Internal symmetry in this case refers to cases where a part of the protein structure can be superimposed (approximately) on another. It's not a strict mathematical symmetry, more a 'resemblance'
For instance in the 12 transmembrane helix transporters, the first six helices are arranged similarly to the second six, such that if one were to cut the protein in half one would see that the two halves (approximately) superimposed.
Fig 3 of this Nature Micro shows it clearly. http://www.nature.com/articles/nmicrobiol20159 Where the blue helices in 3a could be rotated 180 degrees and then look very similar in arrangment to the yellow helices.