TY - GEN
T1 - The Prediction and Operational Control System of the Cofiring Combined Cycle Power Plant Using Deep Learning Methods to Improve Power Generation Performance
AU - Wiranata, Addien Wahyu
AU - Asfani, Dimas Anton
N1 - Publisher Copyright:
© 2024 IEEE.
PY - 2024
Y1 - 2024
N2 - Cofiring is one step in accelerating the mix of new renewable energy. The utilization of biomass is substituted in Coal is used as fuel in boilers in the process of transitioning thermal plants to biomass. However, the use of substantial amounts of biomass will have an impact on thermal and kinetic changes in turbine equipment so that it affects the output of the generator and speed drop governor. Using cofiring test performance data parameters of 3%, 5%, 100/0, 250/0, 50%, and 100% will be processed using machine learning, namely with the deep learning for time series forecasting (DL- TSF) method to obtain generator output optimization and speed drop governor. The use of deep learning will combine multilayer perceptron, convolutional neural network, and long short-term memory algorithms with adaptive moment estimation (ADAM) optimization and will be validated with mean square error (MSE) for each process. The output of this machine learning will later provide output optimization of generator loading and governor openings above 70% of the capable power capacity. The use of the long short-term memory algorithm will also be able to function as an estimator in the control speed droop governor function to maintain equipment work efficiency. In the future, the use of cofiring through deep learning process control for time series forecasting will be able to increase the efficiency of generator output with various uses of biomass types and increasing percentages so that the distribution of new renewable energy mix can continue to increase.
AB - Cofiring is one step in accelerating the mix of new renewable energy. The utilization of biomass is substituted in Coal is used as fuel in boilers in the process of transitioning thermal plants to biomass. However, the use of substantial amounts of biomass will have an impact on thermal and kinetic changes in turbine equipment so that it affects the output of the generator and speed drop governor. Using cofiring test performance data parameters of 3%, 5%, 100/0, 250/0, 50%, and 100% will be processed using machine learning, namely with the deep learning for time series forecasting (DL- TSF) method to obtain generator output optimization and speed drop governor. The use of deep learning will combine multilayer perceptron, convolutional neural network, and long short-term memory algorithms with adaptive moment estimation (ADAM) optimization and will be validated with mean square error (MSE) for each process. The output of this machine learning will later provide output optimization of generator loading and governor openings above 70% of the capable power capacity. The use of the long short-term memory algorithm will also be able to function as an estimator in the control speed droop governor function to maintain equipment work efficiency. In the future, the use of cofiring through deep learning process control for time series forecasting will be able to increase the efficiency of generator output with various uses of biomass types and increasing percentages so that the distribution of new renewable energy mix can continue to increase.
KW - Deep Learning for Time Series Forecasting (DL-TSF)
KW - Speed Drop Governor
KW - Thermal-Biomass Power Plant
UR - http://www.scopus.com/inward/record.url?scp=85210802672&partnerID=8YFLogxK
U2 - 10.1109/ICT-PEP63827.2024.10733418
DO - 10.1109/ICT-PEP63827.2024.10733418
M3 - Conference contribution
AN - SCOPUS:85210802672
T3 - ICT-PEP 2024 - International Conference on Technology and Policy in Energy and Electric Power: Resilient Power Systems: Navigating the Clean Energy Transition, Proceedings
SP - 324
EP - 329
BT - ICT-PEP 2024 - International Conference on Technology and Policy in Energy and Electric Power
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2024 International Conference on Technology and Policy in Energy and Electric Power, ICT-PEP 2024
Y2 - 3 September 2024 through 5 September 2024
ER -