Article 3: The Top 4 Solar Asset Monitoring Challenges—and What You Can Do About Them
The Challenge of Granularity
By Steve Hanawalt
In my first two articles in this eight-part series on solar performance monitoring, we discussed the challenge of working with real-world operating data and the challenge of scale. Solar monitoring applications are tasked to process very high volumes of noisy operating data and, traditionally, have not performed that task very well.
In this third article, we will discuss another common problem with solar monitoring systems—the challenge of granularity.
The Challenge of Granularity
The Data Warehousing Wiki defines granularity as “the level of detail of your data within the data structure.” Granularity in data systems has to do with detail—how detailed do we need to make our performance models to effectively monitor solar power assets? The answer, it turns out, is very detailed.
Why is that? As I shared in the last article, I came into renewables via a long career in the traditional power industry monitoring gas turbines, steam turbines and boilers. I expected monitoring the performance of solar power assets to be easy by comparison. I mean, the solar power asset is so much simpler than the fossil power asset, right? What is it about the solar power plant that makes it such a challenge to successfully monitor?
The granularity of the solar asset class is what makes it a particularly challenging asset to model. By way of comparison, a 500 MW gas plant has 3 electric generators. How many generators would the same size solar power plant have? Over 1.5 million. Add to that the over 100,000 strings, 10,000 combiner boxes, and 300 large central inverters and you can see why solar monitoring applications have to work with equipment at such a granular data level.
Why does data granularity introduce such challenges to reliable and accurate solar performance monitoring? Let’s look at three challenges that keep monitoring software developers up at night.
Granularity and Energy Models
First of all, to successfully characterize the operational performance of solar assets, we need to look deep into the DC array. This is because, if we are going to understand solar plant underperformance, we need to drill all the way down to the generator level: the PV module.
However, PV module DC electricity is ultimately delivered to the off-taker via a series of electrical circuits made up of strings, fuses, conductors and combiner boxes. The performance of each of these assets makes up the operational performance of the plant as a whole. Therefore, our performance model needs to include a model of each and every DC array asset with enough granularity to identify equipment outages and underperformance.
Granularity and Problem Identification
Because the solar asset class includes a high volume of assets that directly generate electricity or transfers electricity, the identification of which assets or groups of assets are causing the performance problem becomes very difficult. This is especially difficult in that most solar power projects do not meter power, voltage or current below the inverter level. If we are seeing inverter-level underperformance, how do we know if the problem is in the modules, the strings, the combiner boxes or the inverter itself?
The solar data and performance model must be sufficiently granular to tease out not only the existence of a problem, but where the problem actually resides. This means very detailed asset performance models are needed and the ability to track performance to a specific asset is required. The performance model needs a detailed asset and metadata registry to solve this problem.
Granularity and Energy Accounting
A follow-on to the energy model and problem identification challenge, is the energy accounting granularity problem. If we’re successful in developing granular performance models and use those models to identify where the fault is occurring, we then need to properly account for these energy losses.
Granularity at the energy accounting level is important, because, without it, we will likely double-count losses and not provide users the information they need to recover those losses. Many performance monitoring systems were never designed with this level of detailed energy accounting in mind, and so create more confusion than clarity.
The challenge of insufficient data granularity prevents many solar monitoring software applications from realizing their full potential. Until this fundamental design problem is resolved, users will continue to be frustrated with software monitoring tools that don’t work.
Stay tuned—we believe we have a way through this problem. You can read more about it in article 7 coming up in a few weeks. In my next article, we will discuss another problem common to solar performance monitoring software applications: the challenge of selecting the right performance model.