7 checks when analysis results don't match the PVSyst manual

Table of Contents

• What to consider first when analysis results don't match

• Check 1: Are the meteorological data and site conditions consistent with the assumptions?

• Check 2: Ensure the azimuth and tilt angle inputs are not misaligned with the on-site conditions

• Check 3: Are the selection criteria for modules and inverters correct?

• Check 4: Are there any inconsistencies between the array configuration and the string design?

• Check 5: Are the shadow settings and the reflection of surrounding obstacles neither insufficient nor excessive?

• Check 6: Are loss settings being double-counted or left unrecorded?

• Check 7: Are you misinterpreting how to read the report results?

• Practical steps for reviewing analysis results

• Create reproducible verification procedures using the PVSyst manual

What to consider first when analysis results do not match

Many people who want to understand the seven checks in the PVSyst manual for when analysis results do not match find themselves in a situation where the generation simulation results disagree with internal estimates, past projects, local intuition, or other documents, and they don’t know where to start reviewing. In photovoltaic power generation simulations, small differences in input conditions affect annual energy production, loss rates, performance ratio, and peak behavior. Therefore, rather than judging a result as simply “too high” or “too low” based only on the output, it is important to check in sequence which assumptions are influencing which outcomes.

The reasons why analysis results do not match are not limited to simple input mistakes. Choices of meteorological data, definitions of azimuth, handling of tilt angle, module and inverter specifications, string configuration, treatment of shading, temperature losses, wiring losses, anticipated degradation and soiling, and how reports are read are just some of the multiple factors that can combine to produce discrepancies. In particular, when comparing early-design estimates with detailed simulation results, it is not uncommon for the underlying assumptions being compared to be different in the first place.

When consulting the PVSyst manual, it is important not just to read the function descriptions but to do so while organizing which items in your project influence the analysis results. Even if it looks like you are filling in the input fields on the screen one by one, you are actually deciding the assumptions for energy production step by step. Especially when the results do not match, rather than starting over from scratch, you should adopt an approach of checking by correlating the input conditions with the output results.

Also, there are several types of "mismatch" in analysis results. These include cases where the annual energy production does not match, where the monthly generation trend does not match, where the performance ratio differs from expectations, where only the breakdown of losses looks unreasonable, where inverter clipping is excessive, or where shading losses appear extreme. Lumping these together can cause you to lose sight of the cause. The starting point is to clarify which values are off, compared with what, and by how much.

Check 1: Are the meteorological data and site conditions consistent with the assumptions?

When analysis results don’t match, the first things to check are the weather data and the site conditions. In solar power generation simulations, meteorological conditions such as solar irradiance, air temperature, and wind speed form the foundation of the results. No matter how accurately you enter the equipment parameters, if the location or type of meteorological data differs from what you assumed, the annual power generation can change significantly. In particular, if the solar irradiance data used for internal estimates differs from the meteorological data selected in PVSyst, it is natural that the results will not match.

When assessing site conditions, it is important to verify latitude, longitude, elevation, and time zone. Even if you intend to use a representative nearby location, solar radiation and temperature trends differ in mountainous areas, coastal areas, snowy regions, and urban areas. Changes in elevation alter temperature conditions and affect thermal losses. Along the coast, humidity and cloudiness trends differ, and in mountainous areas terrain can produce differences in solar radiation conditions. If analysis results are lower than expected, you should first check whether the meteorological conditions have been set more severely than intended before examining the equipment side.

When checking meteorological data, you should look not only at the annual solar radiation but also at the monthly distribution. Even if the annual totals are similar, data that are high in summer and low in winter and data that are relatively even throughout the year will produce different monthly electricity generation patterns. In snowy regions or areas with low winter insolation, monthly differences strongly affect results, so judging validity based only on the annual value can lead to oversights. When comparing with past projects or actual results, you also need to distinguish whether the comparison year represents a long-term average (a normal year) or the actual result for a specific year.

