Saurabh Kulkarni

Impact of pulse loads on electric ship power system: With and without flywheel energy storage systems

This paper presents the analysis of pulse load operation on the health of a simplified electric ship power system. Two scenarios of the pulse load operation, with and without an energy storage system have been addressed. The energy storage used is a flywheel as it has a very fast time response in supplying high power demands. The health of the electric ship power system is monitored by observing key indicators in the components critical to the working such as the generator and propulsion motor. The time-domain simulation of two test cases is carried out in the PSCAD/EMTDC software platform. The results underscore the vital importance of the flywheel energy storage system in maintaining the stability of the ship power system in the event of pulse load operation.

Incipient Fault Location Algorithm for Underground Cables

Cable failure process is gradual and is characterized by a series of single-phase sub-cycle incipient faults with high arc voltage. They often go undetected and eventually result in a permanent fault. The objective of this paper is to develop a robust yet practical incipient fault location algorithm taking into account the fault arc voltage. The algorithm is implemented in the time-domain and utilizes power quality monitor data to estimate the distance to the fault in terms of the line impedance. It can be applied to locate both sub-cycle as well as permanent faults. The proposed algorithm is evaluated and proved out using field data collected from utility distribution circuits. The average absolute error in locating incipient faults for three underground cable failures analyzed in this paper was found to be 7.37%, 4.69% and 3.58%, respectively.

Waveform characterization of animal contact, tree contact, and lightning induced faults

In this paper signal processing tools are used to uncover common and unique characteristics of faults resulting from animal contacts, tree contacts and lightning. For each fault type a large number of voltage and current waveform data sets measured at monitoring stations on distribution systems are analyzed. The characteristics include but are not limited to the presence of impulse-like oscillations, the number of phases involved, the duration of fault event, the phase angle, the time of day, the spectral content in the time-frequency and time-scale domains, the rate of rise of voltage or current, and the arc voltage. An individual characteristic alone is insufficient to provide an estimate of the fault type. However, by combining common and unique characteristics extracted from a fault event, it may be possible to estimate the fault type accurately.

Feature analysis and classification methodology for overhead distribution fault events

This paper presents the analysis of unique features and classification method for identifying the root cause of overhead distribution fault events. In particular, the paper focuses on faults caused by animal contacts, tree contacts and lightning-induced disturbances. The proposed methodology is implemented and tested using 148 real-world fault events. The methodology has a classification rate of 90%, which demonstrates a good performance in identifying the cause of the events.

Time-Domain Algorithm for Locating Evolving Faults

Evolving faults are faults beginning in one phase of a distribution circuit and spreading to another phase after a few cycles. This paper develops a time-domain algorithm for estimating the location of such faults. The algorithm is divided into two parts, namely, the single line-to-ground portion of the fault and the line-to-line-to-ground portion of the fault. The arc voltage that exists during faults is taken into consideration while deriving this methodology. For the single line-to-ground portion of the fault, the distance to the fault is estimated in terms of the loop or self-reactance between the monitor and the fault. On the other hand, for the line-to-line-to-ground and line-to-line portion of the fault the distance is estimated in terms of the positive-sequence reactance. The reactance-to-fault estimate is more robust than that of the resistance-to-fault, because it is unaffected by fault resistance. Two evolving fault cases and two line-to-line fault cases are analyzed in detail and the error in the location estimates is found to be below 10% in each case. Ten additional cases are analyzed and linear regression analysis is conducted to demonstrate the accuracy of the fault location estimates.

Waveform characteristics of underground cable failures

In this paper unique characteristics of underground cable faults are extracted from the voltage and current waveforms recorded by power quality monitors. These characteristics are used to classify cable faults on the basis of fault duration, specific cable equipment failure and root-cause behind the fault. They are also used to distinguish underground cable faults from other overhead distribution line faults. Waveform signature analysis, wavelet transforms and arc voltage during the fault event is used for cable fault identification and classification.

Effect of load current on fault location estimates of impedance-based methods

Impedance-based methods use simple load models to estimate fault location. However loads in a practical system do not conform to the simplified load models leading to an adverse impact on accuracy of estimation. The objective of this paper is to analyze the effect of load current on fault location estimates of the Takagi, positive-sequence reactance and loop reactance methods. The derivations of the three methods are presented paying special attention to load modeling. The methods are then used to conduct fault location analysis on a modified version of the IEEE 34-Node Test Feeder. The analysis is repeated using the long and lightly loaded original test feeder. When these feeders operate under no-load conditions, fault location estimates of all three methods are highly accurate. Increase in level of load current on both feeder conditions does not affect the accuracy of the Takagi and positive-sequencereactance methods severely. They can be used to locate faults under light-load and heavy-load conditions on short as well as long feeders. The loop reactance method gives highly erroneous estimates when load current magnitude is increased.

Estimating transient response of simple AC and DC shipboard power systems to pulse load operations

The sources of transients in an electric ship power system include those encountered in conventional terrestrial power systems as well as certain sources unique to the shipboard environment such as the pulse load. This paper explains the modeling of the electric ship power system components including the pulse load for studying the effects of the transients in different frequency ranges. The models of AC and DC power systems are compared for these frequency ranges.

Fault location using impedance-based algorithms on non-homogeneous feeders

Impedance-based algorithms like the positive-sequence reactance and Takagi methods are derived assuming a homogenous line conductor. This assumption is violated in practical distribution feeder circuits. The objective of this paper is to demonstrate that these methods can still be effectively applied to non-homogenous feeders by using the line parameters of the most commonly occurring conductor type in the circuit. This approach is first tested on a modified IEEE Test Feeder by simulating two cases, namely, the reference case and the case with a homogenous feeder. For the positive-sequence reactance method the maximum absolute error in both cases is about 8%, while for the Takagi method it is 6.92% and 9.53% in the homogenous case and reference case, respectively. The proposed approach is then demonstrated using ten fault cases on utility distribution feeders. The average absolute error obtained for the positive-sequence reactance and Takagi estimates is 10.23% and 9.34%, respectively, which corresponds to a median error value of 0.16 miles and 0.15 miles in the location estimates.

Distribution fault location using short-circuit fault current profile approach

Impedance-based algorithms do not consider load current and non-uniform line impedance per unit, thus introducing errors in fault location estimates. To minimize these errors, this paper proposes a short-circuit fault current profile approach to complement impedance-based algorithms. In this approach, circuit model of the distribution feeder is used to place faults at every bus and the corresponding short-circuit fault current is plotted against reactance or distance to fault. When a fault occurs in the distribution feeder, fault current recorded by the relay is extrapolated on the current profile to get location estimates. Since the circuit model is directly used in building the current profile, this approach takes into account load and non-uniform line impedance. The approach is tested using modified IEEE 34 Node Test Feeder and validated against data provided by utilities. Location estimates are within 0.8 miles of the actual fault location when the circuit model closely represents the distribution feeder.