Search

The wake effect is the major obstacle to reaching the maximum power generation for wind farms, since choosing the suitable wake model that satisfies both computational cost and accuracy is a difficult task. Deep Reinforcement Learning (DRL) is a powerful data-driven method that can learn the optimal control policy without modeling the environment. However, the "trial and error" mechanism of DRL may cause high costs during the learning process. To address this issue, we propose an ensemble-based DRL wind farm control framework. Under this framework, a new algorithm called Actor Bagging Deep Deterministic Policy Gradient (AB-DDPG) is proposed, which combines the actor-network bagging method with the Deep Deterministic Policy Gradient. The gradient of the proposed method is proved to be consistent with the DDPG method. The experiment results in WFSim show that AB-DDPG can learn the optimal control policy with lower learning cost and a more robust learning process. • This paper proposes an ensemble-based DRL framework for wind farm control. • An algorithm is proposed under the proposed framework to reduce the learning cost. • The gradient used in the proposed algorithm is proved to be consistent with the DDPG. [ABSTRACT FROM AUTHOR]

• A framework solution for multi-timescale coordinated operation of CHP and wind power. • Day-ahead and real-time uncertainties are handled by stochastic optimization. • The solution is applied to a real-life case in Denmark. • Comparative analysis between coordinated operation and independent operation. • Comparative analysis between stochastic and deterministic methods. This paper proposes a multi-timescale coordinated operation approach which coordinates the combined heat and power (CHP) plant and wind farm operation in a deregulated day-ahead heat market and a real-time balancing electricity market. This approach aims to optimize profits of the CHP-wind farm portfolio by considering energy sale in the day-ahead heat market and penalty in the real-time two-price balancing market. Multiple uncertainties of heat load, wind power generation, day-ahead electricity price, up- and down-regulation prices are taken into account for the optimal operation. A stochastic optimization method is applied to solve the proposed coordinated operation problem. The proposed multi-timescale coordinated operation approach is simulated on a real-life system in Copenhagen Denmark where an energy company owns both a CHP plant and a wind farm. The simulation results verify that the proposed method can achieve a profitable operation in the markets and a highly robust operation against the multiple uncertainties. [ABSTRACT FROM AUTHOR]