Also, temperature data are important. Because solar panels’ output decreases when they become hot, if temperature conditions are set higher the temperature-related losses will be larger. Conversely, using data from cooler regions can produce higher estimated generation. Even if annual solar irradiance is matched, results will not agree if the temperature conditions differ. When checking how meteorological data are handled in the PVSyst manual, a practical point is to verify not only irradiance but also temperature and seasonal variations.

When checking site conditions, also review how close the location set when the project was created is to the actual planned installation site. In particular, if you created the project by copying multiple cases, location information from a previous case may remain. On the screen you tend to focus on system capacity and module settings, but if the site settings are different, the entire analysis results will be skewed. If you feel the results don't match, first return to the project's basic information and verify that the location and weather data are consistent with the case assumptions.

Check 2: Are the entered azimuth and tilt angles consistent with the on-site conditions?

Next to check are the azimuth and tilt angles. In solar power generation, the amount of solar irradiance a panel receives changes depending on the direction it faces and the angle at which it is installed. If the analysis results are higher or lower than expected, input errors in the azimuth or tilt angles may be the cause. In particular, if the azimuth reference or sign convention is handled incorrectly, what was intended to be set as south-facing may actually be in conditions closer to an east–west orientation.

Azimuth can be represented differently depending on the software or documents. In some cases it is expressed with south as the reference and east/west indicated by plus and minus, while in other cases it is expressed with north as the reference and measured clockwise. You need to convert the azimuth notation on design drawings or on-site survey documents to the definition required by PVSyst rather than entering them as-is. If you input them without checking this, an orientation that is correct on the drawings may be treated as a different direction in the simulation.

The same applies to tilt angles: it is important not to confuse roof pitch, racking angle, terrain slope, and the actual module surface angle. In roof-mounted projects, the roof pitch and the module installation angle may coincide, but they differ when a rack is added or when tilt is applied on a flat roof. For ground-mounted projects, you need to treat the slope of the prepared/graded surface and the racking tilt angle separately. If the analysis results do not match, check which angle the entered tilt refers to.

Azimuth and tilt angles affect not only annual power generation but also monthly generation. With a larger tilt angle, winter generation tends to be relatively higher, while with a smaller tilt angle, summer can be advantageous. In east-west installations the generation profile in the morning and evening changes, and the peak times differ from south-facing installations. Even if the annual values alone don't raise concerns, examining monthly or hourly results can reveal configuration errors.

In projects with multiple roof planes or arrays oriented toward multiple azimuths, you need to check the azimuth and tilt angles of each plane individually. If you consolidate them into a single representative value, the results will differ from the actual layout. In particular, for roofs divided east–west, roofs of multiple buildings, or installation surfaces with level differences, the distribution of generation differs from that of a simple south-facing model. Refer to the PVSyst manual and confirm that you have correctly separated and entered the conditions for each installation surface.

Confirmation 3: Are the selection criteria for modules and inverters correct?

One common cause of discrepancies in analysis results is differences in the selection criteria for modules and inverters. Even if you only match the solar panels’ nominal output, differences in temperature coefficient, voltage characteristics, current characteristics, efficiency, and behavior under low irradiance will change the amount of energy generated and how losses occur. Inverters also change the outcome depending on rated capacity, input voltage range, conversion efficiency, maximum input current, number of MPPTs, and so on.

The first thing to confirm is whether the module data you have selected actually matches the model that is intended to be used. If you accidentally choose a similar part number, the output and electrical characteristics can differ subtly. Do not judge solely by the manufacturer name or capacity; you need to cross-check the nominal maximum output, open-circuit voltage, short-circuit current, maximum power operating voltage, maximum power operating current, temperature coefficients, and so on on the datasheet. Because specifications can change depending on differences at the end of the model number, check this carefully.

Next, check the inverter settings. If the inverter capacity is small relative to the system capacity, output limitation is likely to occur at generation peaks. Conversely, if the inverter capacity is too large, efficiency during periods of low load factor may differ from what was expected. If the analysis results show clipping losses larger than anticipated, check the ratio of module capacity to inverter capacity, the string configuration, and the input range settings.

