Synchrophasor Data Recovery for Missing PMU Measurements

The patent combines Nuclear Norm Minimization with Total Variation Minimization into a hybrid algorithm. The Nuclear Norm component handles randomized missing data by minimizing the nuclear norm of the data matrix, while the Total Variation component handles temporally missing data by minimizing the temporal variation of the recovered values. This merging allows the system to recover both types of missing data effectively.

The recovery algorithm is constructed as a composite method integrating two different mathematical approaches: Nuclear Norm Minimization (which assumes low-rank structure) and Total Variation Minimization (which assumes piecewise smooth temporal behavior). This composite algorithm leverages the strengths of both methods to handle diverse missing data patterns that neither method could address alone.

The patent employs computationally efficient algorithms (Nuclear Norm Minimization and Total Variation Minimization) that require significantly less computational power and processing time compared to Hankel Norm Minimization. These lighter-weight algorithms achieve acceptable recovery performance without the excessive resource demands of Hankel Norm methods, making them suitable for real-time or near-real-time PMU data recovery applications.

The patent changes the mathematical parameters and objective functions used in the recovery process. Instead of using Hankel Norm (which involves complex matrix operations and high computational complexity), the patent uses Nuclear Norm and Total Variation parameters that lead to more computationally tractable optimization problems with lower resource requirements while maintaining recovery effectiveness.