Cloud computing is in our energy forecast

Four years ago, I had a conversation with a major regional transmission organization (RTO) about connecting to its control center through a cloud connection rather than using a decryption box mounted on a rented rack in a random data center.  The response was, “The cloud …  We’ve heard of that.  We know it’s coming, but we’re not there yet.”

Fast forward and I always look back at that conversation with incredulity. Were some of us ever so clueless?

I am the CEO of Mosaic Power.  We are a third-party aggregator of residential electric water heaters.  In simplest terms, we pull together thousands of water heaters into a single entity of controllable energy storage by delaying when individual heaters draw or consume power from the grid.  The cloud is a critical part of our business, used to store and process all the data from the water heaters we control, transmitted through a private cellular network.

In other words, we are part of the larger universe of distributed energy resources (DERs). Often loosely and variably defined, DERs encompass a range of technologies or approaches — such as demand response, storage, distributed generation, and electric vehicles — that can supply or consume energy (or both) and also be controlled, which is what makes them resources.

The key technologies enabling the spread of DERs are machine-to-machine connections over the Internet, and cloud computing technology.  Why is cloud computing such a game changer?  Individual companies no longer require their own dedicated systems, and physical space with clean rooms and staff. Machine-to-machine networks typically communicate simple data, such as changes of state or environmental data, meaning that networks can have lower communications costs, thus reducing the cost of reliable connections to DERs.

Innovative products interconnected throughout the bulk electric power system have been using cloud computing for some time. But the organizations that actually control the grid have been reluctant to adopt cloud computing in their operations for a variety of reasons such as data security, end-to-end responsibility for service, and lack of price sensitivity.

Third-gen, dispatchable demand response

So, what exactly is “the cloud,” and why has it suddenly become practically ubiquitous? Generically speaking, the cloud means “computed somewhere else, using someone else’s data servers” — essentially digital infrastructure as a service.  At Mosaic, the someone and somewhere are Amazon Web Services.

Outsourcing data infrastructure has a number of benefits.  First, cloud computing providers offer very high levels of site security, connectivity and redundancy at reasonable costs, especially for small, startup companies. Additionally, expanding your infrastructure is painless. You can pay for a system sized to your current needs and increase to a new server in a matter of seconds as your data needs grow.

For Mosaic and other demand management providers, this flexibility is now propelling the emergence of a third generation of advanced residential demand response programs, which could open the way to 24-7 dispatchability.

The first generation of residential demand response — such as summertime air conditioning cycling programs — generally involved simple one-way turn-off switches that could not communicate back whether the signal was received or the connected device responded. The second generation — generally used to reduce utility demand charges — incorporates two-way communication devices that can confirm on or off, but are limited in data sampling time frames. With advance notice of a high-demand situation, second-gen technology allows customers to shift load from one time of day to another, as is common in regions with significant solar penetration.

Third-generation demand response programs now coming onto the market rely on the use of modeling and prediction made possible by big data storage and processing techniques – a much more accurate way to simulate, dispatch and verify demand events. It could eliminate the concept of discrete, limited-time demand response events, and open a portfolio of continuous energy management services, drawing on dispatchable resources with responsiveness and reliability comparable to — or even better than — conventional generation.

Third-generation demand response could be used to provide reliable, verifiable services, including:

• Energy arbitrage (reducing short, sudden peaks as utility billing is settled at intervals as small as 5 minutes)
• Fast frequency response (shedding load instantly as other resources start spinning too slowly)
• Demand charge management (local high-load reduction)
• Fast regulation (reacting in seconds to increases and decreases in electricity supply)
• Distribution line load reduction

Smart thermostats and water heaters are now using cloud technologies and behavior learning to take this level of demand response even further.  Instead of relying on customers’ awareness and understanding of pricing on the local electric grid, technologies that “learn” or are aware of consumers’ behavior – with no consumer intervention required – are able to merge the needs and preferred comfort levels of consumers with the energy needs of the grid. Thus, consumers receive the best energy prices with minimal impact on their comfort.

Standards and security

Grid control centers across the country are increasingly being asked to interface with companies that rely on cloud computing, and the potential benefits are significant. Between cloud computing, laptops and mobile phones, control centers could conceivably reduce operational costs significantly while increasing computational and communication speed and capacity at the same time.

Of course, equally significant challenges lie ahead. The North American Electric Reliability Corporation (NERC) provides guidelines for any assets that have the potential to negatively impact the reliability of the grid. Recently, NERC has begun creating standards that incorporate having control centers reliant on cloud computing.

Moving from dedicated computer systems to interconnected cloud platforms also raises security issues. The U.S. Department of Energy’s Argonne National Laboratory in Illinois  has been tasked with developing a framework of best practices for cloud-based grid applications.

Key data security issues include securing the data in transit and in connected storage.  In addition to focusing on the physical mechanics of resilience, such as backup power and redundant telecommunications routing for specific technologies, appropriate service level agreements for each function must be developed.  We may also see less focus on new services to leverage and reuse existing capital infrastructure, and more focus on the most efficient ways to use human capital to create new services.

To date, the most popular cloud-based applications at the grid level are simulations.  Increasingly complex simulations are now possible for transmission planning and outage conditions, allowing contingency plans to become increasingly complex and numerous. More types of crises are planned for, with more possible solutions and mitigations outlined in detailed work-around plans.

Just a few years ago, these simulations – and the demand reduction data contained in them – were considered merely informative.  But demand reduction has become increasingly automated, and the potential data acquired has become more frequent and reliable over less expensive networks.  Dependable demand reduction is becoming not just a matter of hot-weather cost reduction, but a critical part of infrastructure planning from New York City to Portland, Oregon to Hawaii.

Electricity use has gotten more tightly integrated into our increasingly complicated daily lives.  Environmental concerns have resulted in the deployment of large numbers of variable generation resources – like wind and solar – which makes balancing electric supply and demand more complicated.

Ongoing cost reduction paired with increasing computing capability and flexibility means cloud computing is in our industry’s forecast for the foreseeable future.  However, there will be new and evolving questions to ask – whether your business or operations control center is in the cloud or still firmly planted on the ground – as more and more of the services in our energy world rely on cloud computing.

Technology providers will need to keep data security tight to protect grid infrastructure; utilities will need to ensure that security is provided across changing data infrastructures. But, technology will not be enough. As with many aspects of our energy transition, getting the cloud right will also require that all stakeholders — technology providers, grid operators and utilities — engage, communicate and collaborate with each other.

Laurie Vaudreuil is CEO of Mosaic Power. She can be reached at [email protected].