Preliminary AMI deployment costs from the U.S. DOE Smart Grid Investment Grant Program

Melissa Chan, Daniel Handal, Thomas Murray and David Feliciano*; and Joseph Paladino**


 

Introduction

 

Many in the electricity industry agree that transforming current U.S. electricity infrastructure to a Smart Grid – automating transmission, distribution, and metering systems – could make electricity delivery more reliable and efficient while saving money for consumers. However, upgrading existing infrastructure will be expensive.  A 2004 Electric Power Research Institute (EPRI) report estimated it would cost $165 billion to develop Smart Grid infrastructure through 2020[1].  In 2011, EPRI updated this estimate to between $338 and $476 billion[2].  A 2008 Brattle Group report estimated that expanding transmission and distribution with Smart Grid capabilities would cost $880 billion[3]. One step towards meeting the need for significant investment is the U.S. Department of Energy (U.S. DOE) American Recovery and Reinvestment Act (ARRA) Smart Grid Investment Grant (SGIG) program. (Figure 1.)

 

Figure 1

 

 

The SGIG program allocates $3.4 billion ($7.9 billion with cost share) among 99 recipients. Awards focus on advanced metering infrastructure (AMI), customer systems, distribution, and transmission. Not only is the SGIG program a step towards having a Smart Grid, it also provides a window into utility Smart Grid deployment experience and costs. All 99 recipients report the number of assets that they purchase and install each quarter of the SGIG program. With such a large pool of recipients, the data shows a broad array of deployment levels, technologies chosen, and installation schedules. We provide a method to assess the costs, considering the range of recipients’ deployment strategies. The AMI costs presented in this paper can be updated over the course of the program, and provide insight into the range of AMI technology deployment costs. Calculated on a per meter cost for residential, commercial, and industrial customers, these costs can be used to revise and update cost estimates of Smart Grid infrastructure development.

 

SGIG program AMI technologies

 

We think of AMI as multiple integrated systems. While a conventional system involves monthly visits to read electromechanical meters, AMI can automate information delivery to the utility and customers (Figure 2).

 

Figure 2

 

The smart meter is the foundation of AMI, enabling automated delivery of electricity consumption to a utility. It can also support two-way communication.  A field area network transmits information to customers, such as a time-of-use rate. A backhaul system collects customer meter data and transmits it to the utility. Within these systems, there are myriad equipment combinations among the 68 SGIG recipients that are implementing AMI projects.

 

Because SGIG recipients include utilities around the country and varying scales of service and customer demographics -- municipal utilities, investor-owned utilities, and cooperative utilities -- their AMI requirements vary. This leads them to choose different solutions driven by factors such as population density, topography, and legacy systems. Current AMI meter communication networks can be broken down into three categories: wireless, power line carrier and dedicated communication lines. A utility that has a high population density (i.e. urban) is more likely to use a wireless mesh network. However, a power line carrier AMI solution is an excellent choice for a rural utility with one customer per square mile, while a dedicated communication line is an effective solution for utilities that offer customers multiple services (internet, television, water, etc.) because all of these services can share the cost of the communication line. Although these examples highlight some factors that drive utilities to one type of solution or another, each utility has some unique design criteria.

 

Since legacy systems, current needs and the deployment vary, no two SGIG recipients are installing identical AMI systems. Some recipients already have a backhaul system and field area network with some smart meters, and are using their grant to install smart meters to their remaining customers. Some recipients have no AMI systems yet, and are only installing smart meters. Recipients install different equipment. For example, although all recipients install smart meters, the cost varies according to brand and additional functions needed. One utility might buy smart meters that collect and transmit hourly electricity data, turn service on and off remotely, and send an alert in the case that it is damaged or tampered with. Another utility might buy the base model instead, and the costs incurred by the utilities will vary depending upon the different features purchased.

 

SGIG AMI asset costs

 

Our analysis examines the range of asset costs incurred by SGIG recipients. We thought that this approach is best, given the breadth of factors affecting AMI choices: utility type, smart meter functions, field area network, and backhaul network options. Through the third quarter of 2011, 68 utilities reported AMI costs. Figures 3 and 4 show our findings, with statistical outliers removed. Because not all utilities reported costs in all AMI technology categories, we show the number of utilities reporting each cost.

