5 Ways to Simulate Solar Power Generation｜Tips to Improve Accuracy

Simulating solar power generation is not a task that can be completed simply by multiplying the installed capacity by a coefficient. The rough estimates used in initial assessments and the forecasts used for internal presentations or adoption decisions require different levels of accuracy and different factors to be considered. Nevertheless, in practice people sometimes conclude the annual kWh solely from the installed capacity, or conversely scrutinize details so closely that the initial decision is delayed.

Considering that the readers searching for "solar power generation calculation" are practitioners, what they really want to know is not the difficult theory itself, but at which stage you should use which simulation method to obtain figures that can be used in your work. Therefore, this article organizes five representative methods for simulating solar power generation and clearly explains the situations each is suited for and the approaches to improve accuracy.

Table of Contents

• Basics of solar power generation simulation

• Method 1 Estimate by system capacity and annual coefficient

• Method 2 Aggregate using monthly solar irradiation conditions

• Method 3 Simulate separately for each installation surface

• Method 4 Apply loss coefficients to approximate actual electricity generation

• Method 5 Adjust using measured data or similar projects

• Tips to improve accuracy

• Common mistakes in simulations

• A clear procedure for practitioners

• Summary

Basics of Solar Power Generation Simulation

When simulating solar power generation, you should first clarify the difference between kW and kWh. kW is the output capacity of the installation, while kWh is the amount of electricity actually generated over a given period. For example, a 10 kW installation only indicates the system size and does not automatically mean 10,000 kWh per year. To estimate annual kWh, you need to consider where the installation is located, its orientation, its tilt angle, and the loss conditions under which it will operate.

The important thing here is not to try to complete the simulation perfectly in a single run. When forecasting solar power generation, it is more suitable for practical work to first grasp the rough outline and then add conditions to improve accuracy. It is difficult to put together all the fine conditions in the initial assessment, and conversely, if you proceed with only rough calculations until the end, you will be troubled by the gap from actual results.

Also, simulations have an objective. The required level of granularity changes depending on whether you want to compare system sizes, determine whether installation is feasible, or forecast self-consumption. In some situations a rough estimate of annual generation is sufficient, while in others it is better to look at monthly generation trends and how they overlap with time of day. In other words, a good simulation is not always about performing the most detailed calculation; it is about judging the accuracy needed at that moment and choosing the appropriate method.

The five methods covered in this article differ in the balance between accuracy and effort. Some allow you to quickly estimate from equipment capacity, while others use measured data to apply more field-oriented corrections. First grasp the overall picture, and as you read on, be aware of which stage you are currently at so it becomes easier to choose between them.

Method 1: Estimate using equipment capacity and annual coefficient

The simplest and most commonly used method is to approximate using the system capacity and an annual coefficient. The idea is simple: Annual generation (kWh) = System capacity (kW) × Indicative annual generation per 1 kW (kWh/kW·year). As long as you know the system capacity, you can quickly obtain an estimate of the annual kWh, making this suitable for preliminary evaluations and comparing multiple options.

For example, with a 10 kW system, if you use about 1,100 kWh per kW per year as a guideline, the annual generation is 11,000 kWh. For 5 kW it would be 5,500 kWh, and for 20 kW it would be 22,000 kWh, so you can intuitively grasp the scale according to system capacity. In practice, once the rough capacity is visible from the roof size and number of panels, this method is often used to organize the overall scale.

The strength of this method is its speed and ease of comparison. When arranging multiple system-size options—such as 3 kW, 5 kW, and 8 kW—first aligning their annual kWh estimates with this method makes the outline of the project clear. It is very effective for internal kickoff meetings and the preliminary stage before a rough proposal.

However, its weaknesses are also clear. It is difficult to adequately reflect individual conditions such as regional differences, installation orientation, shading, and losses. In other words, this method is excellent as a starting point, but relying on it alone for a final decision can cause results to skew either high or low. In particular, for projects where roof surfaces face multiple directions or for sites likely to be affected by shading, it is safer to proceed to the next method.

Even so, it is an extremely important first step in simulating solar power generation. By grasping the broad annual generation up front, it becomes easier to judge how finely to examine the conditions later. For practitioners, it can be said to be the basic method they should have.

Method 2: Stack by monthly solar radiation conditions

If the annual coefficient alone feels too coarse, the next effective approach is to build up calculations based on monthly solar insolation conditions. Rather than looking at the year as a whole, this method calculates generation for each month and sums those values to obtain the annual kWh. Because seasonal differences are directly reflected, it yields a simulation closer to reality than a rough estimate using only an annual coefficient.

