의사결정나무 기반 하천 유량 및 수질 예측모델의 성능 평가

전기일 ( Gee Ill Jun ),조경철 ( Gyeong Cheol Jo ),남건우 ( Gun-woo Nam ),기서진 ( Seo Jin Ki ) 한국환경기술학회 2022 한국환경기술학회지 Vol.23 No.6

This study aimed to investigate the performance of decision tree-based models for water quantity and quality prediction. The models adopted for performance assessment included decision tree (DT), random forest (RF), and extreme gradient boosting (XGB), which was fed by the data sets collected from two monitoring stations in the Nakdong River during 2018-2021. A 7:3 ratio was used to prepare training and testing sets for three prediction models and their hyperparmeters were tuned to improve the accuracy of prediction. We found that XGB which was not sensitive to input data resolution outperformed the other two models, DT and RF. In contrast, the prediction error for DT model decreased progressively in response to increasing monitoring frequency from 7 through 3 to 1 day as well as after applying post-pruning, regardless of dependent variables. When the accuracy of prediction for RF model was assessed as a function of the number of independent variables, more than 4 variables was effective in maintaining its prediction performance as compared to all variables adopted. Therefore, both monitoring frequency and pruning play an important role in reducing the prediction error of decision tree models, in addition to hyperparameter optimization.

하천 유량 및 수질 예측을 위한 머신러닝 알고리즘 성능 비교

전기일 ( Gee Ill Jun ),권도혁 ( Do Hyuck Kwon ),기서진 ( Seo Jin Ki ) 한국수처리학회(구 한국수처리기술연구회) 2020 한국수처리학회지 Vol.28 No.1

This study was conducted to compare the predictive accuracy of three popular machine learning algorithms, i.e., Bayesian Regularized Neural Networks (BRNN), Neural Network (NN), and Support Vector Machines with Radial Basis Function Kernel (SVM), for modeling river water quality and quantity. Input data with 11 parameters collected from three monitoring sites in the lower part of the Nakdong River between January 2008 to December 2018 were input into the three algorithms. The data were divided into two subsets, which included a training data set and test data set in a ratio of 70:30. The results showed that NN displayed better performance in prediction accuracy than BRNN and SVM when optimizing one or two tuning parameters. The prediction accuracy was, on average, higher for chlorophyll a than for discharge, regardless of which machine learning algorithm was used. Identical results were also observed in the training data set. The two variables biochemical oxygen demand (BOD) and chemical oxygen demand (COD) played an important role in predicting chlorophyll a, whereas suspended solids (SS) was designated as the largest contributor for the prediction of discharge in all three algorithms. In terms of error metrics and performance indices such as root mean squared error (RMSE), mean absolute error (MAE), and R², the three machine learning algorithms exhibited improved performance in the test data set, as compared to the training data set. We believe that this preliminary study helps identify suitable methods for enhanced prediction of emerging water quality or quantity parameters such as harmful algal blooms.

과수의 탄소발자국 표지를 위한 LCA 동향 및 해석 : 경남지역 참다래를 중심으로

Markus Deurer,Brent Clothier,허근영(Keun-Young Huh),전기일(Gee-Ill Jun),김인혜(Inhea Kim),김대일(Daeil Kim) 한국원예학회 2011 원예과학기술지 Vol.29 No.5

