Smart Grids with DER at the Edges

A smart grid is an integrated system of systems. The traditional electric grid constituents include equipment for generation, transmission, distribution, and meters. Other systems being added to smart grids are demand-responsive premises networks with attached devices, appliances, and distributed energy resources (DER). DER may include photovoltaic panels (PV), wind turbines, fuel cells, and storage (stationary and mobile). To make the grid smart, these systems must interoperate. This means that systems developed independently, possibly using different and incompatible technologies, must be adapted to communicate and to work together for a functioning smart grid that keeps the lights on. To achieve the goals of a smart grid while managing costs (called “least cost” planning) utilities require a cost-based perspective.

To plan and manage a smart grid requires data to predict equipment capabilities and customer needs on time scales from seconds to years. Very short-term decisions typically involve electrical parameters such as maintaining AC frequency. Multi-year predictions focus on equipment installations and power capacities. The introduction of DER implies that potentially significant grid resources are now being added at the edge of the grid. To accommodate and evaluate these new resources, utility cost-analysis must become more granular focused than traditional central plant analysis and regional load forecasts. Moreover, a utility customer will no longer be represented just by a meter and a load, but may have resources for generation and storage, and even have loads that migrate during the day (i.e., electric vehicles). Utility operators need near real-time tools to monitor the entire grid, including the edges.

Distributed Marginal Price (DMP) Metrics

Integral Analytics (IA)1 has developed a software analysis tool that delivers new metrics to predict the performance and needs of the entire grid. Measurements are gathered from transmission and distribution grid equipment using SCADA (supervisory control and data acquisition) equipment, data from billing systems and smart meters, and several other third party sources. IA creates a logical base map of the grid and operating parameters. Changes

to the grid are projected onto the base map using predicted developments in technologies, standards, demographics, and the introduction of customer programs such as Demand Response, storage, solar, Transactive Energy, and others.

IA uses these data to produce cost-based measurements called Distributed Marginal Price metrics (DMP). DMP is a prediction over a wide range of time scales of node-by-node costs. The software can predict grid requirements in terms of both magnitude and timing. This is unique in the power industry. No one has captured long term and short term data to generate such a prediction of grid equipment needs. Grid reliability can be determined as a consequence of equipment deployment or deferral. Since equipment costs form the basis of the cost optimization analysis, the utility can be assured that future investment decisions are performed at least cost. Further, the DMP metric represents the direct marginal avoided cost of alternative resources such as solar, storage, Demand Response, or energy efficiency. Placing all resources within a common cost-based analytical framework insures that all resources compete on a level playing field.

DMP Versus LMP (Localized Marginal Price)

DMP delivers metrics of costs, benefits, and tradeoffs to grid operators not only on a minute-by-minute basis, but forecasts over many years. These metrics are convertible to annual forward tenders, third party rebates, or other market exchange metrics. DMP can be considered a major expansion of the localized marginal price (LMP) metrics used by transmission grid operators. LMP was introduced for transmission grid operators to determine the cost of power at specific nodes on the grid in order to manage network congestion. LMP includes a fixed element of grid costs and a variable component that reflect the incremental cost of any dispatch required to add or to remove a generation resource in order to manage grid congestion. DMP also has fixed and variable components for both grid and supply. With DMP, the analytical focus lies with KVA and KVAh such that both grid and supply can be optimized jointly. At the edge of the grid, voltage and power factor play more important and nuanced

www.integralanalytics.comIntegral Analytics, Inc. | 513.762.7621

Introduction to Distributed Marginal Prices (DMPs)Kenneth Wacks, Ph.D. | Management & Engineering Consultant | Home, Building & Utility Systems | www.kenwacks.com

1Integral Analytics is a client of Dr. Wacks.

roles, which DMP explicitly addresses. These issues at the distribution level are the analog of transmission-level congestion.

LMP was enabled by technological advances in data acquisition, database management, data analysis, and communications. IA has extended these capabilities to the distribution grid with near real-time data analysis focused on marginal-cost-based metrics. Some the factors that are encompassed in DMP include:

❚ Distance from substation to consumption and resulting efficiency losses.

❚ The value of providing reactive power, voltage support, and other such technical services at different points on the grid.

❚ Capacity and limiting factors along the circuit.

❚ Broader grid reliability measures1.

According to a study from Boston University2, LMP for a transmission system addresses high-voltage transmission operational costs including maintenance, fuel, congestion, and reserves. These represent about 65% of the total delivered cost of electricity consumed in the United States. DMP measures costs within the low-voltage side of the grid, but importantly includes LMP costs such that 100% (vs. 65%) of the total cost to serve is addressed. To integrate DER, DMP provides the local LMP plus the distribution costs that account for the remaining 35% of delivery costs.

