Hybrid Bond Graph (HBG) is a bond graph-based modeling approach which provides an effective tool not only for dynamic modeling but also for fault detection and isolation (FDI) of switching systems. Efficient algorithms exist for fault detection and isolation based on functional redundancy. In a qualitative approach, this redundancy can be captured by a temporal causal graph (TCG), a directed graph that may include temporal information. To avoid the combinatorial explosion when using a bank of TCGs in parallel, causal paths are parameterized by the state of local switches. This paper presents a qualitative approach to diagnose a hybrid system. As an example we use a coupled three tank system with autonomous hybrid behaviors, and illustrate this through a simulation.