8 things to check when PVSyst results seem incorrect

Table of Contents

• Initial approach to take when PVSyst results seem incorrect

• Check item 1: Are the project site and meteorological data consistent?

• Check item 2: Do the azimuth and tilt angle inputs match the on-site conditions?

• Check item 3: Is the combination of module capacity and PCS capacity correct?

• Check item 4: Are the string configuration and voltage range feasible?

• Check item 5: Are the shading settings neither overestimated nor underestimated?

• Check item 6: Are loss factors not overapplied or omitted?

• Check item 7: Have you mistaken the interpretation of monthly results for annual results?

• Check item 8: Have you aligned the comparison conditions before report output?

• Verification of PVSyst results is greatly affected by the accuracy of on-site conditions

What to check first when PVSyst results seem wrong

One thing beginners often stumble over when using PVSyst is that they look at the generated energy and PR shown on the results screen and immediately judge them as "too high" or "too low." Of course, clearly anomalous values can appear, but the first important step is to clarify what those numbers are being compared against. Depending on whether they differ from past projects, from in-house rough calculations, from simulations under different conditions, or from on-site measured values, what you need to check will vary.

For example, even if the annual energy production is higher than expected, it is natural for the production to be larger if you selected a location whose meteorological data shows a high annual average solar irradiance. Conversely, even if the PR appears low, if you have carefully added up temperature losses, shading losses, wiring losses, mismatch losses, and so on, the result may actually be closer to reality. In other words, when reviewing PVSyst results you should not only look at the numbers themselves but also trace the input conditions and the flow of losses that produced those numbers.

What you should look at first is not the final values of the simulation results, but the consistency of the project conditions. Verify that the site, meteorological data, array surface, equipment capacities, string configuration, shading conditions, and loss settings do not contradict each other as a single design proposal. In particular, when creating a new project by copying a past one, old site information or loss settings can remain. Even if it appears to be a new project on the surface, if the internal settings are still from the previous project, it is only natural that the results will be incorrect.

Also, because PVSyst has many input parameters, trying to review all settings at once can actually make it harder to identify the cause. In practice, it is easier to isolate causes by checking in this order: first the site and meteorological data, next the array surface orientation and tilt, then the capacity and strings, then shading and losses, and finally how to read the reports. By separating whether the problem is with the energy production, the PR, the loss diagram, or the monthly profile, it becomes easier to see which settings need to be corrected.

Check Item 1: Are the project location and meteorological data correct?

When PVSyst results seem odd, the first things to check are the project location and the meteorological data. In solar PV simulations, irradiance and temperature have a major impact on energy production. For that reason, even a shift of just tens of kilometers (tens of miles) can change the results in mountainous areas, coastal areas, urban areas, or snowy regions. In particular, when latitude and longitude were entered manually or a nearby representative site was selected, you should verify that the actual installation location matches the meteorological conditions.

A common mistake is that the project location is set correctly while the meteorological data remain for a different site. When copying a past project, users sometimes change only the location information and forget to reload or reselect the weather data. In this case, even if the screen shows the new project name, the assumptions about solar irradiance stay those of the old project, causing the estimated power generation to be unnaturally high or low.

Also, the results vary depending on the type of meteorological data used. Monthly average data, hourly data, long-term average data, data close to a specific year, etc.—the annual power generation and the monthly profile change depending on the nature of the data used. Representative meteorological data may be sufficient for preliminary design estimates, but when using the results for internal briefings or feasibility assessments, it is necessary to clearly state the assumptions about the meteorological data used.

When checking meteorological data, it's important to examine not only the annual global horizontal irradiance but also the monthly distribution. Even if the annual value appears reasonable, the monthly results will seem off if a particular month is unusually high, winter values are too low, or the effects of the rainy season or snowfall don't appear to be reflected. In particular, if PVSyst results show abnormally high generation only in summer, a sharp drop only in winter, or seasonal trends that don't match nearby projects, you should reconsider the selection of meteorological data and the site settings.

Furthermore, differences in elevation and surrounding terrain cannot be ignored. If the installation site is on a mountainside, in a valley, along the coast, or on a plateau, solar irradiance, cloud cover, temperature, and snowfall tendencies can vary even within the same municipality. If the meteorological data in PVSyst represent broad-area values, they may not fully reflect local characteristics. The first step in isolating the cause when results seem odd is to question whether the meteorological conditions are truly representative of the site.

