processing of the monitoring data in the building. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 A flow chart of an intelligent data management system for building fire safety monitoring provided in an embodiment of the present application;

[0044] Figure 2 A flowchart for implementing the position space feature value provided in the embodiment of the present application;

[0045] Figure 3 This is a flowchart for implementing intelligent data management based on classification and labeling of collected data provided in an embodiment of the present application. DETAILED DESCRIPTION

[0046] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the specific implementation method, structure, features and effects of the intelligent data management system for building fire safety monitoring proposed in the present application are described in detail below in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0047] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0048] The specific scheme of the intelligent data management system for building fire safety monitoring provided by the present application is described in detail below with reference to the accompanying drawings.

[0049] See also Figure 1 , which shows a flow chart of an intelligent data management system for building fire safety monitoring provided by an embodiment of the present application. The system includes the following modules: data acquisition module, spatial feature analysis module, and monitoring data processing module. The specific contents of each module are:

[0050] The data acquisition module is used to collect the temperature data of each temperature sensor at all times.

[0051] Specifically, first determine the location of the combustibles in the building, which include fixed combustibles and container combustibles. Then, arrange temperature sensors around each combustible in the building. In this embodiment, the center of mass of the combustible is taken as the center of the circle, and the temperature sensor is arranged around the combustible. The temperature sensors are arranged in a layout manner. In this embodiment, the arrangement unit of the sensors is meter. It should be noted that the arrangement manner of the temperature sensors is not fixed. In other embodiments, the implementer may select a suitable arrangement manner.

[0052] Furthermore, this embodiment collects temperature data monitored by the temperature sensor at each moment at a time interval of three minutes, and obtains all temperature data within 12 hours, wherein the number of temperature sensors is greater than or equal to 2. The time interval for the temperature sensor to collect temperature data is not specifically limited, and the implementer may determine it according to the specific situation.

[0053] Preferably, in order to avoid data loss caused by unstable transmission network during the collection process, the present embodiment uses linear interpolation to fill the data. Linear filling is a common technology in the field of data processing, and its process will not be described in detail.

[0054] The spatial feature analysis module is used to determine the spatial feature value of the temperature data by analyzing the change trend of the data collected by the temperature sensors at adjacent locations and the change characteristics of the temperature data with the spatial distance. The specific process is as follows: Figure 2 As shown, it includes steps 201, 202 and 203:

[0055] 201 , determining the fluctuation credibility of each temperature data based on the difference of the temperature data between the temperature sensors.

[0056] When smoldering of combustibles occurs in a building, the temperature around the combustibles and the temperatures at other locations in the surrounding space gradually decrease. By analyzing the changing relationship between the temperature data of each temperature sensor and the temperature data of its surrounding temperature sensors, combined with the temperature change characteristics of the ultra-early stage of a fire, it is possible that each temperature data is monitoring data in the ultra-early stage of a fire during the monitoring process.

[0057] Furthermore, when there is no smoldering of combustibles, the temperature data of each temperature sensor in the building are not much different, and the size of the temperature data basically fluctuates around the room temperature; however, when there is smoldering of combustibles, the heat radiation generated by the fire radiates outward from the fire source, resulting in large differences in temperature data between different temperature sensors, that is, the credibility of the temperature data of each temperature sensor can be obtained based on the difference in the temperature data of the temperature sensors.

[0058] On the other hand, the greater the distance between the two temperature sensors in the building, the higher the possibility that the ambient temperatures of the two temperature sensors are greatly different. For example, the temperature data of the temperature sensor near the vent is closer to the outdoor temperature than the temperature data of the temperature sensor inside the building. Then, when smoldering of combustibles occurs, there will be a certain difference in the degree of change in the data collected by the two temperature sensors, that is, the fluctuation trends of the data collected by temperature sensors at different locations are not consistent.

[0059] For any temperature sensor, all the temperature data collected by each temperature sensor is used as input, and the Robust Random Cut Forest (RRCF) algorithm is used to output the abnormal points in all the temperature data. It should be noted that due to the all-weather monitoring, the temperature sensor is in a state of continuously collecting temperature data, and the RRCF algorithm is a dynamic detection method that can accurately detect abnormal points. Among them, the RRCF algorithm is a common technology in the field of data processing, and its process will not be repeated.

