Six Checks to Judge Whether PVSyst Results Can Be Trusted

Table of Contents

• Key points to grasp first so you don't take PVSyst results at face value

• Check 1: Can you organize what assumptions the results are based on?

• Check 2: Are the meteodata and the installation site consistent?

• Check 3: Are the shading conditions and 3D scene natural relative to the site?

• Check 4: Are the system configuration and electrical conditions reasonable?

• Check 5: Do the loss settings and the appearance of PR look natural?

• Check 6: Have you cross-checked results by comparing alternatives and using on-site intuition?

• Summary: Linking PVSyst results to practical decision-making

Key points to grasp first so you don't take PVSyst results at face value

PVSyst is a very powerful simulation tool for planning and comparative evaluation of photovoltaic systems. Because it lets you organize and check annual generation, monthly generation, PR, the loss tree, the impact of shading, and the plausibility of the system configuration, it has become indispensable for practitioners. However, precisely because it is convenient, there is a pitfall: it is easy to accept the numbers shown on the screen as correct without sufficient scrutiny. PVSyst is a tool that computes results based on the assumptions you input, and if those assumptions are vague, even plausible-looking results can be difficult to use in practice.

What really matters in practice is not only whether the result is high or low. It is being able to explain under which conditions that result was produced. For example, even if the annual generation looks high, if the meteodata are optimistic, the shading assumptions are lenient, or the loss settings are too light, that figure can easily fall apart later. Conversely, if the annual value is somewhat conservative but the result carefully incorporates on-site conditions and maintenance assumptions, it can be stronger as a basis for investment or design decisions. In other words, whether the results are trustworthy is determined not by how attractive the numbers look but by the consistency of the assumptions.

Also, it is easier to judge the reliability of PVSyst results by navigating among related screens and settings than by reading a single output alone. If annual generation appears higher than expected, check the loss tree and PR breakdown to see whether the reason for the higher value is reasonable. If it is low, inspect where the dominant influences lie: shading, temperature losses, PCS settings, wiring losses, or utilization losses. Although a result may look like a single number, many pieces of information underlie it. In practice, the ability to read those underlying factors is extremely important.

Here we summarize six minimum checkpoints practitioners should verify to judge whether PVSyst results can be trusted. None of them are difficult theory-wise, but they are often overlooked in practice and can greatly affect the reliability of the results. By checking them in order, you move beyond just looking at calculation results and get closer to being able to confidently use those numbers in practical decision-making.

Check 1: Can you organize what assumptions the results are based on?

The first check is very basic but also the most important: confirm whether you can organize what assumptions the result is based on. PVSyst has many settings, and a common practical situation is that after tweaking multiple conditions a bit here and there, it becomes unclear which conditions are fixed and which are provisional. Numbers produced in that state may look tidy but are often hard to trust.

For example, it is common in practice to run simulations where the site location is provisional, the azimuth is an ideal value, the tilt is based on pre-construction assumptions, shading is set simply, and loss coefficients are reused from previous projects. That approach is not inherently wrong, but in such cases you must clearly state whether the output is for rough estimation or for detailed analysis. Treating a rough estimate as a detailed-design figure is risky, and if many provisional assumptions remain when you intend a detailed study, the basis for decisions will be weak. First, you need to be able to explain what level of assumption accuracy this result was created with.

Also, when creating comparative options, clarify what variables are actually being compared. For example, if you want to see the effect of different PCS options but you also change azimuth or shading at the same time, the meaning of the differences becomes ambiguous. If you aim to evaluate meteorological differences but you also alter loss coefficients, it becomes unclear how much of the difference is due to weather. Trustworthy results are not the ones with the highest numbers but those in which it is clear what was compared and what was held constant. If that is ambiguous, you will inevitably struggle when explaining the results later.

Therefore, before looking at PVSyst results, it is effective to briefly organize what decision the calculation is intended to support. Clarify whether it is a candidate-site comparison, a layout comparison, a PCS comparison, or a check of shading conditions, and keep conditions irrelevant to that purpose as consistent as possible. Whether a result can be trusted is more influenced by the clarity of assumption organization than by the complexity of the calculation.

Check 2: Are the meteodata and the installation site consistent?

Next, check the consistency between the meteodata and the installation site. PVSyst results are strongly affected by meteorological conditions, so no matter how carefully you configure the system, if the meteodata do not represent the actual site well, the reliability of the results declines. Moreover, discrepancies in the meteorological basis are not easy to spot visually and may only appear as a subtle mismatch in annual generation or PR. Therefore, before doubting the system configuration, first verify this foundation.