국내외 탄소발자국 연구 동향을 분석하였으며 뉴질랜드 참다래의 전과정 평가 방법의 체계를 해석하였고 경남지역 참다래를 중심으로 과수의 탄소발자국표지 적용에 대하여 제언하였다. 제품의 탄소발자국은 일반적으로 ISO14040 시리즈 PAS 2050 GHG Protocol 등과 같은 국제표준에 따라서 평가된다. 국내 탄소발자국 표지는 한국환경산업기술원이 환경부의 위탁하에 한국 표준을 마련하고 "기후변화대응" "저탄소상품"이란 문구가 표시된 탄소성적표지를 운영하고 있지만 현재까지 농산물은 인증대상에서 제외되어 있다. 반면 영국 스웨덴 덴마크 뉴질랜드 등은 농산물의 전과정 평가를 이미 수행하고 있다. 뉴질랜드 참다래의 전과정 평가는 ISO 표준 방법들을 준용하고 ISO 표준에서 분명한 기준이 없는 경우에는 PAS 2050 표준을 준용한다. 시스템 경계는 뉴질랜드 참다래의 총 온실가스 배출량에 대한 각 단계에서의 영향을 이해하기 위해서 과원 생산에서 소비까지의 모든 전과정 단계들을 포함하였다. 전과정 평가 결과에서 총 온실가스 배출량에 대한 각 단계의 영향은 과원 생산 17% 포장(냉장) 12% 뉴질랜드 항구 선적 1% 선박 운송 41% 영국 Zeebrugge에서의 재포장 3% 대형유통점 6% 소비 및 쓰레기처리 21%로 나타났다. 선박운송과 같은 장거리 운송은 무역에서 장애물이 될 수 있으며 이것을 극복하기 위해서는 과원과 냉장에서 온실가스의 감축이 필요하다. 국내 산업현황 분석에 따르면 최근 참다래 수입량은 매우 빠르게 증가하고 있음에도 불구하고 참다래 재배면적과 생산량은 감소하고 있으며 수출량도 감소하고 있다. 그 이유는 수입 농산물들에 의해서 경제적 이익이 감소하였기 때문인데 탄소발자국 표지는 참다래뿐만 아니라 국내 과수 산업의 새로운 활력을 가져올 수 있다. 과수의 국제 표준에 의한 탄소발자국 도입은 가능한 한 조속히 수행되어야 할 것이다. 본 논문에서 제시된 뉴질랜드 참다래의 전과정 평가 방법 체계는 이를 위해서 유용하게 활용될 수 있다. As part of a feasibility study for introducing carbon labeling of fruit products in Korea we explore the use of carbon footprints for Korean kiwifruit from Gyeongnam region as a case study. In Korea the Korean Environmental Industry and Technology Institute (KEITI) is responsible for the carbon footprint labeling certification and has two types of certification programs: one program focuses on climate change response (carbon footprint labeling analysis) and the other on low-carbon products (reduction of carbon footprints analysis). Currently agricultural products have not yet been included in the program. Carbon labeling could soon be a prerequisite for the international trading of agricultural products. In general the carbon footprints of various agricultural products from New Zealand followed the methodology described in the ISO standards and conformed to the PAS 2050. The carbon footprint assessment focuses on a supply chain and considers the foreground and the background systems. The basic scheme consists of four phases which are the'goal','scope','inventory analysis' and'interpretation' phases. In the case of the carbon footprint of New Zealand kiwifruit the study tried to understand each phase's contribution to total GHG emissions. According to the results shipping orchard and coolstore operation are the main life cycle stages that contribute to the carbon footprint of the kiwifruit supply chain stretching from the orchard in New Zealand to the consumer in the UK. The carbon emission of long-distance transportation such as shipping can be a hot-spot of GHG emissions but can be balanced out by minimizing the carbon footprint of other life cycle phases. For this reason it is important that orchard and coolstore operations reduce the GHG-intensive inputs such as fuel or electricity to minimize GHG emissions and consequently facilitate the industry to compete in international markets. The carbon footprint labeling guided by international standards should be introduced for fruit products in Korea as soon as possible. The already established LCA methodology of NZ kiwifruit can be applied for fruit products as a case study.

병원폐기물의 처리현황과 대책에 관한 연구

전기일,김형갑 진주산업대학교 환경문제연구소 1998 環境硏究論文集 Vol.2 No.-

국민들의 의료기관이용이 증가함에 따라 의료활동에 수반되어 발생하는 폐기물 발생량이 증가하고 있으며 이러한 병원폐기물의 관리는 건강과 환경의 위험을 최소화하는 방법에 의하여 폐기물의 취급, 저장, 운송, 처리의 체계적 접근이 필요하다. 그래서 본 논문에서는 우리나라 병원폐기물의 관리현황과 실태, 적출물 관리 중심으로 문제점을 파악하여 그 개선방안을 제시하였다.

진주시 도시폐기물의 최적관리 방안

전기일 진주산업대학교 환경문제연구소 1998 環境硏究論文集 Vol.1 No.-

최근 산업의 고도화 및 대량생산과 소비의 증대에 따라 복잡 다양한 폐기물의 발생량의 증가로 처리 처분의 문제점으로 귀결되어 폐기물의 적정처리에 대한 논란이 야기되어지고 있는 실정이며 또한 발생원에서의 소량배출을 우선 요구되고 있는 실정이기도 하다. 본 연구에서는 진주시에 있어서 생태도시로서 거듭나기 위한 시민, 기업, 지방자치단체의 친환경성 검토에 이르게 되어 환경문제 해결의 실마리를 제시하고자 한다.