Also, it is important to ensure that the combination of modules and inverter is electrically appropriate. If the number of modules in series per string is too high, the open-circuit voltage can rise at low temperatures and may exceed the inverter’s input voltage range. If the number of modules in series is too low, the operating voltage can drop at high temperatures and may fall outside the MPPT range. These conditions will appear not only in the annual energy yield but also in warnings and the loss breakdown. When results don’t match, don’t just look at capacity—check the voltage and current compatibility.

If you create module data or inverter data as user-defined entries, you must also check the units and coefficients of the input values. Errors in the sign of the temperature coefficient, percent notation, decimal notation, the units of voltage values, and so on can greatly affect the results. Even when substituting a similar model because there is no closely matching model in the database, you need to understand how much that substitution difference will affect the results.

When using the PVSyst manual, it is important not only to operate the equipment selection screen but also to understand how each item affects power generation. If the analysis results do not match expectations, mistakes in equipment selection are relatively easy to find, whereas if something is overlooked the cause can remain unknown until the end. In particular, when creating a project by duplicating a previous one, making a habit of checking whether the previous project's equipment settings remain can reduce simple mistakes.

Check 4: Are there any inconsistencies between the array configuration and the string design?

Even if the modules and inverters are correctly selected, the analysis results will not match if there are inconsistencies in the array configuration or string design. The number of modules in series per string, the number of parallel strings, assignment to MPPTs, the capacity of each array, and connection conditions by orientation all affect both energy yield and losses. Especially in projects with multi-orientation roofs or distributed layouts, inputting the array configuration becomes more complex and can easily lead to discrepancies in the results.

The first things to check are the total number of modules and the system capacity. Verify that the number of modules assumed in the drawings and design documents matches the number entered in PVSyst. Check that the DC capacity calculated from module output and quantity aligns with internal documents and the estimate conditions. If the annual energy generation is higher than expected, the input capacity may simply be too large. Conversely, if it is lower, it may be due to an insufficient number of modules or some arrays not being entered.

Next, verify the number of modules in series per string and the number of parallel strings. If the number of modules in series differs, the operating voltage changes. If the number of parallel strings differs, the input current and the load on each MPPT change. Even if the array appears to fall within the inverter’s input range, temperature conditions can cause efficiency to drop during certain time periods. String design should not rely solely on on-paper capacity matching; it must be considered together with the actual electrical operating range.

When multiple orientations and tilts are mixed, also check whether strings with different conditions are connected to the same MPPT. Grouping east-facing and west-facing, south-facing and north-leaning, or shaded and unshaded surfaces on the same MPPT can cause differences in generation characteristics to appear as losses. If the analysis shows large mismatch losses or electrical losses, this connection arrangement may be the cause.

For array configurations, it is also important to ensure that the grouping used in the simulation matches the grouping used in actual construction. If the drawings are divided into multiple areas but PVSyst treats them as a single representative array, differences in shading, orientation, and temperature conditions may not be adequately reflected. Conversely, dividing them more finely than necessary can result in some settings being omitted. When results do not match, it is effective to compare the array configuration diagram and the simulation settings side by side.

Also, check whether any warnings or notices are displayed for the configuration you entered. If you ignore warnings and proceed with the analysis, you may still obtain results, but they could represent conditions that are difficult to adopt in practice. The PVSyst manual advises reading the results while verifying the meanings of warnings and the relationships among input items. If the analysis results do not match expectations, it is important to review not only the figures in the report but also the warnings shown during the setup stage.

Check 5: Are the shadow settings and the incorporation of surrounding obstacles appropriate?

One factor that can create large discrepancies when analysis results do not match is the shadow settings. The impact on power generation from surrounding buildings, trees, utility poles, spacing between racking rows, rooftop equipment, parapets, and mountain shadows varies greatly depending on the project. If shadow settings are insufficient, estimated power generation tends to come out higher, whereas if they are set excessively the estimates tend to come out lower. The way shadows are handled particularly influences results for rooftop installations in urban areas and ground-mounted installations in mountainous regions.

