Publications

Results 22801–22900 of 99,299

Search results

Jump to search filters

Photovoltaic Inverter Momentary Cessation: Recovery Process is Key

Conference Record of the IEEE Photovoltaic Specialists Conference

Pierre, Brian J.; Elkhatib, Mohamed E.; Hoke, Andy

Momentary cessation refers to an inverter control mode. When the inverter terminal voltage falls below (or exceeds) a certain level, the inverter ceases to output any current, but attempts to maintain (or quickly regain) phase-locked loop synchronization to allow for quick reinjection of current when the voltage recovers to a certain point. This paper presents a photovoltaic (PV) momentary cessation model developed in PSS/E. Simulations are presented for a high voltage transmission line fault contingency in the Hawaiian island of Oahu power system on a validated PSS/E model, modified to include a custom distributed PV inverter model, and different near-future distributed PV penetration levels. Simulations for the island power system include different penetration levels of PV, and different recovery times (ramp rates and delays) after momentary cessation. The results indicate that during low voltage events, such as faults, momentary cessation can produce severe under frequency events, causing significant load shed and shortly thereafter, in some cases, over frequency events that cause generation to trip offline. The problem is exacerbated with higher penetration levels of PV. If momentary cessation is used (as is typically the case for distribution-connected resources), the recovery process after momentary cessation should be carefully considered to minimize impacts to bulk power system stability.

More Details
TYPE Conference Poster YEAR 2019
DOIOSTIScopus

Bulk Power System Dynamics with Varying Levels of Synchronous Generators and Grid-Forming Power Inverters

Conference Record of the IEEE Photovoltaic Specialists Conference

Pierre, Brian J.; Villegas Pico, Hugo N.; Elliott, Ryan T.; Flicker, Jack D.; Lin, Yashen; Johnson, Brian B.; Eto, Joseph H.; Lasseter, Robert H.; Ellis, Abraham

Inverters using phase-locked loops for control depend on voltages generated by synchronous machines to operate. This might be problematic if much of the conventional generation fleet is displaced by inverters. To solve this problem, grid-forming control for inverters has been proposed as being capable of autonomously regulating grid voltages and frequency. Presently, the performance of bulk power systems with massive penetration of grid-forming inverters has not been thoroughly studied as to elucidate benefits. Hence, this paper presents inverter models with two grid-forming strategies: virtual oscillator control and droop control. The two models are specifically developed to be used in positive-sequence simulation packages and have been implemented in PSLF. The implementations are used to study the performance of bulk power grids incorporating inverters with gridforming capability. Specifically, simulations are conducted on a modified IEEE 39-bus test system and the microWECC test system with varying levels of synchronous and inverter-based generation. The dynamic performance of the tested systems with gridforming inverters during contingency events is better than cases with only synchronous generation.

More Details
TYPE Conference Poster YEAR 2019
DOIOSTIScopus

Segmentation certainty through uncertainty: Uncertainty-refined binary volumetric segmentation under multifactor domain shift

IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops

Martinez, Carianne; Potter, Kevin M.; Smith, Matthew D.; Donahue, Emily; Collins, Lincoln N.; Korbin, John P.; Roberts, Scott A.

Deep learning segmentation models are known to be sensitive to the scale, contrast, and distribution of pixel values when applied to Computed Tomography (CT) images. For material samples, scans are often obtained from a variety of scanning equipment and resolutions resulting in domain shift. The ability of segmentation models to generalize to examples from these shifted domains relies on how well the distribution of the training data represents the overall distribution of the target data. We present a method to overcome the challenges presented by domain shifts. Our results indicate that we can leverage a deep learning model trained on one domain to accurately segment similar materials at different resolutions by refining binary predictions using uncertainty quantification (UQ). We apply this technique to a set of unlabeled CT scans of woven composite materials with clear qualitative improvement of binary segmentations over the original deep learning predictions. In contrast to prior work, our technique enables refined segmentations without the expense of the additional training time and parameters associated with deep learning models used to address domain shift.

More Details
TYPE Conference Poster YEAR 2019
OSTIScopus
Results 22801–22900 of 99,299
Results 22801–22900 of 99,299