The decentralization of the grid could lead to a future where localized microgrids power much of the country with clean energy.

Microgrids, though, are not synonymous with clean power. Many of them rely on fossil fuels like diesel generators or natural-gas-powered fuel cells and microturbines. Currently, those technologies tend to cost less and offer greater dispatchability than solar and batteries.

If the young markets for localized generation via microgrids are left to their own devices, we can expect a largely gas-driven microgrid future, University of California at San Diego researchers found in a new study published in the journal Energy Policy.

In many cases, these localized heating and generation assets can be more efficient than central generation transmitted over long distances. They don't suffer line losses, and they allow operational efficiencies for a campus or factory by combining heat and power from one generator rather than procuring them separately from the utility. That reduces systemwide greenhouse gas emissions, but still locks in gas generation on private land for the long haul.

"Decentralization could radically reduce customer energy costs, but without the right policy framework, it could create large numbers of small decentralized sources of gas-based carbon emissions that will be difficult to control if policymakers want to achieve deep cuts in greenhouse gas emissions," the authors write.

Making model microgrids

The realm of microgrid research is new enough that there isn't a ton of literature on it yet. As such, lead author Ryan Hanna and his peers had to build a model to test the performance of different configurations of microgrids over the course of a year.

They tested three types of microgrid: a commercial building like a big-box store, a critical asset like a hospital, and a larger campus like a university. They got building usage data from the Department of Energy and ran it through the Distributed Energy Resource Customer Adoption Model developed by Lawrence Berkeley National Lab, which calculates the cost and benefit of distributed technologies based on their capital expense, operation and maintenance costs, and ability to supply load (among many other variables).

The result is a new model to optimize the configuration of distributed assets that best fit the needs of the three archetypal microgrids. The model developed for this study assesses the economic value of microgrids for full-time self-supply; future iterations will layer in other uses, like resilience (more on this later).

When the researchers ran the study for a climate and power market based on Southern California, each of the three archetypal microgrids used solar PV firmed by batteries to supply peak load, but they relied on gas generation for baseload power. In the critical asset and campus microgrids, most of the gas generation comes in the form of combined heat and power plants, serving thermal needs as well as electrical.

The economics of Southern California's energy landscape drive this trend: plentiful sunlight, cheap gas and expensive grid electricity. The authors find a robust business case for self-supply microgrids as opposed to buying all power and heat from utilities. Only when they tested extremely high carbon pricing and gas prices did microgrid adoption become uneconomical.

"The case for microgrids is a case for natural gas fired locally that also generates significant thermal energy," the authors write. "Though smaller microgrids by contrast rely relatively less on gas and more on renewables, across all microgrids gas generators supply the majority of on-site electric and thermal energy."

That could be a problem for jurisdictions that are pursuing deep decarbonization over the long term.

California, for instance, has a legally mandated goal to reduce carbon emissions to 40 percent below 1990 levels by 2030, with the long-term goal of 80 percent below 1990 levels. A switch from central gas generation to localized gas generation would cut carbon emissions in the near term, but could lock in emissions in the long term that would be hard for the state to address.

That point is critical to keep in mind. Observers of the grid transition often mention decentralization and decarbonization in the same breath, but the two are not causally linked. It's the role of policymakers to structure the roll-out of microgrids -- through levers such as carbon pricing, interconnection policy, tariff structures, etc. -- to serve their societal goals.

Policy designed to encourage resilience through microgrids will produce different deployments than would a policy to create a low-carbon society through microgrids.

The model versus reality

At this point, it's worth analyzing a few places where the microgrids modeled in the study diverge from the microgrids we see in reality.

Some of the model constraints have direct, real-world analogues. For instance, the authors employed a space constraint for PV capacity, reflecting that PV deployments are limited by the roof or parking lot space actually available within the geographical perimeter of the site.

The model also sizes its microgrid configurations based on a resource-adequacy constraint: They must have enough localized capacity to meet critical loads in island mode. The thing is, the island mode backup application doesn't appear in the optimization, because the model only looked at microgrids serving as full-time local generators for their hosts.

The study concludes that a robust business case exists for private microgrids as a means of saving money compared to traditional grid energy procurement. If that were the case, we would expect to see a rich array of private self-supply microgrids saving cash for savvy businesses and universities. 

That reality has not yet come to pass.

"The pure self-supply business case for microgrids is only starting to be proven out," said Andrew Mulherkar, a microgrid analyst at GTM Research. Making a microgrid capable of islanding adds significant expense and complexity compared to simple localized generation, so customers don't tend to do it unless they are especially sensitive to loss of power.

More complicated than it looks on paper

If you hang around microgrid discussions at industry conferences, nearly everyone talks about them in the context of ensuring electrical reliability for critical services like data centers, public safety headquarters and military bases.

We're in a situation where the fundamental economics may have shifted, but the market itself hasn't caught up to the change. Here's what Mark Feasel, vice president of smart grid at Schneider Electric, told me for a previous story:

“Ten years ago, resiliency was the only reason you would buy a microgrid, because the energy would cost too much to create -- it would never be cheaper than a grid,” he said. “Now with PV and [combined heat and power] with natural gas, in many states you can generate energy cheaper than you can buy it.”

Indeed, Hanna told me, the UCSD campus uses its microgrid, which includes 30 megawatts of natural gas co-generation, to self-generate about 85 percent of annual electricity and almost all heating and cooling, saving millions in utility bills annually. That's striking, although it's worth noting the campus has built up the microgrid over time with ample grant funding as an ongoing grid edge research experiment; it's not representative of a typical commercial institution building one purely for operational savings.

Customers largely don't know about microgrids or what they can do. On the supply side, these products come highly customized to every user, so there isn't a quick and easy "off the shelf" option. They demand substantial upfront capital in exchange for long-term savings, which are sensitive to rate structures and DER regulatory policies currently undergoing a period of rapid change.

If the microgrid business case works in the more limited self-supply context modeled here, we can expect those assets to be even more valuable when resilience and reliability get factored in to future studies.

"If you have a set of customers that are already ROI positive for a microgrid, if you add the reliability value stream, they'll be even more positive," Hanna said. "For customers at ROI neutral, this might be the value stream that pushes them into the region where microgrid investment is benefical."

The discrepancy between the report's findings and the scant real-world market for self-supply microgrids serves as a reminder that cultural and regulatory friction can hold back economic decisions that look simple on paper.