Multi-source heterogeneous monitoring data operation and maintenance management method of integrated environmental protection management and control platform

Abstract

This invention relates to the field of data governance technology, and particularly to a method for the operation and maintenance of multi-source heterogeneous monitoring data in an integrated environmental protection control and management platform. The method involves loading multiple emission source device tag numbers and multiple sets of emission source production parameters for a target pollutant in a target monitoring area; loading environmental monitoring parameters and target pollutant monitoring concentrations for the target monitoring area; performing emission prediction based on the multiple emission source device tag numbers and multiple sets of emission source production parameters to obtain the emission amounts of multiple target pollutants; fusing the emission amounts of multiple target pollutants based on the environmental monitoring parameters to obtain a fitted concentration of the target pollutant; and assigning a reliable label to the target pollutant monitoring concentration if the concentration deviation between the monitored concentration and the fitted concentration is less than or equal to a concentration deviation threshold; otherwise, assigning an anomaly label to the monitored concentration. This invention avoids the impact of erroneous data caused by sensor disturbances on decision-making.

Description

Technical Field

[0001] This invention relates to the field of data governance technology, and in particular to a method for the operation and maintenance of multi-source heterogeneous monitoring data in an integrated environmental protection management and control platform. Background Technology

[0002] Traditional enterprise pollutant monitoring relies heavily on sensors to collect data multiple times and average the results to reduce errors. This method supports environmental management and inspection responses through simple data storage. However, this approach has a significant drawback: it relies too heavily on the stability of the sensors. In actual pollutant monitoring, sensors may experience disturbances that cause abnormal data collection, thus compromising the accuracy of the final data. Summary of the Invention

[0003] This invention addresses the problem in existing technologies where sensors may cause abnormal data collection due to disturbances, by providing a method for the operation and maintenance of multi-source heterogeneous monitoring data in an integrated environmental management and control platform.

[0004] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0005] This invention provides a method for the operation and maintenance of multi-source heterogeneous monitoring data in an integrated environmental protection management and control platform, comprising: loading multiple emission source device tag numbers and multiple sets of emission source production parameters for a target pollutant in a target monitoring area; loading environmental monitoring parameters and target pollutant monitoring concentrations for the target monitoring area; performing emission prediction based on the multiple emission source device tag numbers and the multiple sets of emission source production parameters to obtain the emission amounts of multiple target pollutants; fusing the emission amounts of the multiple target pollutants based on the environmental monitoring parameters to obtain a fitted concentration of the target pollutant; and, if the concentration deviation between the monitored concentration of the target pollutant and the fitted concentration of the target pollutant is less than or equal to a concentration deviation threshold, assigning a reliable identifier to the monitored concentration of the target pollutant; otherwise, assigning an anomaly identifier to the monitored concentration of the target pollutant.

[0006] Optionally, loading multiple emission source device tag numbers of the target pollutant in the target monitoring area includes: obtaining an initial set of device tag numbers of the emission pollutants having the target pollutant; traversing the initial set of device tag numbers and statistically analyzing the detection support set of the target pollutant in the target monitoring area; extracting initial device tag numbers from the initial set of device tag numbers whose detection support is greater than or equal to a support threshold from the detection support set, and setting them as the multiple emission source device tag numbers.

[0007] The process of traversing the initial device tag set and calculating the detection support set of the target pollutant in the target monitoring area includes: extracting a first device tag from the initial device tag set; loading a historical detection location set of the target pollutant where the first device tag is in operation and the other initial device tag is in shutdown state; calculating the proportion of occurrences of the target monitoring area in the historical detection location set of the target pollutant, setting it as the detection support of the first device tag, and adding it to the detection support set.

[0008] The process of loading environmental monitoring parameters for the target monitoring area includes: loading several target pollutant monitoring data for the target monitoring area, wherein any one of the target pollutant monitoring data includes the historical concentration of the target pollutant, multiple sets of emission source production record parameters and environmental record parameters for the multiple emission source equipment tag numbers; performing cluster analysis on the target pollutant monitoring data based on the multiple sets of emission source production record parameters to obtain clustering results; traversing the clustering results of the target pollutant monitoring data, performing grey relational analysis on the historical concentration of the target pollutant and the environmental record parameters to obtain environmental attributes with a grey relational degree greater than or equal to the relational degree threshold, and adding them to the target pollutant monitoring attributes of the target monitoring area; and loading the sensing data of the target pollutant monitoring attributes of the target monitoring area as the environmental monitoring parameters.

[0009] Specifically, based on the multiple sets of emission source production record parameters, cluster analysis is performed on the monitoring data of the several target pollutants to obtain clustering results of the target pollutant monitoring data. This includes: obtaining multiple sets of first emission source production record parameters for the multiple emission source equipment tag numbers of the first target pollutant monitoring data; obtaining multiple sets of second emission source production record parameters for the multiple emission source equipment tag numbers of the second target pollutant monitoring data; comparing the multiple sets of first emission source production record parameters and the multiple sets of second emission source production record parameters with the same emission source production parameters to obtain multiple production parameter similarities; when the multiple production parameter similarities are all greater than or equal to the production parameter similarity threshold, the first target pollutant monitoring data and the second target pollutant monitoring data are added to the same category; otherwise, the first target pollutant monitoring data and the second target pollutant monitoring data are added to a different category.

[0010] Specifically, the process involves comparing the first and second sets of production record parameters of the multiple emission sources with the production parameters of the same emission source to obtain multiple production parameter similarities. This includes: obtaining the first and second sets of production record parameters of the first emission source; loading a predefined production parameter attribute deviation threshold; and calculating the percentage of attributes whose production parameter deviations between the first and second sets of production record parameters are less than or equal to the corresponding production parameter attribute deviation threshold, setting this percentage as the first emission source production parameter similarity, and adding it to the multiple production parameter similarities.