First, what you want to check is whether the area considered for shadows matches the actual site conditions. Check that all obstacles that can be confirmed from site photos and drawings are reflected, and conversely that no unnecessary obstacles remain in the simulation. When duplicating a project, shadow models or obstacle conditions from the previous project may remain. If you do not notice this, you may end up analyzing completely different site conditions.

Next, check the height, distance, and relative positions of obstacles. The impact of shading increases the closer an obstacle is, and becomes more pronounced during periods of low solar altitude. Simply entering a slightly greater height can increase shading losses in winter and during the morning and evening. If the distance or azimuth is off, the times of day or seasons when shading occurs will change. If shading losses are larger than expected, you need to review not only the obstacle heights but the entire layout.

In ground-mounted projects, checking inter-row shading is also important. The time periods during which a front row casts shadows on a rear row change depending on the spacing between arrays, tilt angle, racking height, and terrain gradient. Closing the spacing between rows to use land more effectively increases installed capacity, but it can also increase shading losses. If analysis results do not match, it is necessary to separate the increase in energy production due to the higher capacity from the increase in losses caused by inter-row shading.

For roof-mounted projects, shadows from parapets, HVAC equipment, roof penthouses, and adjacent buildings often cause problems. Obstacles that look small on drawings can cast long shadows during seasons with low solar elevation. In particular, if winter power generation is lower than expected, shading settings may be having an impact. Check monthly generation; if output drops sharply in a particular season, suspect seasonal variations in shading to more easily identify the cause.

When configuring shadows, also be mindful of the difference between near shadows and distant shadows. Shadows from nearby obstacles and those from terrain or the horizon differ in how they are input and in how their effects appear. In regions where distant mountains or terrain affect morning and evening solar radiation, results may not match if you only consider the usual surrounding obstacles. While checking the PVSyst manual for how shadows are handled, organize which kinds of shadows are being represented by which methods.

Check 6: Are loss provisions being double-counted or not recorded?

Another factor that easily creates differences in analysis results is the loss settings. In solar power generation simulations, many loss elements are handled, such as temperature loss, wiring loss, mismatch loss, soiling, degradation, incidence-angle loss, inverter loss, transformer loss, and availability. If these settings are overestimated the generated power will be lower, and if they are left unset it will be higher. Furthermore, if the same loss is counted twice under different items, the results will be harsher than expected.

First, I want to confirm which losses are set in PVSyst and which losses are being assumed in the other documentation. If the in-house estimated generation already includes soiling, degradation, and downtime rate, adding the same losses in PVSyst will result in double-counting. Conversely, even if you intended to include losses in the estimate, if they are not set in PVSyst the results may come out higher. When comparing, it is essential to align the loss assumptions.

In terms of temperature losses, the installation method and ventilation conditions are important. When modules are installed close to the roof versus on ground-mounted racks with good airflow, the rate at which module temperature rises differs. If ventilation conditions are set optimistically, temperature losses are smaller and power generation is higher. Conversely, if conditions are set more conservatively than reality, power generation will be lower. For high-temperature regions or projects where modules are nearly flush with the roof, temperature loss settings need to be checked carefully.

Wiring losses are another item that is easy to overlook. Losses vary depending on cable length, cross-sectional area, how the DC and AC sides are treated, and the scope up to transformers and receiving equipment. In the early stages of design, simplified loss rates are used, while in detailed design they are often adjusted to be closer to the actual wiring lengths, so it is natural for results to change between stages. If the analysis results are lower than the estimate, check whether the wiring losses have been set too conservatively.

Handling soiling and snow is also important. While a fixed percentage is typically set for normal soiling losses, if separate snow or seasonal losses are also assumed, winter generation may end up being underestimated. In environments prone to soiling—such as farmland, factories, coastal areas, and alongside major roads—losses should be taken into account, but entering large values without justification makes the results difficult to explain. Loss settings need to be considered together with maintenance conditions, cleaning frequency, and the local environment.

