How Big is the Smart Grid?

Posted on September 10th, 2009 by admin in Smart Grid

To use a favorite cliché, size does matter, for some interesting reasons.

The Original Grid (OG), in all its aging glory, consists of a remarkably large number of physical objects. These physical objects in the OG have been counted to a reasonable extent. However, it’s very important to plan for the number of elements in the Smart Grid (SG) for two reasons, one of which is surprising.

First. the SG will overlay the OG with more physical objects, almost all of which will be sensors/actuators, and network communication system elements such as WiFi or cellular modems. All of these physical and logical entities require planning and operational management by the utility firms. The physical stuff in the OG is mostly large, and there is a lot of it. The electronic network and sensors of the SG are much smaller and there’s even more of it. Second, the surprising and very important point is that all of these sensors will require a unique logical address in the Internet Protocol (IP), and these IP addresses are already rapidly being consumed by the non-grid world.

We will explain the very real scope of the problem and what can be done to solve it.

Scope of the Grids

Let’s see the scope of these two grids by examining the table below, and then consider the address limits again.

Scoping IP Addresses for Smart Grid

Scoping IP Addresses for Smart Grid

The table is organized to roughly enumerate the numbers of large objects in the OG and then to estimate the numbers of logical elements required to fully monitor and control the OG using the SG overlay. So we list the real objects and also “attach” SG sensors, add it all up and see where the numbers lead us.

First, according to Erik Udstuen, a general manager at GE Fanuc Intelligent Platforms, for a typical OG utility with 5M meters, there would be only 30,000 sensors for its whole OG. We note none of the OG meters has a sensor, so the sensors are for conventional utility components, as implemented in SCADA systems (“Supervisory Control and Distribution Automation”), by example. Of course, the number of sensors will rapidly grow with the addition of SG functions and Smart Meters.

Next, let’s approzimate the overall number of big objects in the OG in the USA. According to the list shown above, the entire OG has at least 600k large objects. If sensors are attached as indicated in the table, the whole SG could operate an additional 2.8M sensors each with a unique IP address.

Finally, focus on the edge of the OG, where customers attach things to power. Our estimate that there will be one Smart Meter per household and 100M households is straightforward. Then we assume one sensor per Smart Meter. We also make some crude approximations to the potential numbers of Appliances and Plug-In Vehicles. These are all low. For example, we assume only 4 appliances per household to get 400M smart appliances, each with 1 sensor.

If we add the number of envisioned sensors for the OG with SG overlay and the edge of OG sensors, we get about 505M sensors, which is a low estimate of what is likely to evolve. Keep in mind these numbers apply to USA alone. The rest of the developed world adds it’s own similar growth.

Why the Numbers Matter Now

Here’s why the numbers are so important: the world wide internet is literally running out of distinct addresses for connected objects, even without consideration of the Smart Grid. These “IP addresses” must uniquely identify every physical object that communicates on the internet, such as PCs, routers, and cell phones, cable modems, set top boxes. From a grid viewpoint, now include appliances, like a coffee maker, attached to the OG but being programmable through the web to have our coffee ready upon arrival home.

Given the relentless growth of the web, the “Internet of Things” is literally limited by a lack of sufficient IP addresses for world wide growth.

Using the currently installed internet protocol IPv4, only about 4B objects can attach to the internet. This is fewer than the expected number of cell phones and PCs to attach, so the scope of the problem really hits home.

Help is on the Way!

Actually, help has already arrived. Network engineers have known for many years the currently installed internet protocol, IPv4 would finally be unable to address all the attachments desired for the internet, even without the SG. So they developed a new protocol and tested it. The current version, IPv6 is ready for installation and is being selectively deployed commercially. IPv6 offers many important improvements over IPv4, most important for us being the number of address bits.

See the summary table:

Comparing Protocol Address Spaces

Comparing Protocol Address Spaces

The address space for IPv6 is truly stunning. Using IPv6, every grain of sand on earth can be addressed!

The catch is this: any small to medium firm will have to spend about $2M to install new equipment and software to process IPv6. This will be a process not familiar with utility IT staffs. So the reasons to use IPv6 by a specific utility firm, from its parochial point of view, need be substantial. It’s like the Social Security dilemma: the government knows it will run out of money, but refuses to deal with the problem in advance of the pain.

The electric utility industry won’t need to deal with millions of installed SG devices in the next 5 years, but it certainly will over time. But if they use IPv4 now and transition later, the project will be much more complex. Meanwhile, the industry will have foregone many other communication and networking advantages from IPv6 (such as very strong security).

We will follow this issue with interest.

Leave a Reply

More News