Five Basics You Should Know Before Using PVSyst’s Optimization Features

Table of Contents

• What PVSyst’s optimization features are used for

• Basic 1: Separate items to be varied in optimization from items to be fixed

• Basic 2: Improve the precision of input conditions before optimizing

• Basic 3: Do not judge by a single evaluation metric

• Basic 4: Incorporate on-site constraints before optimizing

• Basic 5: Optimization results are usable only after sensitivity checks

• Summary: Applying PVSyst’s optimization features in practice

What PVSyst’s optimization features are used for

Many practitioners who search for PVSyst are trying to improve their power generation forecasts even slightly, quickly compare differences between design conditions, or refine the assumptions behind profitability. The optimization features serve as a way to efficiently compare multiple design candidates and identify which conditions influence the results. They are not merely tools to raise numerical values, but should be used as instruments to improve the accuracy of design decisions.

A practical point to note is that the term “optimization” often gives the impression that the best solution will be produced automatically. In reality, the conclusions drawn depend greatly on which conditions are treated as variables, which constraints are applied, and which metrics are prioritized. In other words, the design philosophy before entering optimization tends to have a greater influence on results than the optimization tool itself.

Also, in photovoltaic system design it is not always sufficient to look only at power generation. There are many aspects to consider—site conditions, equipment layout, shade effects, maintenance access, future operability, and the validity of electrical configurations. When using PVSyst’s optimization features, you should not separate these elements; instead, clarify which aspects will be refined by simulation and which will be addressed in detailed design or on-site decisions.

This article explains five basics you should know before using PVSyst’s optimization features. It is organized to be useful both to those trying optimization for the first time and to those already using it but unsure how to interpret results, focusing on practical pitfalls and key points. By being clear about what to assume, what to question, and how to translate results into decisions before any optimization操作, you can greatly change how you use the outcomes.

Basic 1: Separate items to be varied in optimization from items to be fixed

The first basic is to clearly separate what will be varied in the optimization and what will be fixed. If you start analyses with this unclear, the meaning of results becomes hard to interpret. For example, if you simultaneously make large changes to mounting angle, azimuth, row spacing, equipment capacity, and electrical configuration, it becomes difficult to determine which element improved the outcome. Optimization becomes easier to judge the more organized the choice of variables is; greater freedom does not automatically make optimization more useful.

In practice, start by fixing conditions that absolutely cannot be changed. Site shape, slope orientation, transport access, connection conditions, provision of maintenance aisles, and clearance from surrounding structures are not easily altered later. If you ignore these constraints and run optimization, top-ranked options may look good numerically but be practically infeasible. Before starting optimization, identify items in the design conditions that cannot be changed and fix them as prerequisites.

Next, limit the number of items to be varied—fewer variables are easier to handle in practice. For example, in an initial stage compare only mounting angle and row spacing, and then in a subsequent stage refine the electrical configuration. Dividing the study into two or three stages helps to identify which factors cause differences in results and facilitates explanation and internal consensus among designers.

It is also important to set realistic ranges for the variables. Even if a wide range can be explored theoretically, including values that are impossible in practice can derail the optimization direction. Narrow down candidate angles, upper and lower bounds for spacings, and the number of configuration options in advance according to the design intent. PVSyst’s optimization features are most effective when used to prioritize a reasonable candidate set, not to expand choices without limit.

Adopting this mindset alone will change how optimization results appear. Instead of judging by numbers alone, you should be able to explain what was fixed and which specific conditions were compared. In practice, being able to verbally justify the rationale for the chosen proposal is extremely important. In that sense, organizing variables before optimization should be seen not only as preparation for calculation but as building the foundation for design decisions.

Basic 2: Improve the precision of input conditions before optimizing

The second basic is to improve the precision of input conditions before starting optimization. This might sound obvious, but it is often where the biggest practical differences arise. No matter how attractive the optimization results look, if inputs such as solar irradiation conditions, loss assumptions, layout conditions, or shading inputs are coarse, the results will be no more than indicative. Because the optimization function compares based on the provided assumptions, the vaguer those assumptions are, the vaguer the conclusions will be.

Be particularly careful not to treat meteorological or installation conditions as uniform representative values. That may be acceptable in preliminary studies, but if you proceed to optimization, you need to at least identify what will be the major constraints for that specific site. Seasonal solar patterns, morning and evening shading, the visual impact of row spacing on sloping ground, and the existence of nearby objects can all influence the optimization direction. Performing numerical comparisons with these aspects unclear can produce differences that are meaningless.

