KR-102963110-B1 - MAJOR POLLUTANT MANAGEMENT SYSTEM AND METHOD THEREOF

KR102963110B1KR 102963110 B1KR102963110 B1KR 102963110B1KR-102963110-B1

Abstract

The present invention discloses a pollutant management system and method. According to a specific embodiment of the present invention, the management range of operational factors of a workplace that manages the pollutant emission concentration emitted from the workplace can be objectively presented, the reliability of the presented management range of operational factors of the workplace is improved, and the pollutant emission concentration can be reduced. Furthermore, by selecting a major operational factor based on the PI of each operational factor among at least one operational factor of the workplace, selecting a selected operational factor among at least one major operational factor based on the sensitivity of the selected major operational factor, and controlling the selected operational factor by adjusting the management range of the selected operational factor, the number of operational factors for managing the pollutant emission concentration is reduced, thereby reducing the computational complexity of pollutant management. Consequently, the pollutant emission concentration can be managed in real time, and human error in pollutant emission management can be fundamentally eliminated, thereby having the effect of preventing environmental pollution caused by pollutants exceeding regulations in advance.

Inventors

강필구
이형주
박재홍
김선정
박지의
전태완
감종훈
민승기
이광훈
최승희

Assignees

대한민국(환경부 국립환경과학원장)
포항공과대학교 산학협력단

Dates

Publication Date: 20260511
Application Date: 20240502

Claims (11)

