A stationary pollution source monitoring data analysis method and system and a storage medium

Abstract

The application relates to the technical field of data analysis, and discloses a fixed pollution source monitoring data analysis method and system and a storage medium. The method comprises the following steps: collecting fixed pollution source monitoring data based on a monitoring device, calculating the proximity values between different monitoring data, classifying the monitoring data into several categories based on the proximity values, obtaining first monitoring data under a normal state, obtaining corresponding category monitoring data based on the classification result, taking the category monitoring data collected at each time point as target monitoring data, selecting several category monitoring data closest to the target monitoring data collection time point as first data, calculating the abnormal threshold value of each category based on the first data and the target monitoring data, obtaining real-time monitoring data, calculating the deviation value of the corresponding real-time monitoring data and the corresponding first data in each category of the real-time monitoring data, and judging whether the current monitoring data is abnormal based on the deviation value and the corresponding abnormal threshold value. The application can improve the sensitivity of data analysis.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data analysis, in particular to a fixed pollution source monitoring data analysis method and system and a storage medium. BACKGROUND

[0002] The monitoring data of fixed pollution sources, such as factory emissions, sewage treatment plants, etc., usually includes the concentration, emission amount, etc. of various pollutants. These data may be affected by factors such as equipment failure, operation food, environmental changes, etc., resulting in abnormal emissions. Therefore, timely and accurate monitoring and identification of abnormal emissions are crucial for environmental protection and enterprise operation.

[0003] A similar prior art is Chinese patent application No. CN115901550A, which provides a pollution source monitoring and analysis system and method based on the Internet of Things, including a real-time three-dimensional space module and an atmospheric pollution source concentration evaluation module. The three-dimensional space module is provided with a wind direction data module, a real-time wind speed data module and a wind direction and wind speed data module. The atmospheric pollution source concentration evaluation module is provided with a chemical plant positioning module, a collection module, a concentration calculation module and a control staggered time module. The collection module is used to collect the concentration of pollution gas emitted by the chemical plant on the same day. The concentration calculation module is used to calculate the concentration index of the pollution gas. The control staggered time module is used to prevent the concentration of pollution gas from being too high.

[0004] A similar prior art is Chinese patent application No. CN107064443A, which provides an industrial water pollution source monitoring method based on wireless sensing, including planning exploration based on the positions of the sewage discharge sources of each industrial enterprise near the water body; checking each industrial enterprise and its sewage discharge source, and making corresponding identification; disposing a sewage detector at the sewage discharge source position, and recording the model of the sewage detector and the sewage discharge source of the enterprise corresponding thereto; applying Kingview to develop a "regional industrial water pollution source monitoring system", and setting a water pollution source over-limit threshold in the system.

[0005] However, the existing analysis method analyzes based on multiple monitoring data. Due to the large number and complexity of the monitoring data, it is easy to cause insufficient sensitivity of abnormal emission detection. Therefore, the present application provides a fixed pollution source monitoring data analysis method, system and storage medium. SUMMARY

[0006] The present application provides a fixed pollution source monitoring data analysis method, system and storage medium for improving the sensitivity and accuracy of fixed pollution source monitoring data analysis.

[0007] In a first aspect, the present application provides a fixed pollution source monitoring data analysis method, which comprises:

[0008] Step S1, collecting fixed pollution source monitoring data based on a monitoring device, the monitoring data including pollution concentration data, environmental data, device operation parameters and factory production data;

[0009] Step S2, calculating the proximity value between different monitoring data, classifying the monitoring data into several categories based on the proximity value, and storing the classification result in a database;

[0010] Step S3, obtaining first monitoring data in a normal state, obtaining corresponding category monitoring data based on the classification result, taking the category monitoring data collected at each time point as target monitoring data, selecting several category monitoring data closest to the target monitoring data in collection time point as first data, and calculating an abnormal threshold of each category based on the first data and the target monitoring data;

[0011] Step S4, obtaining real-time monitoring data, calculating the deviation value of the corresponding real-time monitoring data and the corresponding first data for each category in the real-time monitoring data, and judging whether the current monitoring data is abnormal based on the deviation value and the corresponding abnormal threshold.

