Impact of PV module degradation on inverter clipping losses

Read full document

Overview

In the PV industry, it is standard practice to perform detailed energy yield simulations for the first year of a power plant's operation. However, as PV modules age, their performance naturally declines - a process known as degradation. Recent research by Solargis, presented at the European PVPMC workshop, highlights how this long-term degradation influences inverter clipping losses (CLs) and overall energy yield predictions over a typical project lifetime.

Methodology

To understand the interaction between aging and power limitation, we analyzed several sites using DC/AC ratios of 1.25 and 1.50. To capture the nuance of solar variability, the simulations utilized two types of satellite-based data: Typical Meteorological Year (TMY) and Multi-year Time Series (TS). These were processed at both 15-minute time steps and 1-minute time steps to evaluate the impact of temporal resolution on the results.

Degradation was modeled as 0.8% at the end of the first year and 0.5% for each subsequent year. The research compared three calculation methods:

1. Full TS simulation: Degradation applied directly to each step of a 20+ year sub-hourly time series.

2. TMY repetition: Repeating a single TMY year for the project lifetime with degradation applied annually.

3. Post-processing: Applying a linear degradation factor to the Long-Term Average (LTA) yield - the current industry standard.

Key research findings

The study revealed that the relationship between degradation and energy loss is not as straightforward as once thought.

Non-linear Reduction of Clipping

While PV module degradation is linear, the resulting decrease in clipping losses is non-linear. As the modules degrade, the DC array effectively "shrinks" relative to the inverter capacity. This causes clipping losses to drop significantly faster in the early years of a project than in the later years.

The Role of Temporal Resolution

The research found that using a 15-minute time step (or hourly data) often underestimates clipping losses compared to a 1-minute time step. This is because high-frequency fluctuations in irradiance are smoothed out in lower-resolution data, even though they would have triggered an inverter's power limitation in reality.

Figure 1: CLs throughout expected plant lifetime with different input data and different DC/AC ratios. Note the differences between the flat profile of TMY-based analysis and variable profile of TS-based analysis.

Why standard industry practice may be inaccurate

Many practitioners use a "post-processing" approach: they calculate the first-year yield and then subtract a flat annual degradation percentage (e.g. -0.5% per year) for the rest of the project life.

By simply subtracting a flat degradation percentage from a first-year simulation, designers fail to account for the fact that a degrading system actually "reclaims" some energy that would have otherwise been clipped. Failing to model this interaction can lead to an overestimation of future yield losses - and an underestimation of actual energy production - by as much as 3.6%.

Recommendations for PV professionals

To ensure the highest accuracy in financial and technical assessments, Solargis recommends the following:

• Use sub-hourly data: Utilize data with 15-minute or 1-minute time step for clipping analysis.

• Integrate degradation: Run simulations that incorporate degradation directly into the simulation chain, calculating is specifically for each day of the analyzed period.

• Site-specific analysis: Avoid global assumptions, as the impact varies based on local climate and DC/AC ratios.