Explain 5 reasons for differences in energy yield in the PVSyst manual

Table of Contents

• What readers of the PVSyst manual want to know: the true cause of differences in energy yield

• Differences in power generation result not only from calculation errors but also from differences in conditions.

• Reason 1 The way meteorological data are chosen is different

• Reason 2: The assumptions about orientation, tilt, and array configuration are different

• Reason 3: The way shadows are handled and the accuracy of the 3D scene differ.

• Reason 4: The way loss settings are configured is different

• Reason 5 The output items and the periods being compared are different

• Basic procedure for checking differences in power generation

• Approaches to make it easier to explain differences during project review

• Points to note for mastering the PVSyst manual in professional practice

• Summary

The true nature of the differences in energy production that readers of the PVSyst manual want to know

Many people who consult the PVSyst manual are not simply trying to learn how to use the software. In fact, they want to be able to explain why, for the same solar power plant project, the estimated energy generation differs between the design company, the person responsible for generation simulations, materials prepared for financial institutions, internal review documents, and results from other software.

In simulations of photovoltaic power output, even slight changes to the input conditions can affect the annual generation, monthly generation, performance ratio, breakdown of losses, and projected revenue from electricity sales. While PVSyst allows detailed configuration of conditions, it is also a tool whose results can vary easily depending on which screen you entered what, which data you selected, and which losses you entered manually.

The important point here is not to immediately conclude that “one of them is wrong” just because there is a difference in energy production. Most differences in energy production arise more from assumptions, meteorological data, the treatment of shading, loss settings, and differences in the items being compared than from differences in the formulas themselves. In other words, the purpose of reading the PVSyst manual is not to memorize where the buttons are, but to be able to break down differences in results into differences in conditions.

Especially in practical work, there are many situations where differences in power generation must be explained. For example, if the PVSyst result is lower than the rough estimate from the initial assessment, you need to explain to the client and within the company why it is lower. Conversely, if it is higher than past estimates, you must clarify which conditions changed, otherwise you may lose credibility in later stages.

In this article, from the viewpoints that should be checked in the PVSyst manual, we explain five reasons why differences in energy production occur. Rather than focusing on the operational procedures themselves, we emphasize points that practitioners are prone to overlook when verifying discrepancies.

Differences in power output arise not only from calculation errors but also from differences in conditions

When PVSyst results differ, many people first suspect input errors. Of course, input mistakes such as capacity, number of modules, number of power conditioners, units, azimuth, and tilt angle can cause differences in generated energy. However, what is common in practice is not clear mistakes but differences in judgment and assumptions among the people in charge.

For example, even at the same location, the annual solar irradiation changes depending on which meteorological data are used. When the irradiation changes, the amount of power generated will naturally change as well. Also, even with the same module capacity, the way comparison results appear changes depending on whether you base them on DC capacity or AC capacity and on how you treat oversizing.

Furthermore, the way shadows are handled also makes a big difference. Results calculated under simplified shadow conditions and results obtained by creating a 3D scene that reflects near-field shading differ in energy production, particularly in the morning and evening and during winter. Energy production changes depending on how much you include shadows from the terrain, surrounding buildings, spacing between racking, and between module rows.

Loss settings are another element that’s easy to overlook. Cable losses, mismatch losses, soiling losses, temperature losses, conversion losses, downtime rate, degradation over time, and so on may each seem small, but when they add up they can have a large impact on annual energy production. If one estimate uses standard values and another uses conservative values, it’s natural that the results won’t match.

Also, there are cases where the items being compared are different. If you are looking at different figures—generation at the generator terminals, generation at the grid interconnection point, the amount of electricity available for sale, theoretical values before loss deductions, or effective values after loss deductions—the generation amounts can appear to differ. Comparing numbers alone without checking the meaning of each output item in the PVSyst manual causes unnecessary confusion.

To correctly assess differences in power output, it is important to first think of the "reasons a difference occurred" by dividing them into three categories: input conditions, calculation conditions, and output items. Input conditions are meteorological data, location, system capacity, azimuth, tilt, shading, losses, and so on. Calculation conditions are the simulation model and correction methods, time resolution, and the range over which conditions are reflected. Output items are what the numbers being compared ultimately mean.

Reason 1: Different methods of selecting meteorological data

The most fundamental reason for differences in power generation is differences in meteorological data. Simulations for photovoltaic power generation depend strongly on weather conditions such as solar irradiance, temperature, and wind speed. In particular, because solar irradiance is directly linked to power output, changing the meteorological data used can cause the annual power generation to vary significantly.

