Unfortunately not and I didn’t much research on fractional VARs in the last months, as I concentrated more on univariate fractional and multivariate cofractional modeling. So, if you have any questions about this, you can ask me any time.
In my opinion, the adoption of fractional VARs could also be realized by a two-step procedure, where you first estimate the integration order (particularly via a multivariate approach) and then secondly estimate a classical VAR on the fractional differences…this would be my suggestion.