How to Interpret PVSyst Results That Seem Too Low: 4 Approaches｜How to Identify Causes

Table of Contents

• What it means when PVSyst results seem too low

• First check assumptions, not annual energy

• Approach 1: Check irradiance and meteorological data

• Approach 2: Check array losses and temperature losses

• Approach 3: Check wiring losses and equipment losses

• Approach 4: Check site conditions such as shading, soiling, and snow

• Typical causes for PVSyst results appearing low

• Readings to avoid when results are low

• Points to check in comparison reports

• How to think when explaining low results

• Summary

What it means when PVSyst results seem too low

When you run a PVSyst simulation for a solar power plant, you may encounter situations where the annual energy is lower than expected, the PR is low, a particular loss item is large, or your results look worse compared with another company's report.

The important point is not to simply judge PVSyst results as “low” or “high.” PVSyst outputs result from a stack of assumptions: input conditions, meteorological data, modules, PCS, wiring, shading, soiling, temperature, output limits, grid-side conditions, and more. So a low-looking result may simply reflect conservative inputs, or it may be caused by an input error that introduced unnecessary losses.

When PVSyst results seem too low, you should not only look at the annual energy but step through where the losses occur. In particular, separate whether the cause is low irradiance, large losses at the array, losses after the PCS, or strong site effects such as shading or soiling.

In PVSyst reports you can follow the reasons for low results in detail by checking not only the final annual energy but Specific production, Performance Ratio, Normalized productions, the Loss diagram, Monthly values, and so on. When reading PVSyst, do not stop at the final figure; adopt the mindset of decomposing where generation drops.

First check assumptions, not annual energy

When PVSyst results feel too low, many people first look at annual energy or PR. Of course annual energy is important as the final deliverable, but it is not the first place to look when investigating causes.

The first thing to check is the project assumptions. Confirm that plant capacity, module capacity, PCS capacity, array configuration, tilt angle, azimuth, installation site, meteorological data, ground reflectance, grid connection point, presence of output limits, and similar settings are correct.

For example, if the DC capacity is entered smaller than expected, annual energy will naturally be low. If PCS capacity is small and many clipping events occur, generation will be suppressed. If azimuth or tilt differs from reality, the way the array receives irradiance changes and can strongly affect results. If the meteorological data location is wrong or altitude and temperature conditions differ, those can also lead to low results.

When reading PVSyst, it is important to read inputs together with outputs rather than just viewing the results screen. Instead of concluding “PVSyst produced a low value,” ask “which condition is producing this low result?”

When comparing to other companies’ reports in particular, simply comparing annual energy is meaningless. If meteorological data, loss conditions, the DC/AC ratio, PCS rating, output limits, wiring losses, transformer losses, auxiliary losses, and shading conditions are not aligned, different results are to be expected. For PVSyst comparisons, first confirm that you are comparing on the same basis.

Approach 1: Check irradiance and meteorological data

When PVSyst results seem too low, the first thing to check is irradiance. The foundation for PV generation simulation is the meteorological data. No matter how high-performance the modules or PCS are, if input irradiance is low, generation will be low.

In PVSyst check values such as Global horizontal irradiation, Global incident in collector plane, and Effective global after losses. By checking whether horizontal-plane irradiance is low, whether irradiance converted to the tilted plane is low, or whether irradiance is reduced after shading and IAM, you can pinpoint where the issue lies.

For example, if horizontal-plane irradiance itself is low, the choice of meteorological data may be the cause. Annual irradiance varies depending on whether you use Meteonorm, SolarGIS, satellite data, or nearby station data. In mountainous, coastal, snowy, or fog-prone regions the data source can substantially affect results.

If horizontal irradiance is reasonable but tilted-plane irradiance is low, check tilt angle, azimuth, and the transposition model. Cases where tilt is set less favorably than reality, azimuth is off, or east–west and low-tilt layouts are not correctly reflected can lead to lower-than-expected irradiance on the panel surface.

If the tilted-plane irradiance is sufficient but Effective global after losses is low, suspect near-field shading, far-field shading, IAM, soiling, or snow. By separating irradiance stages you can distinguish “the irradiance coming from the sky is low,” “irradiance is being reduced before reaching the panel surface,” or “irradiance reaches the panel surface but is not being effectively utilized.”

It is also important to check monthly values. Annual totals alone do not show which seasons are driving the low result. If only winter is extremely low, snow or low solar altitude may be the reason. If summer is low, temperature losses, output limits, or PCS clipping may be at work. If low irradiance strongly appears during the rainy season or winter, check the regional characteristics of the meteorological data.