Checklist Item 2: Do the entered azimuth and tilt angles match the site conditions?

When PVSyst results show energy production that is higher or lower than expected, the next things to check are the azimuth and tilt angles. The orientation and tilt of the solar panels are directly related to the solar irradiance they receive. Even a slight change in the mounting surface can alter annual energy production and seasonal generation patterns, so input errors can affect the overall results.

A common issue with azimuth is a misunderstanding of the reference direction. On design drawings or site notes, expressions such as south-facing, east-facing, or west-facing are sometimes used, but when entering data into PVSyst you must correctly convert them to match the software’s angle convention. If a value entered thinking it is south-facing actually corresponds to an east–west orientation, the monthly trends in energy yield and the time-of-day generation patterns will change significantly. In particular, east–west layouts, single-slope roofs, roofs with multiple planes, and slope installations are prone to azimuth mix-ups.

Tilt angle is equally important. Confusing the angle of ground-mounted racks, roof pitch, the slope of the graded surface, and the method of mounting racks on a slope will cause the actual array surface to differ from the array surface in the simulation. For example, if the roof pitch is understood as a rise/run ratio rather than in degrees, an error in conversion can result in a significantly different tilt angle. Also, if the site ground is tilted, looking only at the angle of the rack itself may not correctly represent the actual angle of the module surface.

When a project has multiple array surfaces, extra caution is required. If south-facing and west-facing surfaces, roof surfaces with different tilts, and staggered mounting frames coexist, entering only a single representative angle will simplify the results more than they actually are. A representative value is acceptable at the rough-estimate stage, but if the results seem off, check whether you need to separate the conditions by surface. If a particular surface has a lot of shading, a different tilt, or a large share of the capacity, its impact on the overall results cannot be ignored.

Also, mistakes in azimuth and tilt angles affect not only annual energy production but also the interpretation of PR and the loss diagram. Because the irradiance on the array plane changes, the system can appear to be less efficient, or conversely, unrealistically more efficient. When PVSyst results look odd, before suspecting the equipment or loss settings, it is important to first check which direction the solar panels are assumed to face and at what tilt they are assumed to be installed.

Checklist Item 3: Is the combination of module capacity and PCS capacity correct?

If the PVSyst results for annual energy production, curtailment, clipping, or PR seem off, check the combination of module capacity and PCS capacity. In solar power systems, the ratio of DC-side module capacity to AC-side conversion equipment capacity strongly affects the results. In designs that assume oversizing, output can become capped during certain periods; if that behavior is larger than expected or not present at all, you should suspect an error in the capacity inputs.

A common mistake is entering the wrong number of modules. Examples include the number on the drawings not matching the number in PVSyst, entering only one side's worth, or conversely entering multiple areas twice. Even if the rated capacity of an individual module is correct, the total capacity changes significantly if the count is different. In particular, if only the layout is updated after a design change and the number of modules in PVSyst is not corrected, the results will no longer match the design drawings.

For PCS capacity, you also need to verify the inputs for the number of units and the rated output. The equipment specifications include several values, such as allowable input-side values, rated output-side values, maximum output, and operating range. If it remains unclear which of these values is being treated as capacity in PVSyst, the interpretation of the DC/AC ratio will be off. When using multiple PCS units, confirm the entered number of units, the allocation to each unit, and that the subarray configuration is correct.

When the results look odd, it is important not to look only at the total capacity but to check whether the module capacity, PCS capacity, string configuration, and sub-array configuration are consistent. For example, even if the total module capacity is correct, a biased allocation to certain PCS can create conditions where only those PCS appear to be overloaded. If clipping or conversion losses in PVSyst’s loss diagram differ from expectations, verify not only the capacity ratios but also which arrays are assumed to be connected to which equipment.

Also, the module specification values themselves may be outdated or still set to a different model. If you reuse past projects, even if you think you selected modules of similar capacity, the temperature coefficients, voltage characteristics, and current values may differ. Even if the difference in energy yield appears small, it can affect string voltage and temperature-related losses, causing the simulation results to seem inconsistent. As a PVSyst workflow, before reviewing the results it is important to make a habit of cross-checking the equipment model, rated capacity, number of modules, and number of units against the drawings and specifications.