[0060] Furthermore, each abnormal point is used as a mutation point where the temperature data fluctuates. Therefore, all temperature data can be divided into multiple local data sequences using the abnormal point as a segmentation point, which is used to analyze the data fluctuations of different temperature sensors in the short term. The variance of all temperature data in the local data sequence adjacent to each abnormal point is calculated as the fluctuation radius a of each abnormal point, and the array consisting of the value v of each abnormal point, the number n of temperature data before each abnormal point, and the fluctuation radius a of each abnormal point is used as the feature array of each abnormal point. , which is used to characterize the data characteristics of each abnormal point.

[0061] The fluctuation credibility of each temperature data collected by each temperature sensor is determined based on the characteristic array of the abnormal points corresponding to the temperature sensors and the temperature data.

[0062] First, calculate the fluctuation coefficient of the temperature data of the i-th temperature sensor at time t:

[0063]

[0064] In the formula, is the fluctuation coefficient of the temperature data of the ith temperature sensor at the tth time, M is the number of all abnormal points of the ith temperature sensor before the tth time, N is the number of temperature sensors installed in the building, , are the feature arrays of the bth abnormal point of the i-th temperature sensor and the j-th temperature sensor before the tth time, is the DTW distance between, is the Euclidean distance between the positions of the i-th temperature sensor and the j-th temperature sensor.

[0065] Secondly, the fluctuation credibility of the temperature data of the ith temperature sensor at the t moment is determined by combining the overall fluctuation degree of the temperature data of the ith temperature sensor before the t moment:

[0066]

[0067] In the formula, is the fluctuation credibility of the temperature data of the i-th temperature sensor at the t-th moment, is the normalization function, is the mean of all abnormal points of the ith temperature sensor before the tth time, It is the mean of all temperature data of the i-th temperature sensor before the t-th time.

[0068] Among them, the greater the difference in the data fluctuation mutation between the temperature data collected by the two temperature sensors, the greater the difference between the temperature data collected by the two temperature sensors, and the more likely it is that the data is monitored under different smoldering conditions of the combustibles; at the same time, the smaller the spatial distance between the two temperature sensors, the lower the probability that the data fluctuation difference between the data collected by the two temperature sensors is caused by different positions, that is, The larger the value is, the greater the possibility that the ith temperature sensor will receive data fluctuations of combustible smoldering at the t moment; and the more obvious the temperature increase trend of the ith temperature sensor in the first t moments is, the larger the temperature data of the first M abnormal points of the ith temperature sensor is, and the more obvious the trend characteristics of data enhancement are; that is, The larger the value of , the greater the data fluctuation degree of the i-th temperature sensor at the t moment due to the smoldering phenomenon of the combustible.

[0069] 202. The correction coefficient of the fluctuation credibility of each temperature data is determined by analyzing the stability of the change of temperature difference in the building under the premise of increasing distance from small to large.

[0070] When combustibles smolder in a building, the temperature of the local space where the combustibles are located gradually rises, and heat energy radiates outward from the point where the smoldering is serious. Within the sensing range of the temperature sensor, the sensing range is the spatial range determined by the collection radius of the temperature sensor that can collect temperature data. For example, the sensing range of the temperature sensor is Meters, it means that the temperature sensor can collect temperature data in a spherical area with a radius of 10 meters and its installation location as the center. As the distance increases, the difference between the temperature data of the temperature sensor farther away from the point where the smoldering phenomenon is serious and the temperature data of the temperature sensor closer to the point where the smoldering phenomenon is serious becomes greater. The fluctuation credibility of the temperature data of each temperature sensor can be corrected according to the relationship between the difference in temperature data of any two temperature sensors within the sensing range and the distance between the two temperature sensors, so as to avoid internal interference of the sensor, such as interference of local current on the data fluctuation of the temperature data.

[0071] Specifically, the position of the cth temperature sensor is taken as the center of the sphere, the temperature data of all temperature sensors on the sphere are obtained according to the radius r, and the mean value of the temperature data of all temperature sensors on the sphere at the tth time is calculated, which is recorded as ; Then calculate the mean of the temperature data of the cth temperature sensor at the tth moment and the temperature data of all temperature sensors on the spherical surface The difference between the temperature data of the cth temperature sensor and the spherical surface at the tth time is expressed as , and obtain the corresponding differences under different radii in turn, where the difference represents the absolute value of the difference; the value range of radius r is , where T is the collection radius of the sensing range of the arranged temperature sensors. It should be noted that the sensing ranges of all temperature sensors in this embodiment are consistent.

