The autonomous operation of the electric power system of a spacecraft is a key enabling functionality for deep-space missions. This functionality is informed by the real-time physical status of its power system, which is determined by processing the inputs obtained from various voltage and current sensors distributed throughout the electric network. In particular, it is important to have an accurate estimate of the current in each branch of the power system to detect overload conditions. However, voltage and current measurements can have an error due to noise and sensor faults. In this work, the focus is on ensuring accurate estimation of the voltages and currents in the presence of such faults. To this end, a method is proposed to optimally design a network of compound current sensors that enhances the ability to reliably identify faulty or biased current sensors. The proposed sensor fault detection method is validated using simulation results.

The capability to detect and identify faults within power electronic converters can be crucial for many applications. This paper presents a supervised machine learning approach to accurately and reliably diagnose faults in a dc-dc converter system. The proposed algorithm employs support vector machine classification, and utilizes features related to the power spectrum of the converter input current for identifying the converter condition. Simulation results are presented that show the statistical performance of the proposed method under random variations.

The electric power system of a deep space vehicle is mission-critical, and needs to respond to faults intelligently and autonomously. Such a system consists of solar arrays, batteries, and loads. Its topology is re-configurable by virtue of a number of controllable switches. In this paper, an algorithm is set forth that decides how to operate the system under both normal and faulted conditions. Operational decisions include shedding loads, switching lines, and controlling battery charging, based on the importance of the loads and the transmission loss. As part of this algorithm, an optimal power flow problem is formulated as a mixed integer linear program. Results of case studies considering different faults in the system are presented.

In this paper, a new method for average-value modeling and simulation of boost converters subject to hysteresis current control is proposed. It incorporates a slew-rate limitation on the inductor current that occurs naturally in the circuit during large system transients. This new method is compared with five existing methods in terms of simulation accuracy and run time. The performance is evaluated based on a variety of scenarios, and the simulation results are compared with the results of a detailed model. The simulation results show that the proposed method represents the detailed model well and is faster and more accurate than existing methods. Hardware validation is also presented. The slew-rate-limitation model of boost converters subject to hysteresis current control accurately captures the salient details of converter performance while retaining the computational efficiency of average-value models. This model can be used for time-domain simulation studies where both numerical efficiency and accuracy are required.

Distributed Heterogeneous Simulation (DHS) is a method of cosimulation that enables fast simulation of large system models on distributed computing resources. Typically, simulation subsystem boundaries are defined at component interfaces. However, increased simulation speeds could be attained by determining subsystem boundaries that allow the largest possible DHS communication interval that maintains the prescribed accuracy. In this paper, a metric is derived to quantify the error associated with applying DHS to the simulation of a linear system. The error metric derivation is based on analysis of the state transition matrix of the linear system. This assumption is not overly restrictive because the approach could be applied to the linearization of a nonlinear system about an equilibrium point. An example system is used to illustrate the impact of subsystem configuration and communication interval on the error metric. In addition, a genetic algorithm-based technique for identifying optimal subsystem boundaries is proposed.

A synchronous machine model with saturation and cross saturation and an arbitrary rotor network representation that uses a voltage-behind-reactance representation for both the stator windings and the field winding of the machine is proposed. This allows the stator windings and the field winding to be represented as branches in a circuit solver, permitting straightforward simulation with connected circuitry. In particular, the model can be simulated with rectifier loads or with rectifier sources applied to the field winding. The model is validated against experimental data, and its utility is demonstrated in an excitation failure case study.

The most critical element of the nation's energy infrastructure is our electricity generation, transmission, and distribution system known as the "power grid." Computer simulation is an effective tool that can be used to identify vulnerabilities and predict the system response for various contingencies. However, because the power grid is a very large-scale nonlinear system, such studies are presently conducted "open loop" using predicted loading conditions months in advance and, due to uncertainties in model parameters, the results do not provide grid operators with accurate "real time" information that can be used to avoid major blackouts such as were experienced on the East Coast in August of 2003. However, the paradigm of Dynamic Data-Driven Applications Systems (DDDAS) provides a fundamentally new framework to rethink the problem of power grid simulation. In DDDAS, simulations and field data become a symbiotic feedback control system and this is refreshingly different from conventional power grid simulation approaches in which data inputs are generally fixed when the simulation is launched. The objective of the research described herein was to utilize the paradigm of DDDAS to develop a marriage between sensing, visualization, and modelling for large-scale simulation with an immediate impact on the power grid. Our research has focused on methodological innovations and advances in sensor systems, mathematical algorithms, and power grid simulation, security, and visualization approaches necessary to achieve a meaningful large-scale real-time simulation that can have a significant impact on reducing the likelihood of major blackouts.