Defining terms is a natural place to begin any endeavor. This post presents my take on the Smart Grid, ideally with just the right balance of authority and controversy to kick off my blog with a lively discussion. Since it is longer than I would like, I'll summarize to save you the trouble of reading it: Because the smart grid doesn't exist, 'Smart Grid' is a key marketing term needed to catalyze its development. As the actual physical entity takes shape, I claim that there will be three progressively increasing levels of sophistication and capability. The simplest is Demand Response, already in place to a large extent. At mid-complexity is 'Smart Grid + Storage' - limited to a few pilot projects, notably Smart Grid City in Boulder CO. At its most developed is 'Smart Grid Enabled Distributed Generation', where 'Distributed' is used in the conventional sense of closer to the load, but also to refer to Transmission Line Superhighways and their associated control that enable generation a much further distance from the load. Of course, this piece is based on many sources, a number of which are in my bookmarks.
The Smart Grid is both critical enabling technology for re-energizing the world and a somewhat empty marketing phrase. Good marketing is essential for realizing a complex concept, so lets take apart 'smart grid' the marketing phrase. Calling it a Smart Grid of course implies that the present grid is dumb. Like all good marketing, this overstates the case. Electricity delivery is far more reliable than internet or mobile phone service, two darlings of the High Tech era. To a large extent this reliability is the result of an extremely well engineered, if un-sexy, control system. The grid is already pretty darn smart. However the end customer, either commercial or residential, is the recipient of electricity devoid of production cost information. The consumer uses the last watt that tips millions of people into a blackout no differently and at no different price than they use power from the utility's excess capacity that effectively has no incremental cost. This permits an extreme level of inefficiency, requiring vastly more generation capacity than a 'smarter' scenario would enable. This 'smart layer' allows better pricing, forecasting, control and in the end more efficient, greener and more cost-effective energy for all players. Since the completed smart grid is an immense and extraordinarily complex infrastructure project that is still largely in the concept stage, a good marketing campaign is critical. People from the President of the United States to the most technologically unsophisticated of end consumer need to be aware of the smart grid and need to "want one of those" for the complete vision to be executed. We knew the internet was hot when we were poking along on 4 KBPS dial-up, sending occasional plain text emails. Ten years later, we stream video on demand and happily pay our $60/month. The same progression is required with smart grid.
If the Smart Grid "lights don't come on until the whole thing is wired", there will be neither the patience nor the capital to complete the project. For adoption to occur there needs to be a valuable technology solution at incremental levels of sophistication and build out.There is some debate as to what those levels are and I offer my opinion. From a technology standpoint, Smart Grid is motivated by a combination of three factors. An inefficient marketplace, a very large step function in marginal cost due to the relative cost of peak load to base load, and finally, the poor base load characteristics for most renewables. The simplest Smart Grid strategy, demand response, focuses on improving market efficiency. At a mid-level of complexity, storage allows peak demand to be met at lower cost. At the most sophisticated level, a large-scale physical infrastructure allows far more efficient incorporation of renewables by enabling generation that is both much further from the load and much closer. Let's look at each of these scenarios.
There is near universal agreement that the present US electricity market is inefficient, meaning that the cost of providing energy is not accurately reflected in its price (energy prices are fixed, energy costs are highly variable). Economists would say that there is a sharp step function in the marginal cost of electricity when supply moves from base load to peak load. Since the consumer pays a constant rate, there is currently no incentive to reduce peak load consumption. "Base Load", somewhat confusingly, is used to refer to generation (not load) resources that are used to meet the minimum energy requirement of the demand cycle. In warm industrial climates, the demand cycle is roughly a year, with minimum at winter nights and maximum at hot weekdays. A key feature of base load is that it is essentially fixed output. Turning base load on and off and even varying it is extremely costly. As a result, when consumption is below base load levels, marginal cost is essentially zero or negative. If not all of the produced power is consumed, it must be shunted and re-routed, at some cost. There is no reason to reduce consumption below base load levels. The situation changes dramatically when consumption rises above base load levels. "Peak Load" refers to generation capability that can be brought on line to meet varying requirements. Peak load generation typically has extremely high fixed and variable costs. The variable cost of operation is higher due to the lower efficiency of variable generation systems, but the fixed costs are truly staggering. Millions of dollars of inventory sit idle and depreciate to meet that one hottest day of the year that drives peak demand. Forecasting future demand presents the most extreme marginal cost scenario. Utilities are often in a position of anticipating that an additional power plant, on the order of a billion dollar investment will be needed to meet as little as hours a year of peak demand anticipated at some point in the future. Despite the varying cost, price is fixed. Customers pay the same whether the utility has tremendous over-capacity and would be producing power even if it weren't consumed, or whether supply is so strained that the utility needs to bring non-compliant generation needs on line, at the cost of massive pollution fines.