delete
A preprocessing unit that collects operational factor data and pollutant data collected from automatic measuring instruments installed at least one workplace and preprocesses the collected operational factor data and pollutant data according to predetermined standards; A pollutant emission derivation unit that outputs at least one predicted pollutant emission amount by learning based on a pre-established learning model into which the above-mentioned preprocessed operational factor data is input; An operational factor selection unit that derives the performance of a learning model based on the error between the predicted pollutant emission amount of the learning model and the actual pollutant emission amount, derives the permutation importance of each operational factor based on the derived performance of the learning model, and selects at least one major operational factor based on the derived permutation importance, wherein the permutation importance is derived from the change in the performance of the learning model according to the variation of each input operational factor, sorts in order of high permutation importance, and then selects a predetermined number of operational factors with high permutation importance as major operational factors; and It includes a pollutant management unit that manages the emission concentration of pollutants at each business site by deriving the sensitivity of each selected major operating factor based on at least one selected major operating factor and the rate of change of pollutant emissions, selecting at least one selected operating factor based on the sensitivity of each major operating factor, and adjusting the management range of each selected operating factor based on the analysis results of the sensitivity of the selected selected operating factor, wherein the sensitivity is derived from the change in the management range of each selected major operating factor and the rate of change of predicted pollutant emissions derived from the learning results of a learning model for each major operating factor, and is equipped to verify the direction and range of change of the management range of the major operating factor. It further includes a control range adjustment module that, when the predicted pollutant emission amount derived from the learning result of the above-mentioned selected operating factor is not less than a predetermined set value, adjusts the control range of the selected operating factor based on the direction of fluctuation and control range of the predicted pollutant emission amount derived from the analysis result of the sensitivity of each selected operating factor, and outputs the predicted pollutant emission amount derived from the learning result of the above-mentioned learning model for each selected operating factor adjusted to the control range. The above management range adjustment module is, A pollutant management system characterized by being configured to adjust the control range of the selected operating factor to an adjustment value (N=N-1) in which a predetermined unit value (1%) is reduced from the reference value (N(%)=100) when the predicted pollutant emission amount output above is not below a predetermined threshold.
In paragraph 2, the pollutant emission amount derivation unit is, A learning module that inputs the above-mentioned preprocessed operational factor data into a pre-established learning model for training, derives predicted pollutant emissions as a result of the training, generates training data with predicted pollutant emissions matching the input operational factor, and divides the generated training data into training data and validation data according to a predetermined ratio; and A pollutant management system characterized by including a training module that derives the average absolute error of at least one derived predicted pollutant emission amount at each time point, sets hyperparameters of the learning model so as to minimize the derived average absolute error, and trains the learning model to derive the predicted pollutant emission amount based on the learning model of the set hyperparameters.
In paragraph 3, the above-mentioned operation factor selection unit is, A PI derivation module that receives validation data of the above learning model and derives the permutation importance of at least one operational factor based on the performance of the above learning model; and A pollutant management system characterized by including a major operating factor selection module that selects at least one major operating factor based on the permutation importance of at least one operating factor.
In paragraph 4, the above-mentioned pollutant management department, A sensitivity derivation module that derives the sensitivity of each major operating factor based on the change in the control range of each major operating factor among at least one selected major operating factor and the rate of change in predicted pollutant emissions derived from the learning result of a learning model for each major operating factor whose control range has changed; and A pollutant management system characterized by including an operating factor management module that manages the pollutant emission concentration of each workplace by controlling at least one selected operating factor to a changed management range when the average value of the predicted pollutant emission amount of the management range adjustment module, derived as a result of learning a learning model for at least one selected operating factor selected based on sensitivity among the at least one major operating factor, is less than a predetermined set value.
delete
In a pollutant management method performed in accordance with the pollutant management system of paragraph 2, At least one processor included in the pollutant management system, (a) A learning step that outputs predicted pollutant emissions as a result of learning a pre-established learning model into which operational factor data of at least one collected business site is input; (b) A key operational factor selection step in which the permutation importance PI of each operational factor is derived based on the performance of the learning model generated through the predicted pollutant emissions and actual pollutant emissions, and the key operational factors among the operational factors of the workplace are derived based on the derived PI of each operational factor, wherein the permutation importance PI is derived from the change in the performance of the learning model according to the variation of each input operational factor, and after sorting in order of high permutation importance, a predetermined number of operational factors with high permutation importance are selected as key operational factors; (c) A sensitivity derivation step for deriving the sensitivity of a major operating factor based on the rate of change of the predicted pollutant emission amount of the learning model according to the variation of the major operating factor derived above, and selecting a selected operating factor among at least one major operating factor based on the sensitivity of the major operating factor; wherein the sensitivity of the major operating factor is derived from the variation of the management range of each selected major operating factor and the rate of change of the predicted pollutant emission amount derived from the learning result of the learning model for each major operating factor, thereby confirming the direction and range of variation of the management range of the major operating factor; and (d) an operating factor management step comprising: adjusting the control range of the selected operating factor based on the directionality and range of variation of the selected operating factor derived from the analysis result of the sensitivity of the major operating factor derived above; and controlling the selected operating factor to the adjusted control range to manage the pollution concentration emission concentration when the predicted pollutant emission amount output as the learning result of the learning model for the major operating factor of the adjusted control range is less than a predetermined threshold (within the regulatory range); The above operational factor management step is, A pollutant management method characterized by being configured to adjust the control range of the selected operating factor to an adjustment value (N=N-1) in which a predetermined unit value (1%) is reduced from the reference value (N(%)=100) when the predicted pollutant emission amount output above is not below a predetermined threshold.
In paragraph 7, the above step (d) is, (d-1) A step of setting the control range of a selected operating factor among the derived major operating factors to a predetermined standard value (N(%)=100); (d-2) A step of outputting a predicted pollutant emission amount as a learning result of a learning model into which a selected operating factor of a set reference value is input; and (d-3) A pollutant management method characterized by including the step of controlling a selected operating factor to a set management range when the predicted pollutant emission amount output above is less than a predetermined threshold.
delete
A computer-readable recording medium characterized by having a program recorded thereon for executing a pollutant management method of either paragraph 7 or paragraph 8 on a computer.
In a computer program stored on a computer-readable recording medium for executing a pollutant management method on a computer in combination with a computer, The above pollutant management method is, (a) A learning step that outputs predicted pollutant emissions as a result of learning a pre-established learning model into which operational factor data of at least one collected business site is input; (b) A key operational factor selection step in which the permutation importance PI of each operational factor is derived based on the performance of the learning model generated through the predicted pollutant emissions and actual pollutant emissions, and the key operational factors among the operational factors of the workplace are derived based on the derived PI of each operational factor, wherein the permutation importance PI is derived from the change in the performance of the learning model according to the variation of each input operational factor, and after sorting in order of high permutation importance, a predetermined number of operational factors with high permutation importance are selected as key operational factors; (c) A sensitivity derivation step for deriving the sensitivity of a major operating factor based on the rate of change of the predicted pollutant emission amount of the learning model according to the variation of the major operating factor derived above, and selecting a selected operating factor among at least one major operating factor based on the sensitivity of the major operating factor; wherein the sensitivity of the major operating factor is derived from the variation of the management range of each selected major operating factor and the rate of change of the predicted pollutant emission amount derived from the learning result of the learning model for each major operating factor, thereby confirming the direction and range of variation of the management range of the major operating factor; and (d) an operating factor management step of adjusting the control range of the selected operating factor based on the directionality and range of variation of the selected operating factor derived from the analysis result of the sensitivity of the major operating factor derived above, and controlling the selected operating factor to the adjusted control range to manage the pollution concentration emission concentration when the predicted pollutant emission amount output as the learning result of the learning model for the major operating factor of the adjusted control range is less than a predetermined threshold (within the regulatory range); wherein An operation program of a pollutant management system characterized by being configured to adjust the control range of a selected operating factor to an adjustment value (N=N-1) in which a predetermined unit value (1%) is reduced from the reference value (N(%)=100) when the predicted pollutant emission amount output above is not below a predetermined threshold.