[0012] In combination with the first aspect, in a first implementation manner of the first aspect of the present application, the proximity value between different monitoring data is calculated, including:

[0013] The monitoring data is normalized, the minimum value and the maximum value of the monitoring data collected in a preset time period are obtained for each normalized monitoring data, the corresponding monitoring data is classified into a first number of data ranges based on the minimum value and the maximum value, each data range is numbered, the corresponding number is obtained for each monitoring data, and the proximity value is calculated for each two monitoring data based on the number.

[0014] In combination with the first aspect, in a second implementation manner of the first aspect of the present application, the proximity value is calculated for each two monitoring data based on the number, including:

[0015] The corresponding number is obtained for each two monitoring data, and the numbers of two corresponding monitoring data form a set of number data.

[0016] The second matrix is initialized, the dimension of the second matrix is the first number by the first number, the position of each set of number data in the second matrix is obtained, and the value of the corresponding position is added by a preset first value, after using all sets of number data, the final second matrix is obtained.

[0017] The first number of the position where the value in the second matrix is not zero is obtained, the first number of rows where all the values in each row of the second matrix are zero is obtained, the first number of columns where all the values in each column of the second matrix are zero is obtained, a second number is obtained by subtracting the first number of rows from the first number, a third number is obtained by subtracting the first number of columns from the first number, a first result value is obtained by multiplying the second number by the third number, the approximation value s is calculated using the first formula, and the first formula is:

[0018] .

[0019] In a third implementation manner of the first aspect, the monitoring data is divided into a plurality of categories based on the approximation values, and the plurality of categories include:

[0020] A first threshold value is preset, each monitoring data is initially taken as a separate category, two categories with the maximum approximation values are obtained, and it is determined whether the approximation value between the two categories is greater than the first threshold value. If yes, the two categories are merged into the same category, the approximation values between the merged category and other categories are calculated, and the step is repeated until the approximation values between the categories are less than or equal to the first threshold value.

[0021] In a fourth implementation manner of the first aspect, the approximation values between the merged category and other categories are calculated, and the calculation includes:

[0022] A plurality of approximation values between a plurality of monitoring data in the merged category and other categories are obtained, a sum of the plurality of approximation values is taken as a second value, a number of the plurality of monitoring data is obtained as a fourth number, and a result value obtained by dividing the second value by the fourth number is taken as the approximation value between the merged category and other categories.

[0023] In a fifth implementation manner of the first aspect, the abnormal threshold value of each category is calculated based on the first data and the target monitoring data, and the calculation includes:

[0024] A first matrix is generated by combining the plurality of first data, for each monitoring data in the first data, an average value corresponding to the monitoring data is calculated, the average values of all the monitoring data are combined as first center data, difference data between the target monitoring data and the first center data is calculated, a mapping point of the target monitoring data on the first matrix is calculated based on the difference data, and a deviation value between the target monitoring data and the mapping point is also calculated.

[0025] A plurality of deviation values are obtained for a plurality of target monitoring data of each category, a maximum value of the deviation values is obtained, and a result value obtained by multiplying the maximum deviation value by a preset first proportion is taken as the abnormal threshold value of the corresponding category.

[0026] In a sixth implementation form of the first aspect, the mapping point of the target monitoring data on the first matrix is calculated based on the difference data, and the mapping point comprises:

[0027] The deviation data of the target monitoring data and the first center data is calculated, a transposed matrix of the first matrix is obtained to obtain a first transposed matrix, the first transposed matrix is multiplied by the first matrix to obtain a first intermediate matrix, an inverse matrix of the first intermediate matrix is calculated to obtain a first inverse matrix, the first inverse matrix is multiplied by the first transposed matrix and then multiplied by the deviation data to obtain a mapping weight, and the first matrix is multiplied by the mapping weight and then added to the first center data to obtain the corresponding mapping point.

[0028] In a seventh implementation form of the first aspect, the current monitoring data is determined to be abnormal based on the deviation value and the corresponding abnormal threshold, and the determination comprises:

[0029] The deviation value and the corresponding abnormal threshold of each category are obtained, if all the deviation values are less than the corresponding abnormal threshold, the current monitoring data is determined to be normal, and if there is a deviation value greater than or equal to the corresponding abnormal threshold, the current monitoring data is determined to be abnormal.

[0030] In a second aspect, the application provides a stationary pollution source monitoring data analysis system, and the system comprises:

[0031] A collection unit is configured to collect stationary pollution source monitoring data based on a monitoring device, and the monitoring data comprises pollutant concentration data, environmental data, device operation parameters and factory production data.