[0011] Optionally, the clustering results of the target pollutant monitoring data are traversed, and grey relational analysis is performed on the historical concentrations of the target pollutants and the environmental record parameters to obtain environmental attributes with grey relational degrees greater than or equal to the relational degree threshold. This includes: obtaining a first cluster of target pollutant monitoring data from the clustering results of the target pollutant monitoring data; extracting multiple historical concentrations of target pollutants and multiple sets of environmental record parameters; performing grey relational analysis on the multiple sets of environmental record parameters based on the multiple historical concentrations of target pollutants to obtain the relational degrees of several environmental attributes in the first cluster; continuing until the relational degrees of several environmental attributes in the Nth cluster are obtained, where N represents the number of categories; calculating the mean of the relational degrees of the same attribute for the relational degrees of the first cluster up to the Nth cluster to obtain the fitting relational degrees of several environmental attributes; and selecting the environmental attributes with fitting relational degrees of several environmental attributes greater than or equal to the relational degree threshold.

[0012] Optionally, based on the multiple emission source equipment tag numbers and the multiple sets of emission source production parameters, emission prediction is performed to obtain the emission amounts of multiple target pollutants, including: collecting the emission source production record parameter set and the target pollutant theoretical emission set of the first emission source equipment tag number, and training the first emission source emission prediction model; storing the first emission source emission prediction model and the first emission source equipment tag number together and adding them to the target pollutant emission predictor set; based on the multiple emission source equipment tag numbers, retrieving the target pollutant emission predictor set, processing the multiple sets of emission source production parameters, performing emission prediction, and obtaining the emission amounts of multiple target pollutants.

[0013] Optionally, based on the environmental monitoring parameters, the emissions of the multiple target pollutants are fused to obtain a fitted concentration of the target pollutants. This includes: loading the emissions of multiple target pollutants from multiple emission source device tag numbers and historical environmental monitoring parameters of the target monitoring area as input data; and loading the target pollutant detection values ​​of the target monitoring area at the same moment with reliable identifiers as output supervision values; training a target pollutant emission fusion processor using machine learning based on the input data and output supervision values, and storing it in association with the target monitoring area; and scheduling the target pollutant emission fusion processor based on the target monitoring area to fuse the emissions of the multiple target pollutants based on the environmental monitoring parameters to obtain a fitted concentration of the target pollutants.

[0014] By implementing this invention, it is possible to load multiple emission source device tag numbers and multiple sets of emission source production parameters for target pollutants in the target monitoring area, so as to eliminate interference from irrelevant devices, concentrate monitoring resources on key emission sources, reduce unnecessary monitoring costs and data redundancy, and improve monitoring efficiency.

[0015] By implementing this invention, it is possible to load environmental monitoring parameters and target pollutant monitoring concentrations of the target monitoring area, avoid monitoring irrelevant environmental factors, make environmental monitoring more targeted, reduce invalid data collection, reduce data processing difficulty, and at the same time ensure that environmental parameters can effectively reflect the impact on the target pollutant concentrations.

[0016] By implementing this invention, it is possible to perform emission prediction based on the multiple emission source equipment tag numbers and the multiple sets of emission source production parameters, and obtain the emission amounts of multiple target pollutants. By obtaining the emission amount of each emission source, it is possible to clearly understand the contribution of each equipment to regional pollution, providing precise guidance for the formulation of pollution control and treatment measures, and facilitating the implementation of targeted emission reduction measures.

[0017] By implementing this invention, it is possible to fuse the emissions of multiple target pollutants based on the environmental monitoring parameters to obtain the fitted concentration of the target pollutants. The fitted concentration is used as a theoretical calculation value and compared with the actual monitoring concentration, providing a scientific reference standard for subsequent judgment on whether the monitoring concentration is abnormal and reliable.

[0018] By implementing this invention, when the concentration deviation between the monitored concentration of the target pollutant and the fitted concentration of the target pollutant is less than or equal to a concentration deviation threshold, the monitored concentration of the target pollutant is marked as reliable; otherwise, the monitored concentration of the target pollutant is marked as abnormal. The abnormality marking provides staff with a clear indication of the problem, which helps to quickly locate the cause of the data anomaly and take timely measures such as repairing the sensor and eliminating interference, thereby ensuring the stable operation of the monitoring system and improving the reliability of the overall monitoring system.

[0019] In summary, by implementing this invention, the accuracy and reliability of pollutant monitoring data can be effectively improved, and erroneous data caused by sensor disturbances and other problems can be avoided from affecting decision-making. Attached Figure Description

[0020] Figure 1 A flowchart illustrating the operation and maintenance management method for multi-source heterogeneous monitoring data of the integrated environmental management and control platform provided by this invention;

[0021] Figure 2 This is a schematic diagram illustrating the structure of cluster analysis of monitoring data for several target pollutants in the multi-source heterogeneous monitoring data operation and maintenance management method of the integrated environmental protection control and governance platform provided by the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0024] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0025] Example 1, as Figure 1 As shown, this embodiment of the invention provides a method for the operation and maintenance of multi-source heterogeneous monitoring data in an integrated environmental protection management and control platform, including:

[0026] S100: Loads multiple emission source device tag numbers and multiple sets of emission source production parameters for the target pollutant in the target monitoring area;

[0027] S200: Load environmental monitoring parameters and target pollutant monitoring concentrations for the target monitoring area;

[0028] S300: Based on the multiple emission source equipment tag numbers and the multiple sets of emission source production parameters, perform emission prediction to obtain the emission amounts of multiple target pollutants;

[0029] S400: Based on the environmental monitoring parameters, the emissions of the multiple target pollutants are fused to obtain the fitted concentration of the target pollutants;

[0030] S500: When the concentration deviation between the monitored concentration of the target pollutant and the fitted concentration of the target pollutant is less than or equal to the concentration deviation threshold, the monitored concentration of the target pollutant is marked as reliable; otherwise, the monitored concentration of the target pollutant is marked as abnormal.

