In Korea, medical waste is placed in special containers and disposed of separately, and it is impossible to open it or analyze its calorific value before incineration. Due to these characteristics, difficulties are arising in the operation of incinerators exclusively for medical waste due to the characteristics of incinerators whose operating conditions are determined based on the calorific value of the waste. In addition, since incinerators require continuous optimization of operating conditions as equipment ages, operation optimization through prediction of operating conditions is necessary. In this study, the incinerator operation characteristics were analyzed using a Neural Network machine learning model using TMS data from a 20-ton/day commercial medical waste incinerator. In order to exclude abnormal operation conditions such as during the overhaul period, operation data was acquired at points where the combustion chamber temperature was above 850℃, and after learning this, correlation analysis between variables and accuracy of prediction data were analyzed. As a result of this study, it was confirmed that after COVID-19, a strong correlation occurred between variables related to combustion conditions (O2, exhaust gas flow rate, combustion chamber temperature, and combustion exhaust gas flow rate) that were not observed before COVID-19. In the before and after COVID-19 variable prediction results, it was confirmed that the exhaust gas flowrate prediction interval was expanded and the O2 concentration cluster formation trend changed significantly. In summary, it was confirmed that significant changes in incinerator operation characteristics occurred before and after the COVID-19 pandemic due to the high calorific value of COVID-19 medical waste and the same-day-discharge same-day incineration principle.
1. 서론
2. 기계학습 알고리즘 설계
3. 의료폐기물 소각로의 COVID-19 전-후 운전특성 변화
4. 결론