The basic idea is that monthly generation (kWh) = system capacity (kW) × average daily equivalent generation hours (h) × number of days in the month × correction factor. For example, with a 10 kW system, if the average daily equivalent generation hours in a spring month are 4.0 h, the correction factor is 0.82, and the number of days is 30 days, the monthly generation is 10 × 4.0 × 30 × 0.82 = 984 kWh. If in another month the equivalent hours drop to 2.8 h, the generation from the same system will change considerably.

The advantage of this method is that it allows you to capture seasonal differences that are not visible from annual totals alone. Although generation may appear to increase in summer, there is an efficiency loss due to high temperatures, and in winter the hours of sunlight tend to be shorter. Declines during the rainy season and snowfall are also not visible when looking at annual aggregates. With month-by-month stacking, these kinds of fluctuations can be incorporated naturally.

It is also useful when you want to develop a forecast for self-consumption. Even if annual generation is sufficient, if months with high consumption and months with high generation are misaligned, you may not achieve the expected benefits. By conducting month-by-month simulations, it becomes easier to see how well demand and generation overlap. This perspective is especially important for buildings whose heating and cooling loads vary significantly by season, or for facilities with large differences in monthly operations.

On the other hand, because it requires more assumptions than Method 1, the required effort increases. That said, for projects that don't need to look at time-of-day breakdowns from the start, it is a very convenient method as an intermediate between the annual coefficient method and detailed analysis. When you want to improve accuracy one step beyond a rough estimate, choosing this monthly buildup first is a realistic approach.

Method 3: Simulate Separately for Each Installation Surface

The third method is to run simulations separately for each mounting surface. This is particularly effective for projects where the roof is divided into multiple planes or for ground-mounted installations where conditions differ by row. If you calculate the entire installation as a single unit, differences in orientation and angle can be masked, leading to predictions that are coarser than reality.

For example, suppose you install 6 kW on the south-facing roof surface, 2 kW on the east-facing surface, and 2 kW on the west-facing surface. The total capacity is 10 kW, but the south-facing 6 kW and the east- and west-facing 2 kW each have different generation conditions. In this case, rather than applying a single uniform factor to the whole 10 kW, it is more realistic to calculate the annual generation separately for the south, east, and west surfaces and then sum them.

The advantage of this method is that it reveals where generation is being increased and where it is being reduced. For example, you can see relationships such as the south-facing side being more efficient while the east-facing side is somewhat weaker, although mounting panels on the east and west faces increased the total capacity. As a result, it becomes easier to judge how to evaluate things for overall optimization, rather than simply judging by which orientation is superior.

Also, because the effects of shadows often differ for each mounting surface, this method pairs well with shadow correction. Conditions such as the south-facing surface having little shade while only the west-facing surface is shaded in the evening, or only certain rows experiencing extended shadows in winter, are difficult to represent in a single overall calculation. By dividing the surfaces, the effects of shadows can be reflected for each part.

This method requires a bit more work for practitioners, but it is quite effective for projects exceeding 10 kW or for projects with complex roof conditions. It is a standard approach when moving from a rough overall estimate of the entire installation to a simulation that is closer to the actual layout. In particular, at sites where the conditions of the installation surfaces vary significantly, simply using this method can considerably improve prediction accuracy.

Method 4: Incorporate loss factors to align with actual power generation

The fourth method is to include a loss factor to bring the estimate closer to actual power generation. Whatever method is used to derive the generation potential, it is risky to regard that value as the actual generated energy. In practice, losses in conversion equipment, wiring losses, efficiency reductions due to high temperatures, soiling, module variability, and minor shading effects accumulate, so actual output is normally lower than the theoretical value.

The idea is: Actual generation (kWh) = Theoretical generation (kWh) × overall correction factor. For example, even if the theoretical value calculated so far is 12,000 kWh per year, if you take the overall correction factor as 0.82, the actual generation will be 9,840 kWh. If conditions are better it might be 0.87, and if harsher 0.75; by assigning a coefficient according to the site, you can bring the figure closer to reality.

This method is important because it greatly affects the reliability of the figures used in practice. Numbers obtained simply by multiplying equipment capacity by an annual factor may look good, but they tend to be reduced when checked in detail. Conversely, figures that properly include loss factors are less likely to be revised during internal explanations or comparative evaluations. In other words, including losses is not about making the figures more conservative; it is a process to make the figures usable.

What you need to be careful about here is to clarify how much you have accounted for as losses. If the regional reference generation already includes common losses, applying an additional large loss factor will result in underestimation. Conversely, if you are using a higher value close to the theoretical as your baseline, it is more consistent to properly include the loss factor. In other words, what matters is not the numerical value of the coefficient itself, but the assumptions behind it.