When reviewing the PVSyst manual, you should first carefully read the sections on creating, loading, and selecting meteorological data. Even for the same location, results will vary depending on whether you use data from a nearby observation station, satellite-derived data, long-term average data, or measured on-site data.

For example, suppose that in a certain project a past simplified estimate indicated high energy production, but a detailed analysis with PVSyst showed a lower result. In that case, before looking at equipment settings you should check the meteorological data’s annual average solar irradiance, monthly solar irradiance, data period, and whether any site corrections have been applied. Initial studies may use optimistic solar data, while detailed studies may use conservative long‑term average data.

What you should pay attention to with meteorological data is not to look only at the annual total but to examine monthly trends. Even for datasets with similar annual solar radiation, those that are high in summer and low in winter and those with mild seasonal variation will produce different amounts of power once system configuration and shading effects are taken into account. In particular, in snowy regions, mountainous areas, coastal zones, and areas with frequent fog or overcast conditions, monthly differences show up as differences in annual power generation.

Also, temperature data are important. The output of solar modules decreases as temperature rises. Therefore, even if solar irradiance is the same, power generation will vary when temperature conditions differ. In hot regions, temperature-related losses tend to increase, while in cold regions the temperature can be advantageous even if irradiance is lower. If you look only at irradiance when explaining differences in generated power, you will overlook the effects of temperature.

Wind speed data also has an indirect effect. Because wind tends to cool modules, it affects the evaluation of temperature losses. If wind speed data are unreliable or deviate from the actual installation conditions, the results of the temperature model may differ.

When handling meteorological data in PVSyst, it is important to check not only the data source name but also the distance to the site, the elevation difference, the data period, the approach to filling missing data, and the monthly solar irradiation trends. If there is a difference in generated energy, first compare the meteorological data side by side, and make sure you can explain the differences in annual irradiation, monthly irradiation, and average temperature—this will make it easier to account for the discrepancy.

Reason 2: The assumptions about orientation, tilt, and array configuration are different

The second reason is a difference in assumptions about orientation, tilt, and array layout. In photovoltaic systems, the amount of solar radiation received depends on which direction the modules face, at what angle they are installed, and how they are spaced in rows. When checking layout and array settings in the PVSyst manual, you need to verify that these assumptions match those used in other calculations.

Even with the same installed capacity, due south-facing and southwest-facing systems generate power at different times of day. Even if the difference in annual energy production is small, there are differences in morning and evening output, peak times, and monthly generation patterns. Especially when evaluating feed-in tariffs and self-consumption rates, the time-of-day generation profile—not just the annual total—is important, so orientation differences cannot be ignored.

Tilt angle also causes differences in energy yield. A low tilt angle tends to receive more solar radiation in summer, but can result in lower energy yield in winter. A high tilt angle can be advantageous in winter, but it introduces trade-offs with installation density, wind load, and racking design. If the tilt angle set in PVSyst does not match the design drawings or racking specifications, the results will differ.

Also, in array layouts, row spacing, number of rows, module orientation, and mounting height have an impact. Even if two layouts appear to have the same capacity on the drawings, the effects of mutual shading differ between a layout with wider row spacing and one that packs modules tightly into the site. If you create a 3D scene in PVSyst, inter-row shading is reflected, so the energy yield may be lower than what a simple capacity calculation predicts.

What often happens in practice is that, in initial assessments, estimates are made using an "ideal orientation and tilt," and in detailed analyses the layout is changed to match the actual site shape. In such cases, even if energy production decreases, this should be viewed as the result of real layout constraints being reflected, rather than PVSyst's calculations being too strict.

Furthermore, in projects with arrays facing multiple orientations, the differences in energy output become more complex. A simple installation that faces only south and an installation divided into east–west or multiple planes will have different generation profiles even if the total capacity is the same. When configuring multiple sub-arrays in PVSyst, you need to verify that the capacity, orientation, tilt, and connection point of each sub-array are correctly assigned.

When checking differences in azimuth and tilt, it is important to cross-check the drawings, layout tables, PVSyst configuration screens, and the conditions field of the output report. Rather than simply looking at the difference in annual energy production, confirming which array conditions differ by how many degrees, how much the layout has been changed, and whether inter-row shading is included will make it easier to explain the cause of the energy production difference.

Reason 3: The handling of shadows and the accuracy of 3D scenes differ

The third reason is how shadows are treated. As causes of energy yield differences in PVSyst, near shading, far-field terrain, inter-array shading, and shadows from surrounding structures are all very important. Because the impact of shadows varies with season and time of day, looking only at annual energy yield can make it difficult to pinpoint the cause.

