Environmental protection management intelligent analysis system and method based on multi-dimensional index monitoring

Abstract

The invention discloses an environmental protection management intelligent analysis system and method based on multi-dimensional index monitoring in the technical field of environmental protection management, and the method comprises the steps: obtaining the emission and concentration data of pollutants through monitoring a waste gas discharge port, a waste water discharge port and solid wastes of an enterprise, and collecting the energy consumption and equipment operation data at the same time; calculating to obtain a multi-dimensional index coefficient for comprehensively evaluating the environmental protection condition; historical data of the multi-dimensional index coefficients are collected, an environmental protection anomaly detection model is constructed based on the historical data of the multi-dimensional index coefficients, and anomaly detection is carried out on the multi-dimensional index coefficients collected in real time; and meanwhile, an environmental pollution trend prediction model is constructed in combination with historical data and time sequence characteristics of the multi-dimensional index coefficients, and trend prediction is performed on the multi-dimensional index coefficients. According to the method, the environmental protection performance of the enterprise is evaluated more comprehensively by analyzing and calculating the corresponding index coefficients according to multiple dimensions of emission, energy consumption and operation equipment.

Description

Technical Field

[0001] This invention relates to an intelligent analysis system and method for environmental management based on multi-dimensional indicator monitoring, belonging to the field of environmental management technology. Background Technology

[0002] With the increasing severity of environmental pollution, the importance of environmental management has become increasingly prominent. Therefore, there is an urgent need for an environmental management method that can comprehensively monitor based on multi-dimensional indicators and achieve intelligent analysis. This method involves calculating the coefficients of indicators in multiple dimensions, constructing predictive models for data anomaly detection and pollution trends, and developing a rule base that combines the results of anomaly detection and pollution trend prediction.

[0003] Traditional environmental management methods mainly rely on manual monitoring and simple data analysis, which have shortcomings such as incomplete monitoring data, low analysis efficiency, and difficulty in timely detection of potential environmental problems. In addition, when faced with a large amount of monitoring data, manual analysis is difficult to quickly and accurately extract valuable information, and there is a lack of technical problems in combining anomaly detection with pollution trend prediction results to formulate comprehensive decision-making rules. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide an intelligent analysis system and method for environmental management based on multi-dimensional indicator monitoring. By analyzing and calculating the corresponding indicator coefficients for multiple dimensions such as emissions, energy consumption and operating equipment, the system can more comprehensively evaluate the environmental performance of enterprises, avoid the one-sidedness of single indicator evaluation, and provide a more comprehensive and accurate environmental status.

[0005] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0006] In a first aspect, this invention provides an intelligent analysis method for environmental management based on multi-dimensional indicator monitoring, including:

[0007] By monitoring the enterprise's exhaust gas outlets, wastewater discharge outlets, and solid waste, data on pollutant emissions and concentrations are obtained. Simultaneously, energy consumption and equipment operation data are collected to calculate a multi-dimensional index coefficient for comprehensive environmental assessment. This multi-dimensional index coefficient includes a comprehensive pollution emission index coefficient calculated based on the pollutant emission and concentration data, using the following formula:

[0008]

[0009] in, The comprehensive pollution emission index coefficient, For the standardized first The average emissions of these pollutants For the first The proportions corresponding to the emissions of each pollutant For the first Average concentration of various pollutants For the first Weighting coefficients corresponding to the concentrations of various pollutants;

[0010] Historical data of the multi-dimensional index coefficients are collected, and an environmental anomaly detection model is constructed based on the historical data of the multi-dimensional index coefficients to detect anomalies in the real-time collected multi-dimensional index coefficients. At the same time, an environmental pollution trend prediction model is constructed by combining the historical data of the multi-dimensional index coefficients and their time series characteristics to predict the trends of the multi-dimensional index coefficients.

[0011] A decision rule base is established by combining the anomaly detection results and trend prediction results. In response to the identification of abnormal data or the prediction of adverse trends, the rule base is matched with the rules to generate corresponding decision suggestions and output them in a visual form.

[0012] Furthermore, the multi-dimensional index coefficients also include the comprehensive pollution energy consumption index coefficient and the comprehensive treatment facility operation index coefficient, wherein: the comprehensive pollution energy consumption index coefficient is calculated based on the consumption data of polluting energy and clean energy in the energy consumption data, and the comprehensive treatment facility operation index coefficient is calculated based on the average operating rate and average failure frequency of production equipment in the equipment operation data.

[0013] Furthermore, the comprehensive pollution emission index coefficient is calculated based on the emission amount and concentration data of the pollutants, including:

[0014] The emissions of air pollutants, water pollutants, and solid pollutants were standardized using Z-scores, and the proportions of the emissions of the three pollutants to the total emissions were calculated as corresponding weighting coefficients. At the same time, the concentration data of the three pollutants were standardized accordingly, and the weighting coefficients were determined based on the degree of environmental impact of the pollutants. The monthly average emissions and average concentrations of the three pollutants were calculated respectively. The average emissions of the three pollutants were added to their corresponding weighting coefficients, and the average concentrations were added to their corresponding weighting coefficients to obtain the comprehensive pollution emission index coefficient.

[0015] Furthermore, the comprehensive pollution energy consumption index coefficient is calculated based on the consumption data of polluting energy and clean energy in the energy consumption data, including:

[0016] Data is collected using appropriate metering instruments based on energy type. The consumption of various energy sources and the total energy consumption of the enterprise over 30 days are statistically analyzed. Energy is categorized according to its pollution level, including polluting and clean energy. All energy consumption is converted to a uniform unit, and the proportions of polluting and clean energy in the total energy consumption are calculated. The total polluting energy consumption is divided by the total number of days and then multiplied by the proportion. A comprehensive clean energy factor is set based on the environmental friendliness of clean energy. The total clean energy consumption is then divided by the total number of days, multiplied by the proportion and the comprehensive clean energy factor, and finally summed to obtain the comprehensive pollution energy consumption index coefficient. The calculation formula is as follows:

[0017]

[0018] in, The comprehensive pollution energy consumption index coefficient, Total energy source for polluting energy sources This represents the proportion of polluting energy sources. Total energy source for clean energy This represents the proportion of clean energy sources. It is a comprehensive cleaning agent.