When PVSyst results look low, examining irradiance assumptions is the first step. Generation stems from irradiance, so focusing solely on losses without verifying irradiance assumptions can lead to incorrect conclusions.

Approach 2: Check array losses and temperature losses

If irradiance is reasonable but PVSyst generation is still low, next check array-side losses. Array losses are the losses that occur in the process of converting sunlight into DC power at the module. In PVSyst’s Loss diagram these include module quality loss, LID, mismatch, IAM, temperature loss, low-irradiance loss, soiling, shading, and others.

Temperature loss is especially influential. PV module output decreases as cell temperature rises. In high-temperature regions, installations with poor ventilation, rooftop mounting, low racking, or weak rear-side airflow tend to have larger temperature losses. In PVSyst the Thermal loss factor setting affects results. If Uc or Uv values are set conservatively, calculated cell temperature will be higher, producing lower generation.

Note that a large temperature loss does not always indicate an input error. If the site is hot and the module mounting does not dissipate heat well, large temperature loss is natural. Conversely, if the site is ground-mounted with adequate ventilation but you modeled harsh rooftop-like thermal conditions, the settings may be overly conservative.

Check mismatch losses too. Mismatch increases with variations between strings, module characteristic differences, shading patterns, and differences in azimuth or tilt. If you have combined multiple azimuths or tilts into a single sub-array in PVSyst, you may obtain losses that do not reflect reality. In complex array configurations, how you split sub-arrays affects results.

IAM losses are another common cause of low results. IAM is the loss due to increased reflection at the module surface at large incidence angles. IAM becomes notable for low-tilt or east–west layouts, low winter solar altitude, or designs with high morning/evening generation shares. IAM values also vary with glass specifications and module characteristics, so verify the chosen module data.

Low-irradiance loss can be overlooked. In cloudy regions or regions with a high proportion of low-irradiance hours, module efficiency under low irradiance affects generation. Check that the module’s PAN file is correct and that you are not using outdated or approximate data.

When reading array losses, don’t just compare loss rates; check whether each loss matches site conditions. Large temperature loss is natural in hot regions. Large mismatch or shading losses are natural with complex shadows. But if a loss is disproportionately large relative to site conditions, reconsider that setting.

Approach 3: Check wiring losses and equipment losses

When PVSyst results are low, also check DC- and AC-side wiring losses, PCS losses, transformer losses, and auxiliary losses. Each may appear as only a few percent, but combined they significantly affect annual generation.

DC wiring loss occurs on the DC side from modules to combiner boxes and from combiner boxes to the PCS. It varies with cable length, cable size, current, voltage, and string configuration. In wide plants with long runs to PCS or where cable sizes are undersized, losses increase. Conversely, if PCS units are distributed near racks, overestimating DC wiring loss can make results look low.

AC wiring loss is similar. If the distance from PCS to substation or point of interconnection is long, AC-side losses increase. The effect is particularly large for low-voltage or long-distance cabling. In PVSyst check where DC, AC, and MV losses are applied to ensure you are not double-counting the same loss.

Transformer losses are also important. Transformers have load-dependent losses and no-load losses, and losses may occur even during standby. Conservative transformer loss assumptions in PVSyst will affect annual energy and PR. For plants with multiple transformers, verify that the actual configuration matches the PVSyst model.

For PCS losses, check that the PCS efficiency curve is correctly set. PCS do not always operate at peak efficiency. Efficiency changes with low load, high load, temperature, and input voltage range. If the PCS model or approximations are inappropriate, results will differ.

Also check the DC/AC ratio relative to PCS capacity. If PCS capacity is small relative to DC capacity, output will cap during strong irradiance and clipping losses will occur. Clipping may be an intentional design choice and not necessarily bad. But if clipping is larger than expected, review DC/AC ratio, PCS rating, power factor settings, and output limit settings.

Power factor settings are often overlooked. How PVSyst handles power factor can change interpretations of active power limits and apparent power. Even if the grid specifies a power factor, determine whether that acts as an active power limit or as a constraint relative to PCS apparent power capacity. Misunderstanding here can make PVSyst results appear low.

Check auxiliary losses as well. When you include monitoring equipment, air conditioning, trackers, PCS standby power, and auxiliary power for substation equipment, these are deducted from annual generation. Auxiliary loss assumptions vary by methodology and often cause discrepancies when comparing reports.