有機性슬러지의 油溫脫水處理

전기일,박현건,석철원 진주산업대학교 산업과학기술연구소 1994 산업과학기술연구소보 Vol.- No.1

본 硏究에서는 油溫脫水法을 利用하여 食品廢水 슬러지를 飼料化할 경우에 필요한 基礎資料를 얻고자 60mmHg減壓下에서 각각 다른 油溫(90℃, 120℃)을 유지하면서 油溫脫水시켰을 때 각 油溫脫水시간에 따른 油溫脫水前後의 化學的 變化등을 분석한 결과는 다음과 같다. 1. 60mmHg減壓下에서 食品發水 슬러지를 油溫脫水했을 때, 각 油溫脫求 時間(20, 40, 60분)에 대한 含水率은, 油溫 90℃일때는 14.7%, 10.5%, 8.6%이었고, 油溫120℃일 때는 12.0%, 9.3%, 7.8%이었다. 20分問 油溫脫水時 水分 除去律이 각각 81.9%, 85%로 나타나 油溫脫氷 初期에 急速히 脫水됨을 알 수 있었다. 2. 食品發水 슬러지의 祖雲白質 含量은 油溫 90℃에서 40分間 油溫脫水 했을 때가 28.8%로 가장 높았고, 雲白質 飼料의 條件을 滿足하였다. 3. 食品發水 슬러지의 粗脂肪 含量은 油溫脫水 時間이 길어질수록 증가하는 경향을 보였고, 油溫이 90℃일 때 보다 120℃에서 다소 높은 粗脂昉 含量을 나타냈다. 4. 食用油의 酸價와 過酸化物價는 增加하는 경향을 보였지만, 油湯脫水法을 이용하므로써 貪用油의 酸敗를 줄일 수 있었다. This study has been carried out to obtain the basic data when food wastewater sludge is utilized for feeding by dehydration in frying oils for 20min, 40min, 160min, at 700mmHg(90, 120℃). Experimental results were summarized as follows : 1 . The water contents of food wastewater sludge was rapidly decreased at the early stage. And the minimum value of water contents was 7.8%. 2. The maximum crude protein contents of fried food wastewater sludge was 28.8%. The value was satisfied for feed. 3. The crude fat contents of fried food wastewater sludge tended to increase more frying time, and the fried sludge at 120℃ was slightly higher than 90℃. 4. The acid and peroxide value of frying oils tended to increase more frying time, but dehydration in frying oils decreased acidification of frying oils.

진주시 쓰레기 성상분석에 관한 연구

전기일 진주산업대학교 2001 산업과학기술연구소보 Vol.- No.8

The physico-chemical characteristics of municipal solid waste from 3 difforent source in Jinju City were investigated. The source are Apartment, residental house and market area. The results are as follows 1. The average physical composition of waste from Apartment, residental house and market area were food waste 68%, 64%, 46% and paper 14%, 22%, 11% and vinyls 7.6%, 4.9%, 6.4% respectively. 2. The bulk density of waste from Apartment, residental house and market area were 434.4㎏/㎥, 434.2㎏/㎥, 361㎏/㎥ respectively. 3. The average of three major contentes of waste from Apartment, residental house and market area were moisture contentes 52.1%, 51.5%, 52.3% and combustible fraction 42.12%, 41.31%, 39.72% and ash contentes 6.72%, 7.18%, 7.98% respectively. 4. The elemental analysis of C, H, 0, N and S of waste from Apartment, residental house and market area were analysed C contentes 59.78%, 53.82%, 54.47% and H contentes 8.65%, 733%, 7.28% and O contentes 21.53%, 29.78%, 28.53% respectively.

소형 퇴비화 장치 개발

전기일 진주산업대학교 1998 論文集 Vol.37 No.-

The compostable foodwaste of 1㎏ a day were added in the composter without applying additives. Thereafter, composting was proceeded successively in the composter, which had the mass reduction and physico-chemical of the produced compost were analyzed. The results are as follows; During the composting process, Water content maintained in the range of 51.04∼53.45%. Hemicellulose and Lignin contents did not show any tendency, but Cellulose content decreased. During the composting process, NH_3-N and NO_2-N were not detected due to nitrification. The contents of inorganic compounds were not accumulated during the composting process. They were in the range of 1.32∼1.71% P_2O_5, 1.29∼1.48% CaO, 0.41∼0.49% MgO, 0.38∼0.74% K_2O. For 20 days, decomposition rate was 67.5% in wet base and 48% in dry base. Concentration of heavy metals(Cu, Cr, Cd, Pb, Zn, Hg, As) was less than the limiting value of the compost. Maturity of the produced compost was 3 grade with maximum reaching temperautre of 46∼48℃.

진주시 쓰레기봉투속의 성분조사

전기일,정병윤,이춘식 진주산업대학교 2000 산업과학기술연구소보 Vol.- No.7

The purpose of this study was to investigated the physics-chemical characteristic of municipal solid waste generated in the Chinju city. It is important to know the physico-chemical characteristic of municipal solid waste. The average physical composition of municipal solid waste in APT and Detached Dwelling are food waste 68% and 55.5%, paper 13.5% and 16.5%, vinyl & plastic 13.7% and 16.5%, glass 2.73% and 1.7%, metal 2.06% and 1.03%, others 6.6% and 2.1%, respectively.

음식물쓰레기 사료화에 대한 고찰

전기일,박현건 진주산업대학교 환경문제연구소 1998 環境硏究論文集 Vol.2 No.-

본 연구에서는 생활쓰레기 발생량중 가장 비중을 많이 차지하는 음식물쓰레기의 처리방안을 찾기위하여 음식물쓰레기 발생현황과 문제점 및 처리실태를 조사하고 음식물쓰레기의 사료화의 조건과 처리시설 및 발전방향에 대하여 고찰하였다.