[0031] Before performing the above five steps, it is also necessary to perform data cleaning, standardization and time alignment processing on emission source equipment data and environmental monitoring data from different manufacturers and using different communication protocols, so as to unify the data format and time resolution and achieve synchronous alignment of multi-source data.

[0032] In step S100 of this application embodiment, loading the tag numbers of multiple emission source devices for the target pollutant in the target monitoring area includes:

[0033] Obtain an initial set of device tag numbers for the emitted pollutants having the target pollutant;

[0034] Traverse the initial device tag set and statistically analyze the detection support set of the target pollutant in the target monitoring area;

[0035] Extract the initial device tag numbers from the initial device tag number set that have a detection support greater than or equal to the support threshold from the detection support set, and set them as the device tag numbers of the plurality of emission sources.

[0036] In this embodiment of the application, the purpose of step S100 is to accurately screen out the key equipment tag numbers that are truly related to the emission of the target pollutants within the target monitoring area, so as to lay the foundation for subsequent emission prediction, concentration fitting and data credibility judgment.

[0037] To achieve the above objectives, it is first necessary to obtain an initial set of equipment tag numbers that emit pollutants with the target pollutant.

[0038] Specifically, it is necessary to first identify all equipment tag numbers that contain the target pollutant in the pollutant emissions, forming an initial set of equipment tag numbers. For example, if the target pollutant is particulate matter, the initial set should include equipment that may emit particulate matter, such as sintering machines, converters, and material yard conveying equipment.

[0039] For example, basic equipment information can be retrieved from production automation systems such as desulfurization DCS systems and energy systems, and combined with pollutant emission inventories, equipment tag numbers that may emit target pollutants can be quickly screened to form an initial set of equipment tag numbers.

[0040] Next, it is necessary to traverse the initial device tag set and count the detection support set of the target pollutant in the target monitoring area.

[0041] In step S100 of this application embodiment, traversing the initial device tag set and statistically analyzing the detection support set of the target pollutant in the target monitoring area includes:

[0042] Extract the first device tag number from the initial device tag number set;

[0043] Load the historical detection location set of the target pollutant where the first device tag number is in the working state and the other initial device tag numbers are in the shutdown state;

[0044] The proportion of occurrences of the target monitoring area in the historical detection location set of the target pollutant is statistically analyzed and set as the first device tag number detection support, which is then added to the detection support set.

[0045] In this embodiment of the application, the purpose of the above sub-step of step S100 is to accurately calculate the correlation between a single device tag number and the emission of the target pollutant, i.e. the detection support, to exclude interference from other devices and to clarify the independent contribution of a single device to the detection of pollutants in the target monitoring area.

[0046] To achieve the above objectives, the first step is to extract the first equipment tag number from the initial equipment tag number set. That is, from the obtained "initial equipment tag number set", individual equipment is selected as the "first equipment tag number" in turn, and its emission contribution is analyzed one by one. For example, "1# sintering machine" is selected as the first equipment tag number first, and after the analysis is completed, "2# converter" is selected, and so on, until all initial equipment tag number sets are traversed.

[0047] The second step requires loading the historical detection location set of the target pollutant where the first equipment tag number is in an operational state, while other initial equipment tag numbers are in a shutdown state. First, it's necessary to filter out historical time periods where the first equipment tag number is in an 'operational' state and the other initial equipment tag numbers are in a 'shutdown' state. For example, filtering for the period from 8:00 to 10:00 on May 10, 2024, where only sintering machine #1 is operating and other equipment is shut down. Then, within the aforementioned time period, the "historical detection location set" of the target pollutant is retrieved, which is the pollutant location information recorded by monitoring devices associated with the monitoring platform, such as air micro-stations, TSP monitors, and video surveillance capture systems, to form the detection location dataset for this operating condition.

[0048] The third step involves calculating the proportion of occurrences of the target monitoring area relative to the historical detection location set of the target pollutant. This proportion is set as the detection support for the first device tag number and added to the detection support set. Specifically, this involves calculating the percentage of occurrences of the aforementioned historical detection location set within the target monitoring area, which is used as the detection support for the first device tag number. The specific calculation method is: Detection Support = (Number of occurrences of pollutant detection locations within the target monitoring area / Total number of occurrences of pollutant detection locations under this operating condition) × 100%

[0049] For example, if the pollutant is detected 100 times when only sintering machine #1 is running, and 85 of those times are within the target monitoring area, then the detection support of sintering machine #1 is 85%, and this value is added to the "detection support set".

[0050] Finally, it is necessary to extract the initial device tag numbers from the initial device tag number set that have a detection support greater than or equal to the support threshold from the detection support set, and set them as the device tag numbers of the multiple emission sources.

[0051] Specifically, a support threshold, such as 60%, can be set according to industry standards or the actual needs of the enterprise. Then, equipment tag numbers with "detected support ≥ support threshold" can be extracted from the initial equipment tag number set as the final "multiple emission source equipment tag numbers". For example, only equipment such as sintering machines and converters with detected support ≥ 60% can be retained, while equipment with support below the support threshold can be removed.

[0052] In step S200 of this application embodiment, it is necessary to load the environmental monitoring parameters of the target monitoring area and the monitoring concentration of the target pollutants;

[0053] The environmental monitoring parameters for the target monitoring area are loaded, including:

[0054] Load several target pollutant monitoring data of the target monitoring area, wherein any one of the target pollutant monitoring data includes the historical concentration of the target pollutant, and multiple sets of emission source production record parameters and environmental record parameters of the multiple emission source equipment tag numbers;

[0055] Based on the production record parameters of the multiple emission sources, cluster analysis is performed on the monitoring data of the target pollutants to obtain the clustering results of the monitoring data of the target pollutants.

[0056] The clustering results of the target pollutant monitoring data are traversed, and gray relational analysis is performed on the historical concentration of the target pollutant and the environmental record parameters to obtain environmental attributes with gray relational degree greater than or equal to the relational degree threshold. These attributes are then added to the target pollutant monitoring attributes in the target monitoring area.

