Smart meters are a cornerstone of the Smart Grid, says the U.S. Department of Energy’s 2009 Smart Grid System Report.1 They provide electricity consumption measurements and time-of-use information to utilities and consumers. They also receive market signals, help coordinate consumer equipment, and adjust consumption.
Failure to widely and successfully deploy such advanced metering technologies and related services to every home will result in a widening gap between energy supply and demand. It also will undermine essential Smart Grid capabilities and, by extension, the good of the public.
On the other hand, according to Edison Electric Institute president Tom Kuhn, successful deployment of an advanced metering infrastructure requires the adoption of new communications, encryption, and semiconductor solutions in the development of a new class of highly reliable, secure, and cost-effective smart meters.2 One of the most significant challenges facing Smart Grid development is the cost, with estimates for just the electric utility advanced metering capability ranging up to $27 billion.
Another key if not more important component of the Smart Grid is the energy consumer. The challenge here has to do with the transformation of consumption habits and attitudes of consumers effectively to “choose” energy efficiency and facilitate “choice” by improving the visibility and communication of critical information.
The good news is that never before has society been so actively concerned about energy efficiency, nor has there ever been so much industry participation in the development of energy-efficient products, as well as consumer-oriented energy-management and automation products such as smart devices, Internet portals, and services alike. So, the trend and momentum is here. Yet only 6% of the 1.6 billion meters installed worldwide are smart meters with two-way communications capability.
For instance, “smart devices” can facilitate reductions or shifts in energy consumption when customers are motivated by dynamic pricing signals or other economic incentives. The flow of information such as pricing signals does not necessarily have to be through the utility-owned smart meter. Alternatively, it can be accessed via consumer Internet appliances such as a broadband router/gateway. For utilities and their customers, each method has pros and cons when it comes to information security (or privacy), grid security, interoperability between various devices, and commercial availability.
The bottom line, though, is that both techniques can work individually or in combination. Recent trials focused on gauging the effectiveness of smart appliances and devices integrated with smart meters and dynamic pricing schemes have shown promising results, achieving significant decreases in peak demand and by extension reduced cost of energy.
However, most of this preliminary work has been based on simple decision-making algorithms to effectively control the on/off states of appliances as a function of pricing signals received. It also has been based on more granular schemes such as sub-metering at the device level, where customers are empowered to make more dynamic and well-informed decisions about their energy usage. The work is still in the early stages of development, largely due to a lack of cost-effective measurement devices for consumer appliances.
Today, such devices are available. Major manufacturers are developing smart appliances, propelled not just by the federal stimulus funding plan but also by the fundamental realization that today’s consumer wants to participate and requires the visibility to do it. From the utilities’ standpoint, with such visibility, consumers can develop a better appreciation and understanding of the true cost of operating appliances and adapt their usage habits. Or, they can rely on more sophisticated applications such as Web-based services launched by Microsoft and Google to ultimately control and reduce energy consumption and associated cost, which is a win-win outcome for absolutely everyone.
The designers of these complex products need to deal with issues such as reliability, cost, and scalability. That’s why semiconductor products specialized for energy measurement are becoming key enablers not just for first-time integrators of energy measurement in appliances, but also for well-established smart-meter developers.
THE ROAD AHEAD
Even with $4 billion in stimulus funding supporting projects in 2010, deployment costs still will be a barrier for small to mid-sized utilities, considering that the increasing costs of smart-meter hardware are amplifying the historically challenging return-on-investment scenarios.
These costs are primarily due to a complex set of communications-related feature sets such as load profile storage, encryption, multiple secure two-way communications interfaces to the grid and in-home appliances, and displays as well as remote disconnect relays and additional printed-circuit boards (PCBs). Also, don’t forget about the public perception of incompetence or worse when rollouts do not go smoothly (see “Smart Meter Rollouts And Standards Stir Controversy”).
There is an additional cost associated with an inherent fear of falling short of developing sufficiently programmable, upgradable, and future-proof solutions to address still-evolving Smart Grid applications and security and interoperability standards. Addressing these concerns can drive up the cost as more memory and communications processing than is practically usable in the short term is embedded in the design just to be safe.
