Simulation-based software tools that can infer system behaviors from a deep model of the system have the potential for automatically building the semantic representations required to support intelligent tutoring in fault diagnosis. The Intelligent Maintenance Training System (IMTS) is such a resource, designed for use in training troubleshooting skills and in conducting research into intelligent instruction. The IMTS incorporates a generalized model of an expert diagnostician, termed Profile, to evaluate student performance and to recommend improved fault isolation strategies. The equipment expert uses domain-independent editing tools to construct the simulation using previously defined generic objects. As the diagrams are interactively assembled, an underlying representation of system content and structure is automatically produced, allowing the graphical simulation to change in response to student actions during training sessions. The first application of the IMTS will be as a trainer of fault isolation skills for the Bladefolding system of the SH-3H helicopter. In this application, the IMTS is coupled to the Generalized Maintenance Training System (GMTS), a videodisc-based simulator that displays high-resolution color views of the controls and indicators as they are manipulated by the student. Directions for future research are indicated and 21 references are provided. (Author/MES)