Tech View: Laying the Groundwork for the Coming Smart Grid

by: Don Bowen, February 16, 2010


Today’s electric grid is dated, overburdened, and lacks the ability to readily accommodate revolutionary technologies such as electric vehicles and wind farm installations. Customers are charged one flat rate, even though the actual cost of energy can fluctuate wildly during the day. Excess generation capacity goes idle in one region as other parts of the country suffer shortages and brown-outs. Usage surges at peak periods, stressing capacity, with little opportunity to shift demand to non-peak periods.

Utilities are often without comprehensive real-time information or the ability to act on it. They can’t remotely read meters, monitor power lines, easily shuttle excess power from one region to another, or consistently predict failures or shortages in time to re-route power. They communicate one-way with customers only through monthly printed bills.

What’s missing are built-in sensors and controls and a modern telecommunications network to relay information and commands. But this is about to change with a new generation of monitoring and control devices—compact, low-power, and capable of two-way communication. Supported by $4.5 billion in grants included in last February’s stimulus package, 40 million devices are about to be installed throughout the grid.

. . . utilities could simply take advantage of the vast existing infrastructure built and managed by AT&T and other providers.

As a result, utilities will know more about what is going on. Sensors on power lines will quickly detect breaks and monitor temperatures to know when lines are overheating and power needs to be rerouted. Sensors at transformers and capacitor banks will warn when failures have occurred or may be imminent. Two-way communications means utilities can relay demand-response commands to customers to promote reductions in energy consumption. They’ll be able to remotely re-route power if needed. Many fixes may need no human intervention since smart sensors can direct machines to automatically respond to certain conditions.

A modern two-way communications network connecting every element of the grid is needed to carry sensor data and provide connectivity and control. Rather than building new infrastructure for all this—an expensive proposition—utilities could simply take advantage of the vast existing infrastructure built and managed by AT&T and other providers. (AT&T alone will spend over $17 billion on capital outlays in 2010.)  Utilities could contract for secure private dedicated space on the AT&T network, or simply purchase existing wireline and wireless voice and data services as needed.

AT&T already provides cellular connectivity to many smart meters in the field. Working with partners, including SmartSynch, a supplier of smart meter systems, and utilities like Texas-New Mexico Power Co., AT&T is using its advanced cellular network to connect directly (and indirectly via mesh collector systems) with tens of thousands of smart meters installed in households around the country.


Smart grid in the home

The home is where people may be affected most by the smart grid. Meters made smart with embedded intelligence will relay detailed usage and cost information so homeowners for the first time will know what they use, when they use it, and what it costs. Smart meters can be programmed for variable pricing, so utilities can change rates, perhaps in hourly, 15-minute, or even 5-minute intervals, charging more during peak periods and less other times; consumers will have an incentive to run appliances during non-peak times. Smart appliances can even respond automatically. Direct communication between home and utility means a new level of transparency and cooperation to reduce peak loads, saving money and the environment.

Smart Grid Home Area Network

And reducing demand during peak times is a key goal since it is the power demands during peak periods that are driving the calls for new capacity; if demand can be distributed throughout the day, building new generating plants can be delayed, perhaps even indefinitely since renewable power sources are coming online.

Today’s grid is largely supplied by nuclear, fossil-burning, and hydro-electric generating sources that put out continuous, predictable flows of electricity. In contrast, renewable sources such as solar and wind generate output that can vary wildly, and cannot always be easily accommodated. The old grid has essentially no way to store excess power when these sources are abundant or reclaim it when such sources are offline. The inclusion of battery storage, such as a uninterruptible power supply (UPS) for a commercial building, could facilitate temporary absorption of such excess power and the return of power back to the grid during times when it is needed.

Electric vehicles are of particular interest for several reasons. In numbers, they place a potentially overwhelming load on the current grid (fast-charging an electric vehicle can take as much power as several large homes with every appliance on). A handful of such vehicles charging simultaneously in one cul-de-sac could easily draw more power than the local distribution transformer can accommodate. Consequently, ways to facilitate and control charging operations are being explored, including time shifting of charging operations to the late evening when energy demand is normally lower. But it’s not all bad news. Electric car batteries might potentially serve as vast reservoirs of energy storage, absorbing it when it’s plentiful and giving back to the grid when it’s needed—an advantageous trait when managing renewable energy sources on the grid.

Homeowners for the first time will know what they use, when they use it, and what it costs.

Appliances can also be smart. Manufacturers are ready to begin inserting chips to collect and report usage data so homeowners will know the power consumption of each device. (Owners of plasma TVs may be in for a surprise.)  And dishwashers and any other device will be programmable to run only when electricity is cheapest. The smart meter, when alerted by the utility, might send a message to the refrigerator for it to go into energy-saving mode, delaying freezer defrost cycles.

All these smart devices will need a way to talk to one other. In the home, communication will likely take place over a home area network (HAN) in which each appliance is a node, and the network is connected to an energy controller.


Smart devices will communicate over a home area network. The most reliable solution may be a mesh network.


AT&T Research is actively investigating what the HAN topology should look like. The most reliable solution may be a mesh network that incorporates machine-to-machine (M2M) communications, allowing devices to communicate automatically, with minimal human intervention. Devices might periodically report information such as thermostat readings every 15 minutes or so or issue an alert, such as when a freezer initiates a defrost cycle.

