Automatic alarm method and system based on TVOC monitoring data

Abstract

The invention provides an automatic alarm method and system based on TVOC monitoring data, and relates to the technical field of data monitoring, and the method mainly comprises the steps: carrying out the time sequence modeling through the historical station monitoring data of a fixed station, and obtaining a long-term STL time sequence model and a short-term STL time sequence model of different time spans; parameter curves of a long-term STL time sequence model and a short-term STL time sequence model are analyzed, and adaptive reference parameters are selected as reference data for alarm judgment, so that the system can send out corresponding alarm information according to different judgment results, automatic alarm of TVOC monitoring data is realized, and in addition, the TVOC monitoring data can be automatically monitored. And different alarm information can be sent based on the reference parameters, so that the accuracy of the alarm information is ensured, the monitoring data is analyzed, and intelligentization is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of data monitoring, and in particular to an automatic alarm method and system based on TVOC monitoring data. Background Art

[0002] TVOC monitoring stations can quickly and accurately detect the types and concentrations of VOCs in the environment. The types and concentrations of VOCs obtained through monitoring can help to promptly discover and prevent environmental pollution problems, and provide data support for the effective control of VOC pollution. However, ordinary VOC monitoring instruments are easily affected by temperature, humidity and other gases, and the test results are easily disturbed. It is difficult to directly determine whether the high values ​​are caused by local pollution or equipment monitoring abnormalities. In addition, the amount of monitoring data is huge, and manual analysis is time-consuming and labor-intensive and lacks scientific basis. Summary of the invention

[0003] The purpose of the present invention is to provide an automatic alarm method and system based on TVOC monitoring data. The method mainly performs time series modeling through historical site monitoring data of fixed sites to obtain long-term STL time series models and short-term STL time series models with different time spans, analyzes the parameter curves of the long-term STL time series model and the short-term STL time series model, and selects suitable benchmark parameters as benchmark data for alarm judgment, so that the system can issue corresponding alarm information according to different judgment results, realize automatic alarm of TVOC monitoring data, and different alarm information can be issued based on the benchmark parameters, which not only ensures the accuracy of the alarm information, but also analyzes the monitoring data to realize intelligence.

[0004] In order to solve the above technical problems, the present invention adopts the following solutions: An automatic alarm method based on TVOC monitoring data, the method specifically comprises the following steps: S1: Obtain current site monitoring data of fixed sites; S2: obtaining a current long-term STL time series model and a current short-term STL time series model of a fixed site, wherein the long-term STL time series model and the short-term STL time series model are time series models with different time spans, and the time series model includes a trend curve, a seasonal curve, and a residual curve; S3: Extract parameters of the trend curve in the current long-term STL time series model and the trend curve, seasonal curve, and residual curve in the current short-term STL time series model to obtain current benchmark parameters; S4: Make an alarm determination based on the TVOC concentration value in the current site monitoring data and the current benchmark parameters, and issue a corresponding alarm message based on the determination result.

[0005] Furthermore, in S2, the process of obtaining the current long-term STL time series model and the current short-term STL time series model of the fixed site is specifically as follows: At the current time point, the TVOC concentration values ​​in the historical site monitoring data of the fixed site are modeled for time series with different time spans, and are decomposed into trend, seasonality, and residual under different time spans to obtain corresponding trend curves, seasonal curves, and residual curves. The trend curves, seasonal curves, and residual curves all take the time point as the X-axis and the TVOC concentration value as the Y-axis.