In practice, you may use data from the representative point closest to the candidate site or reuse data from a nearby past project. This is a natural approach for initial studies, but whether that point truly represents the current site is another matter. Irradiance and temperature conditions vary with coastal versus inland location, elevation differences, whether the surrounding terrain is open or enclosed by mountains, etc. If the meteodata are off, the apparent temperature losses and the monthly output profile will also be skewed.

You should also check monthly distributions, not just annual totals. Even if annual irradiance is similar, differences in monthly patterns can make one option look worse in winter or show differences in summer. Mistaking these differences for design-derived effects can lead to unnecessary setting changes. To distinguish meteorological differences from design differences, it is effective to check monthly generation biases together with the meteodata assumptions.

Therefore, when judging whether results are trustworthy, first confirm how the site location was chosen and how the meteodata correspond to it, and consider to what extent that point represents the actual site. If necessary, compare other nearby or representative datasets to see whether the numbers fluctuate dramatically. Remember that PVSyst results are, before being an equipment calculation, an expression of the meteorological assumptions.

Check 3: Are the shading conditions and 3D scene natural relative to the site?

The way you model shading conditions and the 3D scene also greatly affects result reliability. PVSyst lets you set Near Shading and 3D scenes in detail, but whether those inputs appropriately represent the site without omission or excess can change impressions of annual generation and PR substantially. Both over-modeling and under-modeling obstacles are problematic; the most important thing is whether the main causes of shading are represented naturally. Trustworthy results come not from visually sophisticated 3D scenes but from scenes that realistically capture the necessary elements.

For example, if a project has significant self-shading between front and back rows but the row spacing is modeled too optimistically, annual generation will look optimistic. Conversely, if you model obstacles that are not actually problematic, shading losses may appear unnaturally large and generation will look excessively low. If building or tree heights, positional relationships, or slope orientations are entered ambiguously, the impacts during winter or at sunrise and sunset can deviate from reality. Because shading assumptions are visually persuasive, they are especially prone to hidden mismatches.

Further, shading loss often cannot be judged by area ratios alone. The electrical impact varies depending on which row, which string, or which block receives partial shading. A visually small shadow can have a large effect if it concentrates on a particular grouping. Conversely, a visually large shadow may have little annual impact if it occurs only during limited times or seasons. To judge whether PVSyst results are trustworthy, you need to check not only the total amount of shading but also the character of how it occurs.

Therefore, when reviewing the 3D scene and shading conditions, confirm whether the elements likely to be problematic on-site are properly reflected in the model. Check that buildings, slopes, front-and-back rows, trees, aisles, and edge clearances are not grossly different from on-site intuition. Trust in PVSyst shading assumptions should be based on whether the main on-site conditions are naturally represented, not on how finely the model is crafted.

Check 4: Are the system configuration and electrical conditions reasonable?

The naturalness of system configuration and electrical conditions is also indispensable when judging result reliability. In PVSyst, you can input module counts, PCS capacity, string configuration, and DC/AC ratio to produce results. However, a calculation succeeding does not mean the design is free of practical issues. Even if the numbers look clean, unless the configuration can be realized on-site without difficulty and is electrically reasonable, the significance of the results will be weakened.

For example, if an option increases annual generation by heavily overloading the DC side, check whether the PCS-side power limitation becomes unreasonably large. Conversely, there are options that leave excessive PCS margin and fail to use the module potential. For string configuration, ensure shaded and unshaded rows are not grouped in the same string, and that zones with different azimuths or tilts are not forced into the same grouping. Such configuration mismatches are hard to spot from annual generation alone but can greatly affect result reliability.

Also, don’t overlook wiring and zoning coherence. A layout that is awkward in array arrangement and forces unrealistic aisle or maintenance conditions tends to have optimistic assumptions for wiring losses and utilization losses. In practice, designs that are tidy and naturally realizable are easier to explain and less likely to collapse later. PVSyst is not only a tool to compare equipment performance but also to check whether the equipment configuration integrates naturally with site conditions.

Therefore, when judging result reliability, look at modules, PCS, strings, wiring, and zoning as a single system and confirm there are no unreasonable compromises. Ask not only whether the numbers hold but whether the design is natural. Whether PVSyst results can be trusted depends more on whether the overall system coheres sensibly than only on the correctness of individual input values.

Check 5: Do the loss settings and the appearance of PR look natural?

