http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
재실자 중심 최적 제어를 위한 재실 패턴 예측모델 개발
최영재(Young Jae Choi),김태원(Tae Won Kim),변재윤(Jae Yoon Byun),문진우(Jin Woo Moon) 대한설비공학회 2022 대한설비공학회 학술발표대회논문집 Vol.2022 No.6
Recently, occupant-centric control (OCC) has been attracting attention for energy saving in buildings, and occupancy forecasting is a key element of this. Thus, the purpose of this study is to develop a time series-based intelligent model that predicts occupancy patterns and to check the predictability for various prediction horizons. The forecasting model was developed using a long short-term memory (LSTM) neural network which predicts the number of occupancy after 15 minutes, 60 minutes, and 180 minutes. To evaluate the performance of the developed model, predictions were conducted with test data. mean absolute error (MAE) and root mean squared error (RMSE) were calculated based on the errors between the actual number of occupancy and the predicted value. As a result, the MAE and RMSE of the 15-minute forecasting model was 1.54 and 2.25, respectively showing the remarkable performance. Although the prediction accuracy decreases as the prediction period increases, the 60-minute and 180-minute forecasting model presented superior performance compared to the previous studies with MAE = 2.65, RMSE = 4.18, and MAE = 4.62, RMSE = 6.90, respectively. Therefore, it was possible to confirm the applicability of the developed occupancy pattern forecasting model, and it is expected that it can be used for optimal intelligent building control through performance improvement in the future.
재실자 중심 최적 제어를 위한 재실 패턴 예측모델 개발
최영재(Young Jae Choi),김태원(Tae Won Kim),변재윤(Jae Yoon Byun),문진우(Jin Woo Moon) 대한설비공학회 2022 대한설비공학회 학술발표대회논문집 Vol.2022 No.6
Recently, occupant-centric control (OCC) has been attracting attention for energy saving in buildings, and occupancy forecasting is a key element of this. Thus, the purpose of this study is to develop a time series-based intelligent model that predicts occupancy patterns and to check the predictability for various prediction horizons. The forecasting model was developed using a long short-term memory (LSTM) neural network which predicts the number of occupancy after 15 minutes, 60 minutes, and 180 minutes. To evaluate the performance of the developed model, predictions were conducted with test data. mean absolute error (MAE) and root mean squared error (RMSE) were calculated based on the errors between the actual number of occupancy and the predicted value. As a result, the MAE and RMSE of the 15-minute forecasting model was 1.54 and 2.25, respectively showing the remarkable performance. Although the prediction accuracy decreases as the prediction period increases, the 60-minute and 180-minute forecasting model presented superior performance compared to the previous studies with MAE = 2.65, RMSE = 4.18, and MAE = 4.62, RMSE = 6.90, respectively. Therefore, it was possible to confirm the applicability of the developed occupancy pattern forecasting model, and it is expected that it can be used for optimal intelligent building control through performance improvement in the future.
황인국(In Guk Hwang),변재윤(Jae Yoon Byun),김경미(Kyung Mi Kim),정미남(Mi Nam Chung),유선미(Seon Mi Yoo) 한국식품영양과학회 2014 한국식품영양과학회지 Vol.43 No.6
본 연구에서는 비타민 C 분석법을 검증하고 국내산 고구마 22품종과 조리방법에 따른 고구마의 비타민 C 함량을 분석하였다. 비타민 C 분석법을 검증하기 위해 직선성, 검출한계, 정량한계, 정밀성 및 정확성을 확인하였다. 그 결과 직선성의 상관계수 값이 0.9999이었으며, 검출한계는 0.03 μg/mL, 정량한계는 0.10 μg/mL, 정밀성의 상대표준편차는 5%이하, 정확성인 회수율은 95% 이상으로 우수하였다. 고구마 품종별 AA, DHA 및 TA 함량은 각각 37.76(신율미)~89.25(주황미), 23.37(신자미)~63.94(신율미) 및 68.52(신자미)~115.95(주황미) mg/100 g 범위로 품종에 따라 큰 차이를 보였다. 고구마의 평균 AA, DHA 및 TA 함량은 각각 56.98±12.53, 36.46±9.03 및 93.44±12.00 mg/100 g이었으며, 대부분 품종의 AA 함량은 40~70 mg/100 g 범위에, DHA함량은 20~40 mg/100 g 범위에, TA 함량은 70~90 mg/100 g 범위에 존재하였다. 그리고 육질색 종류에 따른 평균 TA 함량은 일반고구마와 주황색고구마가 자색고구마에 비해 유의적으로 높은 것으로 나타났다. Steaming, baking 및 frying 처리에 따른 AA, DHA 및 TA 함량은 조리 처리 후 10.61~58.41, 2.57~52.81 및 14.54~49.92% 범위로 유의적으로 감소하였고, baking 처리가 steaming 및 frying 처리에 비해 함량 감소량이 큰 것으로 나타났다. 고구마의 비타민 C 함량은 품종 및 조리방법에 따라 변이가 큰 것으로 나타났으며, 추후 연구의 기초자료로 활용이 가능할 것으로 기대된다. This study was carried out to investigate the amounts of vitamin C in 22 sweet potato cultivars cultivated in Korea as well as evaluate the effects of cooking methods on vitamin C contents. Methods for determining vitamin C was validated by determining linearity, specificity, limit of detection (LOD), limit of quantification (LOQ), precision, and accuracy using HPLC. Results showed high linearity in the calibration curve with a coefficient of correlation (R2) of 0.9999. The LOD and LOQ values for ascorbic acid (AA) were 0.03 and 0.10 μg/mL, respectively. The relative standard deviations (RSDs) for intra- and inter-day precision of AA were less than 5%. The recovery rates of AA and dehydroascorbic acid (DHA) were in the range from 98.21~98.64 and 98.28~100.68%, respectively. Depending on cultivar, contents of AA, DHA, and total ascorbic acid (TA) in sweet potatoes varied in the range from 37.76 (Sinyulmi)~89.25 (Juhwangmin), 23.37 (Sinjami)~63.94 (Sinyulmi), and 68.52 (Sinjami)~115.95 (Juhwangmin) mg/100 g, respectively, and their average levels were 56.98±12.53, 36.46±9.03, and 93.44±12.00 mg/100 g, respectively. The average TA levels were also dependent on flesh color, whish was significantly higher in general sweet potato and orange sweet potato than in purple sweet potato. Steaming, baking, and frying processes significantly reduced AA (10.61~58.41%), DHA (2.57~52.81%), and TA (14.54~49.92%) contents in sweet potatoes. The highest reduction of AA, DHA, and TA contents was observed after baking, followed by steaming and frying. We expect that the basic information provided by this study will be useful to plant breeders and food scientists.