[0019] Furthermore, the comprehensive management facility operation index coefficient is calculated based on the average operating rate and average failure frequency of the production equipment in the equipment operation data, including:

[0020] The number of devices in the enterprise is determined. An automated system is used to record the actual operating time of each device over 30 days. Based on production process requirements, the required operating time for each device is determined. The actual operating time of each device is divided by its required operating time to obtain the corresponding operating rate. The operating rates of all devices are summed and divided by the total number of devices to obtain the enterprise's average operating rate. Simultaneously, the number of failures and the actual number of times each device is used are recorded. The number of failures is divided by the actual number of times each device is used to obtain the corresponding failure frequency. The failure frequencies of all devices are summed and divided by the total number of devices to obtain the enterprise's average failure frequency. Based on the average failure frequency and average operating rate, a weighted calculation is performed to obtain the comprehensive management facility operation index coefficient. The calculation formula is as follows:

[0021]

[0022] in, This refers to the operational index coefficient of comprehensive governance facilities. Mean frequency of equipment failure. The average operating rate of the equipment. and These are the corresponding weighting coefficients.

[0023] Furthermore, historical data on the multi-dimensional index coefficients are collected, including:

[0024] For the comprehensive pollution emission index coefficient, comprehensive pollution energy consumption index coefficient, and comprehensive treatment facility operation index coefficient, historical data from the past year are collected. The coefficient data corresponding to these three indicators are collected periodically according to a predetermined time cycle. The collected data is cleaned and standardized. For the cleaned data, the mean and standard deviation of the comprehensive pollution emission index coefficient, comprehensive pollution energy consumption index coefficient, and comprehensive treatment facility operation index coefficient are calculated respectively. Based on historical data and practical experience, a threshold range is set for each indicator. The periodically collected data is compared with the set threshold range. If the corresponding coefficient of a certain indicator exceeds the set threshold range, the data point is determined to be abnormal data. The abnormal data points are recorded, including the outlier value and the determination time.

[0025] Furthermore, an environmental anomaly detection model is constructed based on the historical data of the multi-dimensional indicator coefficients to detect anomalies in the real-time collected multi-dimensional indicator coefficients, including:

[0026] Historical data containing the comprehensive pollution emission index coefficient, comprehensive pollution energy consumption index coefficient, and comprehensive treatment facility operation index coefficient are collected. The data is divided and normalized. A multilayer perceptron neural network is selected as the basic model, and a binary cross-entropy loss function is used as the loss function. The historical data divided into training sets is used to train the model. The trained environmental anomaly detection model is applied to the periodically collected new data to detect anomalies in the comprehensive pollution emission index coefficient, comprehensive pollution energy consumption index coefficient, and comprehensive treatment facility operation index coefficient. The model outputs labels and values ​​for the three index coefficients. If the model outputs a value of 0, it indicates that the data detection is normal. If the model outputs a value of 1, it indicates that the data monitoring is abnormal.

[0027] Furthermore, an environmental pollution trend prediction model is constructed by combining the historical data and time-series characteristics of the multi-dimensional indicator coefficients to predict the trends of the multi-dimensional indicator coefficients, including:

[0028] Using historical data containing the comprehensive pollution emission index coefficient, comprehensive pollution energy consumption index coefficient, and comprehensive treatment facility operation index coefficient, the data is determined to have timestamps and sorted in chronological order. Time features, lag features, and moving statistical features are added. Three index coefficients are selected as target variables. Long short-term memory network is selected as the basic model. Mean squared error is used as the loss function, and the Adam optimizer is used. The training data is divided into batches for training. Each time the model parameters are updated, the loss is calculated using samples and the parameters are updated. The trained environmental pollution trend prediction model is deployed to the system, and data is collected regularly to predict future pollution trends.

[0029] Furthermore, a decision rule base is established by combining the anomaly detection results and trend prediction results. In response to the identification of abnormal data or the prediction of an unfavorable trend, the rule base is matched with the rules to generate corresponding decision suggestions, which are then output in a visual format, including:

[0030] Historical anomaly data, including indicator coefficient labels and corresponding values, is obtained from the environmental anomaly detection model. Corresponding processing rules are specified for different types of anomaly data. At the same time, pollution trend prediction results from the environmental pollution trend prediction model are collected, and corresponding countermeasures are specified for different pollution trends. Combining the anomaly detection and trend prediction results, comprehensive decision rules are formulated and stored in the database. The system runs the environmental anomaly detection model and the environmental pollution trend prediction model regularly, matches the model output results with the rules in the rule base, generates decision suggestions based on the matched rules, and displays them to the user in a visual form.

[0031] Secondly, the present invention provides an intelligent analysis system for environmental management based on multi-dimensional indicator monitoring, used to implement the intelligent analysis method for environmental management based on multi-dimensional indicator monitoring as described above, including:

[0032] Multidimensional index calculation module: It is used to obtain data on the emission and concentration of pollutants by monitoring the enterprise's exhaust gas outlets, wastewater discharge outlets and solid waste, while collecting energy consumption and equipment operation data, and calculating multidimensional index coefficients to comprehensively assess the environmental status.

[0033] Anomaly detection and trend prediction module: This module is used to construct an environmental anomaly detection model based on the historical data of the multi-dimensional index coefficients, and to perform anomaly detection on the real-time collected multi-dimensional index coefficients; at the same time, it combines the historical data of the multi-dimensional index coefficients and their time-series characteristics to construct an environmental pollution trend prediction model, and to predict the trend of the multi-dimensional index coefficients.