Consistency between loss settings and the appearance of PR is another major point for judging result credibility. PVSyst allows you to set various losses such as temperature, soiling, wiring, mismatch, and utilization losses. How appropriately these are placed can significantly change impressions of annual generation and PR. In practice, be careful that although individual losses may look reasonable, stacking them all together can result in being excessively conservative.

For example, if you realistically model shading but also layer other loss items that impose similar disadvantages, the result may look needlessly low. Conversely, if shading and temperature conditions are severe but soiling and utilization losses are set too lightly, the overall balance may look plausible while not matching on-site conditions. When reviewing loss settings in PVSyst, check not just each number but whether the overall loss structure looks natural.

Also check how PR presents itself. PR is a single number but encompasses many losses. If shading is severe yet PR is extremely high, some loss settings may be too light. Conversely, if shading and temperature are realistically modeled and PR is slightly low, that figure may actually be reasonable. In practice, do not judge by PR level alone; judge whether it aligns with the loss settings.

Therefore, when checking whether PVSyst results are trustworthy, review the loss tree and PR together to confirm there are no unnatural highs or lows or duplicated unfavorable assumptions. If you have carefully set loss parameters yet PR is oddly high — or if small condition differences make PR swing drastically low — it is worth rechecking the input assumptions. The reliability of results should be assessed by the overall consistency of loss settings, not by the granularity of their detail.

Check 6: Have you cross-checked results by comparing alternatives and using on-site intuition?

The final check is whether you have reverse-checked results by comparing alternatives and using on-site intuition. PVSyst outputs can look very tidy at first glance, and when looking at a single option it is tempting to trust those numbers as-is. However, in practice it is safer to compare the baseline with alternatives where some conditions are slightly altered and see whether the differences are natural. Also, comparing results with on-site impressions helps you detect mismatches in assumptions.

For example, comparing options with slightly wider versus standard row spacing, slightly different azimuths, or different PCS approaches shows which conditions strongly affect results. If differences barely appear, that condition may not be dominant. If small changes produce large differences, that setting is a highly sensitive issue. Using PVSyst in practice means observing these sensitivities to set design priorities.

Cross-checking with on-site intuition is also indispensable. For example, if the site clearly seems to have strong morning shading but the results show almost no impact, there may be discrepancies in building locations, elevation differences, or 3D scene modeling. Conversely, something that looks very unfavorable on paper may not be so bad when considering on-site spacing and distances. Because PVSyst follows its inputs exactly, you need to iterate between the model and site impressions to validate results.

Therefore, when judging whether results are trustworthy, do not be satisfied with a single annual generation or PR figure. Always compare alternatives and check them against on-site conditions. If you can trace which setting caused the numerical difference and confirm it matches on-site intuition, the result becomes much more trustworthy. To use PVSyst effectively in practice, cultivate the habit of doubting and verifying results, not only looking at them.

Summary: Linking PVSyst results to practical decision-making

What the six checks above have in common is that PVSyst results should not be treated as mere final numbers. Only by including organization of assumptions, meteodata consistency, the naturalness of shading and 3D scenes, the reasonableness of system configuration, the consistency of loss settings and PR, and alternative comparison with on-site intuition does a result become a number you can use for practical decisions. None of these checks require advanced theoretical knowledge; they demand how thoroughly you can trace the background behind the numbers.

What really matters to practitioners is not producing the highest generation or PR. The truly valuable outcome is being able to explain why a number came out as it did and ensuring the assumptions do not contradict site realities or design intent. Do not judge simply because a number is low or high; assess how naturally that number arises from realistic conditions. PVSyst should be used not as a tool to compete for idealized values but as a tool to organize convincing design assumptions.

Also, to truly improve result reliability, do not stop at desk simulations. It is necessary to confirm that site boundaries, slopes, aisles, buildings, trees, maintenance routes, and existing conditions visible on-site are connected to PVSyst settings. If on-site understanding is weak, even the most precise simulation is of limited value. Conversely, if on-site understanding and calculation results are linked, the persuasive power of the numbers increases dramatically.

In that sense, when you want to make site position checks and coordinate acquisition more reliable, it is also effective to use high-precision GNSS positioning devices that attach to an iPhone, such as LRTK. If you can more easily organize on-site position information and site conditions, the azimuth, layout, shading, and aisle conditions you set in PVSyst become closer to reality. If PVSyst improves desk-level comparison precision and LRTK supports precise on-site understanding, simulation results cease to be mere numbers and become practical, site-rooted decision-making material. The ability to discern whether PVSyst results can be trusted not only improves the accuracy of generation forecasts but also develops practical capability to connect desk work and field work.