[0006] Furthermore, in S3, the process of extracting parameters of the trend curve in the current long-term STL time series model and the trend curve, seasonal curve, and residual curve in the current short-term STL time series model to obtain the current benchmark parameters is as follows: Parameter extraction is performed on the Y-axis of the trend curve in the current long-term STL time series model, and the TVOC concentration value at the preset quantile is extracted and used as the long-term trend benchmark parameter; Parameter extraction is performed on the Y-axis of the trend curve in the current short-term STL time series model, and the TVOC concentration value at the preset quantile is extracted and used as the first short-term trend benchmark parameter; Parameter extraction is performed on the X-axis of the trend curve in the current short-term STL time series model, and the time point at the last position is extracted and used as the second short-term trend benchmark parameter; Parameter extraction is performed on the X-axis of the seasonal curve in the current short-term STL time series model, and the time point at the last position is extracted and used as the short-term seasonal benchmark parameter; Parameter extraction is performed on the Y-axis of the residual curve in the current short-term STL time series model, and the TVOC concentration value at the preset quantile is extracted and used as the short-term residual benchmark parameter; Finally, the extracted long-term trend benchmark parameter, the first short-term trend benchmark parameter, the second short-term trend benchmark parameter, the short-term seasonal benchmark parameter, and the short-term residual benchmark parameter are used as the current benchmark parameters.

[0007] Furthermore, the S4 specifically includes the following steps: S41: Perform a first alarm determination based on the TVOC concentration value in the current site monitoring data and the second short-term trend benchmark parameter, the short-term seasonal benchmark parameter, and the short-term residual benchmark parameter. If the current TVOC concentration value meets the first determination condition, a corresponding first alarm message is issued; S42: Perform a second alarm determination based on the TVOC concentration value in the current site monitoring data and the long-term trend benchmark parameter, and issue a corresponding second alarm message if the current TVOC concentration value meets the second determination condition; S43: Perform a third alarm determination based on the TVOC concentration value in the current site monitoring data and the first short-term trend benchmark parameter, and issue a corresponding third alarm message if the current TVOC concentration value meets the third determination condition.

[0008] Furthermore, the first judgment condition is that the value obtained by subtracting the second short-term trend reference parameter and the short-term seasonal reference parameter in sequence from the current TVOC concentration value is greater than the short-term residual reference parameter.

[0009] Furthermore, the second judgment condition is that: the current TVOC concentration value is greater than the long-term trend reference parameter and the current fixed site has satisfied the TVOC concentration value greater than the long-term trend reference parameter for a continuous preset time period.

[0010] Furthermore, the third determination condition is that: the current TVOC concentration value is greater than the short-term trend reference parameter and the current fixed site has satisfied the TVOC concentration value greater than the short-term trend reference parameter for a continuous preset time period.

[0011] Furthermore, the S4 further includes the following steps: S44: Determine whether the current TVOC concentration value satisfies the first determination condition, the second determination condition and the third determination condition. If so, compare the short-term residual benchmark parameter, the long-term trend benchmark parameter and the short-term trend benchmark parameter to obtain the benchmark parameter with the largest value, and select to issue the alarm information corresponding to the benchmark parameter with the largest value, that is, when the short-term residual benchmark parameter is the benchmark parameter with the largest value, issue the first alarm information; when the long-term trend benchmark parameter is the benchmark parameter with the largest value, issue the second alarm information; when the short-term trend benchmark parameter is the benchmark parameter with the largest value, issue the third alarm information.

[0012] Furthermore, the time spans of the trend curve, seasonal curve, and residual curve on the X-axis in the long-term STL time series model are the same, the time spans of the trend curve, seasonal curve, and residual curve on the X-axis in the short-term STL time series model are the same, and the time spans of the trend curve, seasonal curve, and residual curve on the X-axis in the long-term STL time series model are greater than the time spans of the trend curve, seasonal curve, and residual curve on the X-axis in the short-term STL time series model.

[0013] An automatic alarm system based on TVOC monitoring data, comprising: Data acquisition module: used to obtain site monitoring data of fixed sites; Time series modeling module: used to obtain the current long-term STL time series model and the current short-term STL time series model of the fixed site, wherein the long-term STL time series model and the short-term STL time series model are time series models with different time spans, and the time series model includes a trend curve, a seasonal curve, and a residual curve; Benchmark parameter extraction module: used to extract parameters of the trend curve in the current long-term STL time series model and the trend curve, seasonal curve, and residual curve in the current short-term STL time series model to obtain the current benchmark parameters; Alarm module: used to make alarm decisions based on the TVOC concentration value and current benchmark parameters in the current site monitoring data, and issue corresponding alarm information based on the decision results.

