Enterprise esg environmental index evaluation method based on internet of things data
By coding and correlating the equipment, processes, and ESG environmental indicators throughout the entire production process of an enterprise, the problem of ambiguous correlation between production factors and environmental indicators has been solved, enabling precise tracing and efficient handling of environmental indicator anomalies.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- FUZHOU PLANNING DESIGN & RES INST
- Filing Date
- 2026-04-27
- Publication Date
- 2026-07-03
AI Technical Summary
In existing technologies, the correlation between production factors and environmental indicators is vague, making it impossible to accurately match specific production processes. The judgment of abnormal environmental indicators lacks scientific rigor, the source tracing efficiency is low, and it is impossible to distinguish between real anomalies and accidental fluctuations, leading to misjudgments, missed judgments, and indiscriminate investigations.
By coding the equipment, processes, and ESG environmental indicators throughout the entire production process of an enterprise, establishing a coding ledger, constructing correlation analysis groups and calculating correlation coefficients, building a three-level correlation mapping table, and combining working condition thresholds and clustering algorithms, the determination and tracing of real anomalies can be achieved.
It achieves standardized binding of production factors with ESG environmental indicators, accurately matches production processes, distinguishes between real anomalies and accidental fluctuations, improves traceability efficiency, provides clear anomaly root cause location reports, and supports operation and maintenance decisions.
Smart Images
Figure CN122089173B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of intelligent industrial environmental management technology, specifically involving a method for assessing enterprise ESG environmental indicators based on Internet of Things (IoT) data. Background Technology
[0002] The Internet of Things (IoT) technology has been widely applied in the field of industrial ESG environmental indicator assessment, becoming an important tool for enterprise environmental management; however, existing assessment technologies still have significant shortcomings, such as unclear correlation between production factors and environmental indicators, lack of scientific basis for anomaly judgment, and low efficiency in source tracing and handling. Therefore, this invention urgently needs to solve the following technical problems:
[0003] There is no standardized correlation system between production factors and environmental indicators, the correspondence between the two is vague, changes in environmental indicators cannot be accurately matched with specific production links, and there is a lack of quantitative correlation analysis basis, which lays a basic obstacle to tracing the source of indicator anomalies.
[0004] The environmental indicator anomaly judgment uses a single threshold standard, which is not adapted to the differences in different production conditions of enterprises, and cannot effectively distinguish between real anomalies and accidental fluctuations, which is prone to misjudgment and omission. Furthermore, it does not standardize the classification of anomaly types.
[0005] The source tracing and investigation after environmental indicators are abnormal lacks specificity, and the investigation is often conducted on all production factors indiscriminately, which is inefficient and cannot accurately pinpoint the specific operational parameters that cause the abnormality. In addition, there is no standardized form of abnormal information push, which is not conducive to rapid on-site handling. To address this, we propose an enterprise ESG environmental indicator assessment method based on IoT data. Summary of the Invention
[0006] The purpose of this invention is to provide a method for evaluating enterprise ESG environmental indicators based on Internet of Things (IoT) data, in order to solve the problems mentioned in the background art.
[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for assessing enterprise ESG environmental indicators based on IoT data, comprising:
[0008] Step 1: Code the equipment, processes, and ESG environmental indicators of the entire production process of the enterprise and establish a coding ledger. Collect corresponding IoT data, build correlation analysis groups and calculate correlation coefficients to screen strong correlation groups; calculate the single-factor contribution of strong correlation groups to screen core correlation groups, build a three-level correlation mapping table of equipment-process-ESG environmental indicators, and establish a dynamic update and traceability mechanism.
[0009] Step 2: Establish a normal threshold ledger for ESG environmental indicators under different working conditions, organize historical abnormal event records of ESG environmental indicators under different working conditions and draw line charts of indicator changes, extract morphological features from the initial segments, and integrate them into an abnormal feature library of enterprise ESG environmental indicators after clustering and classifying abnormal patterns; monitor ESG environmental indicators in real time and determine them as real anomalies.
[0010] Step 3: For ESG environmental indicators confirmed as genuine anomalies, clarify their codes and operating conditions, retrieve the three-level association mapping table of equipment-process-ESG environmental indicators to generate a process and equipment traceability list, compare the process and operating parameters during the abnormal period, filter abnormal processes, equipment and operating parameter sets, integrate relevant codes and parameters to form an ESG environmental indicator anomaly root cause location report, and push it to the operation and maintenance terminal.
[0011] Preferably, the specific process of coding and establishing a coding ledger for equipment, processes, and ESG environmental indicators throughout the entire production process of an enterprise is as follows:
[0012] The equipment coding adopts the form of equipment type code plus serial number. The equipment type code is divided according to the function of the production equipment, and the serial number is assigned according to the equipment deployment order. Each piece of equipment corresponds to a unique code.
[0013] The process coding adopts the form of process flow code plus step code. The flow code corresponds to the entire production process sequence, and the step code corresponds to the sub-step within a single process. Each process and sub-step corresponds to a unique code.
[0014] ESG environmental indicator coding adopts the form of indicator category code plus indicator sub-code. The category code corresponds to the first-level indicators of resource consumption and pollutant emissions, and the sub-code corresponds to the specific third-level environmental indicators. Each environmental indicator has a unique code.
