Improved Sparse Component Analysis for Multi-Point Harmonic Contribution Evaluation Under Incomplete Measurements

The existing methods for multi-point harmonic contribution evaluation usually require some preconditions, such as the node harmonic current is equivalent to the harmonic emission current and the number of measurements is not less than the number of sources. However, these requirements may not be satisfied in practice. In this context, this paper proposes an improved sparse component analysis (SCA) method to solve the underdetermined blind source separation model, which is used to describe the harmonic state under incomplete measurements. The improved SCA adopts a new two-stage technique. First, the mixing matrix containing harmonic transfer impedances is estimated through the synchrosqueezing S-transform. Then, the source signal vector containing harmonic emission currents is reconstructed through the modified subspace projection. By matching the harmonic impedances and currents, the multi-point harmonic contribution evaluation is further achieved. The efficiency of the proposed method is verified by simulations and field case studies. The results demonstrate that the proposed method outperforms the existing methods in terms of accurately estimating the harmonic emission currents and quantifying the harmonic contributions under incomplete measurements.