The Hierarchical Task Network (HTN) formalism is used to express a wide variety of planning problems in terms of decompositions of tasks into subtaks. Many techniques have been proposed to solve such hierarchical planning problems. A particular technique is to encode hierarchical planning problems as classical STRIPS planning problems. One advantage of this technique is to benefit directly from the constant improvements made by STRIPS planners. However, there are still few effective and expressive encodings. In this paper, we present a new HTN to STRIPS encoding allowing to generate concurrent plans. We show experimentally that this encoding outperforms previous approaches on hierarchical IPC benchmarks.