[0014] Beneficial effects of the present invention: The present invention provides an automatic alarm method and system based on TVOC monitoring data. The method mainly performs time series modeling through historical site monitoring data of fixed sites to obtain long-term STL time series models and short-term STL time series models with different time spans, analyzes parameter curves of the long-term STL time series model and the short-term STL time series model, selects suitable benchmark parameters as benchmark data for alarm judgment, so that the system can issue corresponding alarm information according to different judgment results, realize automatic alarm of TVOC monitoring data, and different alarm information can be issued based on the benchmark parameters, which not only ensures the accuracy of the alarm information, but also analyzes the monitoring data to realize intelligence. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of an automatic alarm method based on TVOC monitoring data in Example 1 of the present invention.

[0016] Figure 2 Schematic diagram of the main sequence in the long-term STL time series model in Example 1 of the present invention.

[0017] Figure 3 Schematic diagram of the trend curve in the long-term STL time series model in Example 1 of the present invention.

[0018] Figure 4 Schematic diagram of seasonal curves in the long-term STL time series model in Example 1 of the present invention.

[0019] Figure 5 Schematic diagram of the residual curve in the long-term STL time series model in Example 1 of the present invention.

[0020] Figure 6 This is a display diagram of the monitoring platform in Example 1 of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. The following description of at least one exemplary embodiment is actually only illustrative and is by no means intended to limit the present invention and its application or use. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] The relative arrangement of components and steps, the numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present invention unless specifically stated otherwise.

[0023] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0024] Additionally, descriptions of well-known structures, functions, and configurations may be omitted for clarity and conciseness.One of ordinary skill in the art will recognize that various changes and modifications may be made to the examples described herein without departing from the spirit and scope of the present disclosure.

[0025] Technologies, methods, and apparatus known to ordinary technicians in the relevant field may not be discussed in detail, but where appropriate, such technologies, methods, and apparatus should be considered part of the authorization specification.

[0026] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0027] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments: Example 1 like Figure 1 As shown, an automatic alarm method based on TVOC monitoring data, the method specifically comprises the following steps: S1: Obtain current site monitoring data of fixed sites; S2: obtaining a current long-term STL time series model and a current short-term STL time series model of a fixed site, wherein the long-term STL time series model and the short-term STL time series model are time series models with different time spans, and the time series model includes a trend curve, a seasonal curve, and a residual curve; S3: Extract parameters of the trend curve in the current long-term STL time series model and the trend curve, seasonal curve, and residual curve in the current short-term STL time series model to obtain current benchmark parameters; S4: Make an alarm determination based on the TVOC concentration value in the current site monitoring data and the current benchmark parameters, and issue a corresponding alarm message based on the determination result.

[0028] Preferably, in S2, the process of obtaining the current long-term STL time series model and the current short-term STL time series model of the fixed site is specifically as follows: At the current time point, the TVOC concentration values ​​in the historical site monitoring data of the fixed site are modeled for time series with different time spans, and are decomposed into trend, seasonality, and residual under different time spans to obtain corresponding trend curves, seasonal curves, and residual curves. The trend curves, seasonal curves, and residual curves all take the time point as the X-axis and the TVOC concentration value as the Y-axis.