[0034] Decision generation and output module: When abnormal data is identified or adverse trends are predicted, it matches the rules in the decision rule base to generate corresponding decision suggestions and output them in a visual form.

[0035] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0036] This invention analyzes and calculates corresponding index coefficients across multiple dimensions, including emissions, energy consumption, and operating equipment, to comprehensively assess a company's environmental performance. This avoids the limitations of single-indicator assessments and provides a more comprehensive and accurate picture of the environmental situation. Simultaneously, by collecting historical data on three index coefficients, classifying and normalizing them, and selecting a multilayer perceptron neural network (LSTM) as the base model for environmental anomaly detection, the invention provides companies with intuitive displays of abnormal indicators and values, helping them understand their current environmental status, improving detection accuracy, and enhancing real-time monitoring and environmental management efficiency. Furthermore, by combining time characteristics, lag characteristics, and moving statistical characteristics, this invention selects an LSTM as the base model for environmental pollution trend prediction. The trained environmental pollution trend prediction model is deployed into the system, and data is collected periodically, enabling real-time prediction and future trend forecasting, thus promoting the intelligent development of environmental protection. Attached Figure Description

[0037] The accompanying drawings, which form part of this specification, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0038] Figure 1 This is a flowchart illustrating the intelligent analysis method for environmental management based on multi-dimensional indicator monitoring provided in Embodiment 1 of the present invention. Detailed Implementation

[0039] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0040] The following detailed description is exemplary and intended to provide further detailed explanation of the invention. Unless otherwise specified, all technical terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this invention is for describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention.

[0041] Example 1:

[0042] This embodiment proposes an intelligent analysis method for environmental management based on multi-dimensional indicator monitoring, including the following steps:

[0043] Step 1: Obtain the emission amount and concentration of the corresponding pollutants from the exhaust gas outlet, wastewater outlet, and solid waste. Calculate the comprehensive pollution emission index coefficient by combining the emission amount with the corresponding weighting coefficient and the concentration with the corresponding weighting coefficient. Divide the total energy consumption of clean energy by the total number of days, multiply by the proportion value and the comprehensive clean factor, and add them together to obtain the comprehensive pollution energy consumption index coefficient. Perform weighted calculations based on the average equipment failure frequency and the average equipment operating rate to obtain the comprehensive treatment facility operation index coefficient.

[0044] Step one involves obtaining the emission amounts and concentrations of the corresponding pollutants from the exhaust gas outlet, wastewater outlet, and solid waste, including the following steps:

[0045] Continuous monitoring instruments are installed at the enterprise's exhaust outlets to collect data on the emission concentrations and amounts of gaseous and particulate pollutants, and to calculate the total emission amount and concentration of each pollutant. Simultaneously, sampling points are set at the four corners of the exhaust outlets to periodically sample and analyze the pollutant composition and concentration in the wastewater. Chemical analysis methods are used to analyze water pollutants, and the laboratory analysis results are compared with a standard material library to determine the types of water pollutants. The concentrations and emission amounts of various water pollutants are obtained through laboratory analysis data. Solid waste samples are collected and pretreated. Pollutants are extracted from the pretreated samples, and the concentration of pollutants in the solid waste is obtained based on the mass of the extracted pollutants measured by the instrument and combined with the sample mass.

[0046] This embodiment uses a continuous monitoring instrument to collect the emission concentrations of gaseous and particulate pollutants in the exhaust gas in real time. Simultaneously, the exhaust gas flow rate is recorded for subsequent emission calculations. Based on the collected concentration and flow data, the emission amount of each pollutant is calculated. The calculation results are recorded and analyzed periodically. Sampling points are set at the four corners of the enterprise's wastewater discharge outlet, and wastewater is sampled weekly at predetermined time intervals. Chemical analysis methods are used to analyze the collected wastewater samples, determining the pollutant components and concentrations. The laboratory analysis results are compared with a standard material spectral library to determine the types and concentrations of water pollutants. Based on the laboratory analysis data and wastewater discharge volume, the emission amounts of various water pollutants are calculated. Representative samples are randomly selected from the enterprise's solid waste for collection. The collected solid waste samples are pretreated, and appropriate extraction methods are used to extract pollutants from the pretreated samples. The mass of pollutants in the extract is measured using an instrument. Based on the measured pollutant mass and sample mass, the concentration of pollutants in the solid waste is calculated.

[0047] It should be noted that the calculation of the comprehensive pollution emission index coefficient includes the following steps:

[0048] The emissions of air pollutants, water pollutants, and solid pollutants were standardized using Z-scores, and the proportions of each pollutant's emissions to the total emissions were calculated as corresponding weighting coefficients. Simultaneously, the concentration data of the three pollutants were standardized accordingly, and weighting coefficients were determined based on the pollutants' environmental impact. The monthly average emissions and average concentrations of each pollutant were calculated. The comprehensive pollution emission index coefficient was obtained by adding the average emissions and their corresponding weighting coefficients, as well as the average concentration and its corresponding weighting coefficient. The formula for calculating the comprehensive pollution emission index coefficient is as follows:

[0049]

[0050] in, The comprehensive pollution emission index coefficient, For the standardized first The average emissions of these pollutants For the first The proportions corresponding to the emissions of each pollutant For the first Average concentration of various pollutants For the first The weighting coefficients corresponding to the concentrations of various pollutants. Where i = 1 represents air pollutants, i = 2 represents water pollutants, and i = 3 represents solid pollutants. The value is 0.4. and The values ​​are all 0.3.