Some costs and benefits not factored into LMP that are significant at the distribution level include:

❚ Costs not factored into LMP: reactive power, line losses in the medium and low-voltage networks, capacity, and limiting factors.

❚ Benefits not factored into LMP: flexible demand response, optimal distributed generation (DG) placement, intermittency mitigation, optimal storage dispatching, and others.

These new opportunities can mitigate the lack of a buffer inventory between supply and demand, which plagues the existing grid, by supporting ancillary services. Through the use of cost-based optimization analysis, electric grid operators can gain value from the optimal placement and interactive operations of certain DER combinations. These optimal portfolios can only be found by first quantifying DMP.

Thus, customers with power generation capabilities (called prosumers) will have a market for selling power into the grid. Third party solar providers will receive higher utility incentives for optimally placed DG. Furthermore, since many will have smart inverters for converting DC solar or storage with control over the power factor (phase between AC voltage and current), these prosumers will be offered an opportunity to be paid for delivering reactive power that helps stabilize grid voltages and minimize line losses. Alternatively, equipment with poor power factors will have a market incentive to be improved by including power-factor correction capabilities. Furthermore, a smart distribution grid with price incentives could automatically manage the load on transformers, thereby extending the life of this equipment. Some distribution grids may include temporary storage to absorb excess generation or may interact with storage facilities at customer premises. A pricing scheme to encourage supplying or depleting these reserves may prove useful and efficient, and DMP is that cost-based metric.

DMP is also a signal of cost, importance, and urgency. DMP may be directly reported as a price for power, may simply be a cost estimate on which to base decisions, or may serve as an ancillary-service cost trigger for use in dispatching flexible demand or storage. In these cases, one to five-minute power flow modeling is necessary and is feasible. The magnitude of the DMP directly reflects the cost or operational urgency. Alternatively, DMP is a long-run cost signal where multi-year forecasts of DMP are estimated. Here, DMP can be used as the foundational platform on which future investment decisions are made. Since DMP is cost based, the result is a least-cost portfolio plan. DMP is derived from analogous least cost optimization methods as are currently used for Integrated

1 Reliability broadly denotes many aspects of electric grid operations. Here, reliability refers to capacity, limiting factors, voltage, power factor, or other normal operating issues. The term does not address emergency or abnormal conditions such as forced outages or other short term extreme conditions.

2 Michael C. Caramanis “It Is Time for Power Market Reform to Allow for Retail Customer Participation and Distribution Network Marginal Pricing?” Boston University Clean Energy and Environmental Sustainability Initiative, March 15th, 2012.

DMPsE X P L A I N I N G

Resource Planning (IRP, also called Least Cost Planning) and LMP calculations.

DMP Applications

IA customers include some of the largest utilities in the country. Many have been mandated by state regulators to expand IRP to include a planning component at the distribution level, called Distribution Resource Planning (DRP). The metrics offered by DMP enable utilities to create an optimal resource plan by measuring costs at a granular level.

The challenges facing utility managers are unprecedented because of fundamental technological and business changes. Utilities are poised to lose revenue from traditional power generation to solar, wind, and storage. DMP enables utilities to evaluate and to quantify the benefits of investments in the distribution grid to accommodate the inevitable growth of distributed energy resources and to improve grid resiliency.

The range of quantitative data offered by DMP is summarized in Figure 1. Actual prices calculated by DMP are illustrated in Figure 2.

In summary, DMP are signals that:

❚ Reflect real prices and costs useful for market tenders (offers to buy and sell).

❚ Measure relative importance of events, which are useful for operations and planning.

❚ Accommodate new grid resources being added at the edge of the grid, some by consumers.

Examples of DMP applications include:

❚ Distribution level planning (DRP versus IRP), forecasting known and unknown changes from a base-case map, and indicating potential investments.

❚ Granular cost analysis at the bulk and distribution level including load flow reliability.

❚ Real-time signals to manage and control DER deployments through incentives (long term tenders).

❚ Incentives based on cost-effectiveness methods approved by regulators.

The  range  of  quantitative  data  offered  by  DMP  is  summarized  in  Figure  1.    Actual  prices  calculated  by  DMP  are  illustrated  in  Figure  2.  