[0029] Preferably, in S3, the process of extracting parameters of the trend curve in the current long-term STL time series model and the trend curve, seasonal curve, and residual curve in the current short-term STL time series model to obtain the current benchmark parameters is specifically as follows: Parameter extraction is performed on the Y-axis of the trend curve in the current long-term STL time series model, and the TVOC concentration value at the preset quantile is extracted and used as the long-term trend benchmark parameter; Parameter extraction is performed on the Y-axis of the trend curve in the current short-term STL time series model, and the TVOC concentration value at the preset quantile is extracted and used as the first short-term trend benchmark parameter; Parameter extraction is performed on the X-axis of the trend curve in the current short-term STL time series model, and the time point at the last position is extracted and used as the second short-term trend benchmark parameter; Parameter extraction is performed on the X-axis of the seasonal curve in the current short-term STL time series model, and the time point at the last position is extracted and used as the short-term seasonal benchmark parameter; Parameter extraction is performed on the Y-axis of the residual curve in the current short-term STL time series model, and the TVOC concentration value at the preset quantile is extracted and used as the short-term residual benchmark parameter; Finally, the extracted long-term trend benchmark parameter, the first short-term trend benchmark parameter, the second short-term trend benchmark parameter, the short-term seasonal benchmark parameter, and the short-term residual benchmark parameter are used as the current benchmark parameters.

[0030] Preferably, the S4 specifically includes the following steps: S41: Perform a first alarm determination based on the TVOC concentration value in the current site monitoring data and the second short-term trend benchmark parameter, the short-term seasonal benchmark parameter, and the short-term residual benchmark parameter. If the current TVOC concentration value meets the first determination condition, a corresponding first alarm message is issued; S42: Perform a second alarm determination based on the TVOC concentration value in the current site monitoring data and the long-term trend benchmark parameter, and issue a corresponding second alarm message if the current TVOC concentration value meets the second determination condition; S43: Perform a third alarm determination based on the TVOC concentration value in the current site monitoring data and the first short-term trend benchmark parameter, and issue a corresponding third alarm message if the current TVOC concentration value meets the third determination condition.

[0031] Preferably, the first judgment condition is that the value obtained by successively subtracting the second short-term trend reference parameter and the short-term seasonal reference parameter from the current TVOC concentration value is greater than the short-term residual reference parameter.

[0032] Preferably, the second judgment condition is: the current TVOC concentration value is greater than the long-term trend reference parameter and the current fixed site has satisfied the TVOC concentration value greater than the long-term trend reference parameter for a continuous preset time period.

[0033] Preferably, the third determination condition is: the current TVOC concentration value is greater than the short-term trend reference parameter and the current fixed site has satisfied the TVOC concentration value greater than the short-term trend reference parameter for a continuous preset time period.

[0034] Furthermore, the S4 further includes the following steps: S44: Determine whether the current TVOC concentration value satisfies the first determination condition, the second determination condition and the third determination condition. If so, compare the short-term residual benchmark parameter, the long-term trend benchmark parameter and the short-term trend benchmark parameter to obtain the benchmark parameter with the largest value, and select to issue the alarm information corresponding to the benchmark parameter with the largest value, that is, when the short-term residual benchmark parameter is the benchmark parameter with the largest value, issue the first alarm information; when the long-term trend benchmark parameter is the benchmark parameter with the largest value, issue the second alarm information; when the short-term trend benchmark parameter is the benchmark parameter with the largest value, issue the third alarm information.

[0035] Preferably, the time spans of the trend curve, seasonal curve, and residual curve on the X-axis in the long-term STL time series model are the same, the time spans of the trend curve, seasonal curve, and residual curve on the X-axis in the short-term STL time series model are the same, and the time spans of the trend curve, seasonal curve, and residual curve on the X-axis in the long-term STL time series model are greater than the time spans of the trend curve, seasonal curve, and residual curve on the X-axis in the short-term STL time series model.

[0036] Based on the above principles, the present invention is further described: In the present invention, the TVOC monitoring data refers to the TVOC data obtained by monitoring at a fixed site through a VOC monitoring instrument. The TVOC data mainly reflects the indicators of volatile organic gases in the air through TVOC concentration values, including harmful substances such as formaldehyde, benzene, toluene, xylene, styrene, etc. Therefore, the smaller the TVOC concentration value, the better the air quality, and vice versa.

