Following the hydrofluoric acid leak accident in Gumi, domestic interest in chemical accidents has surged. In preventing chemical accidents, Quantitative Risk Assessment (QRA) is utilized to derive quantitative results regarding the risks of chemical processes, storage, transportation, and other activities. The accuracy of frequency analysis in QRA depends on the underlying data, so it is crucial to secure high-quality data. Although there are statistical data collected by the National Institute of Chemical Safety in Korea, these are not direct frequency data, making it challenging to use them directly for frequency analysis. Therefore, this study researched a modification factor to convert IOGP (The Internaltional Association of Oil &Gas Producers) frequency data into domestic frequency data. First, the data to be used for deriving the modification factor were selected, and the domestic modification factor was presented and its applicability reviewed. Given that South Korea has a significant proportion of the electrical and electronics industries, and workplaces are enclosed, it is expected that frequency values will differ from other industries, so factors specific to these conditions were also estimated. Moreover, due to the characteristics of IOGP frequency data, only frequency values for specific equipment sizes and leak hole sizes are known, but the actual sizes of equipment and leak holes vary significantly. Thus, the form of IOGP data is not suitable for frequency analysis. Consequently, to analyze frequencies for various equipment, a leakage frequency function was derived through nonlinear regression analysis. The results provided functions and parameters suitable for equipment-specific data, and using the modification factor and frequency values obtained from the leakage frequency function, it will be possible to derive accident frequencies appropriate to the process situation.|구미 불화수소 누출 사고 이후 화학 사고에 대한 국내의 관심도가 급증하였다. 화학 사고를 예방하는 데 있어서 화학물질의 공정, 저장, 운송 등의 위험에 대한 정량적 결과를 도출하는 정량적 위험성 평가(Quantitative Risk Assessment, QRA)가 활용되고 있다. QRA에서 사고 빈도 분석은 기반이 되는 데이터에 따라 정확성이 달라지므로 양질의 데이터를 충분히 확보하는 것이 중요하다. 국내에는 화학물질안전원에서 수집하고 있는 통계 자료가 있으나, 직접적인 빈도 데이터가 아니기 때문에 사고 빈도 분석에 바로 활용하기에는 어려움이 있다. 따라서 본 연구에서는 IOGP(The International Association of Oil & Gas Producers)의 빈도 데이터를 변환해 국내의 빈도 데이터를 도출할 수 있는 modification factor를 연구했다. 먼저 modification factor 도출에 활용할 데이터를 선정하였으며, 국내 modification factor를 제시하고 그 활용 가능성을 검토하였다. 또한 우리나라의 경우 전기 · 전자 산업의 비중이 크고, 사업장이 밀폐되어 있어 다른 산업과는 빈도 값에 차이가 있을 것으로 예상되어 해당 factor와 실내에 대한 factor도 함께 산정하였다. 또한 IOGP 빈도 데이터의 특성상, 특정 설비 크기와 누출공 크기에서의 빈도 값만 알 수 있는데, 실제 설비의 크기 및 누출공의 크기는 그보다 훨씬 다양하므로, IOGP의 데이터 형태는 빈도 분석에 적합하지 않다. 따라서, 다양한 설비에 대한 빈도 분석을 위해 비선형 회귀분석을 통해 누출 빈도 함수를 도출하였다. 도출 결과 설비별 데이터에 적합한 함수와 파라미터를 얻었으며, 앞서 구한 modification factor와 누출 빈도 함수식으로 얻은 빈도 값을 통해 공정 상황에 맞는 사고 빈도 도출이 가능할 것이다.
Alternative Abstract
Following the hydrofluoric acid leak accident in Gumi, domestic interest in chemical accidents has surged. In preventing chemical accidents, Quantitative Risk Assessment (QRA) is utilized to derive quantitative results regarding the risks of chemical processes, storage, transportation, and other activities. The accuracy of frequency analysis in QRA depends on the underlying data, so it is crucial to secure high-quality data. Although there are statistical data collected by the National Institute of Chemical Safety in Korea, these are not direct frequency data, making it challenging to use them directly for frequency analysis. Therefore, this study researched a modification factor to convert IOGP (The Internaltional Association of Oil &Gas Producers) frequency data into domestic frequency data. First, the data to be used for deriving the modification factor were selected, and the domestic modification factor was presented and its applicability reviewed. Given that South Korea has a significant proportion of the electrical and electronics industries, and workplaces are enclosed, it is expected that frequency values will differ from other industries, so factors specific to these conditions were also estimated. Moreover, due to the characteristics of IOGP frequency data, only frequency values for specific equipment sizes and leak hole sizes are known, but the actual sizes of equipment and leak holes vary significantly. Thus, the form of IOGP data is not suitable for frequency analysis. Consequently, to analyze frequencies for various equipment, a leakage frequency function was derived through nonlinear regression analysis. The results provided functions and parameters suitable for equipment-specific data, and using the modification factor and frequency values obtained from the leakage frequency function, it will be possible to derive accident frequencies appropriate to the process situation.
