Diseño de un algoritmo de compresión y restauración de datos basado en el análisis frecuencial para mejorar los índices NMSE, RTE y COR de las señales

Abstract

The growth of smart electrical grids and the exponential increase in their data, along with the implementation of large-scale sensors to gather data from electrical signals, necessitates the optimization of storage and processing of these signals through a compression and restoration algorithm using different mother wavelets. The research aims to develop an algorithm that evaluates various mother wavelets and their different levels to determine the balance between compression, efficiency, and accuracy. Effectiveness will be assessed using metrics such as NMSE, RTE, and COR. The methodology encompasses data processing, selection and testing of at least five mother wavelets and ten different compression levels, and the implementation of an iterative algorithm which compresses and restores transient signals such as voltage drops (SAG), voltage increases (SWELL), voltage spikes, and steady signals. Ultimately, the research seeks to identify the most efficient mother wavelet and thus optimize the Daubechies 4 (db4) wavelets with RTE indices of 0.499, NMSE of 0.1716, and a correlation of 0.7, as well as Symlet 3 (sym3) with RTE of 0.5, NMSE of 0.1715, and a correlation of 0.707, which are deemed appropriate for offering better performance. These simulations were performed using tools available in MATLAB.