Search

To ramp-up the utilization of wind energy, the uncertainty of wind power needs to be properly addressed in its integration into the power grid. The main challenge is to well tackle the risk induced by the uncertain wind power generation in different operation stages of a wind farm. This paper studies a novel bi-level multi-timescale scheduling approach for accommodating the wind power uncertainty via a robust optimization formulation. Wind farm day-ahead and intra-day operations consider different scheduling timescales. A two-stage robust optimization model is developed to plan the day-ahead power generation commitments to the power grid. A multi-stage robust optimization model is proposed to refine intra-day wind farm operation instructions. The overall decision-making process aims to minimize the wind farm operational cost. The proposed models are solved by using computationally tractable optimization methods, which offer a desired performance guarantee. A comprehensive case study based on real datasets to demonstrate the effectiveness and robustness of the proposed approach is carried out. Average cost reductions of 46.648% and 6.883% are achieved by comparing with the deterministic method and the static robust method, respectively. Results verify that the proposed method can provide optimal and robust wind farm operations schedules with a lower cost, higher flexibility, and less conservativeness. Results also demonstrate the feasibility of applying the proposed approach to the real-world wind farm and providing reliable operation guidelines. [ABSTRACT FROM AUTHOR]

• Simultaneously scheduling power productions and maintenance activities of an offshore wind farm is studied. • A two-stage model is developed to offer different time-scale schedule of power productions and maintenance. • Uncertainty of wind power and tide conditions on scheduling is considered. • Robust schedules of the two-stage model tackle uncertainties while optimize operations and maintenance. This paper studies the model for scheduling both offshore wind farm maintenance tasks and power productions in consideration of their distinct scheduling timescales and the wind power uncertainty. A two-stage adaptive robust optimization model is formulated to derive the optimal schedule, which well handles wind turbine maintenance tasks and maximizes wind farm power productions by anticipating the worst wind power output scenario based on a wind power uncertainty set and a set of metocean conditions. In two decision stages, we schedule maintenance tasks daily and power productions hourly respectively. The proposed model is equivalently transformed to a form which can be efficiently solved by the column-and-constraint generation algorithm. Numerical experiments demonstrate the effectiveness and applicability of the proposed model considering wind farms of different large sizes. Results show that the developed model can maximize the total power production of the wind farm and enhance the robustness of obtained maintenance schedules against the wind power uncertainty. [ABSTRACT FROM AUTHOR]

Data-driven predictions of wind turbine power outputs have received numerous discussions with using turbine data from the supervisory control and data acquisition (SCADA) system itself. It is of a high curiosity on studying the feasibility of extending the prediction capability with additionally considering data of other sources and the data-driven modeling principle of doing so. To respond such problem and introduce more innovative prediction techniques, this paper presents a pioneering attempt of studying a two-channel deep network modeling method for wind power predictions which leverage both the wind farm data and farm geoinformation. Novelty can be profiled as follows: 1) To accommodate multi-source data in the input, a new high-dimensional input form of two components, a tensor and a graph, is developed; 2) To advance predictions with such input, a deep graph attention convolutional recurrent (GACR) method, which develops one novel deep network channel with stacking multiple graph convolution and long short term memory (GCN-LSTM) layers for engineering high-level latent features from high-dimensional inputs and another classical feature selection channel for directly engaging valuable wind turbine attributes, is proposed. Comprehensive computational experiments are conducted to verify the value of such modeling development by comparing it with a set of competitive benchmarking models. An ablation study is also conducted to explain the necessity and value of network modules in the proposed method. A new state-of-the-art prediction performance is achieved. • A deep learning wind power prediction method leveraging multi-source data is developed. • A double channel feature engineering is developed for parallel latent feature learning and raw feature selection. • A graph attention convolutional recurrent network is designed for more accurate wind power predictions. • A new state-of-the-art short-term wind power prediction performance is achieved by our proposed method. [ABSTRACT FROM AUTHOR]