Big Data for the New Solar Century…

God forbid, one Monday morning, the Garbage Engineer drives up to your driveway and spills stuff out everywhere, while hoisting the huge bin from ground to truck.  Next, he just drives away.

Cleaning all that trash up would be way harder than if it had just, well, stayed tightly packed in the can. God forbid then, what if someone slips on the mess? Sues you? It was your garbage and your driveway. Right? Sorry, digressing…   Cleaning that mess would take up valuable time, which could have been spent learning how to flyfish.  At home, and online unfortunately, like you do everything else.

Your solar plant data is the same. You already know solar is the new oil and the new energy. Its your business, so you know it best. Data, not so much.

I’d argue your business being successful, depends on data. Solar company owners compete, based on the score of what their profits are, versus expected to be, when the investment was created. So yes, data quality and maintenance are critical. It’s used to determine what you share with your investors, with your banks, and what’s left over for your team at the end of every quarter. As Solar economics are pretty well known to be quite tightly managed, I’d argue that also tightly managing your data is the key to greater profits.

What I’ve noticed is that in most solar companies, O&M people get to choose data providers, even though the impact of effective data utilization is felt most by finance which needs clean data to be able to continue to bring in new investors.

This means the right data partner you use, needs to fully understand data quality, stability and security at the minimum, and – and bonus points – combine it with Machine Learning to help you actually increase your profits and reduce costs.

To run its very best, a Solar company needs Data which is scrupulously clean, so if its anything like the falling garbage this morning, it may need to be decontaminated in order to be useful. Machine Learning built on this data will be where the meaningful improvements can happen.

Deep Machine Learning via Digital Twin technology help you fix issues proactively, focusing first on what affects revenue the most, arming your O&M team with smarter information.

Digital Twin technology on an AI platform along with RealTime weather and historical satellite data help forecast production precisely. This way, you’re not paying regulatory penalties because you’re helping the utility manage its ups and downs in demand.

Of course, monitoring all plant and portfolio details in RealTime with visualized reporting, means there’s no lag between issues occurring and good decisions quickly being made to resolve them. Being on the cloud means this can be done from wherever you are, even home.

Focusing here on just the data portion, on which all plant improvement rests, there’s a few simple rules to remember:

1.    Capture raw data from its source – not summaries and interpretations.

2.    Don’t just capture the data, but validate it, cross-check it against other values and conditions, look out for missing values.

3.    Act on issues found during capture, by not only alerting but also fixing the data in a reasonable manner by intelligent interpolation.

4.    Save the time-series of not just the instrument data, but the meta-data as well – things like installed capacity of any string.

5.    Persist the processed data. Process it in a meaningful way. Don’t let it just sit there. Reprocess it. If there’s a new algorithm which can help you – reprocess again. Add processing to persist smarter aggregate interpretations of data.

6.    Assume that you will miss something today and will need to reprocess everything – design for it and don’t be afraid to reprocess.

7.    Finally, make sure that the created data is distributed and used by all concerned – if its not in use today, it is unlikely to be useful tomorrow.

Lastly all of the above needs to happen in a secure format with strictest privacy maintained.

Bring in the right Data expertise, and let’s look at the garbage problem you had this morning again and substitute data specialists. It’ll be like having someone who knows how to pack the trash can right, put it in the right place to try to ensure garbage is moved to truck in a smooth arc.

Ultimately, looking above, a Solar plant doesn’t need much upkeep – just monitoring all the process, looking out for any issues, and letting you see what the plant is doing 24/7. Easy right?  It’s so easy it keeps all plant managers busy every minute of the day in making countless decisions to most effectively run the plants. That is your expertise. Your clients all trust in your ability to make the millions of decisions to do all of above really well…

Adding Data Expertise can help you make all those decisions more effectively, so your plant functions much better and you reap the benefits.

Author: Kitty Chachra, CRO at Quadrical Ai

Technical assist from Sharat Singh, CEO & Chief Architect at Quadrical Ai

Originally published on January 16, 2021