[0072] For each data collection moment, the difference between the temperature data of the cth temperature sensor and the spherical surface is sorted in order of radius r from small to large, and the first-order difference sequence of the sorting result is used as the correction sequence of the cth temperature sensor at each data collection moment. If the combustible material that is close to the cth temperature sensor smolders at time t, the difference value should increase with the increase of radius r, and the elements in the correction sequence should all be greater than or equal to 0; if the temperature data collected by the cth temperature sensor is interfered by the internal interference of the sensor, there may be elements less than 0 in the correction sequence. Therefore, the more elements less than 0 in the correction sequence, the greater the impact of the internal interference of the sensor on the temperature data fluctuation, and the greater the degree of correction should be.

[0073] The correction coefficient of each temperature sensor at each data collection moment is determined based on the positive and negative values ​​of the elements in the correction sequence of each temperature sensor at each data collection moment:

[0074]

[0075] In the formula, is the correction coefficient of the cth temperature sensor at time t, is the number of elements less than 0 in the correction sequence of the cth temperature sensor at the tth time, is the number of elements in the correction sequence of the cth temperature sensor at time t.

[0076] 203 , using the correction coefficient to correct the fluctuation credibility of each temperature data, and obtain the position space characteristic value at each data collection moment.

[0077] In a building, the degree of temperature rise at the same smoldering time varies depending on the type of combustible material. Therefore, in this application, the ratio of the correction coefficient of each temperature sensor at different data collection times to the maximum value of the correction coefficient of each temperature sensor at all collection times is used as an evaluation of the degree to which the temperature data fluctuation at different data collection times should be corrected, the fluctuation credibility is corrected, and the position space characteristic value at each data collection time is obtained.

[0078] Specifically, the calculation formula for the position space eigenvalue of the cth temperature sensor at the tth time is as follows:

[0079]

[0080] In the formula, is the position spatial eigenvalue of the cth temperature sensor at time t, is the correction coefficient of the cth temperature sensor at time t, is the maximum value of the correction coefficient of the cth temperature sensor at all data collection moments, is the fluctuation credibility of the temperature data of the cth temperature sensor at the tth moment.

[0081] Among them, the more elements in the correction sequence corresponding to the cth temperature sensor at the tth time are less than 0, the more likely it is that the fluctuation of the temperature data at the tth time is caused by the internal interference of the sensor, and the greater the impact on the temperature data caused by the smoldering phenomenon of the combustibles, that is, The larger the value, the more reliable the fluctuation of temperature data. The greater the degree of correction.

[0082] The monitoring data processing module is used to complete the intelligent management of temperature data based on the position spatial characteristic value and temperature data of each temperature sensor at each data collection moment.

[0083] Inside a building, the smoldering of combustibles will not stop immediately after it occurs, that is, the temperature rise caused by the smoldering phenomenon will appear as continuous high-temperature data in the temperature data collected by the temperature sensor. Therefore, in this application, the time interval between any two adjacent abnormal points before each data collection moment is counted. If two adjacent abnormal points are caused by a smoldering phenomenon, then the temperature data between the two adjacent abnormal points are all high-temperature data, and the time interval is relatively large. At the same time, the larger the positional spatial characteristic value of the temperature sensor at a certain data collection moment, the greater the possibility that the temperature data fluctuation at that moment is caused by the smoldering phenomenon.

[0084] Specifically, the calculation formula for the correlation characteristic value between the temperature data of the cth temperature sensor at the tth moment and the early signs of fire is as follows:

[0085]

[0086] In the formula, is the correlation characteristic value between the temperature data of the cth temperature sensor at the tth moment and the early signs of fire, , are the mean and variance of the temperature data between all adjacent abnormal points of the cth temperature sensor before the tth time, is the minimum time interval between all adjacent abnormal points of the cth temperature sensor before the tth time. is the position space eigenvalue of the cth temperature sensor at the tth moment, is the minimum time interval between the tth moment and the corresponding moment of the abnormal point, is a parameter adjustment factor used to prevent the denominator from being 0. The empirical value of is 0.001.

