Getting smart: first things first

Phil Carson | Jan 26, 2010


As I take the plunge into the world of intelligent utilities, I'm not able to instantly answer every question posed by readers. But as questions come my way, I'll get answers and share them.

And as I seek out thought leaders, I'm going to be agnostic on where they draw their paycheck -- if I find an expert at a utility, an industry association, an academic or (heaven forefend) a vendor, I'm going for it. I trust readers to de-spin and share their thoughts.

Over past weeks, I've heard variations on a theme: Why smart meters? Why now? What's the next step? What's the business case?

The shortest possible answers: Smart meters and advanced meter infrastructure (AMI) provide two-way communications serving both utility and consumer. They are the gateway that enables useful data for both parties and must be in place first. The next step for utilities is to use that data to craft business cases for time-of-use pricing and other demand management options. Pilots and full-scale deployments will yield data on consumer acceptance. 

I recently read a cogent piece on smart meters and home area networks by Skip Ashton, senior vice president for engineering at Ember Corp. and chair of the ZigBee Architecture Review Committee. Ember designs, manufactures and writes software for Zigbee-based chips that will provide the smarts in home area networks and in-home appliances. So, naturally, he has a vested interest in smart meter and AMI rollouts that will enable a market for his company's clients' products.

Nonetheless, when I reached him by phone yesterday I found him refreshingly frank and his take on this topic worthy of your time today.

Households consume one-fifth of the nation's energy each year, with 60 percent of that consumption in the form of electricity, according to Ashton. Half that residential use typically is consumed by heating and air-conditioning, thus the low-hanging fruit at the heart of early demand management programs.

The "peak energy demand dilemma," as Ashton put it, is that about 10 percent of electric generating capacity exists only to be utilized about 1 percent of the time.

Enter AMI and the American Recovery and Reinvestment Act (ARRA) stimulus funding. Utilities applying to their regulators for approval and to the U.S. Department of Energy (DOE) for funding originally cited savings due to automated meter reading, and remote connect/disconnect of service as their rationale. But they also realized that by reaching into homes and small businesses they could begin shaving peak load, buying time to figure out new generation sources and efficient capital allocation. AMI subsequently claimed a big chunk of those stimulus funds.

"The joke is that the stimulus funds were 'meter stimulus funds,'" Ashton told me.

In his view, smart meter deployment and its less visible, data backhaul component is just the first step in an evolving business case to realize the benefits of two-way communication between a utility and its customers. Some consumers are recoiling from the thought of a utility reaching into the home to adjust air-conditioning cycling, for example.

"Some people are petrified at this thought," Ashton said. "And utilities do not want to have that 'big brother' image."

Thus, in Ashton's view, opt-in programs for utility control over air conditioning, for instance, will maintain a real-time opt-out provision, should the consumer change his or her mind.

Why meters first? Smart meters provide the means, the gateway, for consumer data to reach the utility, and enable consumer-facing home area networks (HANs) and in-home devices such as energy information displays, according to Ashton.

HANs lag behind smart meter deployments by necessity, in Ashton's view. He points out that HANs largely remain in pilot study status right now. Wearing his ZigBee hat, Ashton said that ZigBee's wireless open standard allows for innovation on the widest possible scale. That grows the market for competition, which implies cost-effective devices.

As for the in-home energy use display, "we have a chicken-or-the-egg problem," according to Ashton. Vendors are building prototypes for trials on the scale of, say, 50,000 units. But an intelligent utility such as Southern California Edison, for instance, has 14 million customers.

"The market has yet to decide how this will roll out," Ashton said. "Meanwhile, there'll be some healthy churn and innovation."

What is the consumer reaction in trials of HANs and energy use displays and how does that relate to reality?

"That's the $64,000 question," Ashton said. "And it's an open question. What is live consumer behavior going to look like on a scale of, say, one million homes?"

Pilot programs this coming summer may provide answers. And though studies show that awareness of energy use in the home leads to behavioral changes and conservation of 5 percent to 10 percent of energy use, will simple displays reinforce that behavior over time? Or will consumers become jaded and backslide on wise use?

That's where time-of-use pricing and automated parameters for the operation of major appliances come into play, Ashton said. The interactivity of smart meters and AMI should provide further incentives to pay attention.

"Efficiency at scale," Ashton said. "That's what it comes down to."

This scenario of evolving business cases gives some people the willies. Let me know what you think is wrong with this picture.

Phil Carson     
Intelligent Utility Daily

Related Topics


Why Start With Meters


This was an excellent introduction to a broad subject.  When I am asked this question, I follow the thread of your answer, but also emphasize another point.  

