Session: 08-10. Gas Turbine Outage Optimization

Paper Number: 125809

125809 - A Surrogate Model Approach to Predict O&M Cost for Combined Cycle Power Plant 

For carbon-neutral 2050, the proportion of renewable energy is increasing, and the problem of power grid instability is emerging due to the load variability of renewable energy. For grid stabilization, the need for flexible operation of gas turbine combined cycle (GTCC) power plants that can handle peak loads are increasing. In GTCC power generation, O&M cost is the next most important after fuel and initial investment cost, and it is necessary to predict O&M cost according to changes in the operating environment for efficient and stable operation of GTCC power plant. However, various uncertainties exist when estimating O&M costs because actual design information is limitedly disclosed due to maintenance contracts between operators and manufacturers, and it is difficult to obtain accurate variables according to the operating condition.

In this paper, the correlation was analyzed considering the uncertainty of the factors affecting the O&M cost of the GTCC power plant through sensitivity analysis, and the main factors were defined. In addition, a surrogate model was proposed that can quickly calculate fixed and variable costs according to changing operating scenarios such as shortening start-up time, enhancing ramp rate, or lower minimum load. Through the developed model, short-term and long-term O&M costs can be estimated due to changes in the operating environment of the GTCC power plant currently being operated. In addition, long-term O&M costs for new power generation facilities such as hybrid GTCC(GT+ESS) can be estimated relatively easily and quickly, which can help establish long-term investment plans.

Presenting Author: WOOSUNG CHOI KEPRI, KEPCO

Presenting Author Biography: Woosung Choi received a B.S. degree from Inha University(Summa Cum Laude), Incheon, South Korea, in 2003, an M.S. degree from KAIST, Daejeon, Republic of Korea, in 2005, a Ph.D. degree from Seoul National University, Seoul, South Korea, in 2018. Currently, he is a principal researcher of KEPCO Research Institute, Daejeon, South Korea. He is in charge of life and risk assessment for power plants and the development of a national standard for the power generation industry (Korea Electric Power Industry Code). His research interests include prognostics and health management, multi-fidelity analysis with FEA, and deep learning analysis.

In particular, he has developed various types of condition diagnosis and prediction solutions using machine learning and deep learning in the energy field. Some of these solutions are being applied to power plant sites, and some of these technologies are being sold and utilized by small and medium-sized businesses.

In the energy field, he has practical experience in international cooperation, such as planning and conducting much international joint research. He directly planned and conducted international joint research in the energy field with Fraunhofer ('19 ~) and Siemens ('20) in Germany and EPRI(`17~`18), SwRI(~'12) in the United States. He also received an assignment from Malaysia TNBR (`16) and successfully applied the research results to the power plant site.

Though a researcher cherishes the value of engineering that contributes to the world, he thinks of people rather than technology. While studying system integrity evaluation, he is working as a board member of AI Friends (AI Research Society) at Daedeok Science Complex.

Authors:

Woosung Choi KEPRI, KEPCO
Bobby M. Webb EPRI
David R. Noble EPRI