Scientists at AT&T Research are looking at the competing protocols for the HAN. Current candidates include Power Line Carrier (PLC) solutions such as IEEE P1901 and ITU PLC uses existing electrical wires to carry information and commands. Its big advantage is that it goes wherever the power lines go; however, there are competing PLC protocols and it’s not yet clear which one might prevail.

A possible wireless solution lies with ZigBee, a low-speed wireless protocol based on the IEEE 802.15.4 standard. Unlike its better known “big brother” Wi-Fi, ZigBee isn’t meant for transmitting voice or video. But it’s perfectly suited for low-bandwidth, ultra-low power M2M data exchanges like the smart grid will commonly use, and is designed to facilitate mesh networks.

Mesh radio networks can enhance reliability; if one link fails, each device can communicate with another, either directly or indirectly. Devices far from the signal source, such as a washer/dryer in the basement, could receive a signal relayed from a device (such as a refrigerator) closer to the energy controller.

The smart meter . . . makes the HAN an access point on the grid, much as Wi-Fi makes the PC an end point on the Internet.

Plus ZigBee efficiently passes small amounts of data, such as a thermostat set point or pool pump status. ZigBee radios are simple, inexpensive, and require very little power, coming on only to transmit data or periodically received commands, and then going back to sleep. Wi-Fi by contrast is designed to carry streaming data such as voice and video, and implementations typically are more complex and require more power than those for ZigBee.

The HAN may connect to the smart grid via a networked smart meter or through a broadband interface (such as U-verse), allowing the transfer of information and (with the permission of customers) enabling the utility to reach into homes and turn down air conditioning or turn off pool pumps when the grid is under stress. Essentially, the smart meter (or smart energy controller) makes the HAN an access point on the grid, much as WiFi makes the PC an end point on the Internet.


Standards needed

Making everything work together requires standards, especially when it consists of 3,000 separate utilities used to operating on their own. The smart grid will be a system of systems, each needing to interoperate. Connecting the refrigerator, to the smart meter-access point, to the electrical substations and even power-generating systems, requires interoperability at each point. In the home, appliances from different manufacturers will need a common protocol to communicate on the HAN and with the smart meter.

The lack of standards is holding up smart appliances. Manufacturers are ready to begin embedding sensors and communications capabilities, but many are waiting until a common standard is identified; supporting multiple protocols is just too costly. Charged with encouraging interoperability and fostering additional smart grid standards is NIST (National Institute of Standards and Technology).  AT&T Research is working with NIST, associated standards boards, and manufacturer organizations to help create the reliable and secure network environment needed to unleash the full potential of the smart grid, with an emphasis on promoting decisions based on solid science. AT&T chairs an IEEE standards subcommittee for smart grid loads, and also contributes to ZigBee, IEEE P1901, and ITU proceedings, along with many others.

Research is also working with appliance makers to explore creating standard profiles for appliances according to the information required for each.


The complicated home environment

Standards are only one area for improvement; there are others. For one, the home may already be occupied by home automation systems that control devices and appliances. Security systems may already monitor doors and control lights. The PC may be hosting a home entertainment HAN that links the TV, remote, and even cell phone to a third-party’s gateway. Telehealth devices with sensors of their own may compete for control of the lights or other appliances. (For more about telehealth, see featured article Tech View: Telecommunications for Health and Body).

What will it take for these systems to interoperate? If the smart grid tells the lights to dim, but the security (or telehealth) system tells the lights to go on, who gets priority? Should the security system always get priority? Even if it’s unreliable? Mediation is clearly needed, but the question is how best to do it.

Interfaces will also need to be carefully considered. If consumers are to take control of electricity use, systems need to be simple. But simplicity is hard, especially when there are a lot of choices, and not everyone agrees what choices to give the consumer, and which are the most important.

Data ownership and security are two other important issues.

How much control should utilities have to tap into consumers’ power usage and make it more efficient? Some homes and businesses might want a utility to remotely control the thermostat or air conditioner in exchange for cheaper rates, but not everyone. Most plans envision the ability for customers to “opt-out”, though they may eventually pay for this privilege.

Data ownership and security are two other important issues. Consumer data is important to utilities; they can use it to predict future consumption patterns and build a better grid. But the amount and detail of data collected with all the new sensors may reveal much personal information, more than some may be comfortable sharing. Safeguards will be needed.

The next step is for AT&T Research to simulate a home environment where multiple systems for security, home automation, telehealth, as well as the smart grid, play together, and test the interactions under a variety of conditions. Well-formed decisions can be made only by understanding how these systems actually interoperate.

But as the complex smart grid inches closer and 40 million or more devices begin relaying information and control signals both within and without homes, these decisions need to be made soon. Otherwise the electrical system and associated appliances could become less reliable, or even damaged, while consumers could be confused and angered, leading to a smart grid that fails to reach its full potential. AT&T scientists will continue working with fellow engineers, standards bodies, and regulatory agencies to ensure the new smart grid doesn’t leave us in the dark.

The reach of the smart grid . . .

Smart Grid Sidebar