How AI improves operational efficiency in renewable energy power plants

The intrusion of artificial intelligence (AI) in various sectors is one of the hot topics in the technology sector these days. Predictions of how machine learning and similar innovative systems would disrupt and improve operations in industries ranging from security to agriculture and beyond have captured the imagination of consumers worldwide. 

While the role of AI at the consumer role is more obvious with the popular culture accepting the arrival of the likes of Alexa and virtual assistants, some of the biggest potentials of AI lies in the industrial side of things. The renewable energy sector is a perfect example of an industry where smart machines are beginning to make an impact that will disrupt the market and have large-scale benefits. 

According to a recent report from the research from DNV GL, more and more operations in sectors like solar and wind energy will get automated with AI in the coming years. The technological development will boost efficiency in the renewable energy sector overall. The major areas within this sector that are expected to come under AI’s influence are decision making and planning, robotics, inspections, condition monitoring, supply chain operation and certifications. This, of course, is besides the technology for power generation itself. 

The key here is to understand that data plays a crucial role in the renewable energy sector. That commercial developments in sectors such as wind and solar energy are relatively new means most of the power generation setups are already equipped with sensors that can collect data. Because of this, major roles that AI has been playing so far have been in predictive maintenance, resource forecasting and control. 

 “We expect the installation of more sensors, the increase in easier-to-use machine learning tools, and the continuous expansion of data monitoring, processing and analytics capabilities to create new operating efficiencies — and new and disruptive business models,” said Lucy Craig, Director of Technology and Innovation at the Energy Division of DNV GL in the report.

The rapidly growing role of AI

According to Larsh Johnson, Chief Technology Officer at Stem, pointed out that the first thing to note is that customers, more than ever before, are demanding

control over their energy decisions.  At the same time, their choices have expanded from utility-supplied renewable solutions, various on-site generation options, and new ways to participate in energy markets.  

“This began under the last wave of distributed solar growth and is now being seen in demand for customer-sited energy storage,” Johnson said. “Whether the customer chooses standalone energy storage systems to manage their energy bills and contribute to grid stability in areas of high renewables penetration or chooses storage to enhance the economics of their solar system, that customer is pulling Stem’s affiliated AI into the energy markets.”

Customer-sited energy storage primarily assists the utility and grid operator to smooth the ever-increasing solar photovoltaic (PV) generation on distribution grids. Stem, with its AI software, has been able to assist with high PV congestion and provide grid edge visibility as it contracted with the Hawaiian Electric Company to develop “virtual power plants” that consists of AI-driven distributed energy storage.

Giving another example of the application of AI in this sector, Johnson added that a specific aim of Stem’s virtual power plant (VPP) for the Tokyo Electric Power Company is to collect data from the AI to help Japan’s Ministry of Energy, Trade, and Infrastructure to develop more comprehensive plans to balance the grid using distributed storage amid higher penetrations of rooftop solar.

Niccolò Teodori, CEO of Elemize, while pointing out that the value of renewable energy as a commodity is no longer in the product itself but also in its generation, elaborated on the part played by AI in plants. 

“Today, AI can improve power plants’ profitability by improving its operations and reducing losses,” Teodori said. “On one side, it is being used to make maintenance more efficient, but on the other, it is used in predicting generational capacity. This is what I really see because when you have unpredictability in power generation, especially in sectors like solar and wind, it affects trading activities on the energy and ancillary markets. Because when you promise to deliver a specific amount of energy and you are not able to keep the promise, you will end up paying fine and perhaps incurring losses.”

The reasons for the need for such an intervention from AI is also worth noticing. Aiden Livingston, Founder of ThermoAI, indicated that it is important to ensure the existing power systems respond to and meet the changing power requirements in the market. 

 

“How do we make these older systems responsive to the grid demands,” Livingston said. “As more renewables are brought on to the grid, the energy they supply is from intermittent sources like wind and solar, and obviously not synced to customer demand. One of the things we are working on, for instance, is how can we shorten the startup and cooldown cycle because these can normally take hours. We use AI to reduce this to the minimum amount of time so that these plants can be as responsive as possible.”

He added that there are several other potential uses of AI, including the cases of predicting demand that can further operational efficiency of plants. A key aspect to take a note of at this juncture is the role AI will have in security.