[0051] This embodiment collects emission data and corresponding concentration data of air pollutants, water pollutants, and solid pollutants from an enterprise over a 30-day period. For each pollutant, the mean and standard deviation of its emissions are calculated. The Z-score formula is used to standardize each data point. The concentration data of the three pollutants are also standardized in the same way. The proportion of each pollutant's emissions to the total emissions is calculated. Based on the pollutant's environmental impact, a weighting coefficient is assigned to the concentration of each pollutant. The average emissions and average concentrations of the three pollutants are calculated. For each pollutant, its standardized average emissions are multiplied by the corresponding weighting coefficient and then summed. The result from the previous step is multiplied by the corresponding average concentration weight and added to the comprehensive pollution emission index coefficient. The final comprehensive pollution emission index coefficient reflects the enterprise's overall pollution emission level and environmental impact; a higher value indicates more severe pollution, while a lower value indicates relatively lighter pollution.

[0052] It should be noted that the calculation of the comprehensive pollution energy consumption index coefficient includes the following steps:

[0053] Data is collected using appropriate metering instruments based on energy type. The consumption of various energy sources and the total energy consumption of the enterprise over 30 days are statistically analyzed. Energy is categorized according to its pollution level, including polluting energy and clean energy. All energy consumption is converted to a uniform unit, and the proportions of polluting and clean energy in the total energy consumption are calculated. The total polluting energy consumption is divided by the total number of days, then multiplied by the proportion. A comprehensive clean energy factor is set based on the environmental friendliness of clean energy. The total clean energy consumption is divided by the total number of days, multiplied by the proportion and the comprehensive clean energy factor, and then summed to obtain the comprehensive pollution energy consumption index coefficient. The formula for calculating the comprehensive pollution energy consumption index coefficient is as follows:

[0054]

[0055] in, The comprehensive pollution energy consumption index coefficient, Total energy source for polluting energy sources This represents the proportion of polluting energy sources. Total energy source for clean energy This represents the proportion of clean energy sources. It is a comprehensive cleaning agent.

[0056] In this embodiment, appropriate metering instruments are selected based on the energy type used by the enterprise for data collection. Using the selected instruments, the enterprise's consumption of various energy sources over 30 days is collected periodically, recording daily energy consumption data, including the consumption of each type of energy. The collected energy consumption data is then categorized and statistically analyzed according to polluting and clean energy sources. For example, highly polluting energy sources such as coal can be classified as polluting energy, while relatively clean energy sources such as electricity and natural gas can be classified as clean energy. The total consumption of each type of energy source over 30 days, as well as the total consumption of all energy sources, is calculated. All energy consumption is converted to a unified unit, and the proportions of polluting and clean energy sources in the total energy consumption are calculated. The total consumption of polluting energy is divided by the total number of days to obtain the daily average consumption. The daily average consumption is then multiplied by the proportion to obtain the percentage of polluting energy in the overall pollution energy consumption. The contribution of clean energy to overall pollution energy consumption is determined by assigning a comprehensive clean energy factor to each type of clean energy based on its environmental friendliness. The comprehensive clean energy factor for electricity and solar energy is set to 1, for steam to 0.8, and for natural gas to 0.3. The comprehensive clean energy factors of the four types of clean energy are weighted according to energy consumption and then added together to obtain the comprehensive clean energy factor value. Then, the total consumption of clean energy is divided by the total number of days to obtain the average daily consumption. The average daily consumption is then multiplied by the proportion value and the comprehensive clean energy factor to obtain the contribution of clean energy to overall pollution energy consumption. The contributions of polluting energy and clean energy to overall pollution energy consumption are added together to obtain the comprehensive pollution energy consumption index coefficient, which assesses the pollution level of the enterprise's energy consumption. The higher the value, the higher the pollution level of the enterprise's energy consumption; the lower the value, the lower the pollution level of the enterprise's energy consumption.

[0057] Data was collected using metering instruments, and statistics on various energy sources for the enterprise from November 1, 2024 to November 30, 2024 are shown in Table 1:

[0058] Table 1: Statistics and Classification of Enterprise Energy Consumption from November 1st to 30th, 2024

[0059] It should be noted that the calculation of the operational index coefficients of comprehensive governance facilities includes the following steps:

[0060] The number of devices in the enterprise is determined. An automated system is used to record the actual operating time of each device over 30 days. Based on production process requirements, the required operating time for each device is determined. The actual operating time of each device is divided by its required operating time to obtain the corresponding operating rate. The operating rates of all devices are summed and divided by the total number of devices to obtain the enterprise's average operating rate. Simultaneously, the number of failures and the actual number of times each device is started are recorded. The number of failures is divided by the actual number of times each device is started to obtain the corresponding failure frequency. The failure frequencies of all devices are summed and divided by the total number of devices to obtain the enterprise's average failure frequency. A weighted calculation is performed based on the average failure frequency and the average operating rate to obtain the comprehensive governance facility operation index coefficient. The formula for calculating the comprehensive governance facility operation index coefficient is as follows:

[0061]

[0062] in, This refers to the operational index coefficient of comprehensive governance facilities. Mean frequency of equipment failure. The average operating rate of the equipment. and For the corresponding weighting coefficients, in this embodiment, The value is 0.7. The value is 0.3.

[0063] This embodiment involves organizing professional personnel to conduct a comprehensive inventory of all equipment within the enterprise, determining the specific number of equipment, recording the actual operating time of each piece of equipment in real time over 30 days, calculating the actual operating time, and establishing the required operating time standard for each piece of equipment based on production process requirements. For each piece of equipment, the actual operating time is divided by the required operating time to obtain the corresponding operating rate. Detailed records of each equipment failure are kept using an equipment maintenance management system or manual recording, including the time of failure, cause, and repair status. The actual number of times each piece of equipment is started is also recorded, including the number of times it operates normally after startup. For each piece of equipment, the number of failures is divided by the actual number of starts to obtain the corresponding failure frequency. The operating rates of all equipment are summed and then divided by the total number of equipment to obtain the enterprise's average equipment operating rate. The failure frequencies of all equipment are summed and then divided by the total number of equipment to obtain the enterprise's average equipment failure frequency. Based on the enterprise's emphasis on equipment operation and failure management, the weights of the average equipment failure frequency and the average equipment operating rate are determined. According to the determined weights, a weighted calculation is performed based on the average equipment failure frequency and the average equipment operating rate to obtain the comprehensive management facility operation index coefficient.

