Sorry for the late response.
What's interesting here is that, for piping, the 5% critical damping for seismic analysis is now the norm (see ASME III Appx. N)while previously 1% and 2% were specified for an operating basis earthquake (OBE) or 2% and 3% were specified for a safe shutdown earthquake (SSE)(see older Appx. N or USNRC Guide 1.61). I was told a few years ago that, even though these lower percentages were accurate, the lower damping associated with the OBE produced higher loads than the higher damping with the SSE. Systems were overdesigned. Acknowledging that fact, the committee upped the percent critical damping.
But that doesn't help you if you still wnat to use 3% now that all data is with 5%. I searched my references and could not locate a conversion for you.
I queried a few associates and have these two suggestions, both of which are very general and unsupported by theory:
1) simply ratio your accelerations by the ratio of the damping. So to move from 5% to 3%, multiply the 5% accelerations by 5/3. This may not be true at the limits (high and low frequencies).
2) multiply the known accelerations by the square root of the ratio. Again, to move from 5% to 3%, multiply the 5% accelerations by SQRT(5/3). In addition, do not allow the high frequency accelerations drop below the zero period acceleration (ZPA) for the site. (That's not an issue here as you are increasing the acceleration rather than decreasing it.)
One way you can evaluate these approaches is to compare the results with those listed in USNRC Reg. Guide 1.60. That document (and CAESAR II) shows accelerations for the "typical" earthquake with a variety of percent critical damping.
_________________________
Dave Diehl