Decentralized dynamic system for optimal power dispatch in wind farms based on node-dependence nature.

Meeting the power demand from the transmission system operator is an important objective for power dispatch, which introduces a power supply-demand equality constraint coupling all the wind turbines among the wind farm into the optimization problem. For a large-scale wind farm, processing the global equality constraint in a centralized or distributed framework is time-consuming and computationally complex. Here we considered the fast and localized execution issue of the power optimal dispatch problems. A completely decentralized dynamic system was designed to optimize power flow while satisfying the electricity supply constraints. A voltage optimization problem with the global power constraints was decoupled into local wind turbine controllers based on the node-dependence nature, which is an inherent characteristic of wind farms and was fitted to the power sensitivity matrix in this paper. The local optimization problem was solved iteratively using the gradient projection method, and the system converged linearly to the equilibrium point. The simulations for the case studies performed in Simulink demonstrate that the proposed method achieves a near-global optimal performance using only local measurements.

Robust Representation Learning for Power System Short-Term Voltage Stability Assessment Under Diverse Data Loss Conditions.

With the help of neural network-based representation learning, significant progress has been recently made in data-driven online dynamic stability assessment (DSA) of complex electric power systems. However, without sufficient attention to diverse data loss conditions in practice, the existing data-driven DSA solutions' performance could be largely degraded due to practical defective input data. To address this problem, this work develops a robust representation learning approach to enhance DSA performance against multiple input data loss conditions in practice. Specifically, focusing on the short-term voltage stability (SVS) issue, an ensemble representation learning scheme (ERLS) is carefully designed to achieve data loss-tolerant online SVS assessment: 1) based on an efficient data masking technique, various missing data conditions are handled and augmented in a unified manner for lossy learning dataset preparation; 2) the emerging spatial-temporal graph convolutional network (STGCN) is leveraged to derive multiple diversified base learners with strong capability in SVS feature learning and representation; and 3) with massive SVS scenarios deeply grouped into a number of clusters, these STGCN-enabled base learners are distinctly assembled for each cluster via multilinear regression (MLR) to realize ensemble SVS assessment. Such a divide-and-conquer ensemble strategy results in highly robust SVS assessment performance when faced with various severe data loss conditions. Numerical tests on the benchmark Nordic test system illustrate the efficacy of the proposed approach.