In my last blog on this continuation series about onsite power options for data centers, I explored just how important electricity is as the core service to everything in a data center, and how onsite power will become a necessary and beneficial solution to data centers by providing lower costs, lower emissions, significantly accelerated and nearly unlimited power capacity, higher availability, higher capacity and a net solution to the immense, long-term and worsening capacity, availability and costs for grid electricity. I say all this from my direct experiences developing GWs of onsite power solutions in many countries just within the last year, and in this blog I’ll dive deeper into more of the types of onsite generation, the net benefits and efficiencies, and being to touch some of the other topics that I’ll explore more in future posts of this blog series, including net emissions, costs, types of generation, and other key topics.
We have to really consider just how fundamental power is or in this case power capacity to the services that our data centers provide. And if we don’t look at operating our own power, we miss one of the most fundamental and biggest opportunities of our lifetimes in this industry to grow faster and to a much larger scale and at a faster pace. With onsite power, we can also provide new and exciting services to empower the people of the world to do more amazing things and do those even more efficiently. For after all, data centers have been the key electrification tool, dramatically increasing the efficiency of how we had done so many things from commerce, to communications, to travel, knowledge, entertainment and more.
For years people used the excuse that we cannot build and operate power plants as efficiently as an electric utility and yet this is not true. We can buy natural gas generators today and other generator technology types that have as good and often much better efficiency than the average efficiency of the electric grid. And yet we also take out all the steps of transformation and the transmission losses and the cost of those transmission lines and transformers along the way by self generating right on site. Simple cycle onsite generation options are typically 10 to 20% higher net efficiency than that of the electric utility power. And of course there is the option to utilize the waste heat from power generation to create additional power generation—known as combined cycle. And the waste heat can be used to provide all of the data center cooling requirements via steam-driven chillers and/or absorption chillers, known as co-gen. When we combine the two this is known as tri-generation. Doing either raises the fuel to power efficiency to 50% to as high as double the efficiency of grid power. So why is this not common or even heard of?
In this blog post I’ll further explore self-powered data centers, and in future blog posts further discuss the technology options, configuration options and some pros and cons to each from my experiences. For now, let’s dive into some high-level reasons for self-powering and some of the benefits. We can self-generate on a temporary or permanent basis, and each has their pros and cons. Whether temporary or permanent, each can be an accelerator to both initial capacity as well as increasing capacity over time. Each can and should increase data center uptime, and permanent solutions can and should decrease operating expenses over electric utility. Both can export power and even each can reduce emissions over electric grid, but this is highly dependent upon the system, utilization and local grid emissions. The key reason why temporary reasons are used is to provide a bridging power solution until electric utility is available at the desired capacity. Both temp and permanent solutions can be aggregated with the grid or islanded from it. Many industrial loads have used onsite power for over a century and still today for reduced costs and higher uptime; these often include airports, university campuses, steel mills, petrochemical processing and manufacturing, mining, and gas fracking, processing and distribution. And many water treatment facilities often generate all or some of their power needs, and many solid waste sites generate and export using the waste fuels created from their process. And even data centers have been self-powered using different fuel types. My first encounter with this was in roughly 2000, and I developed one in 2001, and others since using multiple fuel sources including natural gas, wind and solar now for over two decades.
Since the largest portion of power generation on the grid by fuel type is from natural gas—at about 40% today and more than double the next largest form of generation which is solar and nuclear each today at about 19%, emissions using natural gas to self-generate is often at parity with grids that have high levels of generation via natural gas, lower than grids with existing coal-fueled generation, and higher than those with high percentages of energy from renewables and/or nuclear. But natural gas fueled onsite power systems are often materially more efficient and also lower emissions than same-fuel grid generation since the equipment is both newer, often more fuel efficient and also with state of the art emissions controls. This is not as much the case with shorter term bridging power solutions. But onsite power systems also have another roughly 10% efficiency gain by circumventing grid transformation and transmission losses. While the US power grid is about 40 to 45% efficient—fuel in to electricity out—whereas onsite power from my projects is often about 45 to 50% efficient, and the grid losses from step and step down transformation and transmission losses equate to an added net efficiency benefit of about 10% to onsite power. There are also ancillary factors that come into play, and there are several cleaner fuels, emissions control options, even carbon sequestration options, as well as integration with onsite renewables, which is how I get my start with onsite power at the ripe age of 10 years old. In future blogs with this series, I will provide more information about emissions, technology options, onsite power options with both existing and new power plants as well as other onsite power solutions, costs and timing.
