Industry interview: The role of virtual power plants on the future energy grid, Ted Ko

Industry interview: The role of virtual power plants on the future energy grid, Ted Ko

Industry interview: The role of virtual power plants on the future energy grid, Ted Ko Energy Storage Journal

 

The decentralization of power is changing the nature of the traditional energy grid system with advances in technology allowing operators to manage supply and demand at the macro-level through the aggregation of multiple sources — so called Virtual Power Plants.

One of, if not the biggest, disrupters to the old model of supply and demand will come from VPP’s ability to coordinate myriad distributed energy sources through artificial intelligence and big data collection to balance the grid amid the growing penetration of renewables in the energy mix.

Pilot projects are the beginning, and VPPS will be the successful conclusion, if you believe many within the industry. So much so that in 10 years-time, utilities and transmission companies will no longer own traditional power generation, and utilities may even be rendered obsolete.

Here, Ted Ko, director of policy at digital storage network operator Stem, tells ESJB what he believes the role of VPPs will be, what changes need to happen to ensure they become business-as-usual, and the changing relationship between supplier, end user, and utility.

 

ESJB: What is the future role of VPPs?

Ted Ko, director of policy at Stem

Ko    The nature of electric demand and distribution is changing greatly. Utilities across the US and in many other industrialized markets must manage decentralizing grids, fed by variable renewable energy, which is shifting their needs to more flexible, localized, real-time energy management, and consequently driving up rates for commercial customers.

As a result, utilities and bulk grid operators increasingly seek dynamic, network-enabled VPPs “as a service,” where local, modular capacity, and resources can be instantly applied to address any locational issue. These dynamic, network-enabled VPPs may in the future include islanding microgrids that operate independently of the electric grid when required to maintain service amid larger grid disturbances.

For an individual commercial or institutional customer, Stem’s artificial intelligence (AI)-driven energy storage reduces their energy costs by reducing their peak demand and providing them with real-time and predictive data and energy management visualization tools.

Stem’s AI can then network that system with others into aggregated VPPs, where unused stored energy can provide over a dozen additional energy services for the utility or grid operator at lower cost than traditional grid infrastructure, and with five minutes’ notice.

 

What policy changes have happened, and what needs to be introduced, to allow VPP to become business as usual for utilities?

First, rate design that better reflects the time-varying value of energy is key to making the economics of energy storage attractive to customer engagement.  Second, utilities and their regulators must allow customers to participate in any multiple configurations of demand reduction programs and wholesale market opportunities, knowing that the meter-level capture of data on a one-second basis inherent to energy storage can ensure ratepayers are getting the maximum benefit of “value stacking.”

Moreover, utility regulators and grid operators must set policies that allow third-party energy storage providers to provide additional grid services and remunerate them fairly for these ancillary benefits.

Finally, in the US, policymakers outside of California must create the rules to allow third-party DER aggregators to participate in wholesale markets equally with front-of-meter resources.

 

 How do opportunities to participate in wholesale energy markets via the Demand Response Auction Mechanism (DRAM) program aid the transition to a new power supply and demand network?

The Demand Response Auction Mechanism (DRAM) program is one new opportunity created in the past two years by the California Public Utilities Commission (CPUC) for utility customers to participate in the California wholesale energy market. Stem anticipates additional wholesale opportunities in California in the coming few years.

California is moving from utility-based Demand Response (DR) programs to wholesale market-based DR resources that provide capacity to the California ISO (CAISO) grid operator, and seeking improved response rates, larger scale, and cost efficiencies.

DRAM makes the transactions more economical, larger in scale, and somewhat more predictable than the smaller previous pilots.

DRAM is a success on many fronts, despite the small current VPP sizes and high transaction costs.  It is proving the technical viability of the first customer-based VPPs in the US.

In just one example from one day, Stem engaged 100 California customer systems in 14 VPPs to provide grid relief during an unprecedented heat wave in August 2017.  Throughout the year, Stem has reconfigured its VPPs to meet our customers’ onsite needs and to respond to approximately 600 wholesale dispatch calls.

Therefore, the frequency of customer participation has also risen from the low tens of dispatches to hundreds of dispatches per year.

Moreover, the DRAM contracts are proving the ability of the VPPs to execute in the real-time market; Stem’s own network of customer-sited storage responded to 150 “real-time,” or five-minute dispatch events for San Diego Gas & Electric between January to May of 2017.

 

How can Stem bridge the gap between supplier, end user, and utilities?

Stem is the global leader in customer-sited intelligent energy storage, and our mission is to build and operate the largest digitally-connected energy storage network. We have 800 systems installed or contracted, representing 180MWh, in 75 jurisdictions within four announced US states.

These networks are growing to deliver value to our eight utility contracts, worth 350 MWh of different services, including local capacity, aggregated demand response, alleviation of solar PV penetration, and so on.  The larger our networks grow, the larger the services we can provide to customers, utilities, and grid operators.

End users demonstrably want more control over their energy decisions and to participate in the markets in new ways.   Stem’s proprietary software rapidly responds to spikes in electricity use, drawing on stored power to automatically reduce demand charge costs without requiring operational changes or manual input from the host customer.  Customers pay a monthly subscription fee and on average save up to two to three times what they pay for the service.

Membership in Stem’s networks then enables the customer to participate in new market opportunities, such as through DRAM or other programs.

Stem is demonstrating to the utility and grid operator how to engage customers and how to aggregate diverse customer locations, system sizes, and customer load shapes and types, where Stem’s real-time adaptive dispatch capability allows us to pick and choose what resources to bring to the grid or the utility when called.

 

Will the grid of the future be equally concerned about big data collection and even AI, than supply methods such as PV/wind/fossil fuels to ensure grid balancing services?

Yes, the future grid will depend on big data collection and the artificial intelligence expertise needed to manage it, and customer-sited assets can offer rapid response and location or time-based flexibility benefits.

Forward-thinking utilities and regulators are turning to innovative, flexible, non-wires alternatives, including coordinating customer-sited distributed energy resources, to cost-effectively address challenges in the distribution network.

Operating a network of multi-purpose, grid-responsive energy storage systems requires sophisticated software that includes distributed decision-making and edge computing, honing predictive capabilities using machine learning and neural networks to continuously improve optimization decisions, and reduce latency.

AI-driven energy storage is transforming all expectations of the electricity system — among commercial customers, utilities, electric grid operators, and policymakers — and putting businesses in the driver’s seat to control their costs and even participate in new ways in the energy markets to maximize their economic opportunities.

The AI behind energy storage is transformative because it turns a commodity battery into a platform for grid-edge matchmaking among market opportunities, akin to a shared ride service or shared tourism residence service, putting more power in the control of the consumer.

Stem captures data on a one-second basis, we can pivot on a one-minute basis, and we store terabytes to the cloud. We have analyzed over 3 million hours of field data from over 300 systems using machine learning, smart analytics, and other data science techniques.

This data is used to continuously improve our predictive algorithms, savings projections, hardware controls, and other performance and system life aspects.  The more data we collect, the smarter our platform gets and the more value is delivered to customers, utilities, and grid operators.

What that means is Stem delivers an AI-based platform as a non-wires alternative and an operating expense that offers deployment flexibility that helps make utility system forecasting and planning easier, and reduces the risk of traditional grid capital expense investments that ultimately may not be required as reduced load growth and distributed energy change the nature of the grid.