[0064] The data for each device in the enterprise from November 1, 2024 to November 30, 2024 was statistically analyzed using an automated system, as shown in Table 2:

[0065] Table 2: Equipment-level data for enterprises from November 1st to 30th, 2024

[0066] Step 2: Collect historical data of three indicators within one year, calculate the mean and standard deviation of the coefficients of the three indicators respectively, and set the threshold range to Mi standard deviations above and below the mean. After preprocessing the collected historical data, use a multilayer perceptron to build an environmental anomaly detection model. Then, use the environmental anomaly detection historical data with timestamps, add time, lag and sliding statistical features, and use the three indicator coefficients as target variables to build an environmental pollution trend prediction model using a long short-term memory network.

[0067] Step two involves collecting historical data for the three indicators over the past year, calculating the mean and standard deviation of the coefficients for each indicator, and setting a threshold range of Mi standard deviations above and below the mean. This includes the following steps:

[0068] Historical data of three indicators over the past year were collected. The mean and standard deviation of the comprehensive pollution emission index coefficient, comprehensive pollution energy consumption index coefficient, and comprehensive treatment facility operation index coefficient obtained from the three indicators were calculated respectively. Based on historical data and practical experience, the corresponding thresholds were set to be within a range of Mi standard deviations above and below the mean. If the corresponding coefficient of a certain indicator exceeds the set range, it is judged as abnormal data.

[0069] Regularly collect coefficient data corresponding to the three indicators, preprocess the collected data, and compare the actual data of each indicator with the set threshold range. If the actual data exceeds the set threshold range, it is judged as abnormal data.

[0070] In this embodiment, historical data from the past year is collected for the comprehensive pollution emission index coefficient, comprehensive pollution energy consumption index coefficient, and comprehensive treatment facility operation index coefficient. The coefficient data for these three indicators are collected periodically, such as weekly. The collected data is cleaned and standardized. The mean and standard deviation of the comprehensive pollution emission index coefficient, comprehensive pollution energy consumption index coefficient, and comprehensive treatment facility operation index coefficient are calculated for the cleaned data. Based on historical data and practical experience, a threshold range of Mi standard deviations above and below the mean is set for each indicator. M1 is the multiple corresponding to the comprehensive pollution emission index coefficient, M2 is the multiple corresponding to the comprehensive pollution energy consumption index coefficient, and M3 is the multiple corresponding to the comprehensive treatment facility operation index coefficient. The periodically collected data is compared with the set threshold range. If the coefficient of a certain indicator exceeds the set threshold range, the data point is determined to be abnormal data, and the abnormal data point is recorded, including the abnormal value and the determination time.

[0071] It should be noted that constructing an environmental anomaly detection model includes the following steps:

[0072] Historical data containing comprehensive pollution emission index coefficients, comprehensive pollution energy consumption index coefficients, and comprehensive treatment facility operation index coefficients were collected. The data was partitioned and normalized. A multilayer perceptron neural network was selected as the base model: an input layer with 3 nodes corresponding to the 3 normalized index coefficients was set; two hidden layers were set, with the first layer having 6 nodes and the second layer having 4 nodes; and an output layer with 3 nodes was set to output the predicted values ​​of the 3 index coefficients, using a binary cross-entropy loss function as the loss function. The model was trained using the historical data divided into training sets. The trained environmental anomaly detection model was applied to periodically collected new data to detect anomalies in the comprehensive pollution emission index coefficients, comprehensive pollution energy consumption index coefficients, and comprehensive treatment facility operation index coefficients. The model outputs labels and values ​​for the three index coefficients. If the model outputs a value of 0, it indicates normal data detection; if the model outputs a value of 1, it indicates abnormal data monitoring.

[0073] This embodiment collects historical datasets containing comprehensive pollution emission index coefficients, comprehensive pollution energy consumption index coefficients, and comprehensive treatment facility operation index coefficients. The collected data undergoes preliminary cleaning and organization, dividing it into training, validation, and test sets. A multilayer perceptron neural network is selected as the basic model, with an architecture consisting of one input layer, two hidden layers, and one output layer. The input layer has three nodes, corresponding to the three index coefficients; the first hidden layer has six nodes, the second hidden layer has four nodes, and the output layer has three nodes, corresponding to the anomaly detection results for each of the three index coefficients. A binary cross-entropy loss function is used as the model's loss function. The model is trained using the training set. Preprocessed new data is input into the trained environmental anomaly detection model for prediction. The model outputs the anomaly detection results for the three index coefficients, with each result being either 0 or 1, representing normal or abnormal, respectively. Based on the model's output, the abnormal state of each index is interpreted.

[0074] It should be noted that constructing an environmental pollution trend prediction model includes the following steps:

[0075] Using the same historical data as the environmental anomaly detection model, including comprehensive pollution emission index coefficients, comprehensive pollution energy consumption index coefficients, and comprehensive treatment facility operation index coefficients, the data is timestamped and sorted chronologically. Time features, lag features, and moving statistical features are added. Three index coefficients are selected as target variables, and a Long Short-Term Memory (LSTM) network is chosen as the base model: An input layer is set with an input shape of (number of samples, 7, 3), where 7 represents the time step, representing 7 consecutive days of historical data for each sample, and 3 represents the number of features, representing the three index coefficients corresponding to each time point. Two LSTM layers are set, each with 32 nodes, representing the LSTM network. The dimensionality of the hidden state of TM is determined; a fully connected layer with 16 nodes is added after the LSTM layer to compress the high-dimensional temporal features extracted by LSTM to a lower dimension; an output layer with 3 nodes is set to output the predicted coefficients of the three target variables; mean squared error is used as the loss function, and the Adam optimizer is used with a learning rate of 0.001. The training data is divided into batches of 32 for training. Each time the model parameters are updated, the loss is calculated and the parameters are updated using 32 samples. The trained environmental pollution trend prediction model is deployed to the system, and data is collected regularly to predict future pollution trends.