[0037] At present, ordinary VOC monitoring instruments are easily affected by temperature, humidity and other gases, which makes the detection results easily interfered. It is difficult to directly judge whether the higher TVOC concentration values ​​obtained by monitoring are caused by local pollution or equipment monitoring abnormalities. In addition, the monitoring data is huge in volume, and manual analysis is time-consuming and labor-intensive and lacks scientific basis.

[0038] Therefore, the present invention provides an automatic alarm method based on TVOC monitoring data. The method mainly performs time series modeling according to different time spans, and can perform training and analysis on the monitoring data of each fixed site. Thus, the current TVOC concentration value obtained by monitoring can be automatically determined based on the time series models of different time spans, and different alarm information can be issued according to different determination results. This not only ensures the accuracy of the alarm, but also can issue different alarm information, so that the monitoring personnel can effectively manage the subsequent work according to the alarm information.

[0039] Specifically, in this embodiment, it is first necessary to collect site monitoring data for training, then historical site monitoring data can be collected, and the site monitoring data of the previous year can be selected to be collected, and then the historical site monitoring data are grouped according to the site number and the TVOC concentration values ​​in each group of site monitoring data are sorted according to time. The sorted TVOC concentration values ​​form a TVOC concentration value-time series, which is convenient for subsequent modeling of the TVOC concentration value-time series.

[0040] Then, for each fixed site, the sorted TVOC concentration values ​​are subjected to the time series model STL. STL is a method commonly used in time series analysis. It mainly decomposes the time series into three main components: trend, seasonality, and residual. The trend refers to the overall movement of the time series, which may continue to increase, decrease, or fluctuate; the seasonality refers to any regular seasonal pattern in the time series, for example, ice cream sales in summer are usually higher than in winter; the residual can also be considered as statistical noise, or the impact of temporary events. This residual also has a separate periodic component, but it is usually classified as a trend component.

[0041] Based on the above-mentioned trend, seasonality and residual, the TVOC concentration value-time series is decomposed and modeled, and the relationship between TVOC concentration value and time can be analyzed according to different trends, seasonality and residuals.

[0042] In order to improve the accuracy of the alarm, the present invention also takes into account long-term and short-term factors, and trains the TVOC concentration value-time series separately under different time spans to obtain a long-term STL time series model and a short-term STL time series model. The long-term STL time series model and the short-term STL time series model are both time series models, and decomposition modeling is performed based on trend, seasonality, and residual, so that the long-term STL time series model and the short-term STL time series model both include trend curves, seasonal curves, and residual curves, and the relationship between TVOC concentration values ​​and time is reflected through different parameter curves.

[0043] The trend curve, seasonal curve, and residual curve in the long-term STL time series model and the short-term STL time series model all use the time point as the X-axis and the TVOC concentration value as the Y-axis. However, the time span of the trend curve, seasonal curve, and residual curve in the long-term STL time series model on the X-axis is greater than the time span of the trend curve, seasonal curve, and residual curve in the short-term STL time series model on the X-axis. Specifically, the time span of the long-term STL time series model can be set to 365 days, and the time span of the short-term STL time series model can be set to 30 days. Thus, the long-term STL time series model and the short-term STL time series model can be used to avoid underreporting and misreporting.

[0044] In this embodiment, taking the structure of the long-term STL time series model as an example, the main sequence in the long-term STL time series model, such as Figure 2 As shown, that is, TVOC concentration value-time series; then based on the above trend, seasonality, and residual characteristics, the TVOC concentration value-time series is decomposed and modeled to obtain the trend curve in the long-term STL time series model, as shown in Figure 3 As shown, that is, the trend of the main sequence; the seasonal curve in the long-term STL time series model is obtained, such as Figure 4 As shown, that is, the seasonality of the main series; the residual curve in the long-term STL time series model is obtained, as shown Figure 5 As shown, that is, the residual of the main series; it can be seen that each curve is a digital model with time as the X-axis and TVOC concentration value as the Y-axis, and the short-term STL time series model is structurally the same as the long-term STL time series model, only the time span on the X-axis is different, which will not be repeated here.

