• 대한토목학회논문집
• /
• 제34권6호
• /
• pp.1779-1786
• /
• 2014

물리현상의 난해성으로 인해 수학적인 관계식이 제시되기 어려운 경우 인공지능 기술에 근거한 다양한 기법이 적용되어 왔다. 수리학 분야의 대표적인 예로 교각주위 국부세굴 문제를 들 수 있다. 본 연구에서는 유전자 알고리즘의 진화된 방법인 GEP 기법을 이용하여 교각주위 국부세굴을 예측하는 방법을 제시하였다. 64개의 실험 자료를 이용하여 GEP 모형을 학습시켜 회귀식을 구축하였으며, 33개의 실험 자료를 이용하여 구축된 모형의 검증을 실시하였다. 평형세굴심 예측을 위하여 차원을 갖는 일반 변수와 표준화된 변수로 GEP 모형을 구축하여 예측 결과를 비교하였는데, 차원을 갖는 변수에 의한 GEP 모형이 세굴심을 더 잘 예측하는 것으로 나타났다. 구축된 GEP 모형을 두 가지 현장 실측자료에 적용하였다. 적용 결과, 실험 자료에 적용한 경우에 비해 예측의 정확도가 낮아지는 것을 확인하였다. 또한, 현장 실측자료를 이용하여 학습시킨 경우 실험 자료를 이용하는 경우 보다 예측 능력이 많이 향상되는 것으로 나타났다. GEP 모형의 적용성을 위해 ANN 모형과의 비교를 수행하였으며, 본 연구에서 사용된 GEP 모형이 교각주위 국부세굴 예측에 대하여 실내 및 현장 모두 ANN 모형보다 우수한 것으로 나타났다.

• 경영과학
• /
• 제17권1호
• /
• pp.117-134
• /
• 2000

Pubilc utility industries including the electric utility industry are facing a new stream of privatization com-petition with the private sector and deregulation. The necewssity to solve now and in the future power supply and demand problems has been increasing through the sophisticated generation expansion plan(GEP) approach con-sidering not only KEPCo's supply-side resources but also outside resources such as non-utility generation(NUG) demand-side management (DSM). Under the environmental situation in the current electric utility industry a new approach is needed to acquire multiple resources competitively. This study presents the development of a modified electric generation expansion analysis system(EGEAS) model with avoided cost based on the existing EGEAS model which is a dynamic program to develope an optimal generation expansion plan for the electric utility. We are trying to find optimal GEP in Korea's case using our modified model and observe the difference for the level of reliabilities such as the reserve margin(RM) loss of load probability(LOLP) and expected unserved energy percent(EUEP) between the existing EGEAS model and our model. In addition we are trying to calculate avoided cost for NUG resources which is a criterion to evaluate herem and test possibility of connection calculation of avoided cost with GEP implementation using our modified model. The results of our case study are as follows. First we were able to find that the generation expansion plan and reliability measures were largely influenced by capacity size and loading status of NUG resources, Second we were able to find that avoided cost which are criteria to evaluate NUG resources could be calculated by using our modified EGEAS model with avoided cost. We also note that avoided costs were calculated by our model in connection with generation expansion plans.

• 한국경영과학회지
• /
• 제17권1호
• /
• pp.117-117
• /
• 1992

Pubilc utility industries including the electric utility industry are facing a new stream of privatization com-petition with the private sector and deregulation. The necewssity to solve now and in the future power supply and demand problems has been increasing through the sophisticated generation expansion plan(GEP) approach con-sidering not only KEPCo's supply-side resources but also outside resources such as non-utility generation(NUG) demand-side management (DSM). Under the environmental situation in the current electric utility industry a new approach is needed to acquire multiple resources competitively. This study presents the development of a modified electric generation expansion analysis system(EGEAS) model with avoided cost based on the existing EGEAS model which is a dynamic program to develope an optimal generation expansion plan for the electric utility. We are trying to find optimal GEP in Korea's case using our modified model and observe the difference for the level of reliabilities such as the reserve margin(RM) loss of load probability(LOLP) and expected unserved energy percent(EUEP) between the existing EGEAS model and our model. In addition we are trying to calculate avoided cost for NUG resources which is a criterion to evaluate herem and test possibility of connection calculation of avoided cost with GEP implementation using our modified model. The results of our case study are as follows. First we were able to find that the generation expansion plan and reliability measures were largely influenced by capacity size and loading status of NUG resources, Second we were able to find that avoided cost which are criteria to evaluate NUG resources could be calculated by using our modified EGEAS model with avoided cost. We also note that avoided costs were calculated by our model in connection with generation expansion plans.