[0032] A classification unit is configured to calculate an approximation value between different monitoring data, divide the monitoring data into several categories based on the approximation value, and store the classification result in a database.

[0033] A calculation unit is configured to obtain first monitoring data in a normal state, obtain corresponding category monitoring data based on the classification result, collect category monitoring data at each time point as target monitoring data, select several category monitoring data closest to the target monitoring data in time as first data, and calculate an abnormal threshold of each category based on the first data and the target monitoring data.

[0034] A judgment unit is configured to obtain real-time monitoring data, calculate a deviation value of corresponding real-time monitoring data and corresponding first data for each category in the real-time monitoring data, and determine whether the current monitoring data is abnormal based on the deviation value and the corresponding abnormal threshold.

[0035] In a third aspect, the application provides a computer readable storage medium, and the computer readable storage medium stores instructions, which, when executed on a computer, cause the computer to perform the stationary pollution source monitoring data analysis method.

[0036] Compared with the prior art, the beneficial effects of the present application are at least as follows:

[0037] In the technical solution provided in the present application, first, fixed pollution source monitoring data is collected based on a monitoring device, and the monitoring data includes not only pollutant concentration data but also environmental data, device operating parameters, and plant production data and other related data, so as to improve the accuracy of judging abnormal emissions by collecting various data related to the pollution source; the closeness values between different monitoring data are calculated, the monitoring data are classified into several categories based on the closeness values, the classification results are stored in a database, different abnormal threshold values are determined for monitoring data of different categories, and subsequent real-time detection of abnormalities and alarm are based on the abnormal threshold values, thereby improving the sensitivity of abnormal monitoring; first monitoring data in a normal state are obtained, corresponding category monitoring data are obtained based on the classification results, the category monitoring data collected at each time point are taken as target monitoring data, a plurality of category monitoring data closest to the target monitoring data in terms of collection time point are selected as first data, and the abnormal threshold values of each category are calculated based on the first data and the target monitoring data, so that by calculating the abnormal threshold values of each category, the monitoring data of different categories are analyzed based on different abnormal threshold values, thereby improving the accuracy of abnormal analysis. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor based on these drawings.

[0039] Figure 1 is an embodiment schematic diagram of a fixed pollution source monitoring data analysis method in the present application;

[0040] Figure 2 is an embodiment schematic diagram of initialization of a second matrix and a final second matrix in the present application;

[0041] Figure 3 is an embodiment schematic diagram of closeness values of each monitoring data in an initial state and after each classification in the present application;

[0042] Figure 4 is an embodiment schematic diagram of a fixed pollution source monitoring data analysis system in the present application. DETAILED DESCRIPTION

[0043] The embodiments of the present application provide a stationary pollution source monitoring data analysis method and system and a storage medium. The terms "first", "second", "third", "fourth" and the like (if any) in the specification and claims of the present application and the above-described drawings are used to distinguish similar objects, and do not have to be used to describe a particular order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" or "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0044] For ease of understanding, the specific processes of the embodiments of the present application are described below. Please refer to Figure 1 One embodiment of the stationary pollution source monitoring data analysis method in the embodiments of the present application includes the following steps.

[0045] Step S1, collecting stationary pollution source monitoring data based on monitoring equipment, the monitoring data including pollutant concentration data, environmental data, equipment operation parameters and factory production data.

[0046] Specifically, the directly measured monitoring data (such as various pollutant concentrations) that can directly reflect the pollution degree is easily disturbed by sensor quality or external environment (such as wind speed, wind direction, etc.) and thus produces false data, so only judging whether the pollution emission is abnormal based on the data that can be directly measured from the sensor may affect the accuracy of subsequent judgment of abnormal emission, therefore, not only the pollutant concentration data, but also the environmental data, equipment operation parameters and factory production data and other related data are collected.

[0047] Step S2, calculating the proximity value between different monitoring data, classifying the monitoring data into several categories based on the proximity value, and storing the classification result in a database.

[0048] Specifically, the monitoring data of stationary pollution sources such as factory emission sources, coal-fired power plants, waste incineration plants and the like usually includes various pollutant concentrations, emission flow rates, process parameters and the like, and these data have multi-source, complexity and dynamics, which easily lead to insufficient abnormal emission detection sensitivity, so it is crucial to improve the abnormal emission monitoring sensitivity by optimizing monitoring data analysis.