If utilities are to improve how efficiently they manage their network, they must start by improving how they measure network operating conditions.  Historically, utilities measured customer usage once per month, which doesn't provide enough detail to address performance that changes in real-time.  

Currently, it is uncertain how rapidly the readings must be performed, but some guidelines are obvious:

1.  Readings must be synchronized.  Otherwise, we don't know how to couple together many different readings.  So each device must have a clock, and the clocks must be aligned with each other.

2.  We need data from more than just customer meters.  To optimize the whole network, we must first optimize each section of the network (however we define a section).  Thus, each component within the section must provide some information about its performance.  That is unrealistic for all elements in the section, such as the cables, but fortunately we can build models to describe such elements.  

3.  We need to measure more than just KWH.

4.  We can't ignore what is happening within the customer's premises, even when those premises are homes. 

As I said earlier, it is uncertain how much data we must collect, but the obvious place is (1) with customer meters and (2) in installing and reading other metering elements within the grid.  We must be able to read and write to each of these devices so we can change our measuring rules as we learn more about our networks.  We also must collect all of this information into common databases to allow us to revisit the data in the future and to make it available to other utility organizations so they can also use it to make smart decisions.

We must also not forget that our industry will need more robust software tools.  The data itself is useless unless we can digest it, model it and use it to change how we operate our business.

Ultimately, the goal is to build devices and systems that can dynamically take action.  Today, far too much work requires human intervention.  However, we can't hope to have a network that dynamically improves its efficiency and automatically reacts to user changes, weather changes, etc. until the systems that run the network have the data they need to make the decisions.

So we start with meters.



Howard A. Scott, Ph.D.



Thanks for your points, Howard.

If you're able and willing, could you extrapolate a bit on several statements you made?

"Each device must have a clock, and the clocks must be aligned with each other." Are you referring to devices connected to the HAN or in the grid? Or, which devices?

"We need to measure more than just KWH." Meaning, at the home? Could you mention other data points needed?

"We need data from more than just customer meters." I assume you mean that the utility will need data from each connected device within the HAN? If so, can you help us with the relevance of that granular data?

"We also must collect all of this information into common databases to allow us to revisit the data in the future and to make it available to other utility organizations so they can also use it to make smart decisions." I can hear privacy advocates screaming... What's required to ensure that this data is anonymous?

"Today, far too much work requires human intervention." Are you referring to human judgment calls required to balance load and generation?

I'd love to hear more, when you're able.


Phil Carson


First Things Indeed

Meters that can measure consumption on smaller intervals and transmit this information to consumers in near-real time are clearly necessary.  Unfortunately, I think there are still some utilities (notably in California) who are perfectly willing to have customers pay for more capable meters, but then they want to turn information into a profit center by requiring customers to pay again for timely access to the information these meters provide.  That's just unacceptable.  Compared with the cost of the meters, the cost of providing timely access is pretty small.  without timely access to information, most of the benefits that might otherwise accrue to customers is lost.

It's not clear why utilities have to reach into homes and businesses at all.  As your colleague Christine Richards points out in a companion article, "There's been talk, and some action, on how to automate homes based on consumer preferences, and this will continue to become more important as smart grid concepts reach the broader population."  It's not that hard.  Send a price that a device equipped with suitable intelligence can listen to. For legacy devices, an intelligent retrofit should cost about the same as the switches do.

Existing residential air conditioning controls are possible only because they can easily be switched off and on without affecting comfort.  Add thermal storage or move up the chain to commercial premises and switching no longer works. Intelligence driven by a price signal is a better solution.

The more interesting problem is how to deal with electric vehicles.  Any solution that has utilities controlling the rate and timing of EV charging is likely to fail for two reasons.  First, the required amount of computing power to control individual devices from a central location is neither cost-effective nor practical.  Second, no centralized control algorithm will be able to capture all of the individual preferences of individual consumers, even if we're talking about only a few thousand vehicles.

We use prices to coordinate the actions of market actors in just about every sector of our market economy.  The most notable exception, and perhaps the most important one, is electricity.

good point

Thanks Jack,

You wrote: "It's not clear why utilities have to reach into homes and businesses at all." 

I'm hoping Howard Scott, who also commented on my piece, will bestow more detail on us and address your question.

As for intelligent devices with automated parameters set to respond to price signals, that sounds right. As one reader told me earlier today, no one wants to get home from work and, amidst family and chores, be an energy day trader. Nor manually turn on the washing machine at 1 am. I should think that piece will easily be solved.


Regards, Phil Carson