The rising adoption of plug-in electric vehicles (PEVs) leads to the alignment of their electricity and their mobility demands. Therefore, transportation and power infrastructures are becoming increasingly interdependent. In this work, we uncover patterns of PEV mobility by integrating for the first time two unique data sets: (i) mobile phone activity of 1.39 million Bay Area residents and (ii) charging activity of PEVs in 580,000 sessions obtained in the same region. We present a method to estimate individual mobility of PEV drivers at fine temporal and spatial resolution integrating survey data with mobile phone data and income information obtained from census. Thereupon, we recommend changes in PEVs charging times of commuters at their work stations that take into account individual travel needs and shave the pronounced peak in power demand. Informed by the tariff of electricity, we calculate the12 monetary gains to incentivize the adoption of the recommendations. These results open avenues for planning for the future of coupled transportation and electricity needs using personalized data.

In future transportation systems, the charging behavior of electric Autonomous Mobility on Demand (AMoD) fleets, i.e., fleets of electric self-driving cars that service on-demand trip requests, will likely challenge power distribution networks (PDNs), causing overloads or voltage drops. In this paper, we show that these challenges can be significantly attenuated if the PDNs' operational constraints and exogenous loads (e.g., from homes or businesses) are accounted for when operating an electric AMoD fleet. We focus on a system-level perspective, assuming full coordination between the AMoD and the PDN operators. From this single entity perspective, we assess potential coordination benefits. Specifically, we extend previous results on an optimization-based modeling approach for electric AMoD systems to jointly control an electric AMoD fleet and a series of PDNs, and analyze the benefit of coordination under load balancing constraints. For a case study of Orange County, CA, we show that the coordination between the electric AMoD fleet and the PDNs eliminates 99% of the overloads and 50% of the voltage drops that the electric AMoD fleet would cause in an uncoordinated setting. Our results show that coordinating electric AMoD and PDNs can help maintain the reliability of PDNs under added electric AMoD charging load, thus significantly mitigating or deferring the need for PDN capacity upgrades.

This paper presents the Plug-Load Appliance Identification Dataset (PLAID), a labelled dataset containing records of the electrical voltage and current of domestic electrical appliances obtained at a high sampling frequency (30 kHz). The dataset contains 1876 records of individually-metered appliances from 17 different appliance types (e.g., refrigerators, microwave ovens, etc.) comprising 330 different makes and models, and collected at 65 different locations in Pittsburgh, Pennsylvania (USA). Additionally, PLAID contains 1314 records of the combined operation of 13 of these appliance types (i.e., measurements obtained when multiple appliances were active simultaneously). Identifying electrical appliances based on electrical measurements is of importance in demand-side management applications for the electrical power grid including automated load control, load scheduling and non-intrusive load monitoring. This paper provides a systematic description of the measurement setup and dataset so that it can be used to develop and benchmark new methods in these and other applications, and so that extensions to it can be developed and incorporated in a consistent manner.

To accelerate adoption of non-residential charging for electric vehicles, sites must maximize utilization of existing electrical infrastructure. In this study we model electric vehicle charging at a workplace using real charging data and evaluate the lifetime of the site’s transformer as the number of charging stations is incrementally increased. We implement and compare a range of control schemes for workplace charging including minimizing the peak load, capping the total load, minimizing bills under different rate structures with time-of-use energy costs and demand charges, and directly minimizing the transformer’s aging. These are compared by the number of vehicles they allow the transformer to support, the transformer’s health, and the operator’s electricity bill. We draw a connection between minimizing the peak load and improving the transformer’s health. We observe that minimizing the electricity bill is the best scheme by both criteria when the bill includes a demand charge; in our experiment it allowed the infrastructure to support over 67% more cars than under uncontrolled charging. To protect the transformer we recommend that demand charges or capacity management be applied to parking lots of charging electric vehicles with high infrastructure utilization, and operators schedule charging to minimize their electricity bills.

