When regenerating FRFs with modal data estimated from real measurements, it is common to add upper and lower residuals to the synthesized FRFs, which are related to non-considered modes located below and above the considered frequency range under analysis.
What are the most efficient ways to compute this residuals? Furthermore, considering I have less measurement than excitation nodes (impact testing), is it possible to determine residuals for the unmeasured positions when synthesizing the corresponding FRFs?