TY - GEN
T1 - Distributed Feedforward Optimization for Control of Multi-Energy Network with Temporal Variations
AU - Xu, Yiqiao
AU - Zhang, Zhengfa
AU - Ding, Zhengtao
AU - Jiang, Shuoying
AU - Parisio, Alessandra
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2024/1/19
Y1 - 2024/1/19
N2 - Multi-Energy Network (MEN) is a promising approach to improve the overall efficiency of energy utilization. Yet, balancing its electrical and thermal power in real-time is challenging due to variable demands. In this paper, we formulate a distributed Time Varying Optimization Problem (TVOP) and solve it in continuous-time to track the unknown time-varying optimal trajectories. First, we apply the principles of output regulation theory to reverse engineer the feedforward laws in the presence of projection. These laws are responsible for proactively canceling the effects of temporal demand variations. Then, a projection-based distributed optimization algorithm, alongside a distributed auxiliary protocol based on weighted-sum consensus, result in a novel scheme we term distributed feedforward optimization. One of the key features of our scheme is its data-driven nature, where temporal variations are captured from Ultra-Short-Term Forecasting (USTF) profiles using an exosystem. Under mild assumptions, the proposed scheme provides a guarantee for asymptotic convergence. Simulation results demonstrate the effectiveness of our scheme under an non-ideal case.
AB - Multi-Energy Network (MEN) is a promising approach to improve the overall efficiency of energy utilization. Yet, balancing its electrical and thermal power in real-time is challenging due to variable demands. In this paper, we formulate a distributed Time Varying Optimization Problem (TVOP) and solve it in continuous-time to track the unknown time-varying optimal trajectories. First, we apply the principles of output regulation theory to reverse engineer the feedforward laws in the presence of projection. These laws are responsible for proactively canceling the effects of temporal demand variations. Then, a projection-based distributed optimization algorithm, alongside a distributed auxiliary protocol based on weighted-sum consensus, result in a novel scheme we term distributed feedforward optimization. One of the key features of our scheme is its data-driven nature, where temporal variations are captured from Ultra-Short-Term Forecasting (USTF) profiles using an exosystem. Under mild assumptions, the proposed scheme provides a guarantee for asymptotic convergence. Simulation results demonstrate the effectiveness of our scheme under an non-ideal case.
UR - http://www.scopus.com/inward/record.url?scp=85184824635&partnerID=8YFLogxK
UR - https://www.mendeley.com/catalogue/8b22bb53-33de-3187-a6e2-d49263eb2be7/
U2 - 10.1109/CDC49753.2023.10383891
DO - 10.1109/CDC49753.2023.10383891
M3 - Conference contribution
AN - SCOPUS:85184824635
SN - 9798350301243
T3 - Proceedings of the IEEE Conference on Decision and Control
SP - 4333
EP - 4338
BT - 2023 62nd IEEE Conference on Decision and Control, CDC 2023
PB - IEEE
T2 - 62nd IEEE Conference on Decision and Control, CDC 2023
Y2 - 13 December 2023 through 15 December 2023
ER -