A large number of mechanisms are proposed to manage potential problems in distribution networks caused by the participation of distributed energy resources (DERs) in the wholesale markets. In this paper, we first introduce a practical and straightforward mechanism, based on capacity limits, which avoids conflicts between the transmission system operator and the distribution system operators (DSOs). Using a large number of real electric vehicle (EV) commercial charging stations we then show how an EV aggregator can forecast the opportunity cost incurred by offering a mid-term capacity limit service to the DSO. This cost is computed based on the estimated profit that the aggregator could gain in the day-ahead and real-time markets. The proposed methodology guarantees robustness against evolving EV uncertainty, both in terms of service delivery and driving requirements. It also allows the use of a variety of time-series forecasting methods without forecasting electricity prices and EV scenarios. The results of our empirical analysis show the exponential increase of opportunity cost and the considerable increase of the prediction intervals as the capacity limit decreases. The produced offering curves can be used as an indication of the underutilization cost of DERs caused by the DSO's limitations.

The increased integration of photovoltaic (PV) systems in distribution grids reduces visibility and situational awareness for utilities, because the PV systems’ power production is usually not monitored by them. To address this problem, a method called Contextually Supervised Source Separation (CSSS) has been recently adapted for real-time estimation of aggregate PV active power generation from aggregate net active and reactive power measurements at a point in a radially configured distribution grid (e.g., substation). In its original version, PV disaggregation is formulated as an optimization problem that fits linear regression models for the aggregate PV active power generation and true substation active power load. This paper extends the previous work by adding regularization terms in the objective function to capture additional contextual information such as smoothness, by adding new constraints, by introducing new regressors such as ambient temperature, and by investigating the use of time-varying regressors. Furthermore, we perform extensive parametric analysis to inform tuning of the objective function weighting factors in a way that maximizes performance and robustness. The proposed PV disaggregation method can be applied to networks with either a single PV system (e.g., MW scale) or many distributed ones (e.g., residential scale) connected downstream of the substation. Simulation studies with real field recorded data show that the enhancements of the proposed method reduce disaggregation error by 58% in winter and 35% in summer compared with previous CSSS-based work. When compared against a commonly used transposition model based approach, the reduction in disaggregation error is more pronounced (78% reduction in winter and 45% in summer). Additional simulations indicate that the proposed algorithm is applicable also for PV systems with time-varying power factors. Overall, our results show that – with appropriate modeling and tuning – it is possible to accurately estimate the aggregated PV active power generation of a distribution feeder with minimal or no additional sensor deployment.

We present an algorithm that estimates a clear sky performance signal from the measured power of a PV system. The algorithm uses only observed power output, and assumes no knowledge of weather, irradiance data, or system configuration metadata. This is a novel approach to understanding the clear sky behavior of an installed PV system, that does not rely on traditional atmospheric and geometric modeling techniques.

We analyze the decision problem of a workplace manager determining the optimal number of electric vehicle (EV) charging stations to install in order to supply employees with adequate charging service and take advantage of EV smart charging opportunities, while also considering the investment costs of installing additional infrastructure. To optimize smart charging revenues, we introduce assignment methods that designate specific charging spots to arriving EVs. This includes two heuristic packing methods and an offline oracle. Under each assignment method, we perform a cost-benefit analysis that determines the optimal number of chargers to install on-site. Our results highlight the importance of considering revenues from demand charge reduction to incentivize investment in additional infrastructure needed to tap into the smart charging potential of EVs, at least under current energy market prices.

New sources for ancillary services are needed, yet the requirements for service provision in most countries are explicitly formulated for traditional generators. This leads to waste of the potential for new technologies to deliver ancillary services. In order to harness this potential, we propose to parameterize the requirements of ancillary services so that reserves can be built by combining the advantageous properties of different technologies. The proposal is exemplified through a laboratory test where it shown that the system needs can be covered through cheaper and smaller reserves.

Photovoltaic (PV) systems are increasing in distribution systems, but utilities lack visibility of the generation of this distributed PV. This paper presents a set of methods for disaggregating the photovoltaic (PV) generation downstream of a measurement device that measures net load using only readily available measurements. We propose two strategies in which we use measurements from the substation as well as a proxy solar irradiance measurement. Using these two measurement points, we first propose a multiple linear regression strategy. We estimate a relationship between the measured reactive power and the load active power consumption, which are then used in real-time disaggregation. Then, we expand this strategy to reconstruct the errors in the estimators, thus separating the solar and load signals from their aggregate. We show that it is possible to disaggregate the generation of a 7.5 megawatt photovoltaic site with a root-mean-squared error of ≈ 450 kilowatts.