Check Item 4: Whether the string configuration and voltage range are appropriate

When PVSyst results look wrong, an often-overlooked factor is the string configuration. Even if the module capacity and the PCS capacity match, if the number of modules in series, the number in parallel, the assignment of input circuits, or the voltage ranges are unnatural, the simulation results will deviate from the actual design intent. In particular, if the open-circuit voltage at low temperatures, the operating voltage at high temperatures, and the PCS’s allowable input range are not correctly taken into account, errors or warnings may appear, or the results may become unrealistic.

When checking the string configuration, first assess whether the number of modules in series is realistic. If the series count is too high, the voltage can become excessively high at low temperatures. Conversely, if it is too low, the operating voltage at high temperatures may be too low and fall outside the PCS's proper operating range. Even if PVSyst does not show a warning, you need to confirm that there is sufficient margin relative to the design criteria and the site conditions.

Next, verify the consistency between the number of parallels and the input circuits. In real-world wiring there are constraints on the number of PCS input circuits, the configuration of junction boxes, cable routing, and the division by roof surfaces or sections. Even if the arrangement is valid in PVSyst calculations, if it does not match the actual wiring plan, the results will not correspond to the design drawings. In particular, when east‑ and west‑facing surfaces or multiple tilted surfaces are combined into the same input, strings with different generation characteristics can coexist, leading to mismatch and affecting operational efficiency.

Errors in string configuration can show up in the loss diagram in ways that are hard to recognize. For example, you might see wiring losses or mismatch losses that are larger than usual, PCS losses that differ from what was assumed, or generation that plateaus under certain conditions. Because simply looking at the annual energy yield makes it difficult to identify the cause, go back to the system configuration screen and check the number of modules in series, the number of strings in parallel, the subarrays, and the input assignments one by one.

Also, caution is required when changing modules during the design process. Even if module capacities are similar, differences in voltage or current specifications can change the appropriate number of modules in series. If you use a string configuration that worked with the previous modules as-is, the new specifications may leave insufficient voltage margin. When PVSyst results look incorrect, rather than simply trying to adjust the energy yield, it is essential to confirm that the electrical design is valid.

Checklist Item 5: Are the shading settings too high or too low?

If PVSyst results show a significant drop in energy production, you should check the shading loss settings. Shadows from nearby buildings, trees, utility poles, the front and rear rows of racking, terrain undulations, and rooftop equipment can have a major impact on energy output. However, because the shading settings allow a high degree of input flexibility, they are easily overestimated or underestimated.

A typical example of excessive shading loss is entering the height or distance of obstructions incorrectly. Mixing up units or confusing dimensions on drawings with on-site measurements can lead to assuming shadows larger than they actually are. Likewise, modeling small rooftop pieces of equipment too large can make annual losses appear greater than they truly are. It is important to account for shading carefully, but if you include it so conservatively that it doesn’t match reality, the project viability assessment will become unduly strict.

Conversely, shading losses can also be underestimated. If nearby buildings on site are omitted, mutual shading from front-row racking is not included, or the low solar altitude in winter is not taken into account, estimated power generation tends to be higher than actual. In particular, for projects that place panels across the entire site, projects where neighboring buildings cast shadows, or projects close to mountains or slopes, how shading is handled can greatly affect the results.

When checking shading settings, it’s important to look not only at the total annual losses but also at monthly shading losses and time-of-day trends. Whether shadows are strong only on winter mornings and evenings, are consistent throughout the year, or are significant in summer will point to different causes. If large shading losses occur during summer daytime, the shape or position of obstructions may be modeled larger than they are in reality. Conversely, if shadows are clearly present on site but the model shows almost no losses, you should check for obstructions omitted from the model or review the shadow calculation settings.

Also, the effects of shading relate not only to simple solar radiation blockage but also to electrical losses at the string level. When partial shading affects some modules, it can impact the output of the entire array. When performing shading analysis in PVSyst, check not only the shape of the shadow but also to what extent the module layout and mapping to the strings are represented. If the results seem odd, it is important not to focus solely on creating a tidy 3D scene, but to verify that the settings reflect the actual power loss.

Check Item 6: Are there any excessive or missing loss coefficients?

One of the most important things when checking PVSyst results is reviewing the loss factors. In photovoltaic simulations, you set various losses such as temperature losses, wiring losses, mismatch losses, soiling losses, equipment losses, degradation, and downtime rate. Entering these appropriately yields results that are close to reality, but if settings are duplicated or, conversely, omitted, the energy production and PR will appear incorrect.

