The TU Dresden people did not "verify" that the device works --- instead, they identified many spurious error sources, and recognized that the final results were essentially null on the level of measurement accuracy, and unable to say whether there's thrust or not.

http://arc.aiaa.org/doi/pdf/10.2514/6.2015-4083 https://tu-dresden.de/ing/maschinenwesen/ilr/rfs/ressourcen/...