[0057] Load the sensing data of the target pollutant's attributes to be monitored in the target monitoring area, and set them as the environmental monitoring parameters.

[0058] In this embodiment of the application, the purpose of step S200 is to accurately screen out key environmental attributes that have a significant impact on the concentration of target pollutants, construct a targeted environmental monitoring parameter system, and provide a scientific basis for subsequent fusion of emissions based on environmental parameters and judgment of the reliability of monitoring concentrations.

[0059] To achieve the above steps, it is first necessary to load several target pollutant monitoring data for the target monitoring area. These target pollutant monitoring data include historical concentrations of the target pollutants, and multiple sets of emission source production record parameters and environmental record parameters for the multiple emission source equipment tag numbers.

[0060] The historical concentration of the target pollutant, such as a particulate matter concentration of 80 μg / m³ at a certain time period. 3 The production record parameters of the multiple emission source equipment tags, such as sintering machine output and converter smelting time, can be obtained through the production control system; the environmental record parameters, such as temperature 25℃, wind speed 3m / s, humidity 60%, etc., are specifically obtained from meteorological stations or air micro-stations.

[0061] Next, based on the production record parameters of the multiple emission sources, cluster analysis needs to be performed on the monitoring data of the target pollutants to obtain the clustering results of the monitoring data of the target pollutants.

[0062] like Figure 2 As shown, in step S200 of this application embodiment, based on the multiple sets of emission source production record parameters, cluster analysis is performed on the monitoring data of the several target pollutants to obtain the clustering results of the target pollutant monitoring data, including:

[0063] Multiple sets of first production record parameters of the multiple emission source equipment tags for obtaining the first target pollutant monitoring data;

[0064] Multiple sets of second production record parameters of the multiple emission source equipment tags for obtaining the second target pollutant monitoring data;

[0065] The first production record parameters of the multiple sets of emission sources and the second production record parameters of the multiple sets of emission sources are compared with the production parameters of the same emission source to obtain the similarity of multiple production parameters.

[0066] When the similarity of the multiple production parameters is greater than or equal to the production parameter similarity threshold, the first target pollutant monitoring data and the second target pollutant monitoring data are added to the same category; otherwise, the first target pollutant monitoring data and the second target pollutant monitoring data are added to the different category.

[0067] In this embodiment of the application, the purpose of the above sub-step in step S200 is to classify the target pollutant monitoring data according to the similarity of production conditions, eliminate the interference of different production loads and process states on the "environmental parameter-pollutant concentration correlation analysis", and lay a unified operating condition basis for subsequent accurate calculation of grey relational degree.

[0068] To achieve the above objectives, the first step is to obtain multiple sets of first production record parameters of the multiple emission source equipment tags of the first target pollutant monitoring data, and to obtain multiple sets of second production record parameters of the multiple emission source equipment tags of the second target pollutant monitoring data.

[0069] Specifically, two sets of data need to be selected from the "Several Target Pollutant Monitoring Data" as the first target pollutant monitoring data and the second target pollutant monitoring data. Then, the production record parameters corresponding to the "Multiple Emission Source Equipment Tag Numbers" for each set are extracted to form the "First Production Record Parameter" and the "Second Production Record Parameter". For example, the first production record parameter is: 500t / h output of sintering machine #1 and 40min smelting time of converter #2; the second production record parameter is: 510t / h output of sintering machine #1 and 38min smelting time of converter #2.

[0070] Here, it is necessary to ensure that the emission source equipment tag numbers of the two sets of production record parameters correspond one-to-one, such as both being No. 1 sintering machine, No. 2 converter, etc., to ensure the effectiveness of the comparison.

[0071] The second step is to compare the first production record parameters of the multiple emission sources with the second production record parameters of the multiple emission sources to obtain the similarity of multiple production parameters.

[0072] In step S200 of this application embodiment, the first set of emission source production record parameters and the second set of emission source production record parameters are compared with the same emission source production parameters to obtain multiple production parameter similarities, including:

[0073] Obtain the first set of production record parameters and the second set of production record parameters for the first emission source;

[0074] Load predefined production parameter attribute deviation thresholds;

[0075] The percentage of attributes whose production parameter deviations between the first group of production record parameters and the second group of production record parameters are less than or equal to the corresponding production parameter attribute deviation threshold is calculated and set as the first emission source production parameter similarity, and added to the plurality of production parameter similarities.

[0076] To achieve this, it is necessary to first obtain the first set of production record parameters and the second set of production record parameters of the first emission source. That is, extract two sets of parameters of the same emission source equipment from the "multiple sets of first production record parameters of multiple emission sources" and the "multiple sets of second production record parameters of multiple emission sources", such as the two sets of parameters of the No. 1 sintering machine.

[0077] The first set of production record parameters consists of the production parameters of the emission source equipment in the monitoring data for the first target pollutant, such as output of 500 t / h, outlet temperature of 180℃, and motor power of 1200 kW; the second set consists of the production parameters of the emission source equipment in the monitoring data for the second target pollutant, such as output of 510 t / h, outlet temperature of 175℃, and motor power of 1180 kW. It is necessary to ensure that the production parameter attribute dimensions of the two sets of production record parameters are consistent to guarantee the effectiveness of the comparison.

[0078] Next, predefined production parameter attribute deviation thresholds need to be loaded. Specifically, based on the emission source equipment type and production process of the "first emission source," the platform's predefined "production parameter attribute deviation thresholds"—that is, the allowable normal fluctuation range of each production parameter of the emission source equipment—need to be loaded. For example:

[0079] Assuming the first emission source is sintering machine #1, the production parameter attribute deviation thresholds can be: output deviation ≤ ±5%, outlet temperature deviation ≤ ±10℃, and motor power deviation ≤ ±8%; the production parameter attribute deviation thresholds can be preset with reference to industry standards or based on expert experience.