The rising adoption of plug-in electric vehicles (PEVs) leads to the temporal alignment of their electricity and mobility demands. However, mobility demand has not yet been considered in electricity planning and management. Here, we present a method to estimate individual mobility of PEV drivers at fine temporal and spatial resolution, by integrating three unique datasets of mobile phone activity of 1.39 million Bay Area residents, census data and the PEV drivers survey data. Through coupling the uncovered patterns of PEV mobility with the charging activity of PEVs in 580,000 session profiles obtained in the same region, we recommend changes in PEV charging times of commuters at their work stations and shave the pronounced peak in power demand. Informed by the tariff of electricity, we calculate the monetary gains to incentivize the adoption of the recommendations. These results open avenues for planning for the future of coupled transportation and electricity needs using personalized data.

This paper is the first part of a two-part series in which we present results from one of the first worldwide experimental demonstrations of frequency regulation in a commercial building test facility. We demonstrate that commercial buildings can track a frequency regulation signal with high accuracy and minimal occupant discomfort in a realistic environment. In addition, we show that buildings can determine the reserve capacity and baseline power a priori, and identify the optimal tradeoff between frequency regulation and energy efficiency. In part I, we introduce the test facility and develop relevant building models. Furthermore, we design a hierarchical controller for the heating, ventilation, and air conditioning system that consists of three levels: 1) a reserve scheduler; 2) a building climate controller; and 3) a fan speed controller for frequency regulation. We formulate the reserve scheduler as a robust optimization problem and introduce several approximations to reduce its complexity. The building climate controller is comprised of a robust model predictive controller and a Kalman filter. The frequency regulation controller consists of a feedback and a feedforward loop, provides fast responses, and is stable. Part I presents building model identification and controller tuning results. Specifically, we find out that with an appropriate formulation of the model identification problem, a two-state model is accurate enough for use in a reserve scheduler that runs day-ahead. In part II, we report results from the operation of the hierarchical controller under frequency regulation.

This paper is the second part of a two-part series presenting the results from an experimental demonstration of frequency regulation in a commercial building test facility. In Part I, we developed relevant building models and designed a hierarchical controller for reserve scheduling, building climate control and frequency regulation. In Part II, we introduce the communication architecture and experiment settings, and present extensive experimental results under frequency regulation. More specifically, we compute the day-ahead reserve capacity of the test facility under different assumptions and conditions. Furthermore, we demonstrate the ability of model predictive control to satisfy comfort constraints under frequency regulation, and show that fan speed control can track the fast-moving RegD signal of the Pennsylvania, Jersey, and Maryland Power Market (PJM) very accurately. In addition, we report the observed effects of frequency regulation on building control and provide suggestions for realworld implementation projects. Our results show that hierarchical control is appropriate for frequency regulation from commercial buildings.

Having investigated the determinants of electricity consumption in relation to household and dwelling characteristics, Chapter 6 provides a closer look at the behavioral patterns of household appliance use and electricity consumption. The OTB sample was used to conduct correlation and factor analysis. This Chapter deals with the Research Question III-2 of this thesis: (Chapter 1, Section 3, pg. 16-17) “ III. What are the behavioral patterns and profiles of energy consumption? The sub-question is: What are the behavioral patterns of electricity consumption? How do they relate to the household characteristics, revealing behavioral profiles? “ The research reported in this Chapter was conducted by Bedir. The data was collected by a questionnaire prepared by Guerra Santin and Bedir, using OTB’s means of data collection. The analysis was done, and the paper was written by Bedir. The co-author (E.C. Kara) commented on methodology of the research. The co-author has given his permission to include the paper in the thesis. This study was published in Energy and Buildings: Bedir, M. Kara, E.C. “Behavioral Patterns and Profiles of Electricity Consumption in Dutch Dwellings” Energy and Buildings, Available online 12 June 2017, http://dx.doi. org/10.1016/j.enbuild.2017.06.015

Real-time photovoltaic (PV) generation information is crucial for distribution system operations such as switching, state-estimation, and voltage management. However, most behind-the-meter solar installations are not monitored. Typically, the only information available to the distribution system operator is the installed capacity of solar behind each meter; though in some cases even the presence of solar may be unknown. We present a method for disaggreagating behind-the-meter solar generation using only information that is already available in most distribution systems: advanced metering infrastructure, substation monitoring, and generation monitoring at a few PV systems nearby the circuit. The proposed method accurately predicts which homes have solar in over 90% of cases, and recovers the 15-min resolution PV generation signals with root mean square errors between 20% and 50% of average daily PV generation both historically and real-time. A sensitivity analysis shows the method to be robust to the number of buildings and time span of data used to fit. However including more than 3 solar proxies can cause false positive of PV systems behind meters. We find that the proposed method performs better at homes that export electricity to the grid more often.