When checking wiring and equipment losses, confirm not only the magnitude but which equipment segments they correspond to. Without organizing whether losses apply to DC wiring, AC wiring, MV wiring, transformers, PCS, or auxiliaries, it is difficult to detect double counting or omissions.

Approach 4: Check site conditions such as shading, soiling, and snow

When PVSyst results seem too low, always check losses related to site conditions. Representative items are shading, soiling, snow, topography, nearby structures, vegetation, and rack spacing. These are highly site-specific and hard to judge solely from default values.

Shading losses include far-field and near-field shading. Far-field shading is caused by mountains, hills, buildings, and forests blocking the sun. Near-field shading comes from front-row racks, nearby equipment, utility poles, fences, inverters, trees, and so on. In PVSyst the way you model shading scenes, reflect terrain, specify rack spacing, configure module layout, and handle electrical shading affects results.

If shading losses are large, first examine monthly and hourly impacts. If the loss is large only in winter mornings and evenings, the cause differs from losses that are large year-round. Winter-only issues suggest low solar altitude or inter-row shading. Morning-only or evening-only losses suggest east–west obstacles or terrain features.

Soiling losses also reduce PVSyst results. Dust, pollen, yellow sand, bird droppings, dust from nearby fields, industrial or road dust, and low rainfall frequency affect soiling. Soiling varies widely by region and operating practices; whether a cleaning schedule exists also matters. Conservative, higher soiling assumptions reduce annual energy.

In snowy regions, snow losses are critical. In Hokkaido, Tohoku, the Japan Sea side, and mountainous areas, winter snow can severely reduce generation. How PVSyst handles snow can cause large differences depending on analysis settings. You may model snow as simple monthly losses or estimate it from meteorological conditions and tilt angle. Results that account for snow will naturally be lower than those that do not.

However, if results are low because snow losses were included and those snow assumptions reflect reality, that is not a problem. In snowy regions failing to include snow losses can produce results that are overly optimistic relative to reality.

Check ground reflectance, or albedo, too. In typical conditions albedo has limited influence, but during snow cover albedo rises and can boost tilted-plane irradiance. How you treat winter albedo when you also include snow losses affects winter generation estimates.

For site-condition losses, it is especially important to explain the rationale. If you specify a percentage for soiling loss, which months and how much snow loss you included, and how far shading modeling extended, you must be able to explain these choices; otherwise it will be difficult to convince clients or internal stakeholders why the results are low.

Typical causes for PVSyst results appearing low

Causes for PVSyst results appearing low can be grouped into four major categories: input errors, conservative assumptions, site characteristics, and differences in comparison conditions.

Common input errors include capacity, azimuth, tilt, number of PCS units, string configuration, cable lengths, loss rates, and meteorological data location. For example, if azimuth reference is misentered, an array that should be nearly south-facing might be treated as east–west, reducing generation. Entering the wrong tilt also changes tilted-plane irradiance.

Conservative assumptions include higher soiling, wiring losses, temperature conditions, auxiliary losses, transformer losses, and snow losses. For financial analyses or long-term forecasts, conservative assumptions are natural. But if comparison targets are based on optimistic assumptions, your results alone will appear low.

Site characteristics include low-irradiance regions, high-temperature regions, snowy regions, locations with mountain shadows, complex terrain, narrow inter-row spacing, or many nearby obstacles. In these cases low results reflect reality, not a PVSyst fault.

Differences in comparison conditions are another frequent cause. Another company’s report might omit soiling; yours includes it. Another may not consider snow; yours does. Another might only include DC wiring losses, while yours includes AC and transformer losses. Such differences naturally produce different results.

When PVSyst results appear low, first confirm whether they are truly low. Check whether the result remains low when aligned to the same comparison assumptions, or whether adjusting for assumption differences brings results into a reasonable range.

Readings to avoid when results are low

What not to do when PVSyst results are low is to change settings intuitively based only on the final annual energy. For example, lowering soiling loss, reducing wiring loss, easing temperature assumptions, or lessening snow loss to raise annual energy without justification undermines the simulation’s credibility.

PVSyst is not a tool for producing desired energy outputs but for estimating generation based on inputs. When results are low, find the cause and verify whether the setting is reasonable rather than steering inputs to increase energy.

Also avoid judging solely by PR. PR is a convenient index for comparing plant performance but is influenced by irradiance, temperature, output limits, snow, soiling, albedo, PCS capacity, and other factors. A low PR does not necessarily mean poor design. Realistic inclusion of snow or output limits can make PR appear low.