[0080] Next, it is necessary to calculate the proportion of production parameter attributes whose deviations between the first group of production record parameters and the second group of production record parameters are less than or equal to the corresponding production parameter attribute deviation thresholds, and set this as the first emission source production parameter similarity, and add it to the multiple production parameter similarities.

[0081] That is, each production parameter attribute of the two sets of production parameters of the first emission source is compared one by one, and the proportion of the number of production parameter attributes with "production parameter attribute deviation ≤ corresponding production parameter attribute deviation threshold" to the total number of production parameter attributes is counted as the production parameter similarity of the emission source equipment.

[0082] In the specific calculation process, for a single production parameter attribute, such as output, the absolute deviation between two sets of production parameter attributes is calculated, such as (|510-500|=10t / h), and it is determined whether the absolute deviation is ≤ the deviation threshold of the production parameter attribute. In the above example, 10t / h≤500×5%=25t / h, so it meets the standard.

[0083] If the first emission source has three production parameter attributes, such as output, outlet temperature, and motor power, and two of these production parameter attributes meet the deviation standards, then the percentage of production parameter attribute deviations meeting the standards is approximately 2 / 3 67%, meaning the similarity of the production parameters of the first emission source is 67%.

[0084] Finally, the above judgment results are summarized into multiple production parameter similarity lists, which serve as one of the bases for subsequent judgments on whether two sets of monitoring data are of the same type.

[0085] The third step is to add the first target pollutant monitoring data and the second target pollutant monitoring data into the same category when the similarity of the multiple production parameters is greater than or equal to the production parameter similarity threshold; otherwise, add the first target pollutant monitoring data and the second target pollutant monitoring data into a different category.

[0086] Specifically, a "production parameter similarity threshold" needs to be set first, such as 80%. This threshold setting must ensure that the operating conditions of all emission source equipment that meet this threshold are similar. Then, it is determined whether all of the "multiple production parameter similarities" meet the production parameter similarity threshold requirements.

[0087] If the similarity of production parameters of all emission source equipment is greater than or equal to the production parameter similarity threshold, such as the similarity of production parameters of No. 1 sintering machine being 85% and the similarity of production parameters of No. 2 converter being 90%, both being greater than or equal to 80%, then the monitoring data of the first target pollutant and the monitoring data of the second target pollutant will be classified as "the same type", representing similar operating conditions.

[0088] If the similarity of production parameters of any emission source equipment is less than the production parameter similarity threshold, such as the similarity of production parameters of No. 1 sintering machine being 70% < 80%, it is classified as "outlier", indicating a large difference in operating conditions.

[0089] Repeat this process to traverse all target pollutant monitoring data, and finally form multiple clusters of monitoring data of the same type of target pollutants under similar operating conditions.

[0090] In step S200 of this application embodiment, the clustering results of the target pollutant monitoring data are traversed, and gray relational analysis is performed on the historical concentration of the target pollutant and the environmental record parameters to obtain environmental attributes with a gray relational degree greater than or equal to the relational degree threshold, including:

[0091] From the clustering results of the target pollutant monitoring data, obtain the first cluster of target pollutant monitoring data, and extract the historical concentrations of multiple target pollutants and multiple sets of environmental record parameters;

[0092] Based on the historical concentrations of the multiple target pollutants, grey relational analysis is performed on the multiple sets of environmental record parameters to obtain the correlation degree of several environmental attributes in the first cluster.

[0093] Until the correlation degree of several environmental attributes of the Nth cluster is obtained, where N represents the number of categories;

[0094] The correlation degree of several environmental attributes in the first cluster up to the correlation degree of several environmental attributes in the Nth cluster is calculated by the mean of the correlation degree of the same attribute, and the fitting correlation degree of several environmental attributes is obtained.

[0095] Select the environmental attributes whose correlation degree is greater than or equal to the correlation degree threshold from the several environmental attributes.

[0096] In step S200 of this application embodiment, the purpose of the above sub-step is to accurately quantify the correlation between environmental attributes and pollutant concentration under unified production conditions, and to ultimately identify environmental attributes that have a significant and stable impact on pollutant concentration through cross-condition average verification, so as to provide a scientific basis for the subsequent construction of a targeted environmental monitoring parameter system.

[0097] To achieve the above objectives, it is first necessary to obtain the first cluster of target pollutant monitoring data from the clustering results of the target pollutant monitoring data, and extract the historical concentrations of multiple target pollutants and multiple sets of environmental record parameters.

[0098] The historical concentration of the target pollutant, such as the particulate matter concentration value of the target pollutant in 100 monitoring data points within the cluster: 80 μg / m³ 3 85μg / m 3 78μg / m 3 The multiple sets of environmental recording parameters correspond to the environmental attribute values ​​of each historical concentration data of the target pollutant, such as wind speed 3m / s, humidity 60%, temperature 25℃, etc., ensuring that the environmental parameters correspond one-to-one with the concentration data.

[0099] Next, based on the historical concentrations of the multiple target pollutants, grey relational analysis is performed on the multiple sets of environmental record parameters to obtain the correlation degree of several environmental attributes in the first cluster; until the correlation degree of several environmental attributes in the Nth cluster is obtained, where N represents the number of categories.

[0100] Specifically, it is necessary to first set the "historical concentration of the target pollutant" as a reference sequence to reflect the trend of pollutant change; and set "multiple sets of values ​​for a certain environmental attribute" as a comparison sequence to reflect the trend of environmental factor change.

[0101] Then, using grey relational algorithms such as the Deng's correlation degree method, the correlation degree values ​​between the comparison sequence and the reference sequence are calculated. For example, the correlation degree between wind speed and particulate matter concentration is 0.85, the correlation degree between humidity and particulate matter concentration is 0.3, and so on. The correlation degree of all environmental attributes such as wind speed, humidity, and temperature in the cluster is calculated one by one to form "correlation degree of several environmental attributes in the first cluster". For example, the cluster contains: wind speed 0.85, humidity 0.3, and temperature 0.6.