[0045] Based on the above long-term STL time series model and short-term STL time series model, the analysis is carried out. For the structure of the long-term STL time series model, the trend term can be represented by long_trend; for the structure of the short-term STL time series model, the trend term can be represented by short_trend, the seasonal term can be represented by short_seasonal, and the residual term can be represented by short_residual; all the above data are time series data.

[0046] Specifically, for each fixed site, a model can be built every hour, that is, the long-term STL time series model and the short-term STL time series model are updated regularly, and parameters such as long_trend, short_trend, short_seasonal, and short_residual can be obtained every hour. Then, the data is further processed, that is, the trend curve in the current long-term STL time series model and the trend curve, seasonal curve, and residual curve in the current short-term STL time series model are extracted. Specifically: The 95% quantile of long_trend is recorded as long_thre, that is, the TVOC concentration value at the 95% quantile on the Y axis of the trend curve in the long-term STL time series model is used as the long-term trend benchmark parameter; The 95% quantile of short_trend is recorded as short_thre, that is, the TVOC concentration value at the 95% quantile on the Y axis of the trend curve in the short-term STL time series model is taken as the first short-term trend benchmark parameter; The last digit of short_trend is recorded as last_trend, that is, the last time point on the X-axis of the trend curve in the short-term STL time series model is taken as the second short-term trend benchmark parameter; The last digit of short_seasonal is recorded as last_seasonal, that is, the last time point on the X-axis of the seasonal curve in the short-term STL time series model is used as the short-term seasonal benchmark parameter; The 95% quantile of short_residual is recorded as residual_thre, that is, the TVOC concentration value at the 95% quantile on the Y-axis of the residual curve in the short-term STL time series model is used as the short-term residual benchmark parameter.

[0047] The 95% percentile refers to the TVOC concentration value at the 95% percentile on the Y-axis found by searching on the Y-axis in sequence; the last digit refers to the time value at the last digit found by searching on the X-axis. In addition, the 95% percentile is a number at a preset percentile, which can be set by the staff according to the actual situation.

[0048] Through the above, the analysis of the time series model at the current time point can be completed. Then, when the current site monitoring data at the current time point is obtained through the current fixed site, the current site monitoring data can be analyzed according to different benchmark parameters at the current time point, thereby determining whether to issue a corresponding alarm message.

[0049] Specifically, the current site monitoring data at the current time point is recorded, that is, the current TVOC concentration value is recorded as monitor_data, and then the current TVOC concentration value is judged according to different first judgment conditions, second judgment conditions, and third judgment conditions. Specifically: 1. If monitor_data - last_trend - last_seasonal>residual_thre, it means that the current TVOC concentration value meets the first judgment condition, then a corresponding first alarm message can be issued, and the first alarm message is recorded as suspected local pollution near the station; 2. If monitor_data>long_thre and this condition is met for three consecutive hours, it means that the current TVOC concentration value meets the second judgment condition, and the corresponding second alarm information can be issued. The second alarm information is recorded as the site monitoring value is greater than the long-term trend; 3. If monitor_data>short_thre and this condition is met for three consecutive hours, it means that the current TVOC concentration value meets the third judgment condition, and the corresponding third alarm information can be issued, and the third alarm information is recorded as the site monitoring value is greater than the short-term trend; 4. If the first, second and third judgment conditions mentioned above are met at the same time, the alarm information corresponding to max(residual_thre, long_thre, short_thre) is selected.

