Six Checks to Avoid Mistakes in PVSyst Loss Rate Settings

When performing energy yield predictions in PVSyst, many causes of results being much lower than expected, or conversely overly optimistic compared to actual performance, come down to how loss rates are entered. PVSyst’s official documentation also notes that while reasonable initial values are provided for loss parameters, it is recommended to carefully redefine them for each project after the first simulation. Moreover, because PVSyst treats incidence angle correction, soiling, thermal effects, module quality, mismatch, wiring, auxiliary consumption, external transformers, and so on as separate items, broadly increasing losses “to be conservative” can actually undermine the consistency of the prediction. ([PVsyst][1])

What truly matters in setting loss rates is not placing large numbers. It is deciding where each loss is calculated, what overlaps with what, and what portion of generation is being evaluated, then placing values according to project conditions. This article organizes and explains six checkpoints that practitioners setting loss rates in PVSyst should keep in mind during both the estimating and detailed design phases. ([PVsyst][1])

Table of Contents

• Why PVSyst loss rate settings determine energy yield predictions

• Check 1 Organize where losses are placed to prevent double counting

• Check 2 Do not decide wiring losses by percentage numbers alone

• Check 3 Match temperature losses to racking and ventilation conditions

• Check 4 Treat soiling losses monthly, not as a fixed annual value

• Check 5 Treat module quality and mismatch as separate items

• Check 6 Clarify the scope of downtime and auxiliary loads

• Practical procedure for reviewing loss rate settings

• Conclusion

Why PVSyst loss rate settings determine energy yield predictions

The difficulty of PVSyst loss settings is that losses may appear to be consolidated in one place but in reality have distinct meanings at different points in the power flow. Some losses act before or after solar irradiance reaches the module surface, some are due to module temperature rise, and some, like wiring resistance, vary with current and operating point. PVSyst handles these as detailed “Array and system losses,” and the results can be checked in the Loss Diagram and time series outputs. Therefore, applying a broad correction carelessly can misplace the physical meaning and the calculation location, making it unclear where generation was actually lost. ([PVsyst][1])

In practice, required precision and accountability differ between stages: a rough estimate during sales, comparisons in basic design, finalized values in detailed design, and reconciliation after operation begins. While initial values may be used in early estimates, the further a project progresses the more one must be able to explain “why that loss rate.” PVSyst itself assumes that after the first simulation each loss should be redefined to match project specifications; loss rates are not meant to remain as initial values until the end, but should be supported by stronger justification as conditions solidify. ([PVsyst][1])

Check 1 Organize where losses are placed to prevent double counting

The first thing to confirm is where losses should be placed. PVSyst treats incidence angle correction, soiling, irradiance, thermal effects, LID, module quality, mismatch, DC wiring, AC wiring, external transformer, auxiliary consumption, and so on as separate items. That means if the same phenomenon is entered into multiple items, the calculation will stack them as separate losses. Common practical mistakes include entering soiling month-by-month and then adding another vague “safety factor” loss, or setting mismatch while also lumping uneven shading and nonuniform soiling into a separate fixed loss. Such entries might look conservative, but they lack explainability and reduce reproducibility in project comparisons. ([PVsyst][1])

PVSyst’s guidance treats soiling as a distinct loss and evaluates the electrical impact of partial shading via near-field shading or module layout. Although some causes of mismatch include uneven soiling or partial shading, each has its own computational domain. Therefore, you need to decide up front, for each project, “what to represent where.” For example, if nearby obstructions are evaluated in 3D or via layout, you should not then add that impact again as a larger fixed loss. Practitioners who avoid mistakes in loss settings remove overlaps in where losses are placed before increasing numbers. ([PVsyst][2])

Another easily overlooked item is auxiliary consumption. PVSyst’s official documentation explicitly states that if auxiliary losses are already included in the conversion efficiency, redefining them separately as an Auxiliary loss will result in double counting. If definitions on the inverter side and settings under Detailed Losses conflict, it can become unclear which side carries the loss; therefore, auxiliary loads, nighttime consumption, and PCS peripheral equipment treatment should be ruled early. ([PVsyst][3])

Check 2 Do not decide wiring losses by percentage numbers alone

Wiring loss is the item most prone to “numbers walking off on their own” in practice. In PVSyst the fundamental parameter for wiring loss is resistance, and the percentage display of loss is merely a convenient input method. The official documentation also explains that percentage loss scales with actual power and therefore is not a reliable standalone parameter. Moreover, the annual wiring energy loss often does not equal the nominal loss percentage specified; typically, over a year it becomes about 60% of the specified nominal percentage. In other words, treating wiring loss as a fixed “2% every year” image can lead to misestimation. ([PVsyst][4])