[0015] Organize the equipment, processes, and core ESG environmental indicators throughout the entire production process, assign unique codes according to rules, use the codes as unique identifiers, organize the mapping relationship between codes and corresponding objects, and construct a coding ledger for equipment, processes, and environmental indicators.
[0016] Preferably, the specific process of collecting corresponding IoT data, constructing correlation analysis groups, and calculating correlation coefficients to screen for strongly correlated groups is as follows:
[0017] Real-time collection of equipment operation data, process parameters, and ESG environmental index data corresponding to each piece of equipment, each process, and each ESG environmental index in the enterprise's production scenario;
[0018] Based on the historical data of various objects, a preset number of time-series data samples of the same duration are collected. According to the principle of pairing two to one and unique subject, three types of correlation analysis groups are constructed: equipment-environment indicators, process-environment indicators, and equipment-process.
[0019] The correlation coefficient between the two types of data within each group was calculated using the Pearson correlation coefficient method, and the data were compiled into a log of the association analysis groups.
[0020] Set a threshold for the correlation coefficient, mark groups that meet the threshold as strongly correlated groups, and remove groups that do not meet the threshold as weakly correlated groups.
[0021] Preferably, the specific process of calculating the single-factor contribution of strongly correlated groups to screen core correlated groups, constructing a three-level correlation mapping table of equipment-process-ESG environmental indicators, and establishing a dynamic update and traceability mechanism is as follows:
[0022] The single-factor contribution of each strongly associated group was calculated using the single-factor variable method. The results were updated to the association analysis group ledger. A single-factor contribution threshold was preset. Strongly associated groups that did not meet the threshold were removed. Those that met the threshold were marked as core association groups and a core association ledger was formed.
[0023] Based on the equipment-process-environment indicator coding ledger and the core association ledger, a three-level association mapping table of equipment-process-ESG environmental indicators is constructed to clarify the three-dimensional core correspondence relationship.
[0024] Real-time monitoring of enterprise production scenarios triggers an update of the relationship if any of the following occurs: equipment addition, scrapping, replacement, process addition or deletion, process adjustment, ESG environmental indicator definition, or monitoring standard change. The system updates the coding ledger, re-collects data, updates the core relationship ledger, and revises the three-level relationship mapping table according to the rules. Each version is retained and information is recorded to form a traceability archive of relationship updates.
[0025] Preferably, the specific process of establishing a ledger of normal thresholds for ESG environmental indicators under different operating conditions, compiling a historical record set of abnormal events of ESG environmental indicators under different operating conditions, and drawing line charts of indicator changes is as follows:
[0026] Retrieve the normal threshold ranges of various ESG environmental indicators under different working conditions of the enterprise, and compile them into a normal threshold ledger of ESG environmental indicators under different working conditions.
[0027] Retrieve historical abnormal event data records of various ESG environmental indicators under different working conditions, and organize them into a set of historical abnormal event records by working condition and indicator code.
[0028] For each abnormal event in the historical abnormal event record set for different work conditions and indicators, a two-dimensional coordinate system of time and indicator value is constructed.
[0029] Substitute the time-series data points corresponding to the abnormal event into a two-dimensional coordinate system, and connect each data point sequentially with line segments according to the time order to obtain a complete line chart of the changes in ESG environmental indicators corresponding to the abnormal event.
[0030] Preferably, the specific process of extracting morphological features from the initial fragments, clustering them to classify abnormal patterns, and then integrating them to form an abnormal feature library of enterprise ESG environmental indicators is as follows:
[0031] Set a uniform feature extraction time window duration, and extract a time series data segment of this duration based on the start time of the anomaly occurrence of each historical anomaly event to generate a line chart of the initial segment of the historical anomaly. If the total duration of the event is shorter than this duration, the full time series data is taken.
[0032] Extract the morphological features of line charts of all historical anomalies under the same working conditions and ESG environmental indicators. Use clustering algorithm to calculate the morphological similarity between the lines of each anomaly event. Classify the anomalies with morphological similarity greater than or equal to the preset similarity threshold into the same anomaly pattern.
[0033] Based on the clustering results, an abnormal pattern and feature comparison table of ESG environmental indicators by work condition and indicator was compiled, and combined with the normal threshold ledger of ESG environmental indicators by work condition, an abnormal feature library of enterprise ESG environmental indicators was formed.
[0034] Preferably, the specific process for real-time monitoring of ESG environmental indicators and determining them as genuine anomalies is as follows:
[0035] Real-time collection of monitoring data of ESG environmental indicators under various working conditions of the enterprise, input into the enterprise's ESG environmental indicator abnormal feature library and compared with the normal threshold range of the corresponding working conditions and indicators; if the threshold is exceeded, the initial abnormality start time is marked and relevant information is recorded.
[0036] Using the initial time of the anomaly as a benchmark, time-series data with a uniform feature extraction time window duration are extracted, and a real-time initial anomaly segment line graph is drawn. Morphological features are extracted and a real-time standardized morphological feature vector is constructed. The cosine similarity method is used to calculate the similarity between the feature vector and the corresponding anomaly pattern feature vector. The maximum value is compared with the preset similarity threshold. If the threshold is met, it is confirmed as a real anomaly. If the threshold is not met, it is judged as an accidental fluctuation and only a warning record is made.