[0049] So first calculate the proximity between different monitoring data, the proximity refers to the correlation between different monitoring data, the closer the proximity to 1, the higher the correlation between the two monitoring data, on the contrary, the closer the proximity to 0, the lower the correlation between the two monitoring data, the specific method of calculating the proximity will be explained in detail later, the greater the proximity, the stronger the corresponding correlation, and the monitoring data is divided into several categories based on the proximity, the specific classification method will be explained in detail later, then different abnormal thresholds are determined for different categories of monitoring data, and the abnormal threshold is used to detect abnormality in real time and alarm in the future, improving the sensitivity of abnormal monitoring.

[0050] Step S3, obtaining the first monitoring data in the normal state, obtaining the corresponding category monitoring data based on the classification result, taking the category monitoring data collected at each time point as the target monitoring data, selecting a plurality of category monitoring data closest to the target monitoring data collection time point as the first data, and calculating the abnormal threshold of each category based on the first data and the target monitoring data;

[0051] Specifically, the monitoring data collected under the normal emission state is taken as the first monitoring data, and the category monitoring data refers to the monitoring data collected at each time point for each category. For example, the first category is SO2 concentration, NOx concentration and PM concentration, and the monitoring data collected at each time point under the normal emission state is obtained. For the category monitoring data collected at each time point, such as the time points collected are t1, t2, t3, …, t12, the category monitoring data collected at t1 is {a1, b1, c1}, and 3 time points are selected. The category monitoring data closest to t1 is selected as the first data, such as the category monitoring data collected at t2, 3 and t4 is the corresponding first data, and the abnormal threshold of the corresponding category is calculated based on the first data and the target monitoring data. The abnormal threshold represents the abnormal index data of the corresponding category, and the abnormality of the monitoring data of the corresponding category can be judged based on the abnormal threshold in the future. The specific method of calculating the abnormal threshold will be explained in detail later.

[0052] Step S4, obtaining real-time monitoring data, calculating the deviation value of the corresponding real-time monitoring data and the corresponding first data for each category in the real-time monitoring data based on the deviation value and the corresponding abnormal threshold, and judging whether the current monitoring data is abnormal.

[0053] Specifically, real-time monitoring data collected in real time is acquired, category monitoring data of each category in the real-time monitoring data is acquired based on the classification result, first data corresponding to the category monitoring data, that is, monitoring data closest to the collection time point of the real-time monitoring data, is also acquired, a deviation value of each category is calculated based on the category monitoring data and the first data using a subsequent calculation deviation value method, and whether the current monitoring data is abnormal is judged based on the deviation value and an abnormal threshold value of the corresponding category calculated previously. The specific method of judging abnormality will be explained in detail later.

[0054] The above method first classifies monitoring data based on the proximity value between the monitoring data, then calculates the corresponding abnormal threshold value for the monitoring data of each category, and judges whether the real-time abnormal data is abnormal based on the abnormal threshold value. If an abnormality occurs in one of the categories, it is judged that the current emission condition is abnormal, and a corresponding abnormal warning is issued. The above method can improve the sensitivity of abnormal emission monitoring.

[0055] In a specific embodiment, a first evaluation value is calculated based on a plurality of first key data, specifically including the following steps:

[0056] The monitoring data is normalized, the minimum value and the maximum value of the monitoring data collected in a preset time period are acquired for each normalized monitoring data, the corresponding monitoring data is divided into a first number of data ranges based on the minimum value and the maximum value, each data range is numbered, the corresponding number is acquired for each monitoring data, and the proximity value is calculated for each two monitoring data based on the number.

[0057] Specifically, different monitoring data can have different measurement units and orders of magnitude, so the monitoring data is first normalized. The specific normalization method can use Min-Max normalization to scale the monitoring data to 0-1. For each monitoring data, the minimum and maximum values of the monitoring data collected in a preset time period are obtained, and the corresponding monitoring data is divided into a first number of data ranges based on the minimum and maximum values. Each data range is numbered, and the corresponding number is obtained for each monitoring data. Taking SO2 concentration as an example, assuming that the minimum value of SO2 concentration collected in 12 hours is 0.02 and the maximum value is 0.92, and assuming that the first number is 5, the monitoring data is divided into 5 data ranges, which are [0.02-0.20], [0.20-0.38], [0.38-0.56], [0.56-0.74], and [0.74-0.92], respectively. The five ranges are numbered 1, 2, 3, 4, and 5, respectively. Assuming that the monitoring data collected in 12 hours is 0.04, 0.02, 0.25, 0.43, 0.52, 0.68, 0.78, 0.92, 0.86, 0.75, 0.52, and 0.40, the corresponding numbers are 1, 1, 2, 3, 3, 4, 5, 5, 5, 5, 3, and 3, respectively. The above method is used to divide the range and number of all monitoring data. The subsequent explanation of how to calculate the proximity value based on the number of each two monitoring data is detailed.

