The TMY Blind Spot: Misrepresented temperature and its effect on PV

Do your simulations account for the actual temperature conditions at your PV plant site? If you use Typical Meteorological Year (TMY) datasets as inputs, the short answer is “not really.”

Although a Typical Meteorological Year (TMY) dataset is not meant to capture extreme events, it may also miss conditions that are frequent enough for many people to consider them "typical".

For a PV project, getting temperature conditions right matters to make sure that the expected losses are accurate and plant components can handle the added heat stress.

By analyzing both Solargis time series and Solargis TMY data for a sample site near Petrolina, Pernambuco, Brazil, we show how TMY misrepresents high temperatures over the long term, and how this can undermine trust in the PV project and introduce unnecessary risks.

高温下における太陽光発電システムの性能

温度は、いくつかの点で太陽光発電システムの性能に影響を与えます。太陽光発電モジュールは、吸収した太陽光のすべてを電気に変換するわけではなく、モジュールの温度が上昇するにつれて効率が低下します。

さらに、温度が規定の動作限界を超えると、インバータは出力を低下させることがあり、これは「熱ディレーティング」として知られています。また、温度が高くなると、変圧器やケーブルの抵抗損失も増加し、システム全体の性能にさらなる影響を及ぼします。

Why TMY fails to represent the full picture 

The TMY (Typical Meteorological Year) dataset is designed to cover the conditions most likely to occur at a site. It is constructed by selecting the most statistically representative intervals from a long historical record and assembling them into a single synthetic year. In this process, the selection criteria prioritize irradiance over other parameters, such as temperature.

High temperatures are fairly frequent in some climates. However, some temperature ranges may not be sufficiently reflected in the TMY dataset although the share of records reaching such values can be quite significant. We call these underrepresented situations the “TMY blind spot.”

The problem is that summarizing many years of data into a single year poses risks to the power plant. In other words, a TMY can fail to represent the days that matter most for your PV plant’s thermal losses: those on which your PV components run hottest. 

The Petrolina case study below shows this in more detail. We see how the "blind spot" is situated for air temperatures above 33ºC-35ºC, where TMY "hides" the highest number of equivalent hours in comparison to other thresholds.

Another side of the “TMY blind spot” worth noting: summarizing many years of data into one year completely erases the possibility of knowing the expected range of variability for each month.

The consequences of unexpected yield losses

Temperature loss is an expected, well-understood effect on PV plants operating in a hot climate. However, if the input dataset overlooks the probability of high-temperature conditions and if the variability range is not correctly reported, the consequences can seriously affect stakeholders’ trust in the project.

From a developer’s perspective, it can mean discovering that yield reports did not reflect actual operating conditions and, worse, that different design decisions might have been made with a more appropriate dataset, including key aspects such as module technology, mounting ventilation, and row spacing. From an investor’s perspective, it can mean losing trust in the project's ability —and the teams involved— to deliver the promised returns.

After a misrepresentation of the conditions determining the PV plant performance is identified, some design inputs may need to be reviewed:

• Sizing of protections. Overheated components can cause string fuses to blow when the manufacturer's rated values are exceeded.
• Anticipated degradation. Thermal degradation mechanisms depend on cell temperature exceeding certain thresholds for sustained periods.

If verified generation systematically falls below the expected level during the months with the worst thermal combination, it can also create unnecessary contractual stress:

• Revision of guaranteed energy. The project may face a review process that reduces the amount of energy that can be marketed in the regulated market.
• Exposure under free-market PPAs. The generator may incur penalties or be required to buy energy on the spot market to cover the shortfall, precisely in the months when spot prices also tend to be higher, because the same regional pattern is affecting other solar generators on the same system.

Time series simulation to avoid data representation issues

When TMY is the main input of the simulation, stress situations under high temperatures can become invisible to the calculated results. These underrepresented situations can create gaps between the expected temperature loss in the yield report and what the plant will actually experience.

Although results using TMY can cancel out and end up relatively close to the time series for the annual average, the TMY always carries the risk of misrepresenting the effect on component stress and derating. Moreover, it completely hides the monthly range of expected losses that help stakeholders balance their expectations and optimize energy contracts accordingly.

The good news is that, beyond raising awareness of the problem, closing that gap is today possible by using simulation software capable of running the full multiyear record, not a single representative year. This is already within reach for developers and certifiers. 

Working with a better representation of the power plant's operating conditions opens the door to obtaining better-designed PV plants. In addition, a more accurate calculation of losses creates an improved framework for designing contracts and adjusting the expectations of investors and owners.

What is missing is neither the data nor the software: it is the methodological update the industry needs to close the calculation gaps that put trust in PV projects at risk.