# An Investigation Into the Effects of a Smart Grid Based Real-Time Pricing Retail Electric Market on the Financial Outlook of Distributed Generation Solar Power Systems in New York State

### Abstract

Recent research has suggested that increases in the development and implementation of “smart grid” technologies could allow for a transition from load-profile pricing to real time spot pricing in retail electric supply markets. The following report attempts to investigate the potential effects such a market design could have on the solar value of consumer owned distributed generation (DG) on-site solar power systems in New York State. The study was conducted utilizing data from January, 1, 2014 – December, 31, 2014 to provide a full year analysis. The results of analysis appear promising as they support the initial hypothesis, that spot pricing based electric markets financially favor DG solar panel system owners. Over the stated period, the analysis suggests that a shift from load profile to spot pricing could result in an overall improved solar value of 16.35% depending on the market design criteria used. More specifically, results also appear to indicate a multifaceted correlation with both daily peak prices and seasonal effects.

### Introduction

Over the past two decades, research and development of smart-grid systems and technologies has garnered tremendous interest. “Smart grid” generally refers to a class of technologies which allow for two-way communications between nodes within a utility grid network. The hypothetical example discussed in the this study is a smart grid system in which all nodes in a relevant network are integrated with the aforementioned technology, and thus able to communicate and adapt their controls in real-time.

Under these conditions, Load Servicing Entities (LSE) act as retail sellers of electricity through the local utility distribution firm. With full appreciation for the complexities of transitioning to such an energy market, it will be assumed throughout this paper that all supply and demand “nodes” of relevance are integrated into a smart grid system in this fashion, and thus real-time pricing, or spot pricing, can be passed through to consumers.

Technologically speaking, one of the major benefits of increased communication between utility network nodes is the potential for far more efficient grid management. Economically, smart-grids utilizing a spot pricing market design would more accurately reflect the cost of generation and transmission of electricity thus creating a more efficient market design, and further incentivizing customers to conserve energy.

Other potential benefits of a smart grid include: improved reliability, higher asset utilization, better integration of plug-in electric vehicles (PEVs), and support for new components and applications. Perhaps the greatest environmental benefit is the potential to reduce annual greenhouse gag (GHG) emissions by 60-211 million metric tons of carbon dioxide equivalent (MMT CO2e) compared to “business-as-usual” by 2030, an amount equal to 2.7-9.6 percent of GHG emissions from electricity generation in 2009.

Looking more specially in regards to DG solar and other DG renewable energy technologies, other unique potential benefits of utilizing such a market design include: the potential for government to inject renewable energy subsidies into the marketplace far more effectively and easily, the ability to eliminate the need for net metering practices, and to increase potential entrepreneurial opportunities in the energy market.

While system cost, average retail pricing, geography, incentive structures, and other input functions are essential to finding a consumer’s expected solar value, when DG solar systems are designed as a grid connected system, the real time grid pricing they are compared with may also have a tremendous affect on customer’s financial payback. Furthermore, the extent to which an average consumer owned DG solar system’s generation profile would coincide with “on peak hours” is likely to carry tremendous implications for the consumers’ estimated solar value relative to purchasing the same solar system under a load profiling pricing structure.

This paper limits the scope of research to purely identifying how spot price fluctuations alone could affect the average economic outlook of installing a DG solar panel system in New York State. Based on historical 2014 retail pricing and average NYS solar system specifications in 2014, provides the necessary information to determine and compare the 2014 average financial outlook for a NYS installed fixed mounted flat solar panel system under both load-profile pricing and spot pricing regimes.

It is assumed that the solar electric production value or solar value associated with a DG solar panel system is directly related to the electric supply price. Assuming an initial time t0 exists:

Where SVlp(t) relates to solar value at time t, P relates to the electric price, and EG, relates to the cumulative energy generated at time t. Under load profile pricing market structures, P remains a constant over a relevant range. Under a spot pricing regime however, P now also becomes a function of time:

It is therefore determined that for a given data range of P(t) and corresponding EG(t), one can compare SVlp(t) to SVsp}(t) thus contrasting the savings of a standard DG solar system under both load profile and spot pricing conditions.

Historical retail prices are those actually realized by NYS rate payers while spot prices are hypothetical and thus must be interpreted from LBMP data. By averaging the LBMP data compared to the monthly retail data, one can assume a given multiplier to account for retail gross margin differences through the following formula:

Where M represents the multiplier, RSP represents the projected retail spot price, and LBMP represents location based marginal pricing, which in current markets function as whole sale spot prices.

There are several limits and important assumptions of the proposed methodology which are noteworthy. Analysis was conducted purely utilizing New York State data. An average of each metric were taken for each service area in the State and thus represent arithmetic mean values for the state as a whole. It is possible that results may differ when analyzed on the service area level, in a different State, or any other different geographical boundary. The analysis was completed utilizing data from a single year, 2014. It is assumed that data from each year should not vary widely enough to reverse the trends observed and that by maintaining a constant LSE marginal price increase any potential effects of this nature will be mitigated. The availability of data likewise required minimal compromises to the fidelity of analysis. LBMP data is available in five minute intervals but solar analysis can only be projected accurately based on hourly intervals. This required that LBMP and corresponding data be averaged over hourly intervals, to maintain a consistent level of accuracy. Most importantly however, the research operates under the assumption of a relatively low adoption rate of DG solar which do not have substantial impacts on current marketplace prices. It is expected that higher rates of DG solar adoption would lead to a flattening of any spot price trend and thus reduce many of the positive impacts which may be observed. As 2015, this trend does hold true considering that solar electric generation (both utility scale and DG) account for just 0.4% of all electric generation in the U.S.