Distributed Marginal Costs (4 Types of DMPs)

Voltage

KVAR

Power Factor

Limiting Factors

Line Losses

Ancillary Services

Plant Following

Wind/ Cloud Firming

Current hour LMP

Asset Protection

Circuit Capacity Deferral

Bank Capacity Deferral

Future Congestion (Trans)

Capacity Premium

10 Year LMP Forecasts

Future Covariance

Grid Side Supply Side

VariableCosts

Fixed Costs /

Capacity

Time

Minutes

Hours

Months

Years

Distributed Marginal Cost (DMC) = Distributed Marginal Price (DMP)But a DMC is not a market traded price. It is a real Cost to Serve, for improved planning and more granular avoided cost calculations.

Copyright 2014 Integral Analytics  

Figure  1  –  Range  of  Distributed  Marginal  Costs  

 

Figure  2  –  Grid  Price  Variations  as  Determined  by  DMP  

In  summary,  DMP  are  signals  that:  

− Reflect  real  prices  and  costs  useful  for  market  tenders  (offers  to  buy  and  sell).  

− Measure  relative  importance  of  events,  which  are  useful  for  operations  and  planning.  

− Accommodate  new  grid  resources  being  added  at  the  edge  of  the  grid,  some  by  consumers.  

Examples  of  DMP  applications  include:  

− Distribution  level  planning  (DRP  versus  IRP),  forecasting  known  and  unknown  changes  from  a  base-­‐case  map,  and  indicating  potential  investments.  

− Granular  cost  analysis  at  the  bulk  and  distribution  level  including  load  flow  reliability.  

− Real-­‐time  signals  to  manage  and  control  DER  deployments  through  incentives  (long  term  tenders).  

Figure 2: Grid Price Variations as Determined by DMPFigure 1: Range of Distributed Marginal Costs

www.integralanalytics.com

DMPs

Why Use DMP

Utilities that use DMP metrics can develop a business plan incorporating DER plus central power generation as revenue sources. DMP facilitates distribution planning by offering a top-down plus a bottoms-up impact forecast of load shapes over a planning horizon. Granular costs can be determined at any point in the grid from transmission to the meter with every impact able to scale up or down consistently, linked to actual assets and capital projects. DMP enables an optimum grid as DER proliferates.

IA sees that the best overall solution for a power grid is a synchronous grid with micro resources balanced in near real time, rather than a series of independent microgrids. DMP complements the free-market aspects of Transactive Energy by adding a layer of reliability while minimizing inefficiency and overspending.

The alternatives to DMP are:

❚ Traditional utility rebates or incentives, which mask the very high avoided-cost areas or opportunities. Because electricity has no, or little, buffer between supply and demand, price spikes offer enormous opportunities for managing energy at the tails of the distribution curve.

❚ Inequities as some users learn how to game the system with ad hoc substitutes for accurate cost signals.

❚ Continued operation of rooftop solar, energy storage systems, volt/VAR optimization equipment, demand response and energy efficiency programs,

and other distributed grid resources as separate silos, with unknown risk.

❚ Sub-optimal management of the grid and grid investments.

❚ More regulatory hearings to increase rates or to introduce fixed charges as revenues from traditional generation shrink.

❚ Lobbying to slow DER.

❚ A decrease in investor confidence in the utility stock.

DMP offers the following key capabilities for power systems:

❚ DMP targets the bulk of avoided costs by adding intelligence about the grid.

❚ DMP provides quantitative data to calculate how much grid-edge resources such as DER could save in power plants not built or feeder lines, transformers, and substations not upgraded or replaced.

❚ DMP offers a platform for monitoring and orchestrating a multitude of distribution grid assets and demand-side resources to create an optimal portfolio.

❚ DMP is a powerful analytic tool to determine which combination of technologies and strategies can provide services with minimum costs and maximum benefits for utilities and customers.

About the Author

Dr. Kenneth Wacks has been a pioneer in establishing the home systems industry and a management consultant to clients worldwide, ranging from startups to large companies. His business focus includes home and building systems, energy management, and digital entertainment networks. The United States Department of Energy appointed him to the 13-member GridWise® Architecture Council to develop smart grid strategies. He chairs the ISO/IEC working group responsible for international home and building system standards including energy management. For further information, please contact Ken at +1 781 662-6211, kenn@alum.mit.edu, or visit kenwacks.com.

About Integral Analytics, Inc.

Integral Analytics (IA) is an analytical software and consulting firm focused on operational, planning, and market research solutions for every aspect of the energy industry. Its proprietary analytical, programming, and statistical methods offer clients more precise valuation, faster and more affordably.