When speaking of input precision, attention often focuses only on weather data, but in practice organizing loss settings is also important. If temperature effects, wiring and conversion losses, soiling and aging assumptions, and shading loss treatments are not standardized, the comparison basis among cases will not be aligned. If one scenario is set conservatively while another is set optimistically, apparent rankings can easily be reversed. Before running optimization, clearly separate the conditions that should be common across comparisons from those that should vary by scenario.

Also, improving input precision is not an end in itself. What matters is to prioritize the items that affect the final decision. Trying to make everything high precision from the start can unnecessarily slow the process. The required precision differs by stage: preliminary, basic design, and pre-tender or internal review stages each have different needs. To use PVSyst’s optimization features effectively, determine the necessary input precision at each stage and prepare sufficiently accurate assumptions for that stage.

When input conditions are well prepared, differences in optimization results become meaningful. Conversely, when assumptions are inconsistent, even very detailed comparison tables will not improve decision quality. Practitioners need comparisons that can withstand decision-making, not perfect numbers. Therefore, spending time improving input conditions before rushing to optimization is ultimately more efficient. Those who are proficient in PVSyst often invest heavily in organizing assumptions before running calculations for this reason.

Basic 3: Do not judge by a single evaluation metric

The third basic is not to judge using only one evaluation metric. Hearing the word “optimization,” many people assume they should pick the case with the highest energy yield. However, in practice the annual energy being maximal does not necessarily mean the option is the best. You should also look at layout efficiency, susceptibility to shading, maintainability, electrical margin, and future operational stability.

For example, a scenario that is slightly superior in annual energy might have row spacing that is too tight for maintenance access or suffer locally severe shading in winter, making it difficult to manage on site. Conversely, a scheme with slightly lower generation but more comfortable spacing, easier construction, and lower operational uncertainty may be more likely to be adopted. When reviewing PVSyst’s optimization results, consider not only the maximum value but also the magnitude of differences and their practical implications.

Having multiple evaluation metrics helps prevent misreading numbers. Looking at energy yield, specific yield (yield per capacity), loss breakdowns, differences by influencing factors, and area efficiency from different angles clarifies each option’s characteristics. If you only look at a simple ranking, it can be hard to see why a case ranks high or where its weaknesses lie. To interpret the background of results, it is essential to line up multiple metrics and check their consistency.

Moreover, the evaluation axes should change with the study stage. In the early stages prioritize overall generation trends and layout feasibility; in detailed stages emphasize shading, loss breakdowns, and operational margins. If you keep pursuing the same metric at every stage you may miss perspectives needed as the design is refined. PVSyst’s optimization features are not devices to provide a single answer; realistically, they serve as aids to narrow down options while balancing multiple metrics.

For practitioners, it is important not to adopt the optimization result as the chosen proposal uncritically but to convert it into explainable decision material. Saying “it had the highest generation” alone may not be persuasive to supervisors or stakeholders. If you can explain that an option remains balanced across multiple metrics, respects constraints, and does not degrade drastically with small changes in conditions, your use of the results becomes more robust.

Basic 4: Incorporate on-site constraints before optimizing

The fourth basic is to incorporate on-site constraints before optimizing. This viewpoint is essential for turning desk comparisons into practical decisions. What PVSyst mainly shows is behavior in generation simulation, but in the field there are always other constraints. If you optimize without considering constructability, regulatory constraints, maintenance vehicle access, drainage, site preparation conditions, and consideration for surrounding areas, the resulting options will not be adoptable.

Especially on complex sites, theoretical optimal layouts often do not match layouts that can actually be implemented. When there are steep slopes or large elevation differences, long narrow sites, many existing structures, or non-uniform orientations, simple comparisons do not apply. In such cases, before entering optimization you must decide which areas are candidates for installation and which should be excluded, and determine how much aisle and clearance to secure. Adding site constraints afterwards leads to repeatedly redoing optimization and is inefficient.

On-site constraints are often seen as factors that reduce generation, but in practice they are conditions that protect quality. Although a dense layout may look advantageous in simulation, it can increase burdens during construction and maintenance, and affect post-commissioning issue handling. If you consider aisles, clearances, and inspectability during optimization, you can avoid major rework later. Anticipating how the site will be used from the early design stage leads to more reasonable optimization outcomes.

