| 학회 |
한국공업화학회 |
| 학술대회 |
2021년 가을 (11/03 ~ 11/05, 대구 엑스코(EXCO)) |
| 권호 |
25권 2호 |
| 발표분야 |
학생우수논문발표(석사과정) |
| 제목 |
Optimal operation of sootblowing system using a modified Q-leaning algorithm to maximize recovery boiler efficiency |
| 초록 |
Here, this work proposed optimal operating sequence of sootblowing system to maximize recovery boiler efficiency using a modified Q-learning algorithm. The suggested modified Q-learning algorithm derived the Q-matrix which is a function that predicts the expected dynamic reward (priority for ash deposits removal) of performing a given action (sootblowing) and a given state (each sootblowing location) based on Markov decision process. The reward matrix is continually updated through reward update matrix considering the decrease heat transfer rate according to ash deposits. To develop reward update matrix, we proposed the mathematical equation of heat transfer rate according to temperature difference, deposit thickness growth rate and thermal conductivity of ash deposits. As a result, the power generation increase by 214 kW through optimal sequence without any retrofitting of boiler. |
| 저자 |
임종훈1, 정수환1, 조형태1, 김태복2, 박한신2, 김정환1
|
| 소속 |
1한국생산기술(연), 2무림피앤피(주) |
| 키워드 |
ash deposits; sootblower; modified Q-learning algorithm
|
| E-Mail |
|
