KNN and Denoising Autoencoder Bad Data Recovery for Power Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for identifying and correcting bad data in electric power systems, particularly in synchrophasor-based state estimation, are inadequate as they fail to detect certain types of bad data, are computationally intensive, and can lead to undesired estimation results due to their sensitivity and reliance on iterative processes.
Innovation Solution
A two-stage machine learning-based filtering approach using a K-Nearest Neighbor (KNN) classifier and denoising autoencoders to identify and correct bad data in real-time, which is data-driven and model-free, allowing for efficient processing and adaptation to various patterns of bad data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If statistical methods are used for bad data identification, then bad data can be detected through residual analysis, but computation time increases due to multiple iterations
Solution Approach 1:
The patent applies preliminary action by performing bad data identification before state estimation through a pre-processing filter that analyzes measurement residuals and identifies suspicious measurements in advance. This preliminary identification allows the main state estimation algorithm to proceed with already-filtered data, avoiding the need for multiple iterative cycles of estimation and residual analysis, thus reducing overall computation time while maintaining detection reliability.
2Productivity
If non-iterative linear state estimation is used, then computation speed is improved, but sensitivity to bad data increases making detection difficult
Solution Approach 1:
The patent segments the state estimation process into two independent parts: a pre-processing filtering stage that identifies bad data using residual analysis, and a main linear state estimation stage that computes system states. This segmentation allows the fast non-iterative LSE to operate on pre-cleaned data, maintaining its computational speed advantage while the separate filtering stage handles bad data detection, thus reducing the overall sensitivity to bad data.
3Measurement precision
If measurement redundancy is used for bad data removal, then bad data identification accuracy is improved, but device complexity increases
Solution Approach 1:
The patent introduces an intermediary bad data filter that acts as a mediator between raw measurements and the state estimation algorithm. This filter uses measurement redundancy information and residual analysis to identify and flag suspicious measurements before they enter the main estimation process. By placing this intermediary filtering layer, the system achieves improved bad data identification accuracy without significantly increasing the complexity of the core state estimation algorithm, as the filter operates independently using simple residual calculations.
Data Source
AI summary
Systems and methods for processing measurement data in an electric power system include acquiring the measurement data by a phasor measurement unit (PMU) coupled to a line of the electric power system, and inputting a plurality of the measurement data within a predetermined time window into a K-nearest neighbor (KNN) for identifying bad data among the plurality of the measurement data, wherein when one of the plurality of measurement data contains a bad datum, the machine learning module sends the bad datum to a denoising autoencoder module for correcting the bad datum, wherein the denoising autoencoder module outputs a corrected part corresponding to the bad datum, and when one of the plurality of measurement data contains no bad datum, the machine learning module bypasses the denoising autoencoder module and outputs the one of the plurality of measurement data as an untouched part.