Abstract—Features of ungrounded stiffly connected power systems have driven the development of strategies for detailed modeling and simulation for transient stability analysis. In this paper, a new generalized strategy is proposed for modeling and simulating ungrounded stiffly connected power systems. The proposed strategy neglects the minimal effects of line capacitances on system dynamics to achieve significant improvement in simulation speed. Since interconnection incompatibility occurs when neglecting capacitances, one component model is reformulated on each resistor or inductor bus to derive the voltage inputs of the interconnected component models. Systematic procedures are developed to build the model of an ungrounded stiffly connected system with a set of Differential-Algebraic-Equations (DAE) in the environment of Matlab/Simulink. The authors contrast their approach with two conventional power system simulation programs and a method reported in literature. By modeling and simulation of an example ungrounded stiffly connected power system, the feasibility of the proposed method is demonstrated. Index Terms—ungrounded, stiffly connected, simulation, stability, generalized, reformulated, interconnection incompatibility. I. INTRODUCTION Ungrounded power systems distribute power in ungrounded configuration to ensure continuous power supply in the presence of single phase to ground faults. Therefore, no equipment is intentionally grounded in ungrounded power systems. In stiffly connected power systems, various components are connected through relatively short transmission lines. The short lines are primarily inductive since line capacitances are very small compared to line resistances and inductances [1]. Shipboard power systems (SPS) in U.S naval ships supply electric power to weapon, navigation, communication and operation systems on naval ships. AC radial SPS found currently on US surface combatant ships is a typical ungrounded stiffly connected power system with some unique L. Qi is with Department of Electrical Engineering, Texas A&M University, College Station, TX 77843, USA (e-mail: [email protected]). K. L. Butler-Purry is with Department of Electrical Engineering, Texas A&M University, College Station, TX 77843, USA (e-mail: [email protected]). This work was supported by the Office of Naval Research under grant N00014-9901-0704. features [2]. It is an isolated power system without interconnection support from neighboring terrestrial power systems. Capacity and inertia of generators in SPS are relatively small. Load demand, most of which is dynamic load demand, is comparable to generation capacity. Due to the characteristics of SPS, the voltages and system frequency of SPS are sensitive to disturbances and may vary significantly from nominal values. In this paper, SPS is taken as an example ungrounded stiffly connected power system to demonstrate the proposed method. Transient simulations provide valuable information for transient stability studies. In conventional transient stability studies, some transients, such as stator and network transients, are neglected. However, in ungrounded stiffly connected systems, various components are strongly coupled through short transmission lines. Due to the strong interactions, the simulation results neglecting stator and network transients could cause large errors in transient stability analysis. Reduced order models in conventional stability studies only provide an approximation for actual dynamics. Therefore, detailed models should be adopted to simulate the accurate dynamics in ungrounded stiffly connected systems such as SPS. Problems of modeling stiffly connected power systems and existing solutions are discussed in section II. Component models are described in section III. Interconnection procedures are illustrated in section IV. In section V, a reduced SPS is modeled and simulated as an example system, and representative simulation results are presented. In section VI, conclusions are drawn. II. PROBLEM FORMULATION It is obvious that simulations with detailed models are slower than simulations with approximated models. The speed of the simulations of stiffly connected power systems is thus reduced significantly. Due to small shunt and mutual capacitances, the natural frequencies of short electric lines are very large. Therefore, extremely small integration time steps are required to derive stable and accurate simulation results of stiffly connected systems. One way to improve the simulation speed is to neglect the small line capacitances. The capacitances of short electric lines are so small that they have little effects on system dynamics [1]. By neglecting Reformulated Model Based Modeling and Simulation of Ungrounded Stiffly Connected Power Systems L. Qi, Student Member, IEEE, and K. L. Butler-Purry, Senior Member, IEEEthe capacitances in line models, the negligible effects are not observable, while the simulation speed is improved greatly. However, without capacitances, the tie lines are modeled with pure resistances and inductances. Thus, ungrounded stiffly connected systems would be completely isolated from ground, and inductor and resistor busses emerge in the systems. An inductor or resistor bus is a kind of bus where only inductive or resistive components are connected. Incompatible interconnection occurs when only voltage-in current-out inductive or resistive models are interconnected. In a delta-connected system, the input voltages are line-to-line voltages. Unlike naturally interconnected capacitor busses, the input voltages of inductor and resistor busses need to be established by artificial ways. Traditionally, reformulated current-in voltage-out resistor models derive the input voltages of resistor busses. The derivation of the input voltages of inductor busses is more complex. As we know, the simulation software for power systems can be classified into two main categories: nodal admittance matrix based circuit simulation programs, such as EMTP/ATP [3] (The Electromagnetic Transients Program/Alternative Transient Program), and differential equation based equation solver programs, such as SimPower (Simulate Power Systems) [4]. In both EMTP/ATP and SimPower, the detailed simulation speed is sacrificed significantly to make the negligible effects of the line capacitances on dynamics observable. As discussed earlier, those effects could be neglected, and simultaneously the simulation speed