There are simplified settings for handling shading and more detailed settings that use a 3D scene. In a simplified assessment, shading may be entered as a fixed loss rate. By creating a 3D scene in PVSyst you can model surrounding buildings, trees, terrain, rows of racking, and other obstructions, and capture their time-dependent shading effects. The more detail you include, the closer the result will be to reality, but the estimated energy production may come out lower than initial rough estimates.

What is particularly important to note is that shading effects do not occur uniformly. When the sun is low in the morning and evening, long shadows are more likely to occur, and because the sun’s altitude is low in winter, the impact of shading becomes greater. Therefore, even if the difference in annual power generation is only a few percent, monthly generation in winter can appear as a much larger difference.

Distant terrain is also an element that is easy to overlook. In regions surrounded by mountains, or where there is high terrain to the east or west, solar radiation immediately after sunrise and before sunset is blocked. Even without large buildings nearby, terrain can shorten the actual solar radiation acquisition time. Whether or not distant terrain is considered in PVSyst can lead to differences in power generation, especially in the morning and evening and during winter.

Near-field shading is influenced by buildings, utility poles, fences, trees, cubicles, power conditioners, existing equipment, and similar objects. The results change depending on whether you include these in the 3D scene, ignore them, or treat them as simplified losses. In practice, small obstructions overlooked during a site survey can become problems later.

Also, the accuracy of the 3D scene itself is important. If the heights, positions, shapes, orientations of obstructions, or ground elevation are inaccurate, shadow calculations will not be accurate either. If the coordinates on the drawings are misaligned with the layout in PVSyst, shadows will be over- or under-estimated. This is especially true on sloped or engineered terrain, where the shadow results change depending on how much of the ground elevation is reflected.

When explaining differences in power generation caused by shading, it's easiest to understand if you first separate the results into: the outcome with no shading considered at all, the outcome when shading is included as a simplified loss, and the outcome when shading is modeled in a 3D scene. If the differences are large, it's important to check monthly generation and loss charts to see which seasons are affected.

When consulting the PVSyst manual, be sure to check not only how to create 3D scenes but also how shading loss results are displayed in the report. In practice, you will be expected not just to note that shadows were included, but to explain which shadows are reducing energy production and by how much.

Reason 4: The way loss settings are configured is different

The fourth reason is differences in loss settings. PVSyst's energy production is calculated by subtracting various losses from the theoretically available solar irradiation. If these loss settings differ by person in charge or by project, the final energy yield will differ.

There are many types of losses. Losses due to module temperature, DC cable losses, AC cable losses, mismatch losses, soiling losses, reflection losses, inverter conversion losses, clipping losses due to overloading, downtime losses, degradation rate, and so on. Although each may seem like a small figure on its own, taken together they can have a large impact on annual energy production.

Soiling loss is particularly likely to cause differences in power generation. Soiling loss varies depending on the region, tilt angle, rainfall, surrounding environment, and maintenance frequency. In areas with a lot of dust and sand, at sites close to factories or roads, or on systems with a low tilt angle where rain does not easily wash away dirt, soiling loss may be estimated to be larger. On the other hand, in initial rough estimates soiling loss is sometimes underestimated.

Cable losses are another factor that often causes discrepancies. Even if a standard loss rate is used for rough estimates, losses in the detailed design are set based on the actual cable lengths and wiring routes. In projects with large sites or long distances to the PCS or transformer/substation equipment, cable losses cannot be ignored. If the power output in the detailed study is lower than in the initial assessment, it may be because the wiring conditions were taken into account.

Mismatch losses also need to be checked. Module-to-module variations, string configuration, shading patterns, and connection conditions can cause outputs not to combine as ideally expected. When multiple orientations or multiple tilt angles are combined on the same system, electrical mismatches can lead to differences in energy production.

For inverter-related matters, not only conversion efficiency but also input voltage range, rated capacity, oversizing, and how clipping is handled are important. In designs that increase DC capacity while constraining the AC side, output can hit a ceiling during periods of strong solar irradiance. If PVSyst results show clipping losses, simply comparing energy production by module capacity alone can lead you to misattribute the cause of the difference.

Downtime losses and the availability rate are also important in practice. The outlook for annual power generation varies depending on how much you allow for scheduled inspections, equipment shutdowns, grid-side output control, and incident response. For investment decisions and revenue projections, it is necessary to distinguish between the technically possible amount of generation and the amount that can actually be sold.

When checking the loss settings in the PVSyst manual, it is important to understand the meaning of each item and which values are automatically calculated and which are entered manually. If there is a difference in energy production, check the loss diagram to see which loss items are large and which values have changed compared with previous simulations. If you can explain the breakdown of losses, you can present it not as a mere fluctuation in energy production but as a difference in design conditions.