[0076] In this embodiment, features related on a weekly cycle are extracted, and a lagged version is created for each indicator, that is, the value at the previous time point is used as an additional feature at the current time point. Statistics of the indicator within a specific time window are calculated, including the mean and standard deviation. Three indicator coefficients are selected as target variables. The dataset is divided into training, validation, and test sets. A Long Short-Term Memory (LSTM) network is selected as the base model, with one input layer of shape (number of samples, 7, 3), indicating the use of features at 7 time steps, with 3 indicator values ​​at each time step. Two LSTM layers are set, each with 32 nodes. A parameter is then set after the LSTM layers. A fully connected layer with 16 nodes is added to further process the output of the LSTM layer. An output layer with 3 nodes is set up to correspond to the predicted values ​​of the three indicators. The mean squared error is used as the loss function, the Adam optimizer is selected, and the learning rate is set to 0.001. The training data is divided into batches of 32 for training to accelerate convergence and reduce memory consumption. The trained environmental pollution trend prediction model is deployed to the system to collect data and make predictions periodically. The preprocessed new data is input into the deployed model for prediction to obtain the predicted values ​​of the three indicators for a period of time in the future.

[0077] Step 3: Establish a decision rule base by combining the environmental anomaly detection model and the environmental pollution trend prediction model. When the system identifies abnormal data or predicts an unfavorable pollution trend, it matches the rules in the rule base and generates corresponding decision suggestions.

[0078] Step 3 includes the following steps: obtaining historical anomaly data from the environmental anomaly detection model, including indicator coefficient labels and corresponding values; specifying corresponding processing rules for different types of anomaly data; collecting pollution trend prediction results from the environmental pollution trend prediction model; specifying corresponding countermeasures for different pollution trends; combining anomaly detection and trend prediction results to formulate comprehensive decision rules; storing the rules in the database; the system periodically runs the environmental anomaly detection model and the environmental pollution trend prediction model; matching the model output results with the rules in the rule base; generating decision suggestions based on the matched rules; and displaying them to the user in a visual form.

[0079] In this embodiment, historical anomaly data, including indicator coefficient labels and corresponding values, is obtained from the environmental anomaly detection model. The types and characteristics of the anomaly data are analyzed to provide a basis for formulating processing rules. Appropriate processing rules are specified based on different types of anomaly data. For example, for anomalies involving excessive pollution emissions, measures such as reducing production and increasing the operating time of treatment equipment may be necessary. These processing rules are stored in the database. An environmental pollution trend prediction model is run to collect future pollution trend prediction results. Corresponding countermeasures are specified based on different pollution trends. For example, for an upward trend in pollution, measures such as strengthening treatment and raising emission standards may be necessary. A comprehensive decision-making rule is formulated by comprehensively considering the output of the environmental anomaly detection model and the prediction results of the environmental pollution trend prediction model. For example, if the anomaly detection model detects excessive pollution emissions and the trend prediction model predicts an upward trend in pollution, stricter countermeasures may be necessary. The formulated comprehensive decision-making rules are stored in the database for matching and execution during regular system operation. The system runs the environmental anomaly detection model and the environmental pollution trend prediction model at set time intervals, matches the model outputs with the rules in the rule base, generates decision suggestions based on the matched rules, and displays the generated decision suggestions to the user in a visual form, such as charts or dashboards.

[0080] It should be noted that before displaying the generated decision recommendations to users, this solution uses a comprehensive evaluation model to assess the overall score of the decision options, achieving a balance between environmental benefits and economic costs in environmental governance. This model, by incorporating a benefit saturation effect and a compliance risk threshold, recommends the decision recommendation with the highest overall score from multiple feasible options to the user. The comprehensive evaluation model expression is as follows:

[0081]

[0082] In the formula: S k The comprehensive score of the k-th decision option; n is the total number of coefficients for the multi-dimensional indicators; w jLet ΔI be the static importance weight of the j-th indicator; j,k Let be the expected improvement of scheme k for the j-th indicator; α be the benefit saturation coefficient; C k R represents the total economic cost of option k; k It serves as a compliance risk adjustment factor.

[0083] In summary, this embodiment calculates the average emissions and average concentrations of three pollutants. For each pollutant, the standardized average emission is multiplied by its corresponding weighting coefficient and then summed. The result from the previous step is multiplied by its corresponding average concentration weight and added to the comprehensive pollution emission index coefficient, ultimately obtaining the comprehensive pollution emission index coefficient. The contributions of polluting energy and clean energy to comprehensive pollution energy consumption are summed to obtain the comprehensive pollution energy consumption index coefficient. Based on the enterprise's emphasis on equipment operation and fault management, the weights of the average equipment failure frequency and the average equipment operating rate are determined. According to the determined weights, a weighted calculation is performed based on the average equipment failure frequency and the average equipment operating rate to obtain the comprehensive treatment facility operation index coefficient. A multilayer perceptron neural network is selected as the basic model, and the model architecture is set, including one input layer, two hidden layers, and one output layer. The number of nodes in the input layer is 3, corresponding to the three index coefficients; the number of nodes in the first hidden layer is 6, the number of nodes in the second hidden layer is 6, the number of nodes in the third hidden layer is 6, the number of nodes in the fourth hidden layer is 6, the number of nodes in the fifth hidden layer is 6, the number of nodes in the sixth ... The hidden layer has 4 nodes; the output layer has 3 nodes. The model is trained using a training set. The preprocessed new data is input into the trained environmental anomaly detection model for prediction. A long short-term memory network is selected as the base model. An input layer is set with an input shape of (7, 3), indicating the use of features at 7 time steps, with 3 index values ​​at each time step. Two LSTM layers are set, each with 32 nodes. A fully connected layer with 16 nodes is set after the LSTM layers to further process the output of the LSTM layers. An output layer with 3 nodes is set, corresponding to the predicted values ​​of the three indicators. The preprocessed new data is input into the deployed model for prediction to obtain the predicted values ​​of the three indicators in the future. The system runs the environmental anomaly detection model and the environmental pollution trend prediction model at set time intervals. The model output results are matched with the rules in the rule base, and decision suggestions are generated based on the matched rules.