At its most basic, the Smart Grid brings price transparency through demand response. The most compelling business case for demand response centers on peak load forecasting. Peak electricity demand is growing faster than supply. Most utilities forecast needing new billion-dollar generation facilities to meet future peak demand. Many have already begun the financing and permitting process. As discussed previously, peak demand is extremely rare and surprisingly easy to reduce. With the right financial incentives, it is easy to secure contracts to reduce demand. The simplest scenario involves raising office temperature a few degrees, only when the need is critical. Typically a third party acts as a middleman and arranges for demand reduction that keeps total load within acceptable parameters. The Utility communicates a request for a certain amount of demand reduction (the company EnerNOC coined the great term 'negawatts'). The Demand Response provider communicates the request to participating consumers. The participating facility raises its thermostat a few degrees. Office workers roll up their shirt sleeves and get paid a bonus for their trouble. The utility is happy to pay for this service rather than for new generating capacity. The consumers have made their own ROI calculations and concluded that selling negawatts is worth more than using the power, and enough is left over for the Demand Response provider to profit as well.
Most observers agree that there is a limit to demand elasticity - there is only so much you can reduce consumption. In the end, businesses produce widgets, not negawatts, and people want to be comfortable. This serves as the lead-in to Smart Grid with Storage. Essentially, load varies on a very short and quite reliable schedule. In warm, industrialized areas peak load is weekday afternoons (when businesses are online and cooling demand is at a maximum), and nighttime loads are a small fraction of peak. The ability to store a modest amount of energy for a short period of time becomes extremely valuable. Typically, this problem separates into two businesses. Storage hardware is one, the other is the necessary 'smarts' to both monetize and control the energy flow (when to store energy, when to draw energy down). This added layer of complexity has the advantage of requiring less behavior modification from end consumers. Phoenix businesses can still keep their offices at 68F in July, they just cool them using ice stored overnight, rather than a compressor running at peak demand time.
The final level of sophistication involves expanding the generation range - generating both closer to and further from the load. It is the most ambitious from the capital investment perspective, but most experts feel that it is an essential component to any significant RPS (Renewable Portfolio Standard - a mandated percent of electicity from renewables). The need for expanded generation range is best understood by considering Wind and Solar renewables.
Efficient use of Wind requires longer distance transmission because the wind blows when it feels like it, and mostly where nobody lives. The best US wind resources are in the sparsely populated Great Planes, as well as far offshore. Furthermore, wind is a highly variable resource, but the larger the areas of aggregation, the smoother the average production. To illustrate: The wind at the McCamey, TX 4 GW facility in the Texas Panhandle ranges from gale force to still, independent of the needs of consumers. By contrast, average wind across a thousand mile radius of the Great Plains is much steadier. These scenarios combine to require transmission over far greater distances than is currently built or maybe even feasible. In addition to this massive physical infrastructure, 'Smart Grid' inter connectivity over a far greater physical distance and number of components than is presently the state of the art is required to draw energy from where the wind is blowing, forecast where it needs to come from next, and deploy peak load resources as needed. The technical challenges to this wider area of control are substantial, and this represents the boldest, longest term vision of Smart Grid.
Efficient use of solar power requires nearly the opposite configuration. Unlike wind, Solar PV's (PhotoVoltaic - Solar Cells) greatest asset is that in warm climates, it tracks peak load (Peak cooling is needed when the sun shines brightest, providing the most Solar PV electricity). This advantage overcomes its (so far) higher cost, but only if solar power is efficiently distributed near the demand. The peak load advantage of Solar is attenuated if not lost when Solar PV is generated at the utility scale, then transmitted with the usual inefficiencies to its destination. Solar PV is remarkably well suited to the suburban sprawl of much of America's sunbelt. Between single story big-box stores and their acres of parking lots, gigawatts of generation await siting. The best use of the PV panels shading the Costco or Walmart parking lot is at that store. Achieving this goal requires significant hardware infrastructure in addition to PV generation, as well as a major 'Smart Grid' component to enable two-way flow of power. Simply stated, the store will still need full AC 30 minutes after the clouds roll in and cut PV production 50% (external supply needs to be bought). Likewise, weekday mid-afternoon is a slow time for retailers, and the store may decide there is more money to be made selling power than keeping its 3 customers at a comfortable temperature. (Excess capacity needs to be sold).
The above progression describes a massive change of a huge, tightly regulated, and fundamentally conservative industry. There are many who believe this Titanic is too big to turn. Others, me included, see utilities largely analogous to Ma Bell in the mid-80's - poised on the brink of a complete, end-to-end overhaul. In 1989, if I had an incredibly tech-savvy editor, I might have been able to connect to their mainframe via modem to write this article. Today, there is no reason I couldn't be posting from my BlackBerry overlooking Cape Cod bay (actually, there is, it's cold and rainy on the bay today). Technology moves at a phenomenal pace, and I'm betting my career that it's coming to Smart Grid next.