[0087] Among them, the closer the c-th temperature sensor is to the corresponding time of the abnormal point at the tth moment, the greater the possibility that the tth moment is in the smoldering stage. The larger the mean of the temperature data between all adjacent abnormal points before the tth moment, the smaller the variance, which means that the temperature data between the abnormal points is more likely to be continuous high temperature data caused by the smoldering phenomenon; and the greater the possibility that the tth moment is in the smoldering stage, the greater the data fluctuation of the temperature data at the tth moment; that is, The larger the value of , the stronger the correlation between the temperature data of the cth temperature sensor at the tth moment and the ultra-early signs of fire.

[0088] Furthermore, based on the above process, the associated characteristic values ​​of the temperature data of each temperature sensor at all data collection moments and the super-early signs of fire are obtained respectively. And all the associated characteristic values ​​are used as input, and the associated characteristic values ​​of all the temperature data of each temperature sensor and the super-early signs of fire are divided into different clusters using the AP (Affinity Propagation Clustering) clustering algorithm, and the mean of all the associated characteristic values ​​in each cluster is calculated as the label value of each cluster, and the label values ​​of all clusters are arranged in order from large to small. Based on the arrangement results, the clusters are marked as the first monitoring danger level, the second monitoring danger level, and up to the Kth monitoring danger level, where K is the number of clusters, to complete the intelligent management of temperature data. The implementation process of this solution based on classification and labeling management of collected data is as follows: Figure 3 As shown in the figure, the AP clustering algorithm is a commonly used technology in the field of data processing, and the process will not be described in detail.

[0089] It should be noted that in this embodiment, only one clustering algorithm for the associated feature values ​​is provided, namely, the AP clustering algorithm. On the premise that the associated feature values ​​can be classified, in other embodiments, other data clustering algorithms may be selected, including but not limited to mean clustering and fuzzy mean clustering. This application does not impose any special restrictions on this.

[0090] It should be noted that the above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

[0091] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. An intelligent data management system for building fire safety monitoring, characterized in that: The system comprises: A data acquisition module, used to collect temperature data of each temperature sensor at all times; The spatial feature analysis module is used to determine the spatial feature value of the temperature data by analyzing the change trend of the data collected by the temperature sensors at adjacent locations and the change characteristics of the temperature data with the spatial distance, specifically: (1) Obtain the abnormal points of the temperature data, divide the temperature data into multiple local data sequences with each abnormal point as a segmentation point, and determine the feature array of the abnormal point based on all temperature values ​​in the local data sequence, the numerical value of the abnormal point, and the number of temperature data; obtain the fluctuation coefficient of each temperature data based on the distance between the feature arrays of the abnormal points of the same order before each temperature data of the temperature sensor and the distance difference between the temperature sensors; determine the fluctuation credibility of each temperature data based on the difference between the numerical value of the abnormal point before the temperature data and the numerical value of the temperature data and the fluctuation coefficient of the temperature data; (2) With a temperature sensor as the center, set several radii, sort the differences between the temperature of the temperature sensor and the temperatures of all temperature sensors within its radius in order from small to large, and then perform first-order differences to obtain a correction sequence; determine the correction coefficient of each temperature sensor at each data collection moment based on the difference between the number of negative numbers in the correction sequence and the total number; (3) At each acquisition moment, the fluctuation credibility of each temperature data is corrected by forward fusion of the fluctuation credibility according to the difference between the correction coefficient and the maximum value of the correction coefficient, and the position space characteristic value at each data acquisition moment is obtained; The monitoring data processing module obtains the associated characteristic values ​​based on the mean and variance of the temperature of all abnormal points before each collection moment, the minimum time interval between adjacent abnormal points, and the spatial characteristic values ​​of the location at the collection moment; all the associated characteristic values ​​are classified into classes with different danger levels to complete the intelligent management of temperature data; The method for obtaining the associated characteristic value includes: In the formula, is the correlation characteristic value between the temperature data of the cth temperature sensor at the tth moment and the early signs of fire, , are the mean and variance of the temperature data between all adjacent abnormal points of the cth temperature sensor before the tth time, is the minimum time interval between all adjacent abnormal points of the cth temperature sensor before the tth time. is the position space eigenvalue of the cth temperature sensor at the tth moment, is the minimum time interval between the tth moment and the corresponding moment of the abnormal point, is the tuning factor.

2. The intelligent data management system for building fire safety monitoring according to claim 1 is characterized in that: The determining the fluctuation credibility of each temperature data based on the difference of the temperature data between the temperature sensors comprises: Determine the fluctuation coefficient of each temperature data collected by each temperature sensor based on the difference of the characteristic arrays between the abnormal points in the temperature data collected by different temperature sensors; The fluctuation credibility of each temperature data is determined based on the overall fluctuation degree of the temperature data of each temperature sensor before each data collection moment and the fluctuation coefficient.