[0084] Example 2:

[0085] The environmental management intelligent analysis system based on multi-dimensional indicator monitoring can realize the environmental management intelligent analysis method based on multi-dimensional indicator monitoring described in Example 1, including:

[0086] Multidimensional index calculation module: It is used to obtain data on the emission and concentration of pollutants by monitoring the enterprise's exhaust gas outlets, wastewater discharge outlets and solid waste, while collecting energy consumption and equipment operation data, and calculating multidimensional index coefficients to comprehensively assess the environmental status.

[0087] Anomaly detection and trend prediction module: This module is used to construct an environmental anomaly detection model based on the historical data of the multi-dimensional index coefficients, and to perform anomaly detection on the real-time collected multi-dimensional index coefficients; at the same time, it combines the historical data of the multi-dimensional index coefficients and their time-series characteristics to construct an environmental pollution trend prediction model, and to predict the trend of the multi-dimensional index coefficients.

[0088] Decision generation and output module: When abnormal data is identified or adverse trends are predicted, it matches the rules in the decision rule base to generate corresponding decision suggestions and output them in a visual form.

[0089] As is known from common technical knowledge, this invention can be implemented through other embodiments that do not depart from its spirit or essential characteristics. Therefore, the disclosed embodiments described above are merely illustrative and not exhaustive. All modifications within the scope of this invention or its equivalents are included in this invention.

[0090] Those skilled in the art will understand that embodiments of the present invention can be provided as 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.

[0091] 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, and 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 processor, 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 illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0092] 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.

[0093] 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.

[0094] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. An intelligent analysis method for environmental management based on multi-dimensional indicator monitoring, characterized by: include: By monitoring the enterprise's exhaust gas outlets, wastewater discharge outlets, and solid waste, data on pollutant emissions and concentrations are obtained. Simultaneously, energy consumption and equipment operation data are collected to calculate a multi-dimensional index coefficient for comprehensive environmental assessment. This multi-dimensional index coefficient includes a comprehensive pollution emission index coefficient calculated based on the pollutant emission and concentration data, using the following formula: ； in, The comprehensive pollution emission index coefficient, For the standardized first The average emissions of these pollutants For the first The proportions corresponding to the emissions of each pollutant For the first Average concentration of various pollutants For the first Weighting coefficients corresponding to the concentrations of various pollutants; Historical data of the multi-dimensional index coefficients are collected, and an environmental anomaly detection model is constructed based on the historical data of the multi-dimensional index coefficients to detect anomalies in the real-time collected multi-dimensional index coefficients. At the same time, an environmental pollution trend prediction model is constructed by combining the historical data of the multi-dimensional index coefficients and their time series characteristics to predict the trends of the multi-dimensional index coefficients. A decision rule base is established by combining the anomaly detection results and trend prediction results. In response to the identification of abnormal data or the prediction of adverse trends, the rule base is matched with the rules to generate corresponding decision suggestions and output them in a visual form.

2. The intelligent analysis method for environmental management based on multi-dimensional indicator monitoring as described in claim 1, characterized in that, The multi-dimensional index coefficients also include the comprehensive pollution energy consumption index coefficient and the comprehensive treatment facility operation index coefficient, wherein: the comprehensive pollution energy consumption index coefficient is calculated based on the consumption data of polluting energy and clean energy in the energy consumption data, and the comprehensive treatment facility operation index coefficient is calculated based on the average operating rate and average failure frequency of production equipment in the equipment operation data.

3. The intelligent analysis method for environmental management based on multi-dimensional indicator monitoring as described in claim 1, characterized in that, The comprehensive pollution emission index coefficient is calculated based on the emission amount and concentration data of the pollutants, including: The emissions of air pollutants, water pollutants, and solid pollutants were standardized using Z-scores, and the proportions of the emissions of the three pollutants to the total emissions were calculated as corresponding weighting coefficients. At the same time, the concentration data of the three pollutants were standardized accordingly, and the weighting coefficients were determined based on the degree of environmental impact of the pollutants. The monthly average emissions and average concentrations of the three pollutants were calculated respectively. The average emissions of the three pollutants were added to their corresponding weighting coefficients, and the average concentrations were added to their corresponding weighting coefficients to obtain the comprehensive pollution emission index coefficient.

4. The intelligent analysis method for environmental management based on multi-dimensional indicator monitoring according to claim 2, characterized in that, The comprehensive pollution energy consumption index coefficient is calculated based on the consumption data of polluting energy and clean energy in the energy consumption data, including: Data is collected using appropriate metering instruments based on energy type. The consumption of various energy sources and the total energy consumption of the enterprise over 30 days are statistically analyzed. Energy is categorized according to its pollution level, including polluting and clean energy. All energy consumption is converted to a uniform unit, and the proportions of polluting and clean energy in the total energy consumption are calculated. The total polluting energy consumption is divided by the total number of days and then multiplied by the proportion. A comprehensive clean energy factor is set based on the environmental friendliness of clean energy. The total clean energy consumption is then divided by the total number of days, multiplied by the proportion and the comprehensive clean energy factor, and finally summed to obtain the comprehensive pollution energy consumption index coefficient. The calculation formula is as follows: ； in, The comprehensive pollution energy consumption index coefficient, Total energy source for polluting energy sources This represents the proportion of polluting energy sources. Total energy source for clean energy This represents the proportion of clean energy sources. It is a comprehensive cleaning agent.