[0102] Then, repeat the aforementioned operation, sequentially retrieving the target pollutant monitoring data from the clustering results of the target pollutant monitoring data for the 2nd cluster, the 3rd cluster, ..., the Nth cluster. Calculate the correlation degree of all environmental attributes within each cluster of target pollutant monitoring data, ultimately obtaining the correlation degrees of several environmental attributes for clusters 1 to N. For example, the correlation degrees of several environmental attributes for cluster 1 are: wind speed 0.85, humidity 0.3; for cluster 2, wind speed 0.82, humidity 0.25; ... for cluster N, wind speed 0.80, humidity 0.28.

[0103] Furthermore, it is necessary to calculate the average correlation of several environmental attributes in the first cluster up to the Nth cluster, to obtain the fitting correlation of several environmental attributes, and select the environmental attributes whose fitting correlation is greater than or equal to the correlation threshold.

[0104] Specifically, for each environmental attribute, the mean correlation degree across all clusters needs to be calculated, which is referred to as the "environmental attribute fitting correlation degree". For example, in the above example, the mean correlation degree of wind speed = (0.85 + 0.82 + 0.80) / 3 ≈ 0.82, and the mean correlation degree of humidity = (0.3 + 0.25 + 0.28) / 3 ≈ 0.28;

[0105] Then, it is necessary to refer to industry standards or the actual needs of enterprises to set a correlation threshold, such as 0.6, and select environmental attributes with "correlation degree of environmental attribute fitting ≥ correlation threshold". For example, the correlation degree of environmental attribute fitting of wind speed is 0.82 ≥ 0.6, so it is selected; the correlation degree of environmental attribute fitting of humidity is 0.28 < 0.6, so it is eliminated. Finally, the key environmental attributes to be monitored for the target pollutant are determined.

[0106] Finally, in step S200 of this application embodiment, it is also necessary to load the sensing data of the target pollutant's monitored attributes in the target monitoring area and set them as the environmental monitoring parameters.

[0107] This involves connecting to devices within the target monitoring area that monitor "attributes to be monitored," such as weather stations measuring wind speed or micro-air stations measuring humidity, and collecting sensor data at a preset frequency, such as wind speed of 3 m / s and humidity of 60%. The collected sensor data is then set as the environmental monitoring parameters.

[0108] In step S300 of this application embodiment, based on the multiple emission source device tag numbers and the multiple sets of emission source production parameters, emission prediction is performed to obtain the emission amounts of multiple target pollutants, including:

[0109] Collect the emission source production record parameter set and the target pollutant theoretical emission set of the equipment tag number of the first emission source, and train the emission prediction model of the first emission source;

[0110] The emission prediction model of the first emission source and the device tag number of the first emission source are associated and stored, and added to the target pollutant emission predictor set;

[0111] Based on the device tag numbers of the multiple emission sources, the target pollutant emission predictor set is retrieved, the multiple sets of emission source production parameters are processed, emission prediction is performed, and the emission amounts of multiple target pollutants are obtained.

[0112] In this embodiment of the application, the purpose of step S300 is to build a dedicated emission prediction model for each key emission source device, and to accurately calculate the emission amount of target pollutants in the target monitoring area by integrating the prediction results of multiple devices.

[0113] To achieve the above objectives, it is necessary to collect the emission source production record parameter set of the first emission source equipment tag number and the theoretical emission amount set of the target pollutant, and train the emission prediction model of the first emission source.

[0114] The emission source production record parameter set includes historical production data such as output, temperature, and operating time.

[0115] The theoretical emission set of the target pollutants is a theoretical value calculated based on material balance and emission coefficient method.

[0116] Considering the task type of the first emission source emission prediction model, the first emission source emission prediction model can be built using a random forest regression model.

[0117] In the parameter settings of the emission prediction model for the first emission source, the number of decision trees is 100; the maximum tree depth is 15 layers; the minimum number of leaf node samples is 5; the feature sampling ratio is 0.7; and the random seed is 42.

[0118] For training the emission prediction model of the first emission source, the training data consists of the emission source production record parameter set of the equipment tag number of the first emission source and the theoretical emission set of the target pollutant. At least 1000 sets of valid samples are collected for training. The maximum number of training rounds is set to 50. If the mean square error of the validation set does not decrease for 5 consecutive rounds, training is terminated early. The mean square error (MSE) of the validation set ≤ 0.01 is used as the convergence criterion. The emission prediction model of the first emission source is trained until convergence, thus obtaining the emission prediction model of the first emission source.

[0119] Furthermore, the emission prediction model of the first emission source and the device tag number of the first emission source need to be stored together and added to the target pollutant emission predictor set.

[0120] The "first emission source emission prediction model" that is about to be trained is associated and stored with the corresponding "first emission source equipment tag number", such as sintering machine #1. Then, the above steps are repeated for "multiple emission source equipment tag numbers", such as converter #2, material yard #3, etc., to train emission prediction models one by one and associate and store them. Finally, they are integrated into a "target pollutant emission predictor set". The target pollutant emission predictor set is a collection of dedicated models for all key emission source equipment, realizing a fast mapping between "equipment tag number and emission prediction model".

[0121] Finally, based on the multiple emission source device tag numbers, the target pollutant emission predictor set needs to be retrieved, the multiple sets of emission source production parameters need to be processed, emission prediction needs to be performed, and the emission amounts of multiple target pollutants need to be obtained.

[0122] That is, based on the "multiple emission source equipment tag numbers" to be predicted, such as No. 1 sintering machine, No. 2 converter, etc., the emission prediction models associated with them are retrieved from the target pollutant emission predictor.

[0123] The "multiple emission source production parameters" of each emission source device are then input into the corresponding emission prediction model to obtain the target pollutant emission of a single device, such as the emission of 80 kg / h from sintering machine #1 and 30 kg / h from converter #2. The emission of all emission source devices is then summed to obtain the "multiple target pollutant emissions", such as the total target pollutant emission of 110 kg / h, which serves as the benchmark data for subsequent concentration fitting.

