July 1, 2009

How big an issue is the stranded asset as utilities push ahead with AMI and smart grid deployments? We are assembling various perspectives on that topic in ongoing reporting and spoke most recently with Eric Miller, chief solutions officer of smart grid firm Trilliant.

          Here's the issue.  Smart grid deployments are driving equipment to be replaced before its recorded asset value -- usually referred to as the “book value” -- reaches zero.  That's usually a no-no for IOUs for several reasons including the importance of the depreciation of assets for shareholders plus how state regulators react to stranded asset issues.

          It can be a tough topic to report on since utilities usually eschew publicizing the dilemma and their technology and service partners are not always keen on chatting about the painful side of big contracts wins.

          Yet, helping customers solve this part of the smart grid deployment dilemma could win vendors some business and some good will among potential client firms and we commend Trilliant for addressing the topic with us.

          Miller has observed that PUCs do not generally see stranded assets as a stopping block in the reach and scope of a smart grid deployment.  Utilities coming before them that demonstrate the final result is a measurable improvement in the service delivered and reduction in the total capital needed to run the utility generally do not have trouble securing their approval, Miller noted.

          He stressed that the utility has to show how the smart grid will lead to net reduction in both operating expenses and total capital to secure approval for a smart grid deployment.

          The most often referenced stranded asset is the treatment of existing meters, said Miller.  Many utilities have made recent investments in remote meter reading and some in automated meter infrastructure.  They need to show that the cost of retrofitting would in fact be higher than outright replacement.

          Trilliant has found that retrofitting costs between $20-25/meter but that small saving is quickly spent with the operational cost of at least two site visits to remove and replace the meter plus the workshop cost for retrofitting.

          Investing in control over a distribution grid can include three key factors: grid reliability improvements; automation of consumer control over power use; extending the effective lifetime of the assets distributing the power, said Miller.  Trilliant begins the discussion with a utility client by establishing what the relative importance of each of those objectives is to the utility.

          Trilliant then recommends a plan that relies on the creation of a multi-tier, interdependent monitoring and management network that lets the utility synchronize the capital investment with the operational changes that will be needed to maximize the value of that investment.

          Miller described the integrated solution as three interdependent networks.  The core network is a broadband Ethernet ring covering the critical distribution assets and connection points between the substation and meter including re-closers, voltage regulators, capacitor banks and feeders.

          The next tier is the neighborhood area network -- commonly called a NAN.  Its purpose is both asset monitoring and demand response.  As such, it covers the meters and any devices installed between a transformer and meters.

          If the utility does not want to immediately implement DR, it does not need to replace all of its meters since this tier needs only to have sufficient devices to support a mesh network.

          The final tier is the home area network (HAN), the communication layer for smart thermostats, in-home displays and consumer installed load control devices.

            Selling to the PUC

Not building a new power plant, the cost of which dwarfs the cost of enabling the exchange of data for grid health and power consumption, is generally how the PUC sees a smart grid rate case.  Many PUCs in fact require that the utility show how it is maximizing the use of existing bulk power supply sources before considering new generation.

          Thus, regulators have not, in Miller's experience, been reluctant to approve the project as long as the utility shows how it will use the data to benefit from the smart grid to align demand and supply from their existing bulk power generation and or purchase agreements.

          Aligning supply and demand requires a significant change in a utility's business processes.  Without the ability to gather much data from the grid, utilities have traditionally taken a conservative approach -- replacing assets on a timed schedule or incurring reliability problems by running assets to failure.  In the former instance the capital cost of operations was higher since some assets were replaced before the end of their useful life.  In the latter, operating costs were higher due to a need to maintain a larger staff to react to the failure event.

          Miller used the example of transformer temperature monitoring to illustrate how the utility would apply information provided by a Trilliant network.  A transformer's useful lifetime is determined in large part by the frequency and duration of peak thermal times.  Trilliant lets the utility monitor each transformer's peak load, its duration and frequency and its temperature during peaks.  Thus the utility can understand peer-asset what the useful lifetime will be.

          Knowing in real time when a transformer's temperature is shortening its effective life would alert an energy management system such as SCADA to activate voltage controllers that would regulate the supply and thus the damaging environment for the transformer.

          Miller believes meter replacement accounts for 5-15% of the total cost of a typical smart grid project -- but if meter replacement is an issue, he urges utilities to start with distribution grid health monitoring and defer smart metering until the utility and its PUC are prepared to consider consumer migration to time-of-use pricing.