Enhanced denosing autoencoder aided bad data filtering for synchrophasor-based state estimation

Synchrophasor-based linear state estimation (LSE) attracts great interests in the past decade due to its high accuracy and straightforward computation and has been the foundation of many wide area monitor system (WAMS) applications. However, an increasing number of data quality issues has been reported, among which bad data can significantly undermine the performance of LSE and many other WAMS applications it supports. In practice, bad data filtering can be very challenging due to a variety of factors including constrained computation time, nonuniform and changing patterns, etc. This paper presents an enhanced denosing autoencoder (DA) aided bad data filtering scheme to pre-process PMU measurements for LSE. Bad data are firstly identified by a classifier module of the proposed DA and then recovered by autoencoder module of the proposed DA. The proposed methodology has two distinct features: 1) it not only identifies bad data based on adaptive learning but also recovers them, including the ones on critical measurements; 2) the algorithm is lightweight and can be deployed on individual PMU level to achieve maximum parallelism and high efficiency, making it suitable for real-time processing. Performance of the proposed methodology is validated and demonstrated through numerical experiments conducted using both simulated and real-world phasor measurements.