### Results and Discussion

Figure 1

Figure 1 illustrates a week’s span of the simulated DG solar system AC power production. Estimates were derived from the National Renewable Energy Labs PVWatts Calculator tool $^5$. The program takes the input measurements based on those shown in Table 1, to generate a predicted power output for each hour annually based on the local geography. Table 2 subsequently describes the key economic output metrics relating to the system.

Raw data was obtained as .csv files from aforementioned sources. Data analysis and computations were completed utilizing Microsoft Excel. Under the above conditions, the Average Monthly Retail Price ($/MWh) was compared to the real time solar panel system AC Power output in hour intervals throughout the year. Utilizing the solar value equation, a corresponding projected solar value was calculated for three scenarios: utilizing Load Profile Retail Pricing, Spot Pricing based on an annual multiplier, and Spot Pricing based on monthly multipliers. #### Comparison of Spot Price Projection Methods The use of a multiplier to project RSP is an important one. Further analysis indicates however that determining the time length upon which the RSP and LBMP averages are taken, greatly influences the multiplier value. As this research scenario represents a hypothetical spot price retail market, there is little precedent to build from. Two sets of spot price retail projections were therefore determined, one where the multiplier used was calculated using yearly averages of RSP and LBMP, and a second set of monthly multipliers, each derived from a given month’s averages. Interestingly, while both multiplier scenarios provide the same yearly average retail price and that which was actually observed under load profile pricing, ($162.10/MWH), the two methodologies yield vastly different results.

Figure 2 represents a week-long snapshot of hourly RSP projections from 7/1/14 – 7/7/14. While the overall price trend between the annual and monthly multipliers is nearly mirrored, the effective amplitude of the monthly multiplier is 63.3% larger than annual multiplier during the month of July.(4.88 vs. 3.08 respectively)

Figure 3 highlights this discrepancy on a year’s timescale. Projected monthly average RSP based on the monthly multiplier, axiomatically matches the observed load profile retails price each month. The monthly average RSP based on the annual multiplier however appears vary substantially for the observed load profile price. This trend is evidence of a pricing skew primarily during January – April.

As a result of the trends observed above, it is hypothesized that monthly multipliers more accurately represent an expected RSP trend as load profile prices are currently set by LSEs on a monthly basis. Monthly multipliers allow LSEs to maintain the equivalent monthly price margins and more accurately represent historically observed price trends.

Figure 2

Figure 3

#### Daily Peak Pricing Effects on Solar Value

Further analysis indicates that RSP has a positive correlation to solar value under both annual and monthly multiplier methods.Figure 4 illustrates a one week time horizon comparison of electric retail pricing methods against solar power system output. To provide a more substantial indication of this translation to solar value, Figure 5 subsequently provides a direct comparison during the same time period.

Retail pricing methods appear to trade off in terms of spot solar values but do not provide a comprehensive documentation of their effects. To provide that information, the summation of spot solar values was taken over the entirety of 2014 for each pricing method. The results of these calculations can be found in Table 3.

Figure 4

Figure 5

#### Seasonal Price Trend Effects on Solar Value

In addition to solar value effects with respect to daily price trends, seasonal effects were also observed. Figure 6 illustrates the expected monthly solar values for each retail pricing method. Importantly Figure 7 provides further indication of the seasonal effects, especially with regard to the RSP monthly multiplier method. During low solar radiation months for the northern hemisphere, (Nov-Mar) solar values remain within + or – 10% between load profile retail prices and monthly multiplier RSP. During higher solar radiation months however, the difference in solar value is clearly more substantial peaking at 33% in September. This trend would therefore appear to indicate a multiplicity of market effects resulting in overall high solar values. Daily peak pricing appears to favor the solar power system production times and this effect is further exaggerated during seasons where solar output is expected to be greater.

Figure 6

Figure 7

### Conclusion and Future Research

It has been conjectured throughout the literature that smart grid infrastructure would allow electric grid retail pricing to transition from load profile methods to sport pricing derived directly from wholesale markets. The study sought to examine the potential affects of such a market transition on the value of DG solar power systems in New York State by examining historical pricing trends from 2014 in relation to the expected solar power system production. Analysis suggests that if RSP were to come into effect, the resulting marketplace could be advantageous to DG solar power system owners, without additional incentive structures or changes in LSE profit margins. While at the time of publication, smart-grid systems are still in their infancy of adoption, a select few LSEs including Commonwealth Edison and Ameren offer real-time pricing options to consumers with installation of smart meters. This opens the possibility for future studies about the current effects of these markets for existing DG systems.