[0050] In summary, the automatic alarm method based on TVOC monitoring data provided by the present invention can realize automatic alarm of TVOC monitoring data, which can be displayed on the monitoring platform, such as Figure 6 As shown, Figure 6There are three kinds of curves, namely, curves with hollow circle marks, curves with triangle marks, and curves with square marks. Among them, the curve with hollow circle marks represents the change of TVOC concentration value on the time axis, and the curve with hollow circle marks is marked with filled circular icons, and the filled circular icons represent alarm events, that is, the TVOC monitoring data is automatically alarmed at the corresponding time, and the curve with triangle marks represents the long-term STL time series model, and the curve with square marks represents the short-term STL time series model. The TVOC monitoring data can be monitored through the time series, which greatly reduces the time and labor costs of manual alarms and increases the scientific basis for alarms. In addition, multi-model fusion utilizes the characteristics of the STL model and considers long-term and short-term factors at the same time, which greatly reduces the probability of false alarms and missed alarms.

[0051] Example 2 An automatic alarm system based on TVOC monitoring data, comprising: Data acquisition module: used to obtain site monitoring data of fixed sites; Time series modeling module: used to obtain the current long-term STL time series model and the current short-term STL time series model of the fixed site, wherein the long-term STL time series model and the short-term STL time series model are time series models with different time spans, and the time series model includes a trend curve, a seasonal curve, and a residual curve; Benchmark parameter extraction module: used to extract parameters of the trend curve in the current long-term STL time series model and the trend curve, seasonal curve, and residual curve in the current short-term STL time series model to obtain the current benchmark parameters; Alarm module: used to make alarm decisions based on the TVOC concentration value and current benchmark parameters in the current site monitoring data, and issue corresponding alarm information based on the decision results.

[0052] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. According to the technical essence of the present invention, within the spirit and principles of the present invention, any simple modification, equivalent replacement and improvement made to the above embodiment still falls within the protection scope of the technical solution of the present invention.

Claims

1. An automatic alarm method based on TVOC monitoring data, characterized in that: The method specifically comprises the following steps: S1: Obtain current site monitoring data of fixed sites; S2: obtaining a current long-term STL time series model and a current short-term STL time series model of a fixed site, wherein the long-term STL time series model and the short-term STL time series model are time series models with different time spans, and the time series model includes a trend curve, a seasonal curve, and a residual curve; S3: Extract parameters of the trend curve in the current long-term STL time series model and the trend curve, seasonal curve, and residual curve in the current short-term STL time series model to obtain current benchmark parameters; S4: Make an alarm determination based on the TVOC concentration value in the current site monitoring data and the current benchmark parameters, and issue a corresponding alarm message based on the determination result.

2. The automatic alarm method based on TVOC monitoring data according to claim 1 is characterized in that: In S2, the process of obtaining the current long-term STL time series model and the current short-term STL time series model of the fixed site is as follows: At the current time point, the TVOC concentration values ​​in the historical site monitoring data of the fixed site are modeled for time series with different time spans, and are decomposed into trend, seasonality, and residual under different time spans to obtain corresponding trend curves, seasonal curves, and residual curves. The trend curves, seasonal curves, and residual curves all take the time point as the X-axis and the TVOC concentration value as the Y-axis.

3. The automatic alarm method based on TVOC monitoring data according to claim 2 is characterized in that: In S3, the process of extracting parameters of the trend curve in the current long-term STL time series model and the trend curve, seasonal curve, and residual curve in the current short-term STL time series model to obtain the current benchmark parameters is as follows: Parameter extraction is performed on the Y-axis of the trend curve in the current long-term STL time series model, and the TVOC concentration value at the preset quantile is extracted and used as the long-term trend benchmark parameter; Parameter extraction is performed on the Y-axis of the trend curve in the current short-term STL time series model, and the TVOC concentration value at the preset quantile is extracted and used as the first short-term trend benchmark parameter; Parameter extraction is performed on the X-axis of the trend curve in the current short-term STL time series model, and the time point at the last position is extracted and used as the second short-term trend benchmark parameter; Parameter extraction is performed on the X-axis of the seasonal curve in the current short-term STL time series model, and the time point at the last position is extracted and used as the short-term seasonal benchmark parameter; Parameter extraction is performed on the Y-axis of the residual curve in the current short-term STL time series model, and the TVOC concentration value at the preset quantile is extracted and used as the short-term residual benchmark parameter; Finally, the extracted long-term trend benchmark parameter, the first short-term trend benchmark parameter, the second short-term trend benchmark parameter, the short-term seasonal benchmark parameter, and the short-term residual benchmark parameter are used as the current benchmark parameters.