Regarding the degradation rate, how you handle it depends on whether you want to look at first-year generation or a long-term average. If you compare the first-year simulation result with the 20-year average generation, it is natural that the figures will not match. When comparing analysis results, align the number of years considered, whether degradation is reflected, warranty conditions, and the definition of generation used in the economic evaluation. When reading the PVSyst manual, it is important not to regard loss items as mere input fields, but to confirm that they are based on the same assumptions as the comparison target.

Check 7: Are you misinterpreting how to read the report results?

Even when the input conditions are correct, misreading how to interpret report results can lead you to feel that "the analysis results don't match." PVSyst reports display multiple indicators such as annual energy production, grid output, performance ratio, normalized production, loss diagram, and monthly results. If you are comparing different numbers, it is natural that the results will appear inconsistent. In particular, it is important not to confuse DC-side generation with AC-side output, or system output with the energy available for sale.

The first thing I want to confirm is the point at which the generation figures being compared are measured. The DC energy produced by the solar panels, the AC energy after passing through the inverter, and the energy at the transformer or the grid interconnection point are not the same. Because there are losses along the way, the figures change in stages. You need to check which stage the generation figure used in internal documents or business plans refers to, and align it with the corresponding item in the PVSyst report.

Care is needed when interpreting the performance ratio. The performance ratio is a useful indicator for evaluating system efficiency, but it varies depending on solar irradiance conditions, temperature conditions, loss settings, and equipment conditions. A high performance ratio does not necessarily mean greater energy yield, nor does a low one necessarily indicate poor design. In high-irradiance regions where temperature losses are large, or in projects with significant shading effects, the way the performance ratio is perceived changes. Rather than judging quality by the performance ratio alone, it should be considered together with the breakdown of losses.

When looking at a loss diagram, check not only which losses are large but also whether those losses are reasonable. If temperature loss is large, check the ambient temperature and installation method; if shading loss is large, check the shading model; if wiring loss is large, check the cable conditions. A loss diagram is like a map for finding the causes. Rather than simply looking at the total loss, tracing at which stages energy is being lost makes it easier to find input errors or incorrect assumptions.

Checking monthly results is also essential. With only the annual energy output, you cannot tell which season the differences occur in. If the discrepancy is large in summer, suspect temperature losses, overloading (overpanelling/oversizing), or inverter limits. If the discrepancy is large in winter, check for the effects of shading, snowfall, low solar altitude, and orientation or tilt. Even if annual values are similar, if the monthly patterns differ, the underlying assumptions may be different. To resolve any inconsistency in the analysis results, it is important to examine monthly trends as well as annual values.

Also, pay attention to the units when transcribing figures from reports into external documents. Confusing kWh, MWh, kWh/kWp, annual values, monthly values, first-year values, average values, etc., can cause large discrepancies in comparison results. Decimal-point values or values per unit of installed capacity are sometimes mistakenly treated as total generation. You need to consult the PVSyst manual, understand the meaning of each indicator in the report, and align the comparison targets and units.

Practical Workflow for Reviewing Analysis Results

When analysis results don't match, it's important to check them in a determined order rather than making ad-hoc fixes to parts that catch your eye. First, clarify what you are comparing against. The points you should examine vary depending on whether you're comparing with an internal estimate, a past project, on-site measured results, or another simulation's results. If you modify only the PVSyst side without understanding the assumptions of the comparison target, you will not be able to identify the cause.

Next, we verify the project's basic conditions. We review the site, weather data, system capacity, modules, inverters, azimuth, tilt, and array configuration. At this stage, we prioritize checking whether the major assumptions are correct rather than focusing on minor losses. This is because when the annual energy production differs significantly, it is often due to a discrepancy in one of the basic conditions. Basic differences—such as different capacity, location, azimuth, or modules—should be ruled out first.

After that, check shading and losses. Inspect, in order, the presence of shading, nearby obstructions, row-to-row shading, distant shading, temperature losses, wiring losses, soiling, degradation, and downtime rate. At this stage, it is important not to change multiple items at once. If you change many settings at once, you will not know which change affected the results. Save the results before making changes so you can compare them with the results after changes; this makes it easier to identify the cause.

