Gareth Brown, CEO of Clir Renewables. Image: Clir Renewables.

Solar capacity in the UK has grown substantially over the past decade, from 27MW in 2009 to over 13GW today. While Feed-in-Tariffs kickstarted this growth in 2010, subsidies have since been lifted, stalling the wider roll-out of utility scale solar arrays. In order to ensure many of these large projects are financially viable, owners and operators have worked within very tight margins when it comes to operational expenditure.

The industry has, as such, focused primarily on bottom-line cost when it comes to operations and maintenance (O&M). This has been (wrongly) justified by the perception that solar assets are simple structures requiring little maintenance. After all, if the process is as easy as identifying and replacing broken modules why invest in advanced asset monitoring?

Unfortunately, this complacent view of solar performance and the resulting lack of investment or innovation in O&M is holding many asset owners back from realising the full potential of their assets. This is clearly a false economy. By choosing a reactive O&M strategy, owners are missing out on the much greater returns that would be unlocked through more advanced asset performance monitoring.

Visual inspection can only go so far in identifying underperformance. While it is easy to assume that the static nature of solar panels equates to less wear-and-tear that can cause underperformance, these assets are, in fact, still exposed to gradual degradation caused by the surrounding environment.

It is often difficult for on-site teams to detect the slow and ongoing deterioration of performance due to factors such as UV exposure, thermal cycling, damp heat, destructive precipitation, and the growth of nearby vegetation – and even more difficult to quantify this impact on energy production.

Additionally, errors in solar asset performance can go beyond the physical module. Grid curtailment is often necessary to balance energy supply and demand, but if the asset derates out of turn or is slow to return to normal operations due to sensor error, owners will lose out on generation (and therefore revenue). Similarly, if irradiance trackers are consistently misaligned, the asset will fail to meet budgeted energy production.

However, the current standard of data monitoring can often obscure whether underperformance is due to environmental or technical issues as outlined above, or simply a dip in irradiance. This is because the data “noise” of natural variations in resource can often mask small but sustained technical performance issues. In order to qualify whether underperformance is due to variation in resource or a fixable error, owners and operators must set asset data in the context of as much environmental and site data as available.

Integrating and analysing solar panel data within the context of its environment is often considered an expensive task, as traditional data approaches require the dedication of significant working hours to sort through and assess each data point produced by a large-scale solar array. The time taken to analyse a solar portfolio is compounded if panels at different sites have been provided by different original equipment manufacturers (OEMs). As data collection methodology can vary significantly between OEMs, it is almost impossible to compare and optimise the whole fleet quickly.

This cost has actively disincentivised asset owners from using data to inform O&M, leading to lower quality data collection from on-site teams. Ironically, this complacency around data collection leads to an increase in the time needed by analysts to relabel and manage data – incurring further costs over time.

Fortunately, recent advances in data analytics such as artificial intelligence and machine learning allow owners and operators to assess and interpret vast quantities of data ten times faster – and at a much lower cost – than traditional methods.

By using AI to fully analyse asset performance in the context of its environment and peers, owners and operators can more effectively target O&M to fix previously invisible problems and ultimately increase energy production – which, naturally, will facilitate returns far beyond what budget O&M could possibly achieve.