No one wants to conduct large-scale rollouts only to find out later that a lack of careful consideration for the evolutionary nature of the Smart Grid can shorten the product’s life cycles. Hasty rollouts also can result in premature obsolescence and the replacement of the units deployed, which would obviously add significantly to the operating cost of the utility’s deployed infrastructure.
But maybe a bigger challenge is the decreased service life and reliability of these complex electronic instruments. The higher number of components and interconnections may result in premature replacement compared to the older, simpler units of two decades back.
SEMICONDUCTOR DESIGN CONSIDERATIONS
The history of the adoption of solid-state technologies in high-volume residential meters tells an interesting story about product lifetimes. It all started a little more than 10 years ago. Today, nearly 85% of the 120 million or so new meters being shipped globally are solid-state. The rest use the old mechanical Ferraris disk.
The first incarnations of mass-market solid-state meters were born out of the necessity to produce large numbers of basic kilowatt-hour meters in rapidly developing markets such as China and India during the second half of the 1990s. At that time, only a few global meter exporters had developed the manufacturing expertise and capacity to produce revenue meters in large volumes, and fewer yet engaged in technology transfer licensing agreements to developing nations for local domestic production.
I visited one such facility at ECE Industries Limited’s facility in Hyderabad, India, in 2004. It once was a thriving mechanical meter manufacturing plant operating under a technology license from a German metering company with a peak production throughput of around 8 million meters a year. This major industrial manufacturing site boasted a beautiful garden oasis, which incidentally was made eternal in many major Bollywood movies of the era. Later during the empty facility tour, my host, an ex-foreman at the plant, explained sadly that operations had ultimately halted due to the rapid emergence of lower-cost solid-state meters with more advanced anti-tamper features.
In China at roughly the same time, first-generation surface-mount metrology ASICs integrated analog-to-digital conversion along with multiplying voltage and current to produce digital pulse outputs to feed an odometer-style kilowatt-hour counter. This development played a transformational role in the creation of a new breed of energy-meter cottage industry in these regions during a time of high demand for electrification and basic revenue metering. A number of competing local OEMs leveraging low-cost PCB manufacturing techniques used them.
The electric power industry’s move away from traditional rate-based regulation toward increased competition in a deregulated marketplace primarily drove the second transformation in electricity meter architectures. As utilities started to require meters to incorporate time-of-use billing and automatic meter reading (AMR), microcontrollers, radios, LCDs, and real-time-clock (RTC) devices were added to metrology ASICs (also called analog front ends, or AFEs), making multi-chip arrangements the standard approach in residential-meter design.
The specific application and metrology expertise, generally higher design skill levels required, changing utility requirements, and globalization, coupled with integration challenges associated with complex electronic systems, helped differentiate higher-caliber meter manufacturers as leading global suppliers and reduced the total number of manufacturers significantly. The increasing importance of the communications infrastructure in metering further helped this trend. Of course, not all suppliers in all regions of the world progressed through these transition cycles concurrently.
ADVENT OF THE SMART GRID
The cycle continues as IC suppliers recognized an opportunity to address the increasing need of meter OEMs to produce large volumes of low-cost, complex, and reliable products with shorter time-to-market requirements. These IC suppliers introduced highly integrated application-specific system-on-a-chip (SoC) products around 2003.
The integration of common functional blocks such as AFEs, RTCs, LCD drivers, and MCUs coupled with refinements in reliability, flexibility, and scalability helped OEMs worldwide develop reliable and low-cost meters much faster than before, using only a single SoC device, and to add communications modules and devices to the product to address region-specific utility infrastructure requirements quickly.
The advent of the Smart Grid concept and the utilities’ increasing requirements for more and more memory, security, and communications in Smart Grid applications further justified the adoption of highly integrated devices with high digital-circuit content, as semiconductor process technology advancements continue to follow Moore’s law. This represents a classic example of the value of semiconductor technology and its transformative contribution to continuing cycles in the evolution of the metering industry.
THE WHOLE IS LARGER THAN ITS PARTS
IC makers can position themselves in terms of partitioning and integration. The differences can be subtle, but they are important. At first glance, Teridian’s metering SoCs appear to use classic block-level integration, including the essential meter. This quick copy-and-paste approach is similar to that employed by discrete analog-to-digital converter (ADC) or microcontroller vendors as they try their hands at SoC designs of a rather generic nature. In that scenario, application-specificity would materialize in the form of reference design and associated code, rather than in the silicon.