[0037] Preferably, for ESG environmental indicators confirmed as genuine anomalies, the specific process of clarifying their codes and operating conditions, and retrieving the three-level association mapping table of equipment-process-ESG environmental indicators to generate a process and equipment traceability list is as follows:
[0038] For ESG environmental indicators that are confirmed to be genuine anomalies, clarify their indicator codes and corresponding operating conditions, retrieve the three-level association mapping table of equipment-process-ESG environmental indicators, determine the associated process codes, and form a process traceability list.
[0039] Collect process parameters for abnormal periods of processes in the list, compare them one by one with preset threshold ranges under the same working conditions, filter out abnormal processes and mark abnormal process codes.
[0040] Based on the abnormal process code, the above three-level association mapping table is retrieved again to determine the associated equipment code of the abnormal process and form an equipment traceability list.
[0041] Preferably, the specific process of comparing process and operating parameters during abnormal periods, screening abnormal procedures, equipment, and sets of operating parameters, integrating relevant codes and parameters to form an ESG environmental indicator anomaly root cause location report, and pushing it to the operation and maintenance terminal is as follows:
[0042] The system collects the operating data of each device in the equipment traceability list during abnormal periods, compares it one by one with the preset threshold range under the same working conditions, filters out abnormal devices and marks them with a unique abnormal device code.
[0043] Based on the abnormal equipment code, the core operating parameters of the abnormal equipment during the abnormal period are extracted, compared with the exclusive preset threshold under the same working conditions, and the parameters exceeding the threshold are selected and organized into a set of abnormal operating parameters.
[0044] The abnormal operation parameter set is associated with the abnormal equipment code, abnormal process code, and abnormal ESG environmental indicator code one by one, and an ESG environmental indicator abnormality root cause location report is generated and sent to the operation and maintenance personnel terminal to guide the on-site abnormality handling.
[0045] Compared with the prior art, the beneficial effects of the present invention are:
[0046] (1) This enterprise ESG environmental indicator assessment method based on IoT data establishes a standardized coding ledger for equipment, processes and ESG environmental indicators throughout the entire production process, combines the Pearson correlation coefficient method and the single-factor variable method to screen core related groups, builds a three-level correlation mapping table and establishes a dynamic update traceability mechanism, realizes the standardized and quantitative effective binding of production factors and ESG environmental indicators, and enables changes in environmental indicators to be effectively matched with specific production links, solves the problem of ambiguous correlation between the two, and provides a reliable quantitative correlation basis for anomaly tracing.
[0047] (2) The enterprise ESG environmental indicator assessment method based on IoT data builds a normal threshold ledger of ESG environmental indicators for different working conditions, extracts the line shape features of historical abnormal indicators and constructs an abnormal feature library through clustering algorithm. It adopts a dual judgment logic of threshold comparison and cosine similarity method, which breaks the limitation of single threshold judgment, adapts to the differences of different production working conditions of enterprises, can effectively distinguish between real abnormalities and accidental fluctuations, avoid misjudgment and omission, and realizes the standardized classification of abnormal types of ESG environmental indicators, which clarifies the feature direction for abnormal source tracing.
[0048] (3) This enterprise ESG environmental indicator assessment method based on IoT data generates a process and equipment traceability list by reverse deducing from real abnormal ESG environmental indicators based on a three-level correlation mapping table. It compares the process parameters, operating parameters and core operating parameters in sequence with preset thresholds under the same working conditions to screen layer by layer. It integrates abnormal information to form a standardized abnormal root cause location report and pushes it to the operation and maintenance terminal. This allows the abnormal traceability to be carried out around the core correlation elements, greatly improving the traceability efficiency, effectively locking the specific operating parameter causes of the abnormal indicators, providing clear decision-making basis for operation and maintenance personnel, and realizing the closed-loop management of the entire process of "discovery-traceability-disposal" of abnormal ESG environmental indicators. Attached Figure Description
[0049] Figure 1 This is a flowchart of the present invention. Detailed Implementation
[0050] 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.
[0051] Example 1;
[0052] Please see Figure 1 This invention provides a method for evaluating enterprise ESG environmental indicators based on Internet of Things (IoT) data, including:
[0053] Step 1: Encode the equipment, processes, and ESG environmental indicators throughout the enterprise's entire production process and establish a coding ledger. Collect corresponding IoT data, construct correlation analysis groups, and calculate correlation coefficients to screen for strongly correlated groups. Calculate the single-factor contribution of strongly correlated groups to screen for core correlated groups. Construct a three-level correlation mapping table of equipment-process-ESG environmental indicators, and establish a dynamic update and traceability mechanism. The specific process is as follows:
[0054] Obtain information on equipment, processes, and ESG environment throughout the entire enterprise production process, and classify and number the equipment, processes, and ESG environment indicators as follows:
[0055] Equipment coding: The coding system uses a "equipment type code + serial number" format. The equipment type code is based on the function of the production equipment, and the serial number is assigned sequentially according to the equipment deployment order. Each piece of equipment has a unique code, marked as follows: i = 1, 2, ..., n; n is the total number of devices;
[0056] Process coding: The process uses a "process flow code + step code" format. The flow code corresponds to the entire production process sequence, and the step code corresponds to a sub-step within a single process. Each process and sub-step has a unique code, denoted as... j = 1, 2, ..., m; m is the total number of processes;
[0057] ESG environmental indicator coding adopts the format of "indicator category code + indicator sub-code". The category code corresponds to primary indicators such as resource consumption and pollutant emissions, while the sub-code corresponds to specific tertiary environmental indicators. Each environmental indicator has a unique code, denoted as . k = 1, 2, ..., p; p is the total number of environmental indicators;
[0058] The entire production process of the enterprise is sorted out, including all production equipment, production processes (including sub-processes) and core ESG environmental indicators. The above objects are assigned unique codes according to the preset coding rules. The codes serve as unique identifiers for each object. Then, the mapping relationship between the codes and the corresponding objects is sorted out to build a coding ledger of equipment-process-environment indicator.
