In this paper, we present an algorithm, called CluE (The Cluster Expander of Compound Graphs), that expands cluster nodes in compound graphs. CluE was designed to work with multiple layout algorithms (e.g., orthogonal layout algorithm, Sugyima layout algorithm, etc.). It keeps the unexpanded nodes in their relative layers with slight changes in their original coordination. This helps the users in maintaining their mental map of the underlying graph. We applied the proposed algorithm for navigating through compound graphs representing the failure mechanisms in embedded systems. We also performed a brief user evaluation study in order to measure the users' acceptance level with regard to the application of CluE for compound graphs. Results show a high acceptance ratio from users having different backgrounds, which indicates the intuitiveness and feasibility of our approach.