A common mistake when adding losses is including the same loss in multiple places. For example, if you apply a separately assumed safety margin or contingency and also enter it among PVSyst’s loss items, the result will be excessively low. Even if your internal estimate sheet already includes soiling and downtime rates, entering the same items in PVSyst can make only the PVSyst result look lower when compared. In this case, the problem is not that PVSyst’s results are wrong, but that the assumptions are not aligned between the comparison targets.

On the other hand, omissions of losses are also a problem. If the wiring distance is long but the wiring loss is left at the standard value, if the effects of soiling or snowfall are hardly taken into account, or if the impacts of shutdowns and maintenance are not considered, the estimated power generation will come out higher. While it may be acceptable to simplify things in initial studies, simulations submitted in practice must be prepared so that you can explain which losses were included under which assumptions.

Temperature losses are also easy to overlook. Module temperature is not determined solely by ambient air temperature; it varies depending on the mounting structure, ventilation conditions, distance from the roof, and the installation environment. Installations close to the roof or in poorly ventilated conditions tend to have higher module temperatures and lower power generation efficiency. If PVSyst results for summer energy yield or PR differ from expectations, check the temperature conditions and the settings related to thermal losses.

Soiling losses should also be assessed differently depending on the region and installation environment. Near farmland, industrial zones, along roads, by the sea, or in areas with low rainfall, soiling can have a greater impact. Conversely, under conditions where regular cleaning or rainfall can be expected, applying excessive soiling losses can lead to underestimating energy production. When using PVSyst, it is important not to apply a uniform loss value but to provide a rationale for the chosen value based on site conditions.

When checking loss settings, read the loss diagram from top to bottom in order and look for where the energy drops significantly; this makes it easier to understand. If only a particular loss is extremely large, check the input values and assumptions for that item. If several losses are slightly larger, check whether the overall assumptions are too conservative. If the results look wrong, it is important to interpret the breakdown of losses rather than directly adjusting the final energy output.

Check Item 7: Are you confusing the interpretation of monthly results with that of annual results?

Causes for feeling that PVSyst results are incorrect include not only input errors but also misinterpretation of the outputs. In particular, confusing monthly results with annual results can make outputs that are actually reasonable appear odd. Solar power generation changes with the seasons — irradiance, solar altitude, temperature, and the way shadows fall — so monthly generation is not constant. Looking at month-by-month variations as well as the annual generation makes it easier to judge the validity of the results.

For example, in summer the solar radiation is high, but because temperatures are higher, temperature losses also increase. In winter, lower temperatures are favorable for module efficiency, but shorter sunshine duration and a lower solar elevation make the system more susceptible to shading. Therefore, it is not necessarily the case that summer is the maximum and winter the minimum. Depending on the region and the installation angle, generation in spring or autumn can appear higher. If you look at monthly results without this knowledge, you may feel that "summer does not grow as much as expected" or that "winter is too low."

Care must be taken when interpreting PR. PR is not simply the amount of generation; it is an indicator of how efficiently the system converted the incident irradiance into electricity. Therefore, months with high generation do not necessarily coincide with months of high PR. In months when losses due to high temperatures are significant, PR can decrease even if irradiance is high. Conversely, in months with low generation, PR can appear relatively high if temperature conditions are favorable.

Also, you need to pay attention to the units in the simulation results. PVSyst reports display several similar figures, such as annual energy production, specific yield, array output, grid output, and energy before and after losses. If you mix up which value you are comparing to your company's estimates or other documents, the results can appear incorrect. For example, if you compare DC-side energy with AC-side output, conversion losses and post-limitation values will not be reflected, and they will not reconcile.

When reviewing monthly results, we check not only energy generation but also solar irradiation, temperature losses, shading losses, and system losses together. If only one month shows low generation, we separate whether that month had low solar irradiation, large shading losses, or large temperature losses. If you look only at the final energy generation without separating causes, you cannot tell whether it is an input error or a result reflecting site conditions. In PVSyst result checks, the practical workflow is to grasp the overall picture from the annual values and investigate causes using the monthly values.

Checklist item 8: Are the comparison conditions aligned before report output?

