What’s Next for Solar: Artificial Intelligence

Share on FacebookTweet about this on TwitterShare on LinkedInEmail this to someone

Artificial IntelligenceThere is arguably no hotter or more contentious topic in technology right now than artificial intelligence or machine learning – using algorithms to enable machines to mimic brain activity and “learn” things for themselves. How smart can we make our machines? Can scientists discover the ultimate algorithm that gives machines the power to understand the world? Does AI pose a rise-of-the-machines risk to the human race?

Still, as the debate rages on, AI is playing a greater role in our lives every day. Algorithms underpin much of life as we know it in the 21st century. It is thanks to algorithms that we are able to determine the quickest route via mobile maps, find love via online dating platforms, tag friends via facial recognition on social media, book travel reservations and speak commands into our phones. And AI is getting more advanced every day, with such innovations as self-driving cars demonstrating broad scale adoption could be only years away.

But what does artificial intelligence mean for the solar industry?

Initially, Internet of Things technology will transform the components on a solar installation into a physical “network” by embedding them with electronics, software, sensors and network connectivity. The components will collect and exchange data, feeding the algorithms essential to AI.

This data will first and foremost revolutionize R&D. No longer will we base our product designs and project solutions on laboratory outcomes and static modeling. Instead, we will be informing a new generation of structural analysis with real time wind and snow load, forces, system performance and other data from live projects.  SunLink is already harnessing this real-time R&D power in its latest product designs to realize more dynamic, lower cost solutions than are possible today.

Very soon, we’ll also be entering an era of machine learning where our solar hardware actually becomes able to troubleshoot without human involvement.

For example, by analyzing all the data collected by the sensors over time, a tracker system might recognize a pattern that when a wind speed of a certain mph hits the system at a certain degree tilt, the system is stressed. Moving forward, any time the tracker sensors picked up on the fact that the wind was approaching that mph, they could then trigger the array to change tilt angle to avoid the stress. When the wind speed drops back down, the system will react by moving back into optimal tilt. There might even be a certain time of the day that a stress condition occurs on a regular basis and with predictability because of precursor events. We will eventually anticipate force and move to mitigate their effects before they happen.

With our systems able to react on their own to stressors and avoid extreme loads, we will have the ability to pull costs out of the system by reducing the metal required, because it will no longer be necessary to design for the max code-level event.

Artificial intelligence is going to figure as prominently in the EnTech-enabled future of solar as it does for every other industry. Smarter solar “machines” are on the way.

Tip: Machine learning is poised to revolutionize R&D and create a new breed of optimized product designs and improved project performance as a result.