Additionally, incorporating on-site constraints refines the set of comparison candidates. By removing infeasible options and concentrating on parameters that vary within feasible ranges, the density of analysis increases. Optimization is not smarter when candidate numbers are larger; it is more meaningful when limited to viable candidates. Practitioners who effectively use PVSyst’s optimization features examine site conditions carefully before simulation and set up the comparison framework accordingly.

This approach is also useful for internal and external explanations. When explaining why only certain candidates were compared or why other options were not chosen, stating that the comparison was conducted with site constraints in mind increases transparency. Excessive pursuit of generation values can lead to design changes requested from the site later. Properly organizing on-site information before optimization is not mere preliminary work but contributes to stabilizing the overall project schedule.

Basic 5: Optimization results are usable only after sensitivity checks

The fifth basic is not to accept optimization results as conclusions without performing sensitivity checks. Whether a top-ranked option from optimization is truly usable can only be judged after verifying that it does not collapse when assumptions are slightly altered. There is always uncertainty in input values, loss assumptions, and operational conditions. If results are overly sensitive to that uncertainty, an apparently optimal option may be difficult to adopt in practice.

Although “sensitivity checks” may sound difficult, the idea is simple in practice. For example, slightly vary key assumptions and observe how rankings and differences change—that alone is meaningful. Check how results move if row spacing is slightly widened, loss assumptions are set more conservatively, or shading treatment is changed. This reveals the robustness of each option. If the difference between first and second place is very small, it may be better to prioritize other practical conditions.

The important mindset here is to prefer robust options over the absolute optimum. In practice, conditions rarely align perfectly with assumptions. Minor adjustments often occur during construction, and field checks may necessitate revisiting assumptions. An option whose performance drops sharply with small changes is harder to deal with than one that maintains performance and feasibility under modest variations. If you want to apply optimization in practice, this robustness perspective is indispensable.

Conducting sensitivity checks also makes it easier to explain optimization results. Rather than saying only that a number was highest, you can explain that trends did not change substantially when key assumptions were altered and that advantages were confirmed from multiple perspectives. This strengthens your position in internal reviews and when explaining to clients. PVSyst’s optimization features deliver more value when used for comparison and verification to enhance the reliability of decisions than for one-off result output.

Ultimately, what matters is not to blindly trust optimization outcomes but to understand their tendencies. Identify which conditions results are sensitive to, where uncertainties lie, and which options are stable; this will change how you use the numbers. Practitioners are expected not to produce neat result tables but to make decisions that are unlikely to require rework. In that sense, sensitivity checks should be integrated into the optimization process from the start, not treated as additional tasks.

Summary: Applying PVSyst’s optimization features in practice

To summarize the basics you should know before using PVSyst’s optimization features: the most important point is not to view optimization as an automatic answer key. The quality of an optimization depends on what you set as variables, which conditions you fix, how much you refine input assumptions, which evaluation axes you use, how you incorporate on-site constraints, and how you check sensitivity. How you proceed with the design matters more for result accuracy than how you use the feature itself.

For practitioners, PVSyst is not merely a tool to calculate generation; it is a tool to compare multiple options and organize the evidence needed for decision-making. Therefore, do not stop at the ranking list—interpret option characteristics, assumptions, feasibility, and robustness. With this perspective, the same features can produce very different decision quality.

Moreover, turning desk simulations into on-site-applicable decisions requires precise design assumptions and reliable on-site information. If site conditions, elevation differences, installable areas, and maintenance access are unclear, optimization accuracy is limited. Bringing simulation assumptions closer to field reality, even if it seems roundabout, is the most reliable shortcut.

In that respect, improving on-site surveying and positional data accuracy complements the use of PVSyst. For example, using an iPhone-mounted high-precision GNSS positioning device such as LRTK makes it easier to improve the accuracy of on-site position checks and records, which helps organize design assumptions. By deepening desk comparisons with PVSyst’s optimization features and simultaneously enhancing the reliability of field-acquired information, you can make stronger, more practical decisions.

When you try optimization in PVSyst, first return to these five basics and focus not on producing results but on creating a state in which you can use results correctly. If you achieve that, the optimization features will become not just convenient tools but practical decision-support that connects design and the field.