Structural Tensor Learning for Event Identification With Limited Labels

Abstract

The increasing uncertainty of distributed energy resources promotes the risks of transient events for power systems. To capture event dynamics, Phasor Measurement Unit (PMU) data is widely utilized due to its high resolutions. Notably, Machine Learning (ML) methods can process PMU data with feature learning techniques to identify events. However, existing ML-based methods face the following challenges due to salient characteristics from both the measurements and the labels: (1) PMU streams have a large size with redundancy and correlations across temporal, spatial, and measurement type dimensions. Nevertheless, existing work cannot effectively uncover the structural correlations to remove redundancy and learn useful features. (2) The number of event labels is limited, but most models focus on learning with labeled data, suffering risks of non-robustness to different system conditions. To overcome the above issues, we propose an approach called Kernelized Tensor Decomposition and Classification with Semi-supervision (KTDC-Se). Firstly, we show that the key is to tensorize data storage, filter information via decomposition, and learn discriminative features via classification. This leads to an efficient exploration of structural correlations via high-dimensional tensors. Secondly, the proposed KTDC-Se can incorporate rich unlabeled data to seek decomposed tensors, invariant to varying operational conditions. Thirdly, we make KTDC-Se a joint model of decomposition and classification so that there are no biased selections of the two steps. Finally, to boost the model accuracy, we add kernels for non-linear feature learning. We demonstrate the KTDC-Se superiority over the state-of-the-art methods for event identification using PMU data.

Value-based sensor and information fusion

“…From aerial [78] to ground robotics [79], there are many instances of RL supporting sensor collections, platform control, as well as mission-driven policies. An example Figure 3 -Decisions to Data is a complex power grid where DL, RL, and AL can coordinate to support effective and efficient operations [80]. A key goal beyond accuracy and timeliness is to utilize the RL agent to support safety [81] and security policies [82].…”

Digital twin meets information fusion: panel summary

“…The use of DT with DDDAS was formulated for aerospace DTs in [65], and as amplified in the work by for Aerospace Digital Twins testbeds. The DDDAS DT concept extends cognitive awareness [66,67], as well as supporting security communications [68], power grid analysis [69,70], predictive maintenance [71], and structural health monitoring. Smirnov et al [72] highlights the use of DT as a simulation for safety in maritime awareness.…”

Human-machine cooperative AI decision-making with heterogeneous data

“…While many definitions might be ascribed to AI and DL, a general definition is that DL does data fitting and AI conducts cognitive inferencing. While machine-level AI has yet to be realized, examples of DL are in many systems designs such as automatic target recognition (ATR) [1,2], health care [3,4,5], and industry [6,7]. There are many concerns being raised as for the interpretability, explainability and usability of these AI/DL techniques; while at the same time policy makers seek methods for verification and validation of trustworthy, accountable, and certifiability of DL systems.…”