Although this level of integration can improve cost and reduce component count, it can fall seriously short of delivering on the complex set of market requirements it targets. The differences with Teridian SoCs include an exclusive focus on energy metering and measurement based on more than 20 years of experience.
While that sounds like marketing-speak, Teridian’s patented “single-converter technology” employs a single 21-bit second-order delta-sigma ADC with up to seven multiplexed analog inputs and a programmable compute engine (CE). This addresses a number of problems associated with multi-channel data-acquisition schemes, such as noise and imbalance.
The ADC also achieves better than 2000:1 dynamic range across the industrial temperature range, which is the best in the industry. The 32-bit CE is a signal processor that hardwires the standard measurement functions, but is upgradable with new custom functions, which would have given that plant in India an edge when it came under fire from lower-priced basic-function competitors.
The fourth generation of Teridian devices replaces current transformers and copper feed-through wiring with low-cost current shunts using a proprietary isolation technology (see the figure). Teridian’s latest-generation metering ICs integrate metering and interface functions. Architectural differentiation includes multiplexing the ADC and a dedicated compute engine for measurement functions.
Most discrete and SoC-based metering solutions have, until now, respected a classical separation between metrology and communications functions. This system partitioning is rooted in legacy AMR industry practices. In that context, many communications infrastructure suppliers operated as standalone business entities and offered connectivity to any meter by means of proprietary communications modules.
However, this practice is beginning to change in large utilities and regions where a common communications technology is required and products developed can reduce hardware cost by integrating functions from either side of the partition, either onto single-board designs or onto more highly integrated SoCs. This reduces the cost and increases the reliability of the smart meter.
Thus we see today’s Smart Grid dynamics catalyzing innovation in design approaches across various industry segments and technology boundaries, particularly in terms of a convergence of metrology and communications down to the device level. By the end of 2009, nearly 50% of smart meters were based on single-board designs in which SoCs have replaced a number of redundant microcontrollers, memory devices, and interfaces. Additionally, in many designs, low-end transceivers have replaced RF/power-line carrier (PLC) modems, leaving the bulk of protocol processing to the SoC as well.
As noted above, both RF and PLC technologies are being exploited for the Smart Grid. In the U.S., solutions for in-home connectivity revolve primary around ZigBee, Homeplug, and Wi-Fi (IEEE 802.15.11) technologies. Grid connectivity (utility networks) may be based on any number of non-standard solutions that do not necessarily need physical-layer (PHY) interoperability. In parallel, there are efforts to define a standard interface between the communication module and devices such as smart meters and appliances. (For more, go to www.usnap.org.)
What is essential in the utilities network is security and higher-layer interoperability as defined in ANSI C12.19 to ensure the integration of various Smart Grid subsystems as a cohesive and secure network of networks. Other technologies are being developed such as SUN PHY (amendment to IEEE 802.15.4) and IEEE 1901, which aims to unify PLC technologies (http://grouper.ieee.org/groups/1901/). These developments make it hard to predict which technology, if any, will dominate in the future. But what is certain is that in the home-area network (HAN), consumers will have more than one option in the market to choose from, and solutions to bridge one communication link to another will be commercially available.
As for security, 128-bit encryption schemes now acceptable in HANs (ZigBee, Homeplug) will likely give way to 256-bit schemes in the utilities’ networks, such as ZigBee Smart Energy 2 and 802.16 WiMAX. Most AMI rollouts require data-transfer rates of 50 to 100 kbits/s. For a list of standards being developed for the U.S. Smart Grid, go to http://smartgrid.ieee.org/standards/ieee-smartgrid-standards-in-development.
1. U.S. Department of Energy 2009 Smart Grid System Report, www.oe.energy.gov/SGSRMain_090707_lowres.pdf
2. “Legislative Proposals to Reduce Greenhouse Gas Emissions,” http://energycommerce.house.gov/images/stories/Documents/Hearings/PDF/Testimony/EAQ/110-eaq-hrg.061908.Kuhn-testimony.pdf