The island of Graciosa in the Azores is small, with just under 5,000 inhabitants and a peak load of just 3 megawatts. But what is taking place there could be an indication of an emerging trend that is about to recalibrate the entire global energy system.
The isolated island in the middle of Atlantic has had to import and burn diesel fuel to meet its energy demands for the past century. It’s an expensive and dirty way to provide energy, and one that islanders were eager to jettison. The solution: construction of a 1-megawatt solar PV plant, a 4.5-megawatt wind facility and a 6-megawatt, 3.2-megawatt-hour energy storage system.
“They could not turn it on until they introduced storage as a buffer to protect against the instabilities [of the renewables] and stabilize the grid,” said Sen Zhang, senior vice president of engineering at Greensmith Energy, a Wärtsilä Company, an energy management software and integration company that was enlisted to integrate and optimize the island’s use of storage, renewable energy and traditional fossil fuels. “You need to have a hybrid control system to control renewables and energy storage and traditional thermal generation together to minimize [fossil fuels] and maximize renewables.”
The challenges Graciosa is tackling are familiar, and not just for far-flung islands. As the world continues to install greater amounts of intermittent renewable energy, the importance of storage is difficult to over-emphasize. GTM Research expects the U.S. energy storage market to double in 2018 alone. Bloomberg New Energy Finance expects the global market to double six times by 2030.
Storage’s near-instant response coupled with its dramatic and ongoing price declines mean that it’s both an increasingly viable resource for ensuring a stable and responsive grid and is more often being paired directly with solar and wind to create hybrid systems.
“There are real technical reasons around the instability and unpredictability of renewables,” said Zhang. The key to unlocking very high penetration of renewables depends on increased use of sophisticated energy management systems.
From slow and dumb to fast and smart
Energy management systems for large-scale generation have always been around, although in the past they didn’t take the form of sophisticated software. As Zhang describes it, in the past, energy management for power generation involved building excess capacity.
“In the old way of energy management, with massive coal and nukes that support baseload, you don’t need speed because none of the assets can move fast anyway,” he said. “So, the solution is to massively overbuild to meet peak or even average demand.”
That approach is obviously a non-starter with hybrid energy systems because wind and solar are not dispatchable, and batteries — while getting cheaper — are still relatively expensive. The capacity of energy storage systems also isn’t so large that they can be used indefinitely. All of which means that energy management systems for hybrid renewable power plants have to be both fast and smart. “It’s like a chess player — you need to see steps ahead about what is going to happen to intelligently optimize your moves,” said Zhang.
For example, Greensmith’s GEMS, a proprietary energy management system, can not only control storage and renewables at increments of 100 milliseconds, it utilizes machine learning and historic and real-time data analytics to calibrate what type of generation is needed at specific times. It also benefits from the knowledge of its parent company, Wärtsilä, which offers everything from internal combustion engine-based power plants to utility-scale solar PV.
“A very fast diesel engine takes five minutes to start, coal plants take half a day, and gas turbines take hours, so you need to turn them on with enough lead time. That is why intelligence is important,” said Zhang.
Data is king
Intelligence that pulls together both weather variables and knowledge of the specific systems being integrated is critical, especially as batteries are a limited resource. Greensmith’s data comes from battery cells, engines, PV plants and any other component of a hybrid energy system and streams that information into the cloud, where it is stored.
“That’s the data foundation for our computations,” said Zhang. “We use the cloud to store six years' worth of data and collect it and refine our algorithms and our knowledge based on the data collected. We are like Google in the energy sector, and data is king.”
The data is then put into a virtual model, which allows for detailed simulations of how a hybrid power plant works. That intelligence is married with computing technologies such as artificial neural networks that are used by industries such as stock trading and social networks that allow for advanced machine learning and predictions.
These data collection and analytics features helped Greensmith to develop a product called the GEMS Analyzer, which can calculate energy output, internal rate of returns and levelized cost of energy for projects before they’re started. “It gives customers confidence that we deliver what we say before they make a decision,” he said.
The future evolution of energy management systems is one that Zhang expects to mirror what has already happened with the operating systems of computers. In the early days, hardware manufacturers required that consumers use their own software. That obviously didn’t last, and Zhang expects the same separation between hardware and software to occur with energy management systems. For instance, Greensmith has already done more than 70 installations with 16 different types of batteries and a dozen different inverter types.
“I think where the market is headed is to be hardware-agnostic, and we are developing hardware-independent software that hides the vendor’s proprietary information from the end user,” he said. “I’ve seen many battery vendors and power electronics vendors come and go, but the whole industry has prospered. I think the way we can keep the momentum and make energy storage more sophisticated with renewables is to have the software totally separate from the equipment vendor.”