3. The intelligent data management system for building fire safety monitoring according to claim 2 is characterized in that: The method for obtaining the fluctuation coefficient of each temperature data includes: The RRCF algorithm is used to detect abnormal points in the temperature data collected by each temperature sensor, and the feature array of each abnormal point is obtained; Calculate the ratio of the DTW distance between each temperature sensor and any other temperature sensor's feature array of any abnormal point of the same order before each moment to the position distance of the two temperature sensors; The cumulative result of the ratio at all abnormal points of all temperature sensors in the building before each moment is taken as the fluctuation coefficient of each temperature data collected by each sensor.

4. The intelligent data management system for building fire safety monitoring according to claim 2 is characterized in that: The method for obtaining the fluctuation credibility of each temperature data includes: Calculate the difference between the mean of all abnormal points before the corresponding collection time of each temperature data collected by each temperature sensor and the mean of all temperature data before the corresponding collection time; The normalized result of the product of the difference and the fluctuation coefficient of each temperature data is used as the fluctuation credibility of each temperature data.

5. The intelligent data management system for building fire safety monitoring according to claim 3 is characterized in that: The method for obtaining the feature array includes: The RRCF algorithm is used to detect abnormal points in the temperature data collected by each temperature sensor; Each abnormal point is used as a segmentation point to divide all temperature data into multiple local data sequences. The variance of all temperature data in a local data sequence adjacent to each abnormal point is calculated as the fluctuation radius of each abnormal point. The array consisting of the value of each abnormal point, the number of temperature data before each abnormal point and the fluctuation radius of each abnormal point is used as the feature array of each abnormal point.

6. The intelligent data management system for building fire safety monitoring according to claim 1 is characterized in that: The method for obtaining the correction coefficient of each temperature sensor at each data collection moment includes: The ratio of the number of elements less than 0 in the correction sequence of each temperature sensor at each data collection moment to the number of elements in the correction sequence is used as the correction coefficient of each temperature sensor at each data collection moment.

7. The intelligent data management system for building fire safety monitoring according to claim 6, characterized in that: The method for obtaining the correction sequence includes: For any data collection moment, taking the position of each temperature sensor as the center of the sphere, obtaining the temperature data of all temperature sensors on the spherical surface at each data collection moment according to the radius r, and calculating the mean of the temperature data of all temperature sensors on the spherical surface at each data collection moment; taking the absolute value of the difference between the temperature data of each temperature sensor at each data collection moment and the mean as the single distance difference of each temperature sensor at each data collection moment; The single distance difference of each temperature sensor at each data collection time under different radii is obtained in turn, all the single distance differences are sorted in order from small to large radius, and the first-order difference sequence of the sorting result is used as the correction sequence of each temperature sensor at each data collection time.

8. The intelligent data management system for building fire safety monitoring according to claim 1 is characterized in that: The method for obtaining the position space characteristic value at each data collection moment includes: Calculate the ratio of the correction coefficient of each temperature sensor at each data collection moment to the maximum value of the correction coefficient of each temperature sensor among all data collection moments; The product of the ratio and the fluctuation credibility of the temperature data collected by each temperature sensor at each data collection time is used as the position space characteristic value at each data collection time.

9. The intelligent data management system for building fire safety monitoring according to claim 1, characterized in that: The method for completing intelligent management of temperature data based on the positional spatial characteristic value and temperature data of each temperature sensor at each data collection moment includes: Determine the correlation characteristic value between the temperature data collected by each temperature sensor at each data collection moment and the super-early signs of fire based on the abnormal point of each temperature sensor before each data collection moment and the position spatial characteristic value of each temperature sensor at each data collection moment; Obtaining correlation feature values ​​between the temperature data collected at all data collection moments of each temperature sensor and the early signs of fire, and dividing all the correlation feature values ​​into multiple clusters using a data clustering algorithm; Calculate the mean of all associated feature values ​​in each cluster as the label value of each cluster, arrange the label values ​​of all clusters in order from large to small, and mark the clusters as the first monitoring danger level, the second monitoring danger level, and up to the Kth monitoring danger level based on the arrangement results; where K is the number of clusters.

Images