Reason 5: The output items and periods being compared are different

The fifth reason is that the output metrics or the periods being compared are different. This may seem very simple, but it frequently occurs in practice. PVSyst reports display multiple generation metrics and loss metrics. Therefore, if you discuss results without confirming which numbers are being compared, it can appear that there are differences in energy generation.

For example, the output of a photovoltaic array and the AC output after passing through the inverter are not the same. Moreover, the energy at the grid interconnection point, and the energy used to calculate feed-in revenue, also differ depending on how losses are deducted and how downtime rates are handled. If you compare generation figures from other documents without understanding what the numbers shown in a PVSyst report mean, you will not be able to identify the causes of the discrepancies.

Also, even when you think you are comparing annual energy production, the reference periods may differ. One document may show first-year generation, another may show average annual generation, and yet another may show generation in a specific year after degradation. In projects that take module aging into account, the figures for the first year and the 20-year average will differ.

You should also be careful with monthly energy generation. When comparing PVSyst's monthly results with measured data or monthly results from other software, you need to confirm whether the year in question is a standard year, a particular year, or an actual measured year. If you compare a simulation using a standard meteorological year with measured values from a specific year, it is natural that the generation will differ due to weather variations.

Furthermore, differences in units can also cause generation amounts to appear different. kWh, MWh, kWh/kWp, kWh/kW, PR, capacity factor — when the metrics differ, the meaning of the comparison changes. For total generation, larger facilities yield larger numbers, but if you compare generation per kWp you can remove differences in system size and make a fair comparison. For projects where installed capacity varies slightly, it is important to check generation per unit of capacity as well as total generation.

Another easily overlooked point is whether the basis is DC capacity or AC capacity. At solar power plants, module capacity and PCS capacity commonly differ. Which one is used as the denominator changes how capacity utilization and power generation efficiency appear. When discussing differences in generated output, you must use the same capacity as the denominator.

When using the PVSyst manual, always verify what each item in the output report indicates. When explaining differences in power generation, it is important to clearly state the names of the items being compared, the units, the period, the capacity basis, and whether degradation is taken into account. Doing so will prevent many unnecessary misunderstandings.

Basic procedure for checking power generation differences

When PVSyst shows a discrepancy in energy production, it's important to check things systematically rather than rely on intuition. The first thing to check is not the total energy production but whether the input conditions match. Confirm that the capacity, module model, PCS configuration, azimuth, tilt, installation site, meteorological data, and loss settings are all aligned.

Next, check the meteorological data. Verify whether the annual solar irradiation, monthly solar irradiation, average temperature, and the data period differ. When there is a large difference in energy output, it is not uncommon for differences in meteorological data to be the cause. In particular, if different datasets were used for the preliminary assessment and the detailed assessment, this should be the first thing to check.

Next, check the layout and shading conditions. Even with the same orientation and tilt, results can vary depending on row spacing, obstructions, distant terrain, and the presence of a 3D scene. If monthly energy production shows large differences in winter, shading effects are suspected. Checking whether morning/evening shadows and terrain shading are being reflected makes it easier to explain the differences.

Next, review the loss diagram. The loss diagram in PVSyst helps you understand at which stages and by how much the energy production is reduced. By looking at which losses are large, you can narrow down the causes of any differences in energy production. For example, whether temperature losses, inverter losses, or shading losses are dominant will change what you need to explain.

Finally, confirm the output items. Check that the values being compared use the same units, the same time period, the same capacity basis, and the same degradation conditions. If you discuss the results without verifying these points, you may end up comparing different numbers even though they actually show the same trend.

When checking differences in power generation, it's also important not to assume a single cause. In practice, differences often result from several factors stacking up—for example, slight variations in weather data, minor differences in shading conditions, and small discrepancies in loss settings. Therefore, breaking down the discrepancies one by one and organizing how much each factor contributes in percentage terms will make your explanation more persuasive.

Approaches to Make It Easier to Explain Differences When Evaluating Projects

When presenting PVSyst results to internal teams or clients/owners, saying only "generation decreased" or "there was a difference because we changed the conditions" is insufficient. You need to explain what changed, why it changed, and how much it affected the results. To do that, it is important to organize the simulation conditions from the outset with an awareness of explaining the differences.

First, consider dividing the study into stages. In the initial study, basic design, detailed design, studies for financial institutions, and the final pre-construction review, the required level of accuracy and the assumptions differ. The initial study may use approximate conditions, while the detailed design reflects conditions based on drawings and equipment specifications. Differences in power output are often a natural change that occurs as the study progresses.

