재생에너지 확대와 발전용 가스수요: 대체효과와 유연성 수요에 대한 실증분석

초록

This study analyzes daily natural gas demand for power generation from January 2023 to December 2025 by introducing a net load volatility indicator, alongside traditional variables such as electricity demand and renewable energy generation, to capture new gas demand patterns driven by the expansion of renewable energy. Applying both traditional regression models and the XGBoost machine learning technique, the results reveal that gas demand for power generation is more significantly affected by net load volatility, such as maximum upward ramping requirements, rather than the mere increase in renewable energy generation. Notably, a distinct phenomenon where gas demand increases non-linearly to cope with intra-day volatility during high electricity demand periods was observed, empirically confirming the value of gas-fired generation as a flexible resource. These findings demonstrate the need for an appropriate compensation mechanism for gas-fired generation as a flexible resource, despite its expected decline in capacity utilization.

키워드