[0059] For enterprise production scenarios, real-time collection of IoT data corresponding to various equipment, processes, and ESG environmental indicators is performed, specifically including:
[0060] Equipment operation data The corresponding code is The operating parameters of the equipment at time t include, but are not limited to: speed, load, running time, power, etc.
[0061] Process parameters The corresponding code is The process parameters at time t include, but are not limited to: reaction temperature, reaction pressure, feed rate, residence time, etc.
[0062] ESG Environmental Indicators Data The corresponding code is The environmental indicators are the monitoring data at time t, including but not limited to unit energy consumption, pollutant concentration, carbon emissions, etc.
[0063] Based on the data collected from various devices Each process and various ESG environmental indicators Collect a preset number of time-series data samples of the same duration, corresponding to historical data.
[0064] The three types of data with time-series samples are used to construct three association analysis groups according to the principle of pairwise pairing and unique subject. Each group contains several specific combinations corresponding to a single subject, as follows:
[0065] Equipment-Environmental Indicator Association Group: composed of equipment running data ESG environmental indicators Monitoring data The composition, each specific combination corresponds to a unique device. - ESG environmental indicators The total number of such association analysis groups is n×p.
[0066] Process-Environmental Indicator Association Group: composed of processes process parameters ESG environmental indicators Monitoring data Each specific combination corresponds to a unique process. - ESG environmental indicators The total number of such association analysis groups is m×p;
[0067] Equipment-process association group: composed of equipment runtime data With process process parameters The composition, each specific combination corresponds to a unique device. - Process The total number of such association analysis groups is n×m;
[0068] For any association analysis group, historical time series data samples of the two corresponding data types in the group are collected, and after normalization and dimensionless processing, the correlation coefficient between the two data types in the group is calculated using the Pearson correlation coefficient method.
[0069] The indexes, corresponding subject information, and calculated correlation coefficients of all association analysis groups are compiled into an association analysis group ledger.
[0070] For any association analysis group in the association analysis group ledger, a preset correlation coefficient threshold is set. If the correlation coefficient of the association analysis group is greater than or equal to the corresponding preset threshold, the association analysis group is marked as a strong association group; otherwise, it is marked as a weak association group and removed.
[0071] For any strongly correlated group, after normalizing and dedimensionalizing the data, its single-factor contribution is calculated using the single-factor variable method, i.e., the parameter to be adjusted is defined as follows: The affected parameters associated with the parameters to be adjusted are: Using the formula: The single-factor contribution of this strongly associated group was obtained. ;
[0072] The single-factor contribution of each strongly correlated group is synchronously updated to the correlation analysis group ledger, forming a complete ledger containing subject code, correlation coefficient, and single-factor contribution. A single-factor contribution threshold is preset, and strongly correlated groups with single-factor contribution less than or equal to the preset threshold are removed. Strongly correlated groups that meet the requirements are retained and marked as core correlation groups, forming a core correlation ledger.
[0073] Based on the equipment-process-environment indicator coding ledger and the core association ledger, a three-level association mapping table of equipment-process-ESG environmental indicators is constructed, clarifying the three-dimensional core correspondence relationship, specifically as follows:
[0074] 1. Equipment Process ESG environmental indicators The unique correspondence between the codes;
[0075] 2. Equipment operating data Process parameters ESG environmental indicator data The matching relationship of IoT data dimensions;
[0076] 3. Correlation coefficients and single-factor contribution values for each core strongly correlated group;
[0077] Real-time monitoring of enterprise production scenarios will trigger an update of the associated relationships if any of the following occurs: equipment addition / scrap / replacement, process addition / deletion / process adjustment, or ESG environmental indicator definition / monitoring standard change. The update operation is as follows:
[0078] New production entities are assigned unique codes according to the original rules, and invalid entities are marked with invalid codes, and the coding ledger is updated synchronously. A preset number of IoT time-series data samples are collected again to supplement data for new entities and remove data for invalid entities. The association analysis groups are reconstructed according to the original method, correlation coefficients are calculated, strong and weak association groups are marked, and core strong association groups are screened, and the core association ledger is updated. Based on the updated two types of ledgers, the three-level association mapping table is revised. After the update is completed, the triggering conditions, update time, and core adjustment content are recorded, and the versions of the coding ledger, core association ledger, and three-level association mapping table before and after the update are retained to form an association relationship update traceability archive.
