The UK solar market has achieved remarkable growth with annual PV capacity increases exceeding 40% over the past four years. Yet, beneath this success lies a quiet problem: most assets operate at 5-10% below their potential. This performance gap represents millions in lost revenue across the sector, driven not by dramatic failures, but by subtle and quiet, systematic inefficiencies that conventional monitoring simply cannot detect.
The challenge facing asset owners has shifted. Installation and commissioning are well-understood processes. The real engineering problem lies in the gap between operational and optimal performance, in building systems that can distinguish between the two.
The Three Hidden Performance Killers
The Gradual Impact of Soiling
The UK’s frequent rainfall provides a natural cleaning mechanism that prevents the extreme soiling losses seen in arid regions. However, the persistent accumulation of pollen, pollution, and agricultural dust still creates measurable impacts.
Conservative estimates place annual soiling losses between 1-4%. For a 20MW asset, even 2% represents approximately £40,000 in annual revenue reduction. The challenge isn’t the magnitude of individual losses, it’s the cumulative effect across time and the difficulty in quantifying when cleaning becomes economically justified.
The engineering question becomes: at what point does the cost of cleaning intervention become less than the cost of continued soiling losses? Without precise measurement, this decision relies on fixed schedules rather than data-driven optimisation.
Degradation and the Mismatch Problem
Normal panel degradation of 0.5% annually is well-documented and factored into financial models. The hidden losses occur through abnormal degradation patterns and their systemic effects.
Potential-induced degradation (PID) represents one of the most severe forms of module degradation. Power losses can reach 8-30% depending on electric field strength, temperature, humidity, and module materials. Standard monitoring systems average data across entire sections, masking early-stage degradation in individual modules until the problem becomes system-wide.
This creates a compounding effect through module mismatch. In series-connected strings, one degraded panel acts as a current bottleneck, forcing all downstream panels to underperform. The result is a systematic inefficiency that bleeds revenue incrementally, day after day, while hiding as minor variations in total output.
From a systems perspective, this represents a classic weak-link problem, yet one that remains invisible to conventional monitoring approaches.
Clipping and Curtailment: System-Level Bottlenecks
The most frustrating losses occur when panels perform optimally but system-level constraints prevent full energy capture.
Inverter clipping occurs on peak production days when DC power generation exceeds inverter AC capacity. Well-designed systems expect 0.5-1% annual losses from clipping. However, without precise performance modelling, it’s impossible to distinguish between acceptable design parameters and problematic losses that may reach 3% or higher.
Grid curtailment presents an even more complex challenge. As UK renewable capacity continues expanding, grid infrastructure increasingly constrains peak-time energy acceptance. Perfect conditions and optimal asset performance become irrelevant when the grid cannot accept the energy produced.
These bottlenecks represent engineering inefficiencies at the system level. Problems that cannot be solved through improved hardware but require better integration and predictive management.
Moving from Reactive to Predictive Operations
The Limitations of Current Approaches
Traditional monitoring relies on reactive, alarm-based systems. Activating only after the problems occur. This approach creates a binary operational state: functioning or failed. The subtle performance variations that precede failure, the gradual losses that never trigger alarms, remain invisible.
Predictive maintenance schedules attempt to address this through regular interventions, but without understanding actual asset condition, these approaches often result in unnecessary maintenance or miss the slowly developing problems entirely.
The engineering challenge lies in creating systems that can detect and quantify the subtle performance variations and causality before they become operational problems.
An Analytics-Based Framework
The solution requires moving beyond simple monitoring to comprehensive performance analytics. This is built upon a foundation of three core components:
Centralised Data Integration: Effective analysis requires comprehensive data collection from all relevant sources: SCADA systems, weather monitoring, grid conditions, and asset management databases. The goal is creating a complete picture of system performance and external influences.
Machine Learning Performance Models: Rather than relying on static performance predictions, machine learning models can be trained on site-specific characteristics, historical performance data, and local weather patterns. This creates a dynamic performance baseline that accounts for seasonal variations, equipment aging, and site-specific conditions.
The model essentially becomes a digital twin that doesn’t degrade, providing a constant benchmark for expected performance under current conditions.
Anomaly Detection and Root Cause Analysis: By continuously comparing expected versus actual performance, the system can identify deviations and classify their likely causes. A gradual, consistent decline suggests soiling; sudden drops in specific strings indicate module or connection issues; DC power consistently exceeding inverter capacity quantifies clipping losses.
The Implementation Reality
The most effective implementations focus on actionable intelligence rather than increased alarm frequency. The goal is not more alerts, but better understanding and comprehension of existing data.
This approach enables maintenance optimisation: Cleaning panels when soiling costs exceed cleaning costs, rather than following fixed schedules. It provides precise fault localisation, reducing inspection time and improving repair accuracy. Most importantly, it creates the documentation necessary for warranty claims and contractor accountability.
Closing the Performance Gap
The 5-10% performance gap across UK solar assets represents a significant optimisation opportunity. Advanced analytics can help recover at least half of these losses by quantifying their sources and providing clear intervention guidance.
However, the broader value lies in shifting operational philosophy from reactive maintenance to predictive performance optimisation. As the UK renewable sector continues to mature, competitive advantage will increasingly depend on operational excellence rather than installation capacity.
The distinction between an asset that merely operates and one that truly performs, lies in understanding the difference between what is happening and what should be happening. This requires moving beyond simple monitoring to comprehensive performance analytics, not because it is technically impressive, but because it solves real engineering problems.
The data and tools exist to make this transition. The question is whether asset owners will embrace the analytical approach necessary to unlock their assets’ full potential.
The renewable energy sector continues evolving from a focus on installation to optimisation. Success increasingly depends on operational excellence and the systematic identification of performance gaps that conventional monitoring cannot detect.
Meet the Author
George Cooper
Founder, Lamina
As the founder of Lamina, George specializes in applying advanced data science and machine learning to solve the renewable energy sector’s most difficult performance challenges.
Drawing on his foundation as a First Class Honours Renewable Energy Engineer, he develops bespoke analytical models that empower asset managers to de-risk their portfolios and unlock hidden revenue.