An Efficient MLKD-ACA Solver for Eddy Current Analysis

In this paper, an efficient multilevel kernel degeneration and adaptive cross approximation (MLKD-ACA) algorithm solver is proposed. The MLKD-ACA algorithm compresses the low-rank off-diagonal matrix blocks in far block interactions which are achieved by subdividing the target object into blocks hierarchically. With the butterfly algorithm, the hybrid method of integral kernel degeneration and ACA algorithms improves the performance of the single level hybrid algorithm. The numerical experiment shows that the MLKD-ACA algorithm solver works well for analyzing the eddy current problems.