1 Aircraft AC Generators: Hybrid System Modeling and Simulation Ashraf Tantawy, Student Member, IEEE, Xenofon Koutsoukos, Senior Member, IEEE, and Gautam Biswas, Senior Member, IEEE Abstract—Integrated Drive Generators (IDGs) are the main source of electrical power for a number of critical systems in aircraft. Fast and accurate fault detection and isolation is a necessary component for safe and reliable operation of the IDG and the aircraft. IDGs are complex systems, and a majority of the existing fault detection and isolation techniques are based on signal analysis and heuristic methods derived from experience. Model-based fault diagnosis techniques are hypothesized to be more general and powerful in designing detection and isolation schemes, but building sufficiently accurate models of complex IDGs is a difficult task. dq0 models have been developed for design and control of generators, but these models are not suitable for fault situations, where the generator may become unbalanced. In this paper, we present a hybrid phase-domain model for the aircraft generator that accurately represents both nominal and parametric faulty behaviors. We present the details of the hybrid modeling approach and simulation results of nominal operation and fault behaviors associated with parametric faults in the aircraft generator. The simulation results show that the developed model is capable of accurately capturing the generator dynamics under a variety of normal and faulty configurations. Index Terms—Hybrid modeling, integrated drive generator, phase-domain model, synchronous AC generator. I. I NTRODUCTION The Integrated Drive Generator (IDG) is the primary source for electrical power in aircraft. The system draws its power from the main engines of the aircraft and comprises a number of subsystems that convert the mechanical energy into AC voltage. Fast and accurate fault detection and isolation is a necessary component for safe and reliable operation of the IDG and the aircraft. The majority of the existing techniques for fault detection and isolation of synchronous generators are model-free [3], [4], [11], [14]. On the other hand, model-based fault diagnosis techniques that utilize structural and analytic information contained in the system model are hypothesized to be more general and powerful in designing detection and isolation schemes [5], [6], [8]. However, building sufficiently accurate models of the IDG electrical subsystem is a difficult task because of the complex nonlinearities and the time- varying spatial relations involved in defining the dynamics of the electromagnetic behavior. In addition, the rectifier subsystem that converts the exciter AC voltage output into DC voltage that excites the main generator field includes switching components that introduce discrete dynamics into the overall system behavior. The authors are with the Institute for Software Integrated Systems and the Department of Electrical Engineering and Computer Science, Vanderbilt Uni- versity, Nashville, TN, 37235, USA (email: ashraf.tantawy@vanderbilt.edu; xenofon.koutsoukos@vanderbilt.edu; gautam.biswas@vanderbilt.edu). The traditional way of analyzing switching circuits relies on averaging or discretization techniques to make analysis of the circuit more tractable [15]. In this paper, we employ hybrid modeling [2], which combines the use of discrete and continu- ous behaviors to develop a hybrid model for the brushless AC generator that accurately captures the system dynamics. The model is used to simulate the machine behavior under both normal and faulty conditions. We explicitly model the exciter and main generators using the phase-domain representation, rather than the dq0 domain representation. dq0 modeling is useful for designing and developing controllers for nominal generator behavior [1], [9], while phase domain representa- tions facilitate the simulation of a wide variety of transient phenomena in the machine, and this permits the generator to be employed in a variety of different load configurations. Unlike other simulation work where the generator is part of a power distribution grid [1], [9], [13], we do not assume that the machine is connected to an infinite bus that is kept at constant balanced 3-phase voltage by other machines. The infinite bus assumption does not apply to the aircraft power distribution system, where the main generator output voltage can vary with input and load configuration changes. We do not assume that the output from the rectifier circuit is a constant DC voltage, but that it is determined by the dynamics of the modeled feedback loop. Also, it is well known that even under normal operating conditions, the rectifier output has ripples, and this can affect the transient behavior of the main generator, especially under fault conditions. We believe that the more detailed models will provide a framework for developing robust model-based diagnosis schemes that exploit analytic redundancy, and potentially reduce the number of sensors required for detection and isolation of faults and degradations in the system. The rest of this paper is organized as follows: Section II summarizes existing approaches for fault detection and diagnosis of synchronous generators, emphasizing the need for exploration of model-based techniques. Section III describes the hybrid model for the synchronous generator part of the IDG. Section IV provides a classification of faults typically found in brushless AC generators. A subset of these faults are injected in the developed model to simulate the faulty system behavior. Section V summarizes simulation results for both normal and faulty behaviors. Finally Section VI concludes the work with future directions to extend the developed model.