[0079] It should be noted that by constructing standardized coding rules for equipment, processes, and ESG environmental indicators and establishing a coding ledger, various production factors and environmental indicators are given unique identification, realizing the basic binding of production factors and ESG environmental indicators, solving the problem of ambiguous correspondence between production links and environmental indicators, providing a unified and standardized coding benchmark for subsequent anomaly monitoring and precise traceability, and improving the traceability accuracy and standardization of assessment methods from the bottom up.
[0080] By combining the Pearson correlation coefficient method with the single-factor variable method to conduct IoT data correlation analysis, strong correlation groups are screened from massive time series data and core correlation groups are accurately identified. This enables the quantitative correlation between production factors and ESG environmental indicators, eliminates weak correlation interference factors, and eliminates the need for indiscriminate analysis of all production factors in subsequent anomaly assessment and source tracing, thus greatly improving the efficiency and scientific nature of assessment and analysis.
[0081] By constructing a three-level correlation mapping table of equipment-process-ESG environmental indicators based on coding ledgers and core correlation ledgers, the three-dimensional core correspondence relationship of coding, data dimensions, coefficients and contribution is clarified, and a core bridge between production operation data and ESG environmental indicator assessment is built. This enables precise binding of production factors and environmental indicators, allowing changes in ESG environmental indicators to directly correspond to specific production links, and providing key technical support for subsequent full-process anomaly assessment.
[0082] By establishing a trigger-based dynamic update and traceability mechanism, the system adapts to various changes in equipment, processes, and ESG environmental indicators standards, synchronously updates various ledgers and revises the three-level association mapping table, and retains full-version archives to achieve traceability of association relationships. This allows the association system to dynamically adapt to changes in enterprise production and regulatory standards, ensuring the long-term applicability of the assessment method, while meeting the traceability requirements of enterprise operations and environmental supervision, and enhancing the practical application value of the assessment method.
[0083] Step Two: Establish a ledger of normal thresholds for ESG environmental indicators under different operating conditions; compile historical records of abnormal events related to ESG environmental indicators under different operating conditions and draw line charts of indicator changes; extract morphological features from initial segments; and integrate these features after clustering to form an abnormal feature library for enterprise ESG environmental indicators; monitor ESG environmental indicators in real time and determine those that are genuine anomalies. The specific process is as follows:
[0084] Based on enterprise production and operation logs, industry environmental supervision logs, and equipment / process maintenance logs, the normal threshold ranges of various ESG environmental indicators under different working conditions are retrieved, and the normal threshold ledgers of ESG environmental indicators under different working conditions are compiled based on the retrieval results.
[0085] Retrieve various ESG environmental indicators under different working conditions throughout the entire enterprise production process. Historical abnormal event data records, categorized by working conditions and indicators. The records are organized hierarchically to form historical anomaly event records categorized by operating condition and indicator. Each record in the record set corresponds to an ESG environmental indicator anomaly set, including: operating condition code, indicator code, anomaly start time, anomaly end time, and all anomaly events. Time sequence data, associated equipment codes, associated process codes, and anomaly handling results;
[0086] For each abnormal event in the historical abnormal event record set for different working conditions and indicators, the time is used as the horizontal axis and the ESG environmental indicator value is used as the vertical axis. Construct a two-dimensional coordinate system of time and index value with the vertical axis as the ordinate;
[0087] The corresponding abnormal event Substitute the time-series data points into a two-dimensional coordinate system, and connect each data point sequentially with line segments according to the time order to obtain the abnormal event (corresponding to...). The corresponding ESG environmental indicators are shown in the complete line graph.
[0088] Set a uniform feature extraction time window duration TQ. Based on the start time of the anomaly of each historical anomaly event, extract time series data segments with a duration of TQ to generate a line chart of the initial segment of the historical anomaly. If the total duration of a historical anomaly event is less than T, then the entire time series data of the event is used as the initial segment line chart.
[0089] For all historical anomaly initial segment line charts under the same working conditions and the same ESG environmental indicators, the core clustering basis is the line shape characteristics. The corresponding shape characteristics are extracted for each line, including: change slope (overall slope, stage slope), number and location of inflection points, fluctuation amplitude, peak position and peak size, total duration of anomaly, and percentage of time exceeding the threshold.
[0090] The K-means clustering algorithm is used to calculate the morphological similarity between the extracted line shape features and abnormal event lines. Abnormal lines with morphological similarity greater than or equal to a preset similarity threshold are classified into the same abnormal pattern.
[0091] Abnormal patterns include: sudden exceedance, continuous rise, abnormal fluctuation, and slow deviation;
[0092] Based on the clustering results, a table comparing anomaly patterns and characteristics of ESG environmental indicators by operating condition and indicator was compiled. The table includes operating condition codes and indicator codes. Abnormal pattern coding and typical abnormal manifestations; combined with the normal threshold ledger for different working conditions, an abnormal feature library of enterprise ESG environmental indicators is formed; the library achieves precise correlation between "working condition - indicator - normal threshold - abnormal pattern - morphological feature";
[0093] Real-time collection of ESG environmental indicator monitoring data under various working conditions of the enterprise, and inputting it into the enterprise's ESG environmental indicator anomaly feature library for real-time comparison with the normal threshold range of the corresponding working conditions and indicators. If the indicator value exceeds the corresponding threshold range, the moment is marked as the initial anomaly start moment, and the indicator value, the extent of exceeding the standard and the corresponding working condition code are recorded simultaneously.