Many of the situations where you feel PVSyst’s results are off occur when comparing multiple design options or simulation results. If the energy production differs between Plan A and Plan B, if the results you produced previously don’t match the current results, or if the values in internal documents don’t match PVSyst’s values, the first thing you should check is whether the comparison conditions are consistent.

Conditions that often cause discrepancies in comparisons are weather data, capacity, loss settings, shading conditions, output limits, and how degradation is treated. For example, if Option A includes shading while Option B does not, the difference in energy output will reflect not only differences between the design options but also differences in the input conditions. Likewise, if only one option includes soiling losses or a downtime rate, the comparison will be unfair. When comparing design options, you should keep everything the same as much as possible except for the conditions you want to change.

Be careful when comparing with past results. If the module model, PCS capacity, loss values, meteorological data, or layout area have changed slightly from a previous simulation, you may not be able to tell which change caused the difference in generated energy. In practice, as a project progresses drawings and equipment specifications are updated, so PVSyst projects tend to be split into multiple versions. When the results seem odd, it is important to confirm which version you are comparing and which drawing it reflects.

Before outputting the report, check not only the final values but also the project name, location, capacity, array surface, equipment configuration, loss conditions, shading settings, and the simulation date. When the report is used for internal sharing or client submission, if you cannot explain the assumptions shown on the report, you will not be able to justify the validity of the estimated energy production. In particular, when presenting multiple options, clearly state in the main text or in supplementary materials where the differences between each option lie; doing so makes it easier to explain numerical differences later.

Also, mistakes occur when transcribing PVSyst results into other documents. If you mix up the units for annual energy production, the units for specific yield, the percent display of PR, the sign of loss rates, or the sum of monthly values, you can end up with strange numbers in internal documents even though the results are correct in PVSyst. If the results seem wrong, you need to check not only the PVSyst settings but also the destination documents and the formulas used.

Reports may look like the final deliverable, but in practice they also serve as materials for verification. Sometimes you only notice something is off when you look at the exported PDF or tables. Therefore, after generating the report it is important to read through the input conditions and the results once. If PVSyst’s results seem incorrect, you can find the cause more quickly by organizing the comparison conditions and performing version control before recreating the report.

Verification of PVSyst results varies greatly with the accuracy of on-site conditions

The points to check when PVSyst results look incorrect are summarized as: project site, meteorological data, azimuth, tilt angle, capacity, string configuration, shading, losses, interpretation of results, and comparison conditions. It is not sufficient to check only one of these; it is important to review them in order as the flow of design conditions. In particular, if the energy output is too high, suspect omitted shading or losses or errors in capacity or meteorological data; if the energy output is too low, check for duplicated losses, excessive shading settings, or input mistakes in azimuth or tilt angle.

PVSyst can handle very detailed conditions, but if the conditions you enter differ from the actual site, the results will also diverge from reality. In other words, simulation accuracy is not determined solely by operations within the software; it is also influenced by the accuracy of preliminary information such as site inspections, drawing checks, surveying, shading assessment, and verification of the installation surface angle. Learning how to use PVSyst is, of course, important, but collecting the correct on-site conditions to enter is equally important.

What field personnel should pay particular attention to is ensuring that information collected on site can be directly reflected in the design parameters. If orientation and tilt, surrounding obstructions, installation area, terrain undulations, and the positions of existing structures remain ambiguous, no matter how carefully you configure settings in PVSyst it will be difficult to justify the validity of the results. Conversely, if the on-site position and geometric information are accurate, analyses of meteorological data and loss conditions become much more convincing.

As a way to improve on-site verification accuracy, utilizing the LRTK GNSS high-precision positioning device that can be attached to an iPhone is useful for pre-design surveys and pre-construction checks of photovoltaic installations. If you can record the planned installation area, racking locations, surrounding obstructions, and terrain features together with high-precision positional information, it becomes easier to verify the assumptions to be entered into PVSyst. Furthermore, by linking and preserving on-site records such as photos and point clouds with position information, you can more easily review site conditions later if the results seem off.

When you feel that PVSyst’s results are incorrect, it is important not to just chase the numbers in the software, but to check one by one what is happening on site, which conditions were entered, and under what assumptions the comparisons are being made. Simulations only become useful decision-making tools in practice when combined with accurate on-site information. Incorporating high-precision positioning such as LRTK into field surveys can increase the reliability of PVSyst input conditions and further enhance the explanatory power of power generation simulations.