When comparing with other reports, it is dangerous to line up PRs without context. PR changes if auxiliary or transformer losses are included, whether energy is referenced at the point of interconnection, or whether it is referenced at the PCS output. Monthly tendencies should not be ignored. The cause of low annual generation may be concentrated in specific months—winter snow, summer temperature losses, spring yellow sand or pollen, or rainy-season low irradiance. Judging only from annual totals makes it easy to miss causes.

Points to check in comparison reports

When results seem low you are often making a comparison—with past internal analyses, other companies’ reports, measured performance, simplified calculations, or assumed revenue figures. Aligning comparison conditions is most important.

First check DC and AC capacities. Even for the same plant these can be defined differently: module nominal capacity, PCS input units, or sum by area. AC capacity can be PCS rated, interconnection capacity, contracted capacity, or effective active power after power factor. Clarify these definitions.

Next check the energy evaluation point. Whether the annual energy is at module output, PCS input, PCS output, after the transformer, at the grid connection, or at the revenue meter changes values. Confirm what E_Grid in the PVSyst report represents and whether it matches the comparison.

Meteorological data matters. At the same location, different meteorological datasets can change annual irradiance by a few percent. Small energy differences may be explained solely by meteorological data choices. When comparing, align horizontal and tilted-plane irradiance and monthly temperature values.

Compare loss conditions item by item: soiling, snow, IAM, temperature, mismatch, DC wiring, AC wiring, PCS, transformer, auxiliaries, shading, and output limits. Be careful: PVSyst item names may be the same while the scope differs.

For example, “wiring losses” could mean module-to-combiner losses, combiner-to-PCS losses, or PCS-to-transformer losses. Transformer losses may include load losses only or also no-load losses—this changes annual totals.

In comparison reports it is clearer to explain differences by breaking down the final energy gap into irradiance differences, array-loss differences, system-loss differences, and other-loss differences. Decomposing differences is key when explaining why PVSyst results are low.

How to think when explaining low results

When PVSyst results are low, simply saying “we used conservative assumptions” is insufficient for internal or client explanations. You must explain which conditions were set, why, and on what basis.

For example, if large snow losses cause the low generation, explain winter meteorological conditions, actual on-site snow behavior, tilt angle, whether snow removal is planned, and past records. For large soiling losses, explain the surrounding environment, cleaning frequency, rainfall, and local dust sources. For large wiring losses, justify with cable lengths, voltage, current, conductor size, and equipment layout.

When explaining, lead with the main cause of the low-seeming result, then supplement with irradiance, array losses, system losses, and site conditions. Starting by listing every small loss confuses the audience about what matters most.

Also note that a low result is not always bad. For business planning or due diligence for lenders, realistic generation that reflects site conditions is often more valuable than overly optimistic figures. A low PVSyst result can be a cautious, risk-aware assessment.

However, conservative assumptions without basis are different from justified conservatism. Unfounded conservatism can make a project’s profitability appear worse than reality. When reading PVSyst distinguish why a conservative setting is necessary and whether the chosen value is within a reasonable range.

When explaining results, organize not only the final annual energy but which losses are abnormally large and which are within typical ranges. This increases credibility. When differences appear versus other analyses, clearly identify the items responsible for the difference.

Summary

When PVSyst results seem too low, do not judge based solely on annual energy. PVSyst outputs result from meteorological data, design conditions, loss assumptions, and site conditions—so check in order where generation is falling.

First check irradiance and meteorological data. If input irradiance is low, generation will be low. Separate horizontal irradiance, tilted-plane irradiance, and effective irradiance, and check monthly trends.

Next check array losses and temperature losses. Verify temperature conditions, IAM, low-irradiance loss, mismatch, and module data selection. Temperature losses vary strongly by region and installation, so ensure consistency with site conditions.

Then check wiring and equipment losses. Verify DC wiring, AC wiring, PCS, transformer, auxiliaries, output limits, and power factor. Watch for double counting of losses and differences in evaluation point.

Finally check site-specific losses such as shading, soiling, and snow. These are highly site-dependent and can strongly affect PVSyst results. In snowy or mountain-shadowed locations, you must be able to explain why results are lower based on site conditions.

Having low PVSyst results is not inherently a problem. The problem is being unable to explain why they are low. If low results reflect site reality, they are valuable for business decisions. If they result from input errors or double counting, correct them promptly.

When reading PVSyst, focus on correctly decomposing the reasons for reduced generation—not on tweaking inputs to improve the result. By dividing checks into irradiance, array losses, system losses, and site conditions you can concretely identify causes even when results appear too low.