Further analysis of this trend using data from other states or nations also remains to be seen. Similarly, one can hypothesize the potential benefits a RSP structure could have on DG battery storage technologies. Such a transition could afford consumers the ability to “trade” electricity along with market fluctuations and optimize trading to their financial advantage. DG solar and DG battery combination systems could also be investigated to compare and contrast the financial potential with other options. The culmination of this research would further illuminate the implications of a smart grid infrastructure on renewable and energy storage markets from a consumers perspective, thus providing important insights for policy markers and utility regulators.

#### Basic Definitions and Industry Nomenclature

• Smart-Grid: A theoretical utility system where all “nodes” are able to communicate in real-time with every other node in the network.
• Node: Any interface point in the utility network whether generation, transmission, or consumption oriented.
• Load-Profile: Using interval metering of electric customers within a profile class, the resulting data is used to develop an accurate representation of a customer’s usage pattern over time. This usage pattern is then used to determine corresponding rate charges for customers within each profile class.
• Locational Based Marginal Pricing (LBMP): A pricing methodology under which the price of Energy at each location in the NYS Transmission System is equivalent to the cost to supply the next increment of Load at that location (i.e., the short-run marginal cost). The short-run marginal cost takes generation Bid Prices and the physical aspects of the NYS Transmission System into account. The short-run marginal cost also considers the impact of Out-of-Merit Generation (as measured by its Bid Price) resulting from the Congestion and Marginal Losses occurring on the NYS Transmission System which are associated with supplying an increment of Load. The term LBMP also means the price of Energy bought or sold in the LBMP Markets at a specific location.
• Real-time or Spot Pricing: Represents the current price of electricity within the marketplace. For the purposes of this paper as retail spot pricing trends in NYS do not yet exist, spot pricing trends are determined by taking the percentage markup between the yearly average LBMP and the yearly average load-profile price and multiplying it by the LBMP price for every time stamp in the calendar year.
• Load Servicing Entity (LSE): Secures energy and transmission service (and related Interconnected Operations Services) to serve the electrical demand and energy requirements of its end-use customers.
Distributed Generation (DG): Refers to electric generation nodes separate from traditional centralized power plants, often consumer owned or operated on-site of a consuming node.
• Solar Value: The monetary value of a given quantity of electric energy generated by a DG solar panel system.
On Peak Hours: The time when customer demand for electricity is highest, usually corresponding to higher electric spot prices.

#### References

1. Joskow, Paul L., and Jean Triole. “RETAIL ELECTRICITY COMPETITION.” NBER WORKING PAPER SERIES 10473rd ser. (2004): n. pag. NBER.org. NATIONAL BUREAU OF ECONOMIC RESEARCH. Web.
\item Environmental Protection Agency, 2011, Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2009. Table 2-13.
2. The Green Grid – Energy Savings and Carbon Emissions Reduction Enabled by a Smart Grid, Technical Update (Electric Power Research Institute (EPRI), June 2008), 1016905.
3. “Average Retail Price of Electricity Monthly.” ELECTRICITY DATA BROWSER. U.S. Energy Information Administration (EIA), n.d. Web.
4. PVWatts Calculator. Computer software. PVWatts Calculator. National Renewable Energy Labs (NREL), n.d. Web. 15 Sept. 2015.
5. Location Based Marginal Price \& Real Time Load Data. N.d. Raw data. NYISO, n.p.
6. Statewide 200kW or Less Residential/Non-Residential Solar Photovoltaic Incentive Program. N.d. Raw data. Data.NY.Gov, n.
7. Net Metering. NationalGridUS.com. National Grid 2015. Web. \\www.nationalgridus.com/niagaramohawk/business/energyeff/4_net-mtrg
8. “For Suppliers and Partners: Load Profiles.” NYSEG.com. New York State Electric and Gas Corporation, n.d. Web. <http://www.nyseg.com/SuppliersAndPartners/electricityescos/loadprofiles.html&gt;.
9. “Glossary of Terms: Load Servicing Entities.” NERC.com. North American Electric Reliability Corporation, n.d. Web. <http://www.nerc.com/files/glossary_of_terms.pdf
10. “Smart Grid.” C2ES.org. Center for Climate and Energy Solutions, n.d. Web. <http://www.c2es.org/technology/factsheet/SmartGrid_edn36
11. “Smart Grid.” U.S. Department of Energy. Office of Electricity Delivery \& Energy Reliability, n.d. Web. 16 Sept. 2015. <http://energy.gov/oe/services/technology-development/smart-grid.
12. Beneficial Complexity: A Field Experiment in Technology, Institutions, and Institutional Change in the Electric Power Industry, Manuscript, September 2009. Joint with David Chassin. Presented at: Research Roundtable, Searle Center for Law, Regulation, and the Economy, Northwestern University, June 2009; International Society for New Institutional Economics conference, June 2009; International Association of Energy Economics conference, June 2009; Workshop in Political Theory and Policy Analysis, Indiana University, September 2008.
13. “What Is U.S. Electricity Generation by Energy Source” U.S. Energy Information Administration, 31 Mar. 2015. Web. <http://www.eia.gov/tools/faqs/faq.cfm?id=427t=3&gt;.