[0094] Based on the initial anomaly start time, time series data of duration TQ are extracted in real time, a real-time initial anomaly segment line graph is plotted, the morphological features of the line are extracted, and a real-time standardized morphological feature vector is constructed.
[0095] The cosine similarity method is used to calculate the similarity between the real-time feature vector and the feature vectors of various abnormal modes of the corresponding working conditions and indicators in the abnormal feature library (the parameters are normalized and dimensionless before calculation) to obtain the similarity value corresponding to each abnormal mode.
[0096] Select the maximum value among all similarity values. If the maximum value is greater than or equal to the preset similarity threshold, then mark the abnormal pattern corresponding to the maximum value as the abnormal type of the indicator and confirm it as a real abnormality.
[0097] If the maximum value is less than the preset similarity threshold, it is determined to be an accidental fluctuation and will not be included in the source tracing scope, but only a warning record will be made.
[0098] It should be noted that by establishing a normal threshold ledger for ESG environmental indicators under different working conditions, the limitations of single threshold assessment are broken, and the actual operating scenarios of different production conditions of enterprises are adapted to make the abnormal judgment of ESG environmental indicators fit the actual production, thereby improving the rationality and scenario adaptability of the initial judgment of abnormal indicators from the bottom up.
[0099] By organizing historical anomaly event records by work condition and indicator and drawing line charts of indicator changes, the system systematically collects all anomaly information and visualizes indicator changes, providing complete and standardized historical data support for morphological feature extraction and anomaly pattern classification, making anomaly analysis traceable.
[0100] By extracting the features of broken lines and using the K-means clustering algorithm to classify abnormal patterns, an abnormal feature library of enterprise ESG environmental indicators is constructed to achieve multi-dimensional and accurate correlation. This enables the standardized and patterned classification of ESG environmental indicator anomalies, clarifies the feature orientation for subsequent anomaly tracing, and improves the systematicness and pertinence of anomaly analysis.
[0101] By combining threshold comparison and cosine similarity method to carry out real-time monitoring and identify real anomalies, the initial anomalies are first identified through threshold screening, and then random fluctuations are eliminated through morphological feature similarity matching. This accurately distinguishes real anomalies from non-tracing fluctuations, effectively avoids misjudgment and omission of anomalies, improves the accuracy of anomaly judgment, and clarifies the anomaly type for accurate source tracing in step three.
[0102] Step 3: For ESG environmental indicators confirmed as genuine anomalies, clarify their codes and operating conditions, retrieve the three-level association mapping table of equipment-process-ESG environmental indicators to generate a process and equipment traceability list, compare the process and operating parameters during the abnormal period, filter the abnormal processes, equipment and operating parameter sets, integrate relevant codes and parameters to form an ESG environmental indicator anomaly root cause location report, and push it to the operation and maintenance terminal. The specific process is as follows:
[0103] For each ESG environmental indicator confirmed as a genuine anomaly, its corresponding indicator code should be clearly defined. Based on the relevant operating conditions, retrieve the three-level correlation mapping table of equipment-process-ESG environmental indicators to determine all associated process codes corresponding to the indicator. Create a process traceability list;
[0104] Real-time collection of each process in the process traceability list Process parameters during abnormal periods For each process Each process parameter is compared one by one with a preset threshold range under the same operating conditions. Processes with at least one process parameter exceeding their corresponding threshold range are identified as abnormal processes and marked with an abnormal process code. ;
[0105] Coded by abnormal process Based on the corresponding criteria, the three-level correlation mapping table of equipment-process-ESG environmental indicators was retrieved again to identify the abnormal processes. Corresponding all associated device codes This will create a list of equipment for traceability.
[0106] Each device in the data acquisition equipment traceability list Operational data during abnormal periods For each device Each operating parameter is compared one by one with a preset threshold range for that parameter under the same operating conditions. Devices with at least one operating parameter exceeding their corresponding threshold range are identified as abnormal devices and marked with a unique abnormal device code. ;
[0107] Coded by abnormal device Based on the corresponding data, the core operating parameters of the abnormal equipment during the abnormal period were extracted, including: rotation speed setpoint, feed flow rate, reaction temperature setpoint, etc.
[0108] For each core operation parameter, compare it with its own preset threshold under the same working conditions, filter out operation parameters that exceed their corresponding thresholds, and compile them into a set of abnormal operation parameters.
[0109] The abnormal operation parameter set, abnormal equipment code, abnormal process code, and abnormal ESG environmental indicator code are matched and linked one by one to form an ESG environmental indicator abnormality root cause location report, which is sent to the operation and maintenance personnel terminal to guide the on-site abnormality handling.
[0110] It should be noted that by relying on the three-level correlation mapping table of equipment-process-ESG environmental indicators, the process and equipment traceability list is generated by reverse derivation from the real abnormal indicators. This allows the traceability of abnormalities to be carried out around the core related elements, avoiding indiscriminate investigation of all production elements and greatly improving the efficiency and pertinence of traceability analysis.
