Hi, magneto_cool:
Your suspect is right as the reported Q factor of spiral inductors on Si substrate with inductor value of several nH is absolute less than 20 even with a lot of Q-boost techniques included(such as thicker metal, etched substrate, high-resistivity substrate etc).
Your information is not enough for us to judge what's wrong with the modeling methodology, you may like to upload your projects here for us to take a look or you may like to try again with the following points in mind:
1. Metal trace with reasonable thickness and conductivity property should be set rather than using PEC.
2. Conductivity/Resistivity of SI substrate should be set correctly.
3. Choose all the objects(Metal trace and Substrate, Passivation etc) and make sure "Solve Inside" is checked if HFSS is used.
4. Fine mesh for the Metal trace should be used by using Mesh options.
Hope the above-mentioned "general rules" can help you a bit.
Best regards,