When reviewing results, we look at annual energy production, monthly energy production, the loss diagram, and the performance ratio as a set. If you only look at annual energy production, you may not notice when multiple factors offset each other. For example, if the capacity setting is large and boosts generation while the loss settings are strict and reduce generation, the final result may appear reasonable at first glance. By checking the loss breakdown together with monthly trends, it becomes easier to find hidden assumption mismatches.

In practice, it is also important to record what was checked. If you document which meteorological data were used, what the azimuth and tilt angles were based on, the assumptions used to set the loss rates, and how extensively the shading model was reflected, it will be easier to explain later. Analysis results are meaningful not only as numbers but when paired with their assumptions. Especially in materials that will be seen by multiple stakeholders—designers, sales representatives, construction personnel, clients, and financial institutions—the explanation of assumptions is important.

Also, even if the results do not match, it does not necessarily mean that the PVSyst settings are wrong. The estimated values being compared may be simplified, the actual values may be affected by the weather in a specific year, or the installation conditions may differ from past projects. What is important is not to immediately decide which is correct, but to be able to explain why the discrepancy exists. If the discrepancy can be explained, it can be used for design and business decisions.

Using the PVSyst Manual to Create Reproducible Verification Procedures

To put the PVSyst manual’s seven checks for when analysis results don’t match into practice, it is important to ensure the same procedure can be followed each time. If each project’s person in charge performs checks ad hoc, mistakes can be overlooked and explanations can vary. Creating a workflow to verify, in order, meteorological data, site conditions, azimuth, tilt, equipment selection, array configuration, shading, losses, and how to read the report increases the reproducibility of the analysis results.

When creating reproducible verification procedures, it is important to preserve the rationale for the input values. Record which location’s weather data was used, whether the azimuth was read from drawings or measured on site, whether the tilt angle refers to the roof pitch or the mounting-frame angle, and whether loss rates are standard values or project-specific conditions. This allows you, when reviewing results later, to check them not as a mere list of numbers but as a history of the decisions made.

Also, it is useful to standardize typical checklist items within the company. For example, if annual energy production deviates significantly from expectations, first check installed capacity, weather data, azimuth, and tilt angle. If monthly output is low only in the winter months, check shading, snow accumulation, tilt angle, and distant shading. If the loss diagram shows large temperature losses, check temperature data, installation method, and ventilation conditions. By mapping symptoms to the corresponding checklist items in this way, you can identify the cause more quickly.

The PVSyst manual is not just for looking up how to operate the software. When analysis results feel questionable, it serves as a reference for checking the meanings of input fields and the interpretation of output indicators. In particular, it is important not to simply fill in the settings screen once and be done, but to return to the assumptions after reviewing the results. Simulation is not only about entering inputs and producing results, but also about validating the plausibility of the assumptions from those results.

When analysis results don't match, instead of rushing to make the numbers line up, verify the reasons for the discrepancy one by one. If you sort out whether the meteorological data is different, the installation angle is different, the equipment specifications differ, the shading is significant, the loss settings are conservative, or the report is being read differently, your ability to explain the results will greatly improve. Simulation results that can be explained are easier to use in any situation—design changes, profitability assessments, client explanations, and internal reviews.

Ultimately, what is important is to treat PVSyst analysis results not as absolute numbers but as material for design decisions based on the underlying assumptions. Solar power generation varies with weather, equipment, installation, maintenance, and aging. For that reason, simulations must make their assumptions explicit and be able to explain which conditions produced which results.

If you have the seven checks in the PVSyst manual for when analysis results don't match, you can address discrepancies by procedure rather than by intuition. By checking, in order, the meteorological data, azimuth, tilt, equipment, array configuration, shading, losses, and report interpretation, you can more easily find not only input errors but also differences in comparison assumptions. Situations where results don't match are not merely troubles but opportunities to understand design conditions more accurately. If you use PVSyst in practice, it's important not only to be able to produce analysis results but also to have verification procedures to explain, review, and communicate those results to stakeholders.