Wavelet transform of selected simulation signals using details as information source

Diagnostic signal is one of the information source on the technical device condition. It is assumed it is a sum of determined component and random component. Most often the determined part, i.e. approximation is being processed, omitting details. Random component gives information on noise and interferences occurring during device operation, but it also contains valuable diagnostic information. These are the subject of the article, and more precisely an answer to a question: whether determined component, i.e. pure signal from details may be recovered and whether the best results are somehow connected with signal obtained from approximation?