[0058] In a specific embodiment, the proximity value is calculated based on the number of each two monitoring data, specifically including the following steps:

[0059] The corresponding number is obtained for each two monitoring data, and the numbers of two corresponding monitoring data form a set of number data.

[0060] Initialize the second matrix, the dimension of the second matrix is the first number multiplied by the first number, obtain the position of each set of number data in the second matrix, and add a preset first value to the corresponding position value. After using all sets of number data, the final second matrix is obtained.

[0061] The first number of the position with a non-zero value in the second matrix is obtained, and the first row number of each row of zeros and the first column number of each column of zeros in the second matrix are also obtained. The second number is obtained by subtracting the first row number from the first number, and the third number is obtained by subtracting the first column number from the first number. The first result value is obtained by multiplying the second number by the third number. The proximity value s is calculated using the first formula: .

[0062] Specifically, taking the two monitoring data of SO2 concentration and temperature as an example, the SO2 concentration is numbered as 1, 1, 2, 3, 3, 4, 5, 5, 5, 5, 3, 3, and the temperature is numbered as 1, 4, 5, 5, 5, 5, 5, 5, 4, 3, 1, 1, assuming that the first number is 5, the initialized second matrix is as shown in Figure 2 As shown in the second matrix, the SO2 concentration and the temperature correspond to each other to obtain 12 sets of numbered data {1, 1}, {1, 4}, {2, 5}, {3, 5}, {3, 5}, {4, 5}, {5, 5}, {5, 5}, {5, 4}, {5, 3}, {3, 1}, {3, 1}, based on the positions of the numbered data in the first data, the original numerical value of the corresponding position is added by a preset first value, assuming that the first value is 1, the final second matrix as shown in Figure 2 is obtained, and the first number of the non-zero position in the second matrix is 9, the first row number is 0, and the first column number is 1, so the second number is 5 and the third number is 4, the closeness s is calculated based on the first formula, assuming that there are m monitoring data in total, m(m-1) / 2 monitoring data pairs need to be calculated.

[0063] The above closeness value calculation can help us obtain the relationship between different monitoring data.

[0064] In a specific embodiment, the monitoring data is divided into several categories based on the closeness value, including the following steps:

[0065] A first threshold is preset, and initially each monitoring data is taken as a separate category, the two categories with the largest closeness value are obtained, and it is judged whether the closeness value between the two categories is greater than the first threshold, if yes, the two categories are merged into the same category, the closeness value between the merged category and other categories is calculated, and the step is repeated until the closeness value between the categories is less than or equal to the first threshold.

[0066] Specifically, assuming that the first threshold is 0.5, and assuming that there are 6 kinds of monitoring data, as shown in Figure 3As shown, the schematic diagram of the proximity values of each monitoring data in the initial state and after each classification, the two categories with the largest proximity values are temperature and pressure in the first classification, and the proximity values of temperature and pressure are 0.9, so temperature and pressure are classified into the same category, the two categories with the largest proximity values are SO2 and NO2 in the second classification, and the proximity values of SO2 and NO2 are 0.8, so SO2 and NO2 are classified into the same category, the two categories with the largest proximity values are {temperature, pressure} and flow in the third classification, so the two categories are classified into the same category, the two categories with the largest proximity values are {SO2, NOx} and PM in the fourth classification, so they are classified into the same category, and the proximity values of {SO2, NOx, PM} and {temperature, pressure, flow} are 0.1 after the fourth classification, which is less than the first threshold 0.5, and the classification is ended.

[0067] In a specific embodiment, the proximity values between the merged category and other categories are calculated, specifically including the following steps:

[0068] The proximity values between the merged category and other categories are obtained, the sum of the proximity values is taken as a second value, the number of the monitoring data is taken as a fourth number, and the result value obtained by dividing the second value by the fourth number is taken as the proximity value between the merged category and other categories.