5. The intelligent analysis method for environmental management based on multi-dimensional indicator monitoring according to claim 2, characterized in that, The comprehensive management facility operation index coefficients are calculated based on the average operating rate and average failure frequency of the production equipment in the equipment operation data, including: The number of devices in the enterprise is determined. An automated system is used to record the actual operating time of each device over 30 days. Based on production process requirements, the required operating time for each device is determined. The actual operating time of each device is divided by its required operating time to obtain the corresponding operating rate. The operating rates of all devices are summed and divided by the total number of devices to obtain the enterprise's average operating rate. Simultaneously, the number of failures and the actual number of times each device is used are recorded. The number of failures is divided by the actual number of times each device is used to obtain the corresponding failure frequency. The failure frequencies of all devices are summed and divided by the total number of devices to obtain the enterprise's average failure frequency. Based on the average failure frequency and average operating rate, a weighted calculation is performed to obtain the comprehensive management facility operation index coefficient. The calculation formula is as follows: ； in, This refers to the operational index coefficient of comprehensive governance facilities. Mean frequency of equipment failure. The average operating rate of the equipment. and These are the corresponding weighting coefficients.

6. The intelligent analysis method for environmental management based on multi-dimensional indicator monitoring according to claim 2, characterized in that, Historical data on the coefficients of the aforementioned multi-dimensional indicators were collected, including: For the comprehensive pollution emission index coefficient, comprehensive pollution energy consumption index coefficient, and comprehensive treatment facility operation index coefficient, historical data from the past year are collected. The coefficient data corresponding to these three indicators are collected periodically according to a predetermined time cycle. The collected data is cleaned and standardized. For the cleaned data, the mean and standard deviation of the comprehensive pollution emission index coefficient, comprehensive pollution energy consumption index coefficient, and comprehensive treatment facility operation index coefficient are calculated respectively. Based on historical data and practical experience, a threshold range is set for each indicator. The periodically collected data is compared with the set threshold range. If the corresponding coefficient of a certain indicator exceeds the set threshold range, the data point is determined to be abnormal data. The abnormal data points are recorded, including the outlier value and the determination time.

7. The intelligent analysis method for environmental management based on multi-dimensional indicator monitoring according to claim 2, characterized in that, An environmental anomaly detection model is constructed based on historical data of the aforementioned multi-dimensional index coefficients to detect anomalies in the real-time collected multi-dimensional index coefficients, including: Historical data containing the comprehensive pollution emission index coefficient, comprehensive pollution energy consumption index coefficient, and comprehensive treatment facility operation index coefficient are collected. The data is divided and normalized. A multilayer perceptron neural network is selected as the basic model, and a binary cross-entropy loss function is used as the loss function. The historical data divided into training sets is used to train the model. The trained environmental anomaly detection model is applied to the periodically collected new data to detect anomalies in the comprehensive pollution emission index coefficient, comprehensive pollution energy consumption index coefficient, and comprehensive treatment facility operation index coefficient. The model outputs labels and values ​​for the three index coefficients. If the model outputs a value of 0, it indicates that the data detection is normal. If the model outputs a value of 1, it indicates that the data monitoring is abnormal.

8. The intelligent analysis method for environmental management based on multi-dimensional indicator monitoring according to claim 2, characterized in that, An environmental pollution trend prediction model is constructed by combining historical data and time-series characteristics of the multi-dimensional indicator coefficients, and the trend prediction of the multi-dimensional indicator coefficients is performed, including: Using historical data containing the comprehensive pollution emission index coefficient, comprehensive pollution energy consumption index coefficient, and comprehensive treatment facility operation index coefficient, the data is determined to have timestamps and sorted in chronological order. Time features, lag features, and moving statistical features are added. Three index coefficients are selected as target variables. Long short-term memory network is selected as the basic model. Mean squared error is used as the loss function, and the Adam optimizer is used. The training data is divided into batches for training. Each time the model parameters are updated, the loss is calculated using samples and the parameters are updated. The trained environmental pollution trend prediction model is deployed to the system, and data is collected regularly to predict future pollution trends.

9. The intelligent analysis method for environmental management based on multi-dimensional indicator monitoring according to claim 1, characterized in that, A decision rule base is established by combining the anomaly detection results and trend prediction results. In response to the identification of abnormal data or the prediction of unfavorable trends, the rules in the decision rule base are matched to generate corresponding decision suggestions, which are then output in a visual format, including: Historical anomaly data, including indicator coefficient labels and corresponding values, is obtained from the environmental anomaly detection model. Corresponding processing rules are specified for different types of anomaly data. At the same time, pollution trend prediction results from the environmental pollution trend prediction model are collected, and corresponding countermeasures are specified for different pollution trends. Combining the anomaly detection and trend prediction results, comprehensive decision rules are formulated and stored in the database. The system runs the environmental anomaly detection model and the environmental pollution trend prediction model regularly, matches the model output results with the rules in the rule base, generates decision suggestions based on the matched rules, and displays them to the user in a visual form.

10. An intelligent environmental management analysis system based on multi-dimensional indicator monitoring, used to implement the intelligent environmental management analysis method based on multi-dimensional indicator monitoring as described in claim 1, characterized in that it includes: Multidimensional index calculation module: It is used to obtain data on the emission and concentration of pollutants by monitoring the enterprise's exhaust gas outlets, wastewater discharge outlets and solid waste, while collecting energy consumption and equipment operation data, and calculating multidimensional index coefficients to comprehensively assess the environmental status. Anomaly detection and trend prediction module: This module is used to construct an environmental anomaly detection model based on the historical data of the multi-dimensional index coefficients, and to perform anomaly detection on the real-time collected multi-dimensional index coefficients; at the same time, it combines the historical data of the multi-dimensional index coefficients and their time-series characteristics to construct an environmental pollution trend prediction model, and to predict the trend of the multi-dimensional index coefficients. Decision generation and output module: When abnormal data is identified or adverse trends are predicted, it matches the rules in the decision rule base to generate corresponding decision suggestions and output them in a visual form.

Images