 

Figure 3

 

 

Figure 4

 

 

As shown in Figure 3, 50 utilities reported meter costs, 41 reported backhaul costs, 37 reported communication costs, and 21 reported other costs. The costs are categorized by utility type to account for possible factors such as population density, topology, demand levels, rate schedules, and other economic characteristics. However, we see that costs do not vary considerably by utility type, such that meter costs are $178 on average with a standard deviation of $84 for all utilities and the mean cost by utility type falls within 2 standard deviations of the total population mean. Communications, backhaul, and other costs are about $13-$15/meter.

 

Aggregating the costs presented in Figure 3, we estimated total AMI deployment cost ($/meter). The estimated total AMI cost is a composite of the meter, communications, backhaul, and other costs reported by SGIG recipients. The average estimated total AMI cost is $220/meter. As shown in Figure 4, 81% of this estimated cost is due to the meter. All AMI projects require smart meters even though, as discussed earlier, AMI requirements vary by utility because of such factors as population density, topology, etc. We are not surprised that smart meters comprise the majority of the estimated total AMI cost because they are the common thread among all the SGIG recipients that reported AMI cost data.

 

Conclusions

 

The U.S. DOE SGIG program is a very small part of the spending needed to develop a nationwide Smart Grid. However, the result of this spending is not trivial. Immediate results of the U.S. DOE SGIG program are the installation and deployment of a portion of the infrastructure needed for a nationwide Smart Grid. Longer term decisions made in the public and private sector to deploy additional assets and electricity services could be based on the lessons learned and insights into the economic, environmental, and social benefits gained from this government sponsored deployment.

 

We are able to estimate AMI asset costs using SGIG reported costs. We used these unit costs to estimate the cost to deploy AMI at SGIG program costs observed through Q3 2011, and see that 81% of installation costs are due to the meter. The SGIG program results provide additional insight into costs by offering greater detail on cost by utility type, communication system type, and customer class per utility. We summarize these costs in our results, but recommend updating them throughout the SGIG program for greater accuracy.

 

Future Work

 

The SGIG program is not only installing smart grid assets; it is also evaluating the effect that these assets have on grid performance and cost. Because the SGIG program addresses all parts of the Smart Grid, it also collects cost data for distribution automation and transmission monitoring equipment. Collectively, these costs can be used to revise the estimated cost of developing a Smart Grid infrastructure. The magnitude of these improvements is estimated by measuring performance before and after installing the equipment. There are several challenges associated with this analysis that we are aware of. Performance is challenging to monitor in many cases because recipients might not have recorded historical performance data. For example, they might not have had a means to measure it, such as a monitor for distribution feeders. Additionally, the magnitude of change may be small because the size of the program is small. Lastly, because the program is only measuring results for a few years, the long term effects will not be known. Regardless of these challenges, we anticipate two things. One, our results will provide ballpark estimates of performance improvements resulting from Smart Grid equipment deployment that can be used to inform future utility Smart Grid deployments. Two, our method can be applied to other contexts, such as Public Utility Commission (PUC) reporting to justify Smart Grid expenses.

 

Acknowledgements

The authors thank David Walls of Navigant for his comments and insight on our findings and paper.

 


* Navigant; 77 South Bedford St., Suite 400, Burlington, MA 01803-5154 Melissa.chan@navigant.com

** U.S. Department of Energy, office of Electricity Delivery and Energy Reliability, 1000 Independence Ave. SW, Washington, DC 20585

[1] EPRI. Power delivery system of the future: a preliminary estimate of costs and benefits, EPRI, Palo Alto, CA: 2004. 1011001.

[2] EPRI. Estimating the costs and benefits of the smart grid: A preliminary estimate of the investment requirements and the resultant benefits of a fully functioning smart grid. EPRI, Palo Alto, CA: 2011. 1022519.

[3] Chupka, M.W., Earle, R., Fox-Penner, P., Hledik, R. Transforming America’s power industry: The investment challenge 2010 – 2030. Edison Electric Institute, Washington D.C.: 2008.

 

Click to view a printable version of this article.