Overplotting scores for Calibration, Test & Validation sets

 

After calculating the PCA for the Shootout Calibra 1 set, I can see the maps of scores for PC1 vs PC2, PC1 vs PC3, PC2 vs PC3,....

But how can I project other sample sets into this PC space?, and in this case  the Shootout Test 1 and Shootout Valida 1.

The idea is to see if the scores for these Test and Valida Set projected into the PCA space of Calibra 1 are into their cloud space, and all the samples are represented, this way we are not extrapolating.