Next, keep a record of any changes to the conditions. Changes that affect energy production—such as updating meteorological data, correcting the azimuth, adjusting the tilt angle to the design value, adding shading conditions, setting cable losses to the as‑built design values, or conservatively revising soiling losses—should be recorded. Without a history, it will be difficult to explain later differences in energy production.

Also, it is important to share the assumptions together with the results. If you only share the annual generation figure, the recipient cannot determine under what conditions that number was calculated. By presenting meteorological data, system capacity, azimuth, tilt, shading, losses, and the comparison period together, the meaning of the generation figure becomes clear.

If a difference in generation appears, consider whether the discrepancy can be explained as the result of improved accuracy rather than treating it as something bad. If incorporating shadows, wiring losses, temperature losses, and downtime risks that were not apparent in the initial study has caused the estimated generation to decrease, that means the outlook has become more realistic. Conversely, revising assumptions and removing overly conservative settings can also lead to an increase in estimated generation.

What's important is not to treat PVSyst's numbers as absolute. PVSyst is a powerful tool for performing detailed simulations, but the validity of its results depends on the validity of the input conditions. Even if you operate it following the manual, if the assumptions do not match reality, the results will also deviate from reality. The ability to explain differences in energy production is both the ability to interpret the tool's results and to verify the input conditions.

Points to Note for Using the PVSyst Manual in Professional Practice

When reading the PVSyst manual, it is important not only to follow the screens in order but also to be aware of which parts of the energy production each setting affects. Understanding which settings affect the amount of solar irradiation captured, which settings affect electrical losses, and which settings affect the report output items will make it easier to find the causes of differences in generated energy.

A common pitfall for beginners is proceeding with the default settings. PVSyst provides standard values and automatic settings in some areas, but these are not necessarily suitable for every project. In particular, loss settings, meteorological data, equipment selection, and shading conditions need to be checked for each project. Even when using standard values, you should be able to explain why those values are appropriate.

Also, be careful when reusing settings from past projects. Even for power plants of a similar scale, appropriate settings can change if the location, weather conditions, orientation, tilt, equipment configuration, cabling distances, or maintenance conditions differ. Copying and using a previous PVSyst file can leave old loss settings or shading conditions intact, causing differences in energy generation and unexplained results.

Pay attention to how numerical data are presented when preparing reports. Rather than emphasizing only annual energy production, include monthly generation, loss diagrams, performance ratio, and the main input conditions so they can be checked, which makes review easier. When differences in generation become a subject of discussion, if the conditions are clearly stated in the report they provide a starting point for identifying the cause.

Furthermore, when comparing multiple scenarios, it is important to manage the conditions you change one at a time. If you compare cases where meteorological data, layout, and losses are all changed at once, you will not know which condition is affecting the difference in energy output. In design comparisons, first create a baseline case, and then separate and check conditions such as only azimuth, only tilt, only losses, and only shading, which makes assessment easier.

The primary purpose of using the PVSyst manual in practice is not simply to learn the correct operations. It is to understand the connection between input conditions and results, and to be able to explain differences in energy production. As projects grow larger, even small differences in energy production can have a significant impact on revenue assessments. That is why it is important to systematically grasp the reasons for differences in energy production.

Summary

When checking the reasons for differences in energy yield in the PVSyst manual, it is important not only to look at the calculation results but to break down and consider the differences in conditions. The main reasons for differences in energy yield can be organized as: choice of meteorological data, assumptions about azimuth, tilt, and array layout, treatment of shading and accuracy of the 3D scene, loss settings, and differences in the output items and periods being compared.

If the meteorological data changes, the solar irradiance and temperature conditions change, resulting in differences in annual energy production. If the azimuth, tilt, or array layout change, the amount and timing of irradiance received by the modules change. If shading conditions are modeled in detail, production can be lower than initial estimates. If loss settings are revised to be realistic, cable losses, soiling losses, temperature losses, downtime losses, and the like accumulate and affect the results. Furthermore, if the units, periods, or capacity baselines of the values being compared differ, the differences can appear larger than they actually are.

In practice, what matters is not treating differences in power generation as mere errors or mistakes, but being able to explain which conditions changed and how. For that, you need to connect and review PVSyst's settings screens, weather data, 3D scene, loss diagram, and output reports.

When there is a discrepancy in energy production, first check the input conditions, then review the meteorological data, layout conditions, shading conditions, loss settings, and output items in that order; checking them in this sequence makes it easier to identify the cause. Ultimately, being able to explain not just the annual energy production number but also the assumptions from which that number was derived is an important practical skill for mastering PVSyst.