Search

In previous literature, constant RLC loads were assumed to impose, on the islanding detection method, the hardest detectable case. For this reason, distributed-generation (DG) islanding studies are usually analyzed and performed by using constant RLC loads. In this paper, different types of loads are taken into account by modeling the load's voltage and frequency dependence. The performance of the over/undervoltage and over/underfrequency protection (OVP/UVP and OFP/UFP) method is examined for the different load models and the results are compared with the constant RLC load case. The analysis is conducted on a constant power-controlled DG designed to operate at unity power factor. A generalized formula, for calculating the nondetection zone of the OVP/UVP and OFP/UFP method, is derived in terms of the load's voltage and frequency dependence parameters. The analysis is further extended on a constant current-controlled DG interface. Index Terms--Distributed generation (DG), inverter, islanding, over/underfrequency, over/undervoltage, static load models.

Although active islanding detection techniques have smaller nondetection zones than passive techniques, active methods could degrade the system power quality and are not as simple and easy to implement as passive methods. The islanding detection strategy, proposed in this paper, combines the advantages of both active and passive islanding detection methods. The distributed-generation (DG) interface was designed so that the DG maintains stable operation while being grid connected and loses its stability once islanded. Thus, the over/undervoltage and over/underfrequency protection method would be sufficient to detect islanding. The main advantage of the proposed technique is that it relies on a simple approach for islanding detection and has negligible nondetection zone. The system was simulated on PSCAD/EMTDC and simulation results are presented to highlight the effectiveness of the proposed technique. Index Terms--Distributed generation (DG), inverter, islanding, over/underfrequency, over/under voltage.

This paper describes the theoretical foundation and prototype implementation of a transient event detector for use in a nonintrusive load monitor (NILM). The NILM determines the operating schedule of the major electrical loads in a building from measurements made at the electric utility service entry. The transient event detector extends the applicability of the NILM to challenging commercial and industrial sites. A spectral preprocessor for use in the transient event detector is introduced first. Then, the transient event detection algorithm is developed. The performance of the algorithm is illustrated with results from a prototype event detector.

A 10-MVA, 13.8-kV and 3,600-rpm superconducting generator at the Massachusetts Institute of Technology was operated at an open circuit voltage of up to 12.9 kV. The test, which revealed that the rotor can remain superconducting through a no-load full voltage fault, resulted in a flash over at the armature winding that short-circuited the machine at 12.2 kV. The flash over occurred prior to the generator's first synchronous run on Cambridge Electric Co.'s power grid.

This paper describes the armature winding built as part of MIT's 10 MVA superconducting generator project. The winding has a helical winding form, is an air-gap winding having no iron teeth in the active region, employs a limited voltage gradient insulation scheme and is cooled by an insulating fluid. Structurally, the winding employs bonded joints between the conductors and structural tubes, which are of filament-wound fiber-reinforced plastic. Preliminary tests indicated that thermal performance of this winding was satisfactory. On the other hand, the winding suffered a catastrophic insulation failure after only a short time at high voltage. The resulting fault gave the structural system of the armature winding (and the rest of the machine) a good test. In this paper we review the major aspects of the design of the winding, describe the thermal tests and then tell of the insulation failure. We can, at best, speculate about the cause of that failure, but we can, on the basis of this experience, make recommendations for further development of machines of this type.

This paper presents a simulator of the manufacturing process of induction motors. Its main purpose is to provide feedback in the form of motor cost and production scheduling during the process of motor design. The paper concentrates on modeling and simulating a factory, modeling production costs, and providing examples of how the design of electric motors can be improved by taking manufacturing information into account via software. The simulator uses principles of system dynamics and technical cost modeling to model the flows of materials and the evolution of product through manufacturing stages across the factory floor. The inputs to the simulator are blueprint data for motor design, and the status of the factory. The output from the simulator is cost and schedule information. The simulator also provides a mechanism whereby design variables can be tested for variations that produce cost improvements.

This paper is, in part, an extended discussion of a paper presented at the 1992 IEEE/PES Winter Meeting (1). That paper raised some issues of modeling which require more extensive attention than is feasible in an ordinary discussion. In this paper, we demonstrate that, regardless of the adequacy of second (one rotor damper branch) or third (two rotor damper branches) order models, a simple form for the rotor can be used. Physical reasoning was used by the authors of (1) to postulate the need for a 'differential leakage' inductance element, representing flux linking both the field winding and damper bars but not the armature winding. We will employ circuit theory to demonstrate that this leakage element does not, by itself, contribute anything to the dynamics of rotor response, and that it is not required to adequately model the frequency response of rotor impedance. This means, (alas!), that at least some of the conclusions of (1) are probably due to an artifact of the curve fitting technique used. Another curve fitting technique and its results are described.

We consider a class of robust, rank-based signal processing filters, particularly with regard to its usefulness in the power electronics field. This class of nonlinear filters is characterized by the inclusion of a sorting element in the signal path. The sorting operation allows these filters to suppress impulsive noise while preserving edges and monotonic sections of signals. This introductory paper concentrates primarily on the median filter, it being the most accessible filter of the class. A working knowledge of issues arising in design and implementation is developed.