If you do not understand this, you might set DC wiring loss at 2% and AC wiring loss at 1% uniformly during the estimating stage and then carry those values straight through to detailed design as fixed numbers. In reality, losses change with cable length, cross-sectional area, circuit configuration, string current, and MPPT arrangement. In large projects string length differences are common, leading electrically not only to wiring resistance issues but also to voltage differences between strings and mismatch concerns. Therefore, while provisional percentage inputs are acceptable early on, as you approach finalization you should transition to estimating resistance based on actual lengths from single-line diagrams and layouts. ([PVsyst][5])

It is also important to define the reference points for wiring losses — i.e., from where to where they apply. PVSyst treats DC and AC sides separately and can express external transformer losses separately if needed. If you manage a “total wiring loss” vaguely, definitions will diverge among internal materials, customer-facing documents, and PVSyst inputs. Before making numbers conservative, clarify whether DC, AC, transformer, and up-to-point-of-delivery are included; doing so substantially increases prediction reliability. ([PVsyst][1])

Check 3 Match temperature losses to racking and ventilation conditions

Temperature loss is another item that's easy to reuse but tends to diverge from actual project conditions. PVSyst uses a thermal model employing a heat loss coefficient U to describe array thermal behavior. The official tutorial represents U as Uc plus Uv×wind speed, but because wind speed data are often insufficiently defined in meteorological data and Uv itself is poorly known, in practice it is recommended not to use wind dependence, set Uv to 0, and include average wind effects in Uc. In other words, although detailed wind settings may look sophisticated, without solid justification they do not necessarily improve accuracy. ([PVsyst][6])

More importantly, connect the meaning of Uc to racking conditions. PVSyst suggests Uc guidance values of 29 W/m²K for fully open ventilation conditions and 15 W/m²K for integrated conditions where the rear is thermally insulated, and notes there is no single universal value for intermediate conditions. This means that for the same nominal output and module, whether the installation is an open ground-mounted rack or near a roof surface, and how much rear ventilation is provided, significantly affects how temperature loss should be set. Even when borrowing Uc values from past projects, confirm that structure, spacing, rear clearance, and airflow characteristics are similar; otherwise the setting can drift. ([PVsyst][6])

PVSyst’s thermal model also proposes adjusting U values to match actual system data. Therefore, avoid adjusting Uc merely to hit a target generation number. Temperature loss should be a number that can be explained based on site conditions; using it as a tuning parameter — “raise Uc because the result was too low” or “lower Uc because it was too high” — will break consistency with other loss items. When you suspect temperature loss, return to racking and ventilation conditions first. ([PVsyst][7])

Check 4 Treat soiling losses monthly, not as a fixed annual value

Soiling losses are among the items most prone to divergence from field reality. PVSyst’s official documentation notes that soiling strongly depends on rainfall and thus can be defined with monthly values; in simulation it is treated as an irradiance loss. Also, because snowfall is not included in meteorological data, sites where snow is important can represent its effect in specific months as partial or full soiling attenuation. Conversely, treating soiling with a single annual fixed value discards seasonal variations such as deterioration during dry seasons, recovery during rainy seasons, and large attenuation during snowfall months. ([PVsyst][8])

In practice, even within the same site soiling behavior differs depending on proximity to the coast, adjacent farmland, number of unpaved roads, dust from recent earthworks, traffic, and the presence or absence of cleaning operations. Nevertheless, if you use an internal standard like “soiling loss 3% every year” verbatim, you cannot express seasonal variability. The important point is to treat soiling not as an insurance number but as a monthly hypothesis reflecting rainfall and cleaning operations. That way, when reconciling actual performance later, it is easier to find which season deviated. ([PVsyst][8])

For sites affected by snow, clarify whether to treat snow as soiling, as downtime, or to incorporate operational snow removal assumptions. Leaving this ambiguous and applying the same winter effect to both soiling loss and downtime again leads to double counting. Monthly soiling settings may seem bothersome, but in practice they are easier to explain than a single annual fixed value. Linking them to site photos, surrounding environment, cleaning plans, and seasonal factors gives the loss numbers grounding. ([PVsyst][8])

Check 5 Treat module quality and mismatch as separate items

Do not confuse module quality loss and mismatch loss. In PVSyst, Module quality loss is a parameter representing the expected actual module performance relative to the nominal specification. Its initial value is set based on module nameplate output tolerances, and with strong positive selection it can even be a negative loss (i.e., a gain). This value is a constant loss applied proportionally across the whole simulation. In other words, it addresses “how actual initial performance compares to nameplate values in the datasheet,” and is distinct from variations caused by wiring, shading, or temperature. ([PVsyst][9])