[0124] In step S400 of this application embodiment, based on the environmental monitoring parameters, the emissions of the multiple target pollutants are fused to obtain the fitted concentration of the target pollutants, including:

[0125] Load the emission amounts of multiple target pollutants from multiple emission source equipment tags and the historical environmental monitoring parameters of the target monitoring area, and set them as input data;

[0126] And load the target pollutant detection value of the target monitoring area with a trusted identifier at the same moment, and set it as the output supervisory value;

[0127] Based on the input data and output supervision values, a target pollutant emission fusion generator is trained using machine learning and stored in association with the target monitoring area.

[0128] Based on the target monitoring area, the target pollutant emission fusion device is scheduled to fuse the emission amounts of multiple target pollutants based on the environmental monitoring parameters to obtain the fitted concentration of the target pollutants.

[0129] In this embodiment of the application, the purpose of step S400 is to construct a concentration fitting model that can integrate emissions from multiple emission sources and environmental factors, accurately calculate the theoretical pollutant concentration in the target monitoring area, and provide a quantitative benchmark for subsequent comparison of measured concentrations with fitted concentrations and judgment of the credibility of monitoring data.

[0130] To achieve the above steps, it is first necessary to load the emission amounts of multiple target pollutants from multiple emission source device tag numbers and the historical environmental monitoring parameters of the target monitoring area, and set them as input data.

[0131] The multiple target pollutant emissions refer to the emissions predicted by each piece of equipment in step S300, such as 80 kg / h for sintering machine #1 and 30 kg / h for converter #2. The historical environmental monitoring parameters of the target monitoring area include environmental data such as wind speed of 3 m / s, humidity of 60%, and temperature of 25°C. The above data are aligned by timestamps to form an input data sample. Each input data entry includes the target pollutant emission amount plus the historical environmental monitoring parameters of the target monitoring area.

[0132] Next, the target pollutant detection values ​​of the target monitoring area at the same moment, with a trusted identifier, need to be loaded and set as the output supervisory values. Specifically, this involves retrieving the "target pollutant detection values ​​of the target monitoring area" at the same moment as the input data, and these detection values ​​must carry a "trusted identifier." This can be achieved by cross-validating the target pollutant detection values ​​of the target monitoring area using multiple monitoring devices and manually reviewing and confirming them. Once confirmed to be correct, the values ​​are deemed trustworthy and assigned a trusted identifier.

[0133] Furthermore, based on the input data and output supervision values, machine learning is needed to train a target pollutant emission fusion processor, which is then associated with and stored in the target monitoring area.

[0134] Considering the task type of the target pollutant emission fusion generator, a gradient boosting regression tree model can be used to build the target pollutant emission fusion generator.

[0135] In the parameter settings of the target pollutant emission fusion generator, the number of weak learners was set to 200; the learning rate was 0.05; the maximum tree depth was 8 layers; the minimum number of sample splits was 10; and the subsampling ratio was 0.8. The training sample data consisted of the target pollutant emissions from the multiple emission source equipment locations and environmental monitoring parameters. At least 1500 valid samples were collected and divided into training and validation sets in a 7:3 ratio for training the target pollutant emission fusion generator.

[0136] For training the target pollutant emission fusion generator, the root mean square error (RMSE) of the validation set (in micrograms per cubic meter) ≤ 5 was used as the core convergence criterion. The maximum number of training rounds was set to 300. If the RMSE of the validation set did not decrease for 10 consecutive rounds, training was terminated early. The target pollutant emission fusion generator was trained using the above method until convergence.

[0137] Finally, based on the target monitoring area, the target pollutant emission fusion device needs to be scheduled to fuse the emission amounts of the multiple target pollutants based on the environmental monitoring parameters to obtain the fitted concentration of the target pollutants.

[0138] That is, based on the "target monitoring area" to be analyzed, retrieve the "target pollutant emission fusion device" associated with it;

[0139] Then, the "emissions of multiple target pollutants" obtained in step S300 and the "environmental monitoring parameters" obtained in step S200 are input into the target pollutant emission fusion unit;

[0140] Based on the learning patterns developed during training, the target pollutant emission fusion processor outputs a "fitted concentration of the target pollutant" that comprehensively considers both emissions and environmental factors, such as 82 μg / m³. 3 This serves as a benchmark for subsequently judging the reliability of the measured concentration.

[0141] In step S500 of this application embodiment, when the concentration deviation between the monitored concentration of the target pollutant and the fitted concentration of the target pollutant is less than or equal to the concentration deviation threshold, the monitored concentration of the target pollutant is marked as reliable; otherwise, the monitored concentration of the target pollutant is marked as abnormal.

[0142] The concentration deviation threshold can be set with reference to the company's historical data, such as 10%.

[0143] Then, the concentration deviation between the monitored concentration of the target pollutant and the fitted concentration of the target pollutant is calculated. The calculation method is: concentration deviation = |monitored concentration of target pollutant - fitted concentration of target pollutant| / fitted concentration of target pollutant × 100%.

[0144] If the concentration deviation is less than or equal to the concentration deviation threshold, a "trusted identifier" is added to the monitored concentration and stored in the trusted database; if the concentration deviation exceeds the concentration deviation threshold, an "abnormal identifier" is added to the monitored concentration, an alarm is triggered, and an abnormal log is recorded.

[0145] Optionally, when the monitoring concentration of the target pollutant is identified as abnormal, an alarm mechanism can be triggered to notify maintenance personnel through the platform interface, SMS, or email; or a maintenance work order can be automatically generated to guide staff to calibrate, repair, or replace the sensor or equipment from which the abnormal data originates.