[0069] Specifically, as shown, temperature and pressure are classified into the same category after the first classification, and the proximity value between {temperature, pressure} and SO2 is calculated as an example, the new proximity value is the sum of the proximity value 0.2 of temperature and SO2 and the proximity value 0.2 of pressure and SO2, which is taken as a second value 0.3, and the second value 0.3 is divided by 2 to obtain 0.15, which is taken as the proximity value between the category of temperature and pressure and SO2, and the proximity values between other categories are calculated using the same method. Figure 3

[0070] In a specific embodiment, the abnormal threshold of each category is calculated based on the first data and the target monitoring data, specifically including the following steps:

[0071] The first data is combined to generate a first matrix, for each monitoring data in the first data, the corresponding average value is calculated, the average values of all monitoring data are combined as a first center data, the difference data between the target monitoring data and the first center data is calculated, the mapping point of the target monitoring data on the first matrix is calculated based on the difference data, and the deviation value between the target monitoring data and the mapping point is also calculated;

[0072] The maximum value of the deviation values is obtained for each category of the target monitoring data, and the result value obtained by multiplying the maximum deviation value by a preset first proportion is taken as the abnormal threshold of the corresponding category.​

[0073] Specifically, assuming that the target detection data of a certain time point of a certain category is {a1, b1, c1}, the corresponding first data assumption is {a2, b2, c2}, {a3, b3, c3} and {a4, b4, c4}, the first center data calculated is {(a2+a3+a4) / 3, (b2+b3+b4) / 3, (c2+c3+c4) / 3}, the difference data of the target detection data and the first center data is {a1-(a2+a3+a4) / 3, b1-(b2+b3+b4) / 3, c1-(c2+c3+c4) / 3}, and then the mapping point of the target monitoring data on the first matrix is calculated based on the difference data. The specific method for calculating the mapping point will be explained in detail later. The mapping point refers to the projection point of the target monitoring data on the first matrix, and is a reference point for judging whether the target monitoring data is abnormal. The deviation value of the target monitoring data and the mapping point is calculated. In the case that all monitoring data is normal monitoring data, the deviation value of the target monitoring data and the mapping point should be small, that is, the difference between the target monitoring data and the mapping point should be small. The deviation value of the target monitoring data and the mapping point is calculated, the Euclidean distance between the target monitoring data and the mapping point is calculated, and the Euclidean distance is taken as the deviation value. Based on this deviation value, the corresponding abnormal threshold is calculated.

[0074] For the monitoring data collected at multiple time points of each category, the corresponding deviation value is calculated, and the maximum value is obtained. The maximum value represents the maximum deviation degree of a certain monitoring data and other adjacent monitoring data under normal circumstances. If the monitoring data at a certain time point is abnormal, the deviation value of the monitoring data at this time point and the adjacent monitoring data should be greater than the maximum value. Therefore, the maximum deviation value is multiplied by a preset first proportion, such as 1.2, and the result value obtained after multiplication is taken as the abnormal threshold corresponding to this category. Based on this abnormal threshold, whether the monitoring data of the corresponding category is abnormal is calculated.

[0075] In a specific embodiment, the mapping point of the target monitoring data on the first matrix is calculated based on the difference data, and specifically includes the following steps:

[0076] The deviation data of the target monitoring data and the first center data is calculated, the transpose matrix of the first matrix is obtained to obtain the first transpose matrix, the first transpose matrix is multiplied by the first matrix to obtain the first intermediate matrix, the inverse matrix of the first intermediate matrix is calculated to obtain the first inverse matrix, the first inverse matrix is multiplied by the first transpose matrix and then multiplied by the deviation data to obtain the mapping weight, and the first matrix is multiplied by the mapping weight and then added to the first center data to obtain the corresponding mapping point.

[0077] Specifically, the mapping point of the target monitoring data in the local space represented by the first matrix is calculated by the above method, the mapping point is an expected value of the target monitoring data in the local space reflected by the first matrix, and represents a projection point of the target monitoring data in the local space. By calculating the mapping point, the deviation data of the target monitoring data and the mapping point can be calculated based on the mapping point, and the deviation data represents the deviation degree of the target monitoring data and the mapping point.

