Real-time analysis of power signals typically incorporates non-principal elements like noise, attenuation components, and disturbances, that usually interfere with the accuracy of extraction for principal components that we call signal essence. This paper introduces a Multi-dimensional Adaptive Signal Essence Capturing (MASEC) real-time analysis framework for power signals. MASEC isolates signals into attenuating components, principal components, and noise for enhancing essence extraction. The research uses a large-scale central weighted window function method based on mean high-order polynomial fitting to extract attenuating components and proposes a biphasic joint high-frequency filtering strategy based on the Tent Chaotic Mapping-Enhanced SSA(TC-SSA) algorithm to filter high-frequency components. Comparative experimental results demonstrate that the MASEC achieves an average 76.3% enhancement in optimization accuracy over mainstream optimization algorithms. Notably, the TC-SSA enhanced MASEC variant exhibits a further 8.6% improvement in solution precision, building upon this performance baseline.