[0146] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0147] Those skilled in the art will understand that embodiments of the present invention can provide methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0148] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0149] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0150] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0151] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0152] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for operation and maintenance management of multi-source heterogeneous monitoring data of an environmental management and control integrated platform, characterized in that, The method comprises the following steps: loading a plurality of emission source device numbers of a target pollutant of a target monitoring area and a plurality of sets of emission source production parameters; loading environmental monitoring parameters and a target pollutant monitoring concentration of the target monitoring area, wherein the environmental monitoring parameters of the target monitoring area comprise: loading a plurality of target pollutant monitoring data of the target monitoring area, wherein any one of the plurality of target pollutant monitoring data comprises a target pollutant historical record concentration, a plurality of sets of emission source production record parameters and environmental record parameters of the plurality of emission source device numbers; based on the plurality of sets of emission source production record parameters, performing cluster analysis on the plurality of target pollutant monitoring data to obtain target pollutant monitoring data clustering results; iterating through the target pollutant monitoring data clustering results, performing grey correlation analysis on the target pollutant historical record concentration and the environmental record parameters to obtain environmental attributes with a grey correlation degree greater than or equal to a correlation degree threshold, and adding the environmental attributes to target pollutant monitoring attributes of the target monitoring area; loading perception data of the target pollutant monitoring attributes of the target monitoring area as the environmental monitoring parameters; based on the plurality of emission source device numbers and the plurality of sets of emission source production parameters, performing emission prediction to obtain a plurality of target pollutant emission amounts, comprising: collecting a set of emission source production record parameters and a set of target pollutant theoretical emission amounts of a first emission source device number, and training a first emission source emission prediction model; storing the first emission source emission prediction model in association with the first emission source device number, and adding the first emission source emission prediction model to a set of target pollutant emission predictors; based on the plurality of emission source device numbers, calling the set of target pollutant emission predictors, processing the plurality of sets of emission source production parameters, performing emission prediction, and obtaining a plurality of target pollutant emission amounts; based on the environmental monitoring parameters, fusing the plurality of target pollutant emission amounts to obtain a target pollutant fitting concentration, comprising: loading a plurality of target pollutant emission amounts of a plurality of emission source device numbers and environmental monitoring historical parameters of a target monitoring area as input data; loading a target monitoring area target pollutant detection value at the same time with a trusted identifier as an output supervision value; based on the input data and the output supervision value, training a target pollutant emission amount fuser using machine learning, and storing the target pollutant emission amount fuser in association with the target monitoring area; based on the target monitoring area, scheduling the target pollutant emission amount fuser, fusing the plurality of target pollutant emission amounts based on the environmental monitoring parameters, and obtaining a target pollutant fitting concentration; when a concentration deviation between the target pollutant monitoring concentration and the target pollutant fitting concentration is less than or equal to a concentration deviation threshold, performing trusted identification on the target pollutant monitoring concentration, otherwise, performing abnormal identification on the target pollutant monitoring concentration.

2. The method of claim 1, wherein, Loading a plurality of emission source device numbers of a target pollutant of a target monitoring area comprises: obtaining an initial device number set of an emission pollutant with the target pollutant; iterating through the initial device number set to count a detection support degree set of the target pollutant in the target monitoring area; Extracting initial device bit numbers with detection support greater than or equal to a support threshold from the initial device bit number set as the plurality of emission source device bit numbers.

3. The method of claim 2, wherein, Traversing the initial device bit number set, a detection support set of the target pollutant in the target monitoring area is counted, including: Extracting a first device bit number from the initial device bit number set; Loading the target pollutant historical detection position set in which the first device bit number is in a working state and other initial device bit numbers are in a shutdown state; Counting the proportion of the number of occurrences of the target monitoring area in the target pollutant historical detection position set as the first device bit number detection support and adding it to the detection support set.

4. The method of claim 1, wherein, Based on the plurality of sets of emission source production record parameters, clustering analysis is performed on the target pollutant monitoring data to obtain a target pollutant monitoring data clustering result, including: Obtaining a plurality of sets of emission source first production record parameters of the plurality of emission source device bit numbers of the first target pollutant monitoring data; Obtaining a plurality of sets of emission source second production record parameters of the plurality of emission source device bit numbers of the second target pollutant monitoring data; Performing same-emission-source production parameter comparison on the plurality of sets of emission source first production record parameters and the plurality of sets of emission source second production record parameters to obtain a plurality of production parameter similarities; When the plurality of production parameter similarities are all greater than or equal to a production parameter similarity threshold, the first target pollutant monitoring data and the second target pollutant monitoring data are added to the same class, otherwise, the first target pollutant monitoring data and the second target pollutant monitoring data are added to different classes.

5. The method of claim 4, wherein, Performing same-emission-source production parameter comparison on the plurality of sets of emission source first production record parameters and the plurality of sets of emission source second production record parameters to obtain a plurality of production parameter similarities, including: Obtaining a first set of production record parameters and a second set of production record parameters of a first emission source; Loading a predefined production parameter attribute deviation threshold; Counting the proportion of attributes of the first set of production record parameters and the second set of production record parameters whose production parameter deviations are less than or equal to the corresponding production parameter attribute deviation threshold as the first emission source production parameter similarity and adding it to the plurality of production parameter similarities.

6. The method of claim 1, wherein, Traversing the target pollutant monitoring data clustering result, performing grey correlation degree analysis on the target pollutant historical record concentration and the environmental record parameter to obtain an environmental attribute with a grey correlation degree greater than or equal to a correlation degree threshold, including: From the target pollutant monitoring data clustering result, obtaining first cluster target pollutant monitoring data, extracting a plurality of target pollutant historical record concentrations and a plurality of sets of environmental record parameters; Based on the plurality of target pollutant historical record concentrations, performing grey correlation degree analysis on the plurality of sets of environmental record parameters to obtain a first cluster of several environmental attribute correlations; Until the Nth cluster of several environmental attribute correlations are obtained, where N represents the number of categories; The same attribute correlation degrees of the first cluster of several environmental attribute correlation degrees to the Nth cluster of several environmental attribute correlation degrees are calculated to obtain several environmental attribute fitting correlation degrees; and the environmental attribute with the environmental attribute fitting correlation degree greater than or equal to a correlation degree threshold is selected.

Images