Description

Major Pollutant Management System and Method Thereof The present invention relates to a pollutant management system and method, and more specifically, to a technology that enables objective management of pollutant emission concentrations by correcting the management range of operating factors emitting pollutants for each factory or facility. The Korea Environment Corporation implements the Chimney Remote Monitoring System (CleanSYS) and makes the data publicly available online. Under this system, millions of data points are generated annually from Continuous Emission Monitoring Systems (TMS), which measure pollutant emission concentrations from chimneys at pre-set intervals for each factory. This vast amount of data is organized using general-purpose software following manual processes, and administrative sanctions are imposed based on the determination of emission limits. Emission management for these pollutants is operated in various forms, such as distributed control systems (DCS), for multiple factories or facilities, and efficient pollutant emission management focused on permissible emission standards involves human decision-making. Consequently, pollution emission management involving human decision-making lacks objectivity. Therefore, it is necessary to develop technology that provides a service for numerically quantifying the management range of operational factors for adjusting pollutant emission concentrations for pollutants whose emissions exceed regulations. The present invention is the result of the Smart Integrated Environmental Management System Establishment Pilot Project (II) conducted in 2023 with funding from the Ministry of Environment and support from the National Institute of Environmental Research (Project No.: NIER-2023-04-02-166). The following drawings attached to this specification illustrate preferred embodiments of the present invention and serve to further enhance understanding of the technical concept of the present invention together with the detailed description of the invention provided below; therefore, the present invention should not be interpreted as being limited only to the matters described in such drawings. FIG. 1 is a configuration diagram of a pollutant management system according to one embodiment. Figure 3 is a figure showing the output data of the preprocessing unit of Figure 1. Figure 3 is a detailed configuration diagram of the pollutant emission amount derivation unit of Figure 1. Figure 4 is a detailed configuration diagram of the operation factor selection section of Figure 1. Figure 5 is a figure showing the amount of pollutants emitted according to the performance of the learning model of Figure 4. Figure 6 is a detailed configuration diagram of the pollutant management unit of Figure 1. Figure 7 is an example diagram showing the sensitivity of the major operating factors of Figure 6. Figure 8 is a flowchart showing the pollutant management process of another embodiment. Embodiments of the present invention are described below with reference to the attached drawings so that those skilled in the art can easily implement them. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. Furthermore, in order to clearly explain the present invention in the drawings, parts unrelated to the explanation have been omitted, and similar parts throughout the specification are denoted by similar reference numerals. One embodiment described below specifically explains a configuration for managing pollutant emission concentrations at a workplace by analyzing the sensitivity of each operational factor based on the rate of change in pollutant emissions due to variations in the control range of the operational factor of the workplace, and adjusting the control range of the operational factor for pollutants exceeding regulations based on the analysis results of the sensitivity of each operational factor. FIG. 1 is a configuration of a pollutant management system according to one embodiment, FIG. 2 is a diagram showing output data of the preprocessing unit of FIG. 1, FIG. 3 is a detailed configuration diagram of the pollutant emission amount derivation unit of FIG. 1, FIG. 4 is a detailed configuration diagram of the major operating factor selection unit of FIG. 1, FIG. 5 is a diagram showing pollutant emission amounts according to learning results by learning model performance of FIG. 4, FIG. 6 is a detailed configuration diagram of the pollutant emission management unit of FIG. 1, and FIG. 7 is an example diagram showing the sensitivity of the major operating factor of FIG. 6. Referring to FIGS. 1 to 7, a pollutant management system of one embodiment is configured to manage pollutant emission concentrations by analyzing the sensitivity of operating factors based on the rate of change of pollutant emission amount according to the change in the management range of operating factors of a workplace, and ad