Wavelet transforms, such as Continuous Wavelet Transform (CWT) create numerous signals based on scale coefficients after decomposition. This entire signal set and their subsets with regard to a chosen function can actually be used to reconstruct redundant representations of the real signal in such a manner that different signal characteristics like anomalies can be detected. This study aims to provide another perspective to wavelet transform studies in terms of the redundancy feature which comes from wavelet type and scale number as well as transform type itself. Consequently, a partial feature construction (PFC) method has been introduced to select the appropriate signal subset which will reconstruct redundant signals for anomaly detection. As a real-world application, ferroresonance data generated by a MATLAB-Simulink model have been used. PFC method has been applied to a subset of decomposed ferroresonance data from CWT, and it has been investigated with different function scenarios. As a result, redundant signals reconstructed by specific functions in PFC have not only allowed detecting ferroresonance phenomenon, but they also made it clearer to observe physical ferroresonance behavior.