[0111] By comparing the process parameters, equipment operating parameters, and core operating parameters in sequence with preset thresholds under the same working conditions, abnormal processes, abnormal equipment, and specific operating parameters exceeding the thresholds are screened and locked down layer by layer. This enables precise tracing from abnormal ESG environmental indicators to abnormal production operating parameters, and clarifies the root cause of the abnormal indicators.
[0112] By associating the abnormal operation parameter set with the abnormal equipment code, abnormal process code, and abnormal ESG environmental indicator code, a standardized abnormal root cause location report is generated and pushed to the operation and maintenance terminal, allowing operation and maintenance personnel to intuitively grasp the core information of the abnormality and providing a clear decision-making basis for the rapid and accurate handling of on-site abnormalities.
[0113] By implementing a closed-loop management system for the entire process of ESG environmental anomalies, from "anomaly detection" to "root cause identification" and "endpoint push," environmental anomalies are deeply integrated with production operations. This allows the handling of environmental anomalies to be implemented at the specific equipment and process level, significantly improving the company's response and handling capabilities for environmental anomalies.
[0114] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A method for assessing enterprise ESG environmental indicators based on IoT data, characterized in that, include: Step 1: Code the equipment, processes, and ESG environmental indicators of the entire production process of the enterprise and establish a coding ledger, collect corresponding IoT data, construct correlation analysis groups and calculate correlation coefficients to screen strong correlation groups; Calculate the single-factor contribution of strongly correlated groups to screen core correlated groups, construct a three-level correlation mapping table of equipment-process-ESG environmental indicators, and establish a dynamic update and traceability mechanism. The specific process of calculating the single-factor contribution of strongly correlated groups to screen core correlated groups, constructing a three-level correlation mapping table of equipment-process-ESG environmental indicators, and establishing a dynamic update and traceability mechanism is as follows: The single-factor contribution of each strongly associated group was calculated using the single-factor variable method. The results were updated to the association analysis group ledger. A single-factor contribution threshold was preset. Strongly associated groups that did not meet the threshold were removed. Those that met the threshold were marked as core association groups and a core association ledger was formed. Based on the equipment-process-environment indicator coding ledger and the core association ledger, a three-level association mapping table of equipment-process-ESG environmental indicators is constructed to clarify the three-dimensional core correspondence relationship. Real-time monitoring of enterprise production scenarios. If any of the following occurs, such as the addition of equipment, scrapping, replacement, addition or deletion of processes, process adjustment, ESG environmental indicator definition, or change of monitoring standards, the relationship update will be triggered. The coding ledger will be updated according to the rules, data will be re-collected, the core relationship ledger will be updated, and the three-level relationship mapping table will be revised. Each version will be retained and information will be recorded to form a relationship update traceability archive. Step 2: Establish a normal threshold ledger for ESG environmental indicators under different working conditions, organize historical abnormal event records of ESG environmental indicators under different working conditions and draw line charts of indicator changes, extract morphological features from the initial segments, and integrate them into an abnormal feature library of enterprise ESG environmental indicators after clustering and classifying abnormal patterns; monitor ESG environmental indicators in real time and determine them as real anomalies. Step 3: For ESG environmental indicators confirmed as genuine anomalies, clarify their codes and operating conditions, retrieve the three-level association mapping table of equipment-process-ESG environmental indicators to generate a process and equipment traceability list, compare the process and operating parameters during the abnormal period, filter abnormal processes, equipment and operating parameter sets, integrate relevant codes and parameters to form an ESG environmental indicator anomaly root cause location report, and push it to the operation and maintenance terminal.
2. The enterprise ESG environmental indicator assessment method based on IoT data according to claim 1, characterized in that: The specific process of coding and establishing a coding ledger for equipment, processes, and ESG environmental indicators throughout the entire production process of an enterprise is as follows: The equipment coding adopts the form of equipment type code plus serial number. The equipment type code is divided according to the function of the production equipment, and the serial number is assigned according to the equipment deployment order. Each piece of equipment corresponds to a unique code. The process coding adopts the form of process flow code plus step code. The flow code corresponds to the entire production process sequence, and the step code corresponds to the sub-step within a single process. Each process and sub-step corresponds to a unique code. ESG environmental indicator coding adopts the form of indicator category code plus indicator sub-code. The category code corresponds to the first-level indicators of resource consumption and pollutant emissions, and the sub-code corresponds to the specific third-level environmental indicators. Each environmental indicator has a unique code. Organize the equipment, processes, and core ESG environmental indicators throughout the entire production process, assign unique codes according to rules, use the codes as unique identifiers, organize the mapping relationship between codes and corresponding objects, and construct a coding ledger for equipment, processes, and environmental indicators.
3. The enterprise ESG environmental indicator assessment method based on IoT data according to claim 1, characterized in that: The specific process of collecting corresponding IoT data, constructing correlation analysis groups, and calculating correlation coefficients to screen for strongly correlated groups is as follows: Real-time collection of equipment operation data, process parameters, and ESG environmental index data corresponding to each piece of equipment, each process, and each ESG environmental index in the enterprise's production scenario; Based on the historical data of various objects, a preset number of time-series data samples of the same duration are collected. According to the principle of pairing two to one and unique subject, three types of correlation analysis groups are constructed: equipment-environment indicators, process-environment indicators, and equipment-process. The correlation coefficient between the two types of data within each group was calculated using the Pearson correlation coefficient method, and the data were compiled into a log of the association analysis groups. Set a threshold for the correlation coefficient, mark groups that meet the threshold as strongly correlated groups, and remove groups that do not meet the threshold as weakly correlated groups.