[0078] In a specific embodiment, whether the current monitoring data is abnormal is judged based on the deviation value and the corresponding abnormal threshold, including:

[0079] The deviation value of each category and the corresponding abnormal threshold are obtained, if all deviation values are less than the corresponding abnormal threshold, it is determined that the current monitoring data is normal, and if there is a deviation value greater than or equal to the corresponding abnormal threshold, it is determined that the current monitoring data is abnormal.

[0080] Specifically, after calculating the deviation value of each category, the abnormal threshold of the corresponding category is obtained, if all deviation values are less than the corresponding abnormal threshold, it is indicated that the monitoring data of all categories is within the normal range, so it is determined that the current monitoring data is normal, that is, there is no abnormal emission at present. Due to the diversity of monitoring data, the monitoring data of some categories may be affected by the external environment, and normal monitoring data may be obtained in abnormal conditions, but the situation that all categories of monitoring data are affected by the external environment at the same time is rare. Therefore, the data is classified, in order to improve the accuracy of judgment, when there is a deviation value greater than or equal to the corresponding abnormal threshold, it is determined that the current monitoring data is abnormal.

[0081] The above describes a fixed pollution source monitoring data analysis method in the embodiments of the application, and the following describes a fixed pollution source monitoring data analysis system in the embodiments of the application, please refer to Figure 4 An embodiment of the fixed pollution source monitoring data analysis system in the embodiments of the application includes:

[0082] The collecting unit is configured to collect fixed pollution source monitoring data based on the monitoring device, and the monitoring data includes pollutant concentration data, environmental data, device operation parameters and factory production data;

[0083] The classification unit is configured to calculate the proximity value between different monitoring data, divide the monitoring data into several categories based on the proximity value, and store the classification result in the database;

[0084] The computing unit is configured to acquire first monitoring data in a normal state, acquire corresponding category monitoring data based on the classification result, collect category monitoring data at each time point as target monitoring data, select a plurality of category monitoring data closest to the target monitoring data collection time point as first data, and calculate an abnormal threshold of each category based on the first data and the target monitoring data.

[0085] The judging unit is configured to acquire real-time monitoring data, calculate a deviation value of corresponding real-time monitoring data and corresponding first data for each category in the real-time monitoring data, and judge whether the current monitoring data is abnormal based on the deviation value and the corresponding abnormal threshold.

[0086] The application also provides a computer readable storage medium, which can be a non-volatile computer readable storage medium or a volatile computer readable storage medium, and the computer readable storage medium stores instructions, which, when executed on a computer, cause the computer to perform the steps of the method for analyzing monitoring data of a stationary pollution source.

[0087] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described system, system and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0088] The integrated unit, if realized in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiments of the application. The foregoing storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0089] The above-described embodiments are only used to illustrate the technical solutions of the application, rather than limit them; although the application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the application.

Claims

1. A method for analyzing stationary source pollution monitoring data, characterized in that, The method comprises: Step S1, collecting fixed pollution source monitoring data based on a monitoring device, the monitoring data comprising pollutant concentration data, environmental data, device operation parameters and factory production data; Step S2, calculating the proximity values between different monitoring data, classifying the monitoring data into several categories based on the proximity values, and storing the classification results in a database; Step S3, obtaining first monitoring data in a normal state, obtaining corresponding category monitoring data based on the classification results, taking the category monitoring data collected at each time point as target monitoring data, selecting several category monitoring data closest to the target monitoring data collection time point as first data, and calculating an abnormal threshold value of each category based on the first data and the target monitoring data, wherein the target monitoring data is category monitoring data at a certain time point, and the first data is several category monitoring data closest to the certain time point; Step S4, obtaining real-time monitoring data, calculating the deviation value of the corresponding real-time monitoring data and the corresponding first data for each category in the real-time monitoring data, and judging whether the current monitoring data is abnormal based on the deviation value and the corresponding abnormal threshold value; The proximity values between different monitoring data are calculated, comprising: normalizing the monitoring data, obtaining the minimum value and the maximum value of the monitoring data collected in a preset time period for each normalized monitoring data, dividing the corresponding monitoring data into a first number of data ranges based on the minimum value and the maximum value, numbering each data range, obtaining the corresponding number for each monitoring data, and calculating the proximity value of each two monitoring data based on the number; The proximity value of each two monitoring data is calculated based on the number, comprising: obtaining the corresponding number for each two monitoring data, and the numbers of two corresponding monitoring data form a set of number data; initializing a second matrix, the dimension of the second matrix being the first number multiplied by the first number, obtaining the position of each set of number data in the second matrix, and adding a preset first value to the corresponding position value, obtaining the final second matrix after using all sets of number data; obtaining the first number of the position with a non-zero value in the second matrix, also obtaining the first row number of each row full of zeros in the second matrix and the first column number of each column full of zeros in the second matrix, obtaining a second number by subtracting the first row number from the first number, obtaining a third number by subtracting the first column number from the first number, obtaining a first result value by multiplying the second number by the third number, and calculating the proximity value s using the first formula: 。 2. The method of claim 1, wherein, The monitoring data is classified into several categories based on the proximity value, comprising: presetting a first threshold value, initially taking each monitoring data as a separate category, obtaining the two categories with the largest proximity value, judging whether the proximity value between the two categories is greater than the first threshold value, if yes, merging the two categories into the same category, calculating the proximity value between the merged category and other categories, repeating the step until the proximity value between the categories is less than or equal to the first threshold value.