The demand for vehicle charging will require large investments in power distribution, transmission, and generation. However, this demand is often also flexible in time, and can be actively managed to reduce the needed investments, and to better integrate renewable electricity. Harnessing this flexibility requires forecasting and controlling electric vehicle (EV) charging at thousands of stations. This paper addresses the problem of forecasting and management of the aggregate flexible demand from tens to thousands of EV supply equipment (EVSEs). First, it presents an equivalent time-variant storage model for flexible demand at an aggregation of EVSEs. The proposed model is generalizable to different markets, and also to different flexible loads. Model parameters representing multiple EVSEs can be easily aggregated by summation, and forecasted using autoregressive models. The forecastability of uncontrolled demand and storage parameters is evaluated using data from 1341 nonresidential EVSEs located in Northern California. The median coefficient of variation is as low as 24% for the forecast of uncontrolled demand at the highest aggregation and 10-15% for the storage parameters. The benefits of aggregation and forecastability are demonstrated using an energy arbitrage scenario. Purchasing energy day ahead is less expensive than in the real-time market, but relies on a uncertain forecast of charging availability. The results show that the forecastability significantly improves for larger aggregations. This helps the aggregator make a better forecast, and decreases the cost of charging in comparison to an uncontrolled case by 60% with respect to an oracle scenario.

Enabling deep penetration of distributed energy resources (DERs) requires comprehensive monitoring and control of the distribution network. Increasing observability beyond the substation and extending it to the edge of the grid is required to achieve this goal. The growing availability of data from measurements from inverters, smart meters, EV chargers, smart thermostats and other devices provides an opportunity to address this problem. Integration of these new data poses many challenges since not all devices are connected to the traditional supervisory control and data acquisition (SCADA) networks and can be novel types of information, collected at various sampling rates and with potentially missing values. Visualization and analytics for distributed energy resources (VADER) system and workflow is introduced as an approach and platform to fuse these different streams of data from utilities and third parties to enable comprehensive situational awareness, including scenario analysis and system state estimation. The system leverages modern large scale computing platforms, machine learning and data analytics and can be used alongside traditional advanced distribution management system (ADMS) systems to provide improved insights for distribution system management in the presence of DERs.

The Internet of Things (IoT) is creating a major shift from the traditional approaches to operating and maintaining all energy systems. With the proliferation of cheap sensors, power grids are generating vast amounts of data and there is an increasing interest around data-driven decision making, or more broadly, energy analytics. Due to the inherent nature of device and data ownership in the power grids, these analytics efforts have been siloed. Energy analytics developers create databases for monolithic applications that lack the infrastructure to quickly transform data between computational resources and persistent data storage, easy dissemination and validation of results, and support integration. We propose a data architecture that supports broader use of data technologies and systems integration guided by publish and subscribe architecture, and polyglot persistence. We provide use cases to discuss our design choices using the electric power grid as the main application domain. However, we believe that the energy industry as a whole can benefit from the proposed architecture.

In EU member states, the consumption of electricity per dwelling has remained more or less constant, although the consumption of large appliances has decreased considerably in the last 2 decades. This stabilization is caused by the increased ownership, usage and consumption levels of Information, Communication and Entertainment (ICE) appliances. This paper aims to analyze electrical appliance use in the Dutch housing stock, and identify behavioral patterns and profiles of electricity consumption. The analysis is conducted by applying descriptive, correlation, and exploratory factor analyses on data collected from 323 dwellings in two neighborhoods in the Netherlands. Our results show that behavioral patterns could be found based on actual occupant behavior of lighting and appliance use, especially depending on household activities like cooking, (personal) cleaning, etc. Behavioral profiles could be determined based on household and dwelling characteristics, i.e. household size, income, education, dwelling type, age, hours of working, outside. The 4 profiles set up in this research are explained as 'family, 'techie,' 'comforty,' and 'conscious.' These profiles showed statistically significantly differences in terms of their electricity consumption levels. (C) 2017 Elsevier B.V. All rights reserved.