4. The automatic alarm method based on TVOC monitoring data according to claim 3 is characterized in that: The S4 specifically includes the following steps: S41: Perform a first alarm determination based on the TVOC concentration value in the current site monitoring data and the second short-term trend benchmark parameter, the short-term seasonal benchmark parameter, and the short-term residual benchmark parameter. If the current TVOC concentration value meets the first determination condition, a corresponding first alarm message is issued; S42: Perform a second alarm determination based on the TVOC concentration value in the current site monitoring data and the long-term trend benchmark parameter, and issue a corresponding second alarm message if the current TVOC concentration value meets the second determination condition; S43: Perform a third alarm determination based on the TVOC concentration value in the current site monitoring data and the first short-term trend benchmark parameter, and issue a corresponding third alarm message if the current TVOC concentration value meets the third determination condition.

5. The automatic alarm method based on TVOC monitoring data according to claim 4 is characterized in that: The first judgment condition is that the value obtained by subtracting the second short-term trend reference parameter and the short-term seasonal reference parameter from the current TVOC concentration value is greater than the short-term residual reference parameter.

6. The automatic alarm method based on TVOC monitoring data according to claim 4 is characterized in that: The second determination condition is that: the current TVOC concentration value is greater than the long-term trend reference parameter and the current fixed site has satisfied the TVOC concentration value greater than the long-term trend reference parameter for a continuous preset time period.

7. The automatic alarm method based on TVOC monitoring data according to claim 4 is characterized in that: The third determination condition is that the current TVOC concentration value is greater than the short-term trend reference parameter and the current fixed site has satisfied the TVOC concentration value greater than the short-term trend reference parameter for a continuous preset time period.

8. The automatic alarm method based on TVOC monitoring data according to claim 4 is characterized in that: The S4 also includes the following steps: S44: Determine whether the current TVOC concentration value satisfies the first determination condition, the second determination condition and the third determination condition. If so, compare the short-term residual benchmark parameter, the long-term trend benchmark parameter and the short-term trend benchmark parameter to obtain the benchmark parameter with the largest value, and select to issue the alarm information corresponding to the benchmark parameter with the largest value, that is, when the short-term residual benchmark parameter is the benchmark parameter with the largest value, issue the first alarm information; when the long-term trend benchmark parameter is the benchmark parameter with the largest value, issue the second alarm information; when the short-term trend benchmark parameter is the benchmark parameter with the largest value, issue the third alarm information.

9. The automatic alarm method based on TVOC monitoring data according to claim 2 is characterized in that: The time spans of the trend curve, seasonal curve and residual curve on the X-axis in the long-term STL time series model are the same, and the time spans of the trend curve, seasonal curve and residual curve on the X-axis in the short-term STL time series model are the same, and the time spans of the trend curve, seasonal curve and residual curve on the X-axis in the long-term STL time series model are greater than the time spans of the trend curve, seasonal curve and residual curve on the X-axis in the short-term STL time series model.

10. An automatic alarm system based on TVOC monitoring data, characterized in that: include: Data acquisition module: used to obtain site monitoring data of fixed sites; Time series modeling module: used to obtain the current long-term STL time series model and the current short-term STL time series model of the fixed site, wherein the long-term STL time series model and the short-term STL time series model are time series models with different time spans, and the time series model includes a trend curve, a seasonal curve, and a residual curve; Benchmark parameter extraction module: used to extract parameters of the trend curve in the current long-term STL time series model and the trend curve, seasonal curve, and residual curve in the current short-term STL time series model to obtain the current benchmark parameters; Alarm module: used to make alarm decisions based on the TVOC concentration value and current benchmark parameters in the current site monitoring data, and issue corresponding alarm information based on the decision results.

Images