3. The method of claim 1, wherein, The proximity value between the merged category and other categories is calculated, comprising: The sum of the plurality of proximity values is calculated as a second value, the number of the plurality of monitoring data is obtained as a fourth number, and the second value is divided by the fourth number to obtain a result value as the proximity value between the merged category and other categories.

4. The method of claim 1, wherein, The abnormal threshold of each category is calculated based on the first data and the target monitoring data, including: The plurality of first data is combined to generate a first matrix, for each monitoring data in the first data, the corresponding average value is calculated, the average values of all monitoring data are combined as first center data, the difference data between the target monitoring data and the first center data is calculated, the mapping point of the target monitoring data on the first matrix is calculated based on the difference data, and the deviation value between the target monitoring data and the mapping point is also calculated. The maximum value of the plurality of deviation values is obtained for each category of the plurality of target monitoring data, and the maximum deviation value is multiplied by a preset first proportion to obtain a result value as the abnormal threshold of the corresponding category.

5. The method of claim 4, wherein, The mapping point of the target monitoring data on the first matrix is calculated based on the difference data, including: The deviation data between the target monitoring data and the first center data is calculated, the transpose matrix of the first matrix is obtained to obtain a first transpose matrix, the first transpose matrix is multiplied by the first matrix to obtain a first intermediate matrix, the inverse matrix of the first intermediate matrix is calculated to obtain a first inverse matrix, the first inverse matrix is multiplied by the first transpose matrix and then multiplied by the deviation data to obtain a mapping weight, and the first matrix is multiplied by the mapping weight and then added to the first center data to obtain the corresponding mapping point.

6. The method of claim 1, wherein, The current monitoring data is determined to be abnormal or normal based on the deviation value and the corresponding abnormal threshold, including: The deviation value and the corresponding abnormal threshold of each category are obtained, if all deviation values are less than the corresponding abnormal threshold, it is determined that the current monitoring data is normal, and if there is a deviation value greater than or equal to the corresponding abnormal threshold, it is determined that the current monitoring data is abnormal.

7. A stationary pollution source monitoring data analysis system for implementing a stationary pollution source monitoring data analysis method according to any one of claims 1 to 6, characterized in that, The system includes: The collection unit is configured to collect stationary source monitoring data based on a monitoring device, and the monitoring data includes pollutant concentration data, environmental data, device operation parameters, and factory production data. The classification unit is configured to calculate proximity values between different monitoring data, classify the monitoring data into a plurality of categories based on the proximity values, and store the classification results in a database. The calculation unit is configured to obtain first monitoring data in a normal state, obtain corresponding category monitoring data based on the classification results, collect category monitoring data at each time point as target monitoring data, select a plurality of category monitoring data closest to the target monitoring data collection time point as first data, and calculate the abnormal threshold of each category based on the first data and the target monitoring data. The judgment unit is configured to obtain real-time monitoring data, calculate the deviation value between the corresponding real-time monitoring data and the corresponding first data for each category in the real-time monitoring data, and determine whether the current monitoring data is abnormal based on the deviation value and the corresponding abnormal threshold.

8. A computer-readable storage medium having stored thereon instructions, the computer-readable storage medium comprising: The instructions are executed by the processor to implement the stationary source monitoring data analysis method of any one of claims 1-6.

Images