4. The enterprise ESG environmental indicator assessment method based on IoT data according to claim 1, characterized in that: The specific process of establishing a ledger of normal thresholds for ESG environmental indicators under different operating conditions, compiling a historical record set of abnormal events for ESG environmental indicators under different operating conditions, and drawing line charts of indicator changes is as follows: Retrieve the normal threshold ranges of various ESG environmental indicators under different working conditions of the enterprise, and compile them into a normal threshold ledger of ESG environmental indicators under different working conditions. Retrieve historical abnormal event data records of various ESG environmental indicators under different working conditions, and organize them into a set of historical abnormal event records by working condition and indicator code. For each abnormal event in the historical abnormal event record set for different work conditions and indicators, a two-dimensional coordinate system of time and indicator value is constructed. Substitute the time-series data points corresponding to the abnormal event into a two-dimensional coordinate system, and connect each data point sequentially with line segments according to the time order to obtain a complete line chart of the changes in ESG environmental indicators corresponding to the abnormal event.
5. The enterprise ESG environmental indicator assessment method based on IoT data according to claim 1, characterized in that: The specific process of extracting morphological features from initial fragments, clustering them to classify abnormal patterns, and then integrating them to form an abnormal feature library of enterprise ESG environmental indicators is as follows: Set a uniform feature extraction time window duration, and extract time series data segments of this duration based on the start time of the anomaly occurrence of each historical anomaly event to generate a line chart of the initial segments of historical anomalies. If the total duration of the event is shorter than this duration, the entire time series data is taken. Extract the morphological features of line charts of all historical anomalies under the same working conditions and ESG environmental indicators. Use clustering algorithm to calculate the morphological similarity between the lines of each anomaly event. Classify the anomalies with morphological similarity greater than or equal to the preset similarity threshold into the same anomaly pattern. Based on the clustering results, an abnormal pattern and feature comparison table of ESG environmental indicators by work condition and indicator was compiled, and combined with the normal threshold ledger of ESG environmental indicators by work condition, an abnormal feature library of enterprise ESG environmental indicators was formed.
6. The enterprise ESG environmental indicator assessment method based on IoT data according to claim 1, characterized in that: The specific process for real-time monitoring of ESG environmental indicators and determining them as genuine anomalies is as follows: Real-time collection of monitoring data of ESG environmental indicators under various working conditions of the enterprise, input into the enterprise's ESG environmental indicator abnormal feature library and compared with the normal threshold range of the corresponding working conditions and indicators; if the threshold is exceeded, the initial abnormality start time is marked and relevant information is recorded. Using the initial time of the anomaly as a benchmark, time-series data with a uniform feature extraction time window duration are extracted, and a real-time initial anomaly segment line graph is drawn. Morphological features are extracted and a real-time standardized morphological feature vector is constructed. The cosine similarity method is used to calculate the similarity between the feature vector and the corresponding anomaly pattern feature vector. The maximum value is compared with the preset similarity threshold. If the threshold is met, it is confirmed as a real anomaly. If the threshold is not met, it is judged as an accidental fluctuation and only a warning record is made.
7. The enterprise ESG environmental indicator assessment method based on IoT data according to claim 1, characterized in that: For ESG environmental indicators confirmed as genuine anomalies, the specific process of clarifying their codes and operating conditions, and retrieving the three-level association mapping table of equipment-process-ESG environmental indicators to generate a process and equipment traceability list is as follows: For ESG environmental indicators that are confirmed to be genuine anomalies, clarify their indicator codes and corresponding operating conditions, retrieve the three-level association mapping table of equipment-process-ESG environmental indicators, determine the associated process codes, and form a process traceability list. Collect process parameters for abnormal periods of processes in the list, compare them one by one with preset threshold ranges under the same working conditions, filter out abnormal processes and mark abnormal process codes. Based on the abnormal process code, the above three-level association mapping table is retrieved again to determine the associated equipment code of the abnormal process and form an equipment traceability list.
8. The enterprise ESG environmental indicator assessment method based on IoT data according to claim 1, characterized in that: The specific process of comparing process and operating parameters during abnormal periods, screening abnormal procedures, equipment, and sets of operating parameters, integrating relevant codes and parameters to form an ESG environmental indicator anomaly root cause location report, and pushing it to the operation and maintenance terminal is as follows: The system collects the operating data of each device in the equipment traceability list during abnormal periods, compares it one by one with the preset threshold range under the same working conditions, filters out abnormal devices and marks them with a unique abnormal device code. Based on the abnormal equipment code, the core operating parameters of the abnormal equipment during the abnormal period are extracted, compared with the exclusive preset threshold under the same working conditions, and the parameters exceeding the threshold are selected and organized into a set of abnormal operating parameters. The abnormal operation parameter set is associated with the abnormal equipment code, abnormal process code, and abnormal ESG environmental indicator code one by one, and an ESG environmental indicator abnormality root cause location report is generated and sent to the operation and maintenance personnel terminal to guide the on-site abnormality handling.