Mismatch loss, on the other hand, is the phenomenon where the combined I-V characteristic’s maximum power point is lower than the simple sum due to variations among modules or strings. PVSyst’s documentation notes that if the variance is under 2% the loss is below 0.5%, but the loss grows rapidly as variance increases, and default values should be adjusted per project. For normal string lengths the default upper bound is typically set to 2%, so adding an extra “just-in-case” loss under another name leads to overly conservative predictions. ([PVsyst][10])

In practice, separate these two: tie module quality loss to procurement specifications, tolerances, selection policy, and acceptance testing strategy; and tie mismatch to string configuration, mixed lots, replacement mixing, nonuniform soiling, temperature differentials, and partial shading risks. Using a safety margin in module quality, another in mismatch, and then adding an unnamed contingency is easy numerically but weak in terms of accountability. To produce predictions trusted in PVSyst, reconnect departmental information to each loss meaning. ([PVsyst][9])

Check 6 Clarify the scope of downtime and auxiliary loads

At the end of the day, treatment of downtime and auxiliary consumption affects the final energy prediction. PVSyst allows defining System unavailability as a time ratio or number of days, during which the system is treated as stopped in the simulation. This is useful for incorporating planned maintenance or fault-related stoppages into the prediction, but it expresses “the system does not generate for that period.” Therefore, decide at the outset whether PCS and substation outages, planned inspections, grid-connection constraints, or externally caused stoppages are to be handled here or treated separately as contractual availability assumptions. ([PVsyst][11])

The same applies to auxiliary consumption. In PVSyst auxiliary consumption is defined for the whole system under Detailed Losses and is included in the simulation only when checked. It covers management energy such as fans, air conditioning, electronic equipment, and lighting, and can be modeled with different daytime/nighttime assumptions. In other words, whether you want to see net output at the point of delivery or the theoretical output after array or inverter conversion will change what you include here. If you include auxiliary and downtime losses without clarifying the evaluation target, internal figures may look reasonable while customer or financier expectations diverge. ([PVsyst][12])

In practice, when preparing a loss-rate table it helps to separate “physical losses,” “operational losses,” “downtime losses,” and “losses beyond the point of delivery.” Classifying before entering values into PVSyst makes comparisons and explanations easier and makes it clear what to update if assumptions change mid-project. The more complex the loss-rate settings, the more dividing the scope rather than increasing numbers improves accuracy. ([PVsyst][1])

Practical procedure for reviewing loss rate settings

When reviewing loss rate settings, instead of immediately changing individual numbers, the quickest approach is to first check the Loss Diagram to see which stage shows the largest drop. PVSyst lets you track each loss in time series and monthly views, so check not only annual totals but also which month and which loss are having an effect. Identify whether wiring, thermal, or seasonal soiling differences are dominant, then review the supporting documents for each item to determine priorities for correction. Trying to detail everything from the start only increases effort; using the Loss Diagram as an entry point and focusing on the largest impacts is practical. ([PVsyst][13])

A clear review sequence is to keep the baseline case based on initial values, then replace values with evidence-based figures in the order: wiring, temperature, soiling, quality, mismatch, downtime. This lets you later explain how much each adjustment affected annual yield. It’s also useful to include sensitivity cases for project comparison and internal approval. For example, show cases with different cleaning frequencies, Uc values split by racking condition, or revised downtime days so the rationale behind numbers is clearer. Because PVSyst separates loss types finely, it is well suited to sensitivity analysis. ([PVsyst][1])

Another practical caution is software version differences. PVSyst release notes mention bug fixes related to IAM and soiling Loss Diagram display when opening older projects in Version 8. When reusing old projects for comparison, don’t just open files — recompute with the same assumptions and confirm the Loss Diagram and main losses are as intended. Especially if past project templates are used as internal standards, inspect not only loss rates but also result interpretation. ([PVsyst][14])

Conclusion

To avoid mistakes in PVSyst loss rate settings, it is more important to organize the meaning and placement of losses than to set large loss values. Do not judge wiring loss solely by a percentage; match temperature loss to racking conditions; consider soiling monthly; separate module quality and mismatch; and clarify the evaluation scope for downtime and auxiliary consumption to improve both prediction accuracy and accountability. Starting from PVSyst’s Loss Diagram makes it easier to see what to review and to narrow the gap with actual performance. ([PVsyst][1])

Keeping records with positional information of site checks, racking layout, maintenance access, surrounding objects, and the ease of cleaning and inspection further strengthens the basis for loss rates. Energy yield prediction differences arise not only from desktop number tuning but also from the accuracy of field condition records. If you want to streamline organizing such site information, using LRTK (an iPhone-mounted GNSS high-precision positioning device) to preserve installation conditions and surroundings for later review can also be effective.