Publications

Abstract not provided.

In this paper, we present a new, light-weight approach for extracting the five single diode parameters (IL, Io, RS, RSH, and nNsVt) for advanced, in-field monitoring of in situ current and voltage (I-V) tracing devices. The proposed procedure uses individual I-V curves, and does not require the irradiance or module temperature measurement to calculate the parameters. It is suitable for operation on a small, single board computer at the point of I-V curve measurement. This allows for analysis to occur in the field, and eliminates the need to transfer large amounts of data to centralized databases. Observers can receive alerts directly from the in-field devices based on the extraction, and analysis of the commonly used single diode equivalent model parameters. This paper defines the approach and evaluates its accuracy by subjecting it to I-V curves with known parameters. Its performance is defined using actual I-V curves generated from an in situ scanning devices installed within an actual photovoltaic production field. The algorithm is able to operate at a high accuracy for multiple module types and performed well on actual curves extracted in the field.

The IEC 61215 and Qualification Plus indoor aging tests are recognized as valuable assessment procedures for identifying photovoltaic (PV) modules that are prone to early-life failures or excessive degradation. However, it is unclear how well the tests match with reality, and if they can predict in-field performance. Therefore, the present work performed indoor-aging thermal cycling tests on pristine-condition modules and evaluated, using in-field current and voltage (I-V) curve scans, modules of the same make and model exposed to the actual environment within a production field. The experiment included the estimate of the overall exposure to thermal cycling in both indoor and outdoor environments, the extraction of the series resistance from the I-V curves, the development of a model based on the indoor results, and finally the testing of the model on outdoor exposure amounts to predict actual changes in resistance. Index Terms - photovoltaic, accelerated aging, series resistance.

Abstract not provided.

Abstract not provided.

The integration of communication-enabled grid-support functions in distributed energy resources (DER) and other smart grid features will increase the U.S. power grid's exposure to cyber-physical attacks. Unwanted changes in DER system data and control signals can damage electrical infrastructure and lead to outages. To protect against these threats, intrusion detection systems (IDSs) can be deployed, but their implementation presents a unique set of challenges in industrial control systems (ICSs), New approaches need to be developed that not only sense cyber anomalies, but also detect undesired physical system behaviors. For DER systems, a combination of cyber security data and power system and control information should be collected by the IDS to provide insight into the nature of an anomalous event. This allows joint forensic analysis to be conducted to reveal any relationships between the observed cyber and physical events. In this paper, we propose a hybrid IDS approach that monitors and evaluates both physical and cyber network data in DER systems, and present a series of scenarios to demonstrate how our approach enables the cyber-physical IDS to achieve more robust identification and mitigation of malicious events on the DER system.

Abstract not provided.

Adoption of plug-in electric vehicles (PEVs) has expanded over the last few years, yet introduction of PEV smart charging has been stalled due to barriers in communication, controls, and an unclear method for determining the value PEVs will bring to the grid. This project will consider the grid impact of a variety of future scenarios, including adoption of different vehicle types, proliferation of extreme fast charging (xFC), expanded adoption of distributed energy resources (DER), and multiple smart charge management approaches. This project will determine how PEV charging at scale should be managed to avoid negative grid impacts, allow for critical strategies and technologies to be developed, and increase the value for PEV owners, building managers, charge network operators, grid services aggregators, and utilities.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

The U.S. DOE Regional Test Center for Solar Technologies program was established to validate photovoltaic (PV) technologies installed in a range of different climates. The program is funded by the Energy Department's SunShot Initiative. The initiative seeks to make solar energy cost competitive with other forms of electricity by the end of the decade. Sandia National Laboratory currently manages four different sites across the country. The National Renewable Energy Laboratory manages a fifth site in Colorado. The entire PV portfolio currently includes 20 industry partners and almost 500 kW of installed systems. The program follows a defined process that outlines tasks, milestones, agreements, and deliverables. The process is broken out into four main parts: 1) planning and design, 2) installation, 3) operations, and 4) decommissioning. This operations manual defines the various elements of each part.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Monitoring of photovoltaic (PV) systems can maintain efficient operations. However, extensive monitoring of large quantities of data can be a cumbersome process. The present work introduces a simple, inexpensive, yet effective data monitoring strategy for detecting faults and determining lost revenues automatically. This was achieved through the deployment of Raspberry Pi (RPI) device at a PV system's combiner box. The RPI was programmed to collect PV data through Modbus communications, and store the data locally in a MySQL database. Then, using a Gaussian Process Regression algorithm the RPI device was able to accurately estimate string level current, voltage, and power values. The device could also detect system faults using a Support Vector Novelty Detection algorithm. Finally, the RPI was programmed to output the potential lost revenue caused by the abnormal condition. The system analytics information was then displayed on a user interface. The interface could be accessed by operations personal to direct maintenance activity so that critical issues can be solved quickly.

Current-voltage (I-V) curve traces of photovoltaic (PV) systems can provide detailed information for diagnosing fault conditions. The present work implemented an in situ, automatic I-V curve tracer system coupled with Support Vector Machine and a Gaussian Process algorithms to classify and estimate abnormal and normal PV performance. The approach successfully identified normal and fault conditions. In addition, the Gaussian Process regression algorithm was used to estimate ideal I-V curves based on a given irradiance and temperature condition. The estimation results were then used to calculate the lost power due to the fault condition.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Commonly used performance models, such as PVsyst, Sandia Array Performance Model (SAPM), and PV LIB, treat the PV array as being constructed of identical modules. Each of the models attempts to account for mismatch losses by applying a simple percent reduction factor to the overall estimated power. The present work attempted to reduce uncertainty of mismatch losses by determining a representative set of performance coefficients for the SAPM that were developed from a characterization of a sample of modules. This approach was compared with current practice, where only a single module’s thermal and electrical properties are testing. However, the results indicate that minimal to no improvements in model predictions were achieved.

Optimal integration of thermal energy storage within commercial building applications requires accurate load predictions. Several methods exist that provide an estimate of a buildings future needs. Methods include component-based models and data-driven algorithms. This work implemented a previously untested algorithm for this application that is called a Laterally Primed Adaptive Resonance Theory (LAPART) artificial neural network (ANN). The LAPART algorithm provided accurate results over a two month period where minimal historical data and a small amount of input types were available. These results are significant, because common practice has often overlooked the implementation of an ANN. ANN have often been perceived to be too complex and require large amounts of data to provide accurate results. The LAPART neural network was implemented in an on-line learning manner. On-line learning refers to the continuous updating of training data as time occurs. For this experiment, training began with a singe day and grew to two months of data. This approach provides a platform for immediate implementation that requires minimal time and effort. The results from the LAPART algorithm were compared with statistical regression and a component-based model. The comparison was based on the predictions linear relationship with the measured data, mean squared error, mean bias error, and cost savings achieved by the respective prediction techniques. The results show that the LAPART algorithm provided a reliable and cost effective means to predict the building load for the next day.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

The Advanced Measurement and Analysis of PV Derate Factors project focuses on improving the accuracy and reducing the uncertainty of PV performance model predictions by addressing a common element of all PV performance models referred to as “derates”. Widespread use of “rules of thumb”, combined with significant uncertainty regarding appropriate values for these factors contribute to uncertainty in projected energy production.

The PV Fault Detection Tool project plans to demonstrate that the FDT can (a) detect catastrophic and degradation faults and (b) identify the type of fault. This will be accomplished by collecting fault signatures using different instruments and integrating this information to establish a logical controller for detecting, diagnosing and classifying each fault.

Abstract not provided.

Abstract not provided.