I did some cursory research on the topic, and basically without information regarding the nature of the test (depth, etc), converting the Richter number to KTs is pretty much guesswork.
The type of rock (granite, sedimentary, etc) makes a big difference in the yield estimate, as does the depth of the explosion, neither of which are known.
The initial measurement from the South Korean seismic station indicated a depth of 0 meters (with unknown error estimate), which means the explosion might have even taken place on the surface.
Satellites would have picked it up though if done on the surface. Also North Koreans are good at digging tunnels. Their engineers were involved in building the Vietnam war ones.