The key to dealing with within part variation for a gage R&R is to understand what question you are trying to answer.
If you are trying to answer "Is this gage capable of measuring this characteristic with sufficient statistical resolution?", then you omit the within part variation by measuring in one location as BevD indicated. It is not the gage's fault you have within part variation, and if the gage passes the gage R&R, it will be accurate enough to detect the variation when used correctly.
If you are trying to answer the question "If I ignore the fact that I have within part variation and allow the operator to measure the part any way they wish, is the measurement technique adequate?" then you will indeed will include the within part variation. It will be a measure of both gage variation and measurement error variation combined - because measuring a characteristic using a technique that ignores within part variation is measurement error. It will be adequate, however, if the within part variation is statistically insignificant. I can think of better ways to determine that rather than burying it in a gage R&R study, however.