In practice, it is useful to keep theoretical values and expected actual power generation separate. When explaining to stakeholders, you can use a two-tier structure — “theoretically this much, but in practice we expect this amount after accounting for losses” — which makes it easier to understand the numbers. Reflecting these losses is essential to making solar power generation simulations accurate.

Method 5: Correct using actual measurements and similar cases

The fifth method is to adjust based on measured values and similar projects. This is highly effective for incorporating site-specific conditions that desktop simulations alone cannot fully capture. In particular, when expanding existing equipment, deploying additional buildings on the same site, or rolling out to facilities with similar uses, measured values serve as a strong basis for decision-making.

The idea is simple: compare the actual power generation of an existing installation with theoretical or reference values, and back-calculate a site-specific correction factor. For example, suppose a 10 kW existing installation had an expected theoretical monthly generation of 1,000 kWh, but the measured value was 780 kWh. In this case, the site-specific correction factor is 0.78. Applying this to a new simulation improves prediction accuracy under conditions similar to that location.

The major strength of this method is that it can collectively capture factors that cannot be fully accounted for on paper. The subtle way shadows fall, temperature conditions that are hard to predict, how easily dirt accumulates, actual operating conditions, and other factors that are difficult to model individually nevertheless show up in measured results. For that reason, if you have records from the same site or similar projects, there is no reason not to use them.

Also, using measured values makes it easier to identify where the simulation can be improved. For example, if values are lower than expected only in summer, the estimate of high-temperature losses may be too optimistic; if they are lower only in winter, the assessment of shading may be insufficient. Measured values are not merely correction material but also provide clues as to where prediction errors exist.

Of course, not every project has measured data. However, by accumulating results from past projects and cases with similar conditions, the accuracy of subsequent simulations steadily improves. For practitioners, this fifth method is an important way to quantify experience and reuse it.

Tips for Improving Accuracy

When improving the accuracy of solar power generation simulations, the most important thing is to refine the input conditions rather than change the method. No matter how sophisticated the formula you use is, if the placement of the system capacity is off the results will be off, and if your assessment of shading is coarse you won't close the gap with actual performance. When you want to raise accuracy, it is more effective to question the quality of the assumptions than the sophistication of the calculation formulas.

The first thing to keep in mind is to set equipment capacity based on the actual layout rather than the theoretical maximum. For rooftops, assume a realistic capacity that accounts for edge clearances, equipment, and inspection/maintenance walkways; for ground installations, include clearances and maintenance access routes. This will make all subsequent figures more stable. If capacity is set too aggressively, it will tend to remain overestimated even after careful corrections.

Next, don’t be satisfied with annual totals alone; break them down once by month and by surface. Even if the annual numbers look similar, breaking them down reveals which months show differences and which surfaces are weak, so you can see where accuracy needs to be improved. In particular, for projects with multi-surface installations or those that prioritize self-consumption, simply breaking the data down significantly reduces oversights.

Furthermore, it is important not only to look at losses in aggregate but also to explicitly document what was included and to what extent. Where did you account for shading, to what degree did you assume temperature losses, and did you include wiring and conversion losses? Without this kind of clarification, coefficients are prone to double-counting or omission. More than the numbers themselves, consistency of assumptions determines accuracy.

And finally, what’s effective is not turning a forecast into a single definitive value. If you present multiple cases—optimistic, baseline, and conservative—it becomes easier to explain even when conditions are uncertain. In the early stages, when on-site information is not yet complete, it is safer in practice to show a range rather than force a single number. The key to improving accuracy is not hitting the exact point from the start, but having a way of thinking that is less prone to fluctuation.

Common mistakes in simulations

One common mistake in solar power generation simulations is determining the energy output solely from the system's installed capacity. For example, it's easy to mechanically estimate that because it's a 10 kW system it will generate about 11,000 kWh per year, but that doesn't necessarily hold true on site. If you proceed without accounting for regional differences, orientation, and shading, you'll find it hard to explain later.

Another common mistake is treating sunshine hours and generation hours as the same thing. Just because daylight lasts longer doesn't mean the system generates at high output throughout that time. Solar irradiance is weaker in the morning and evening, and output naturally drops on cloudy days. That is precisely why the concepts of equivalent full-load hours and the annual capacity factor are necessary.

Dismissing shadows is a common mistake. Even if you think it's fine because only a slight shadow falls, if the shadow appears at the same time every day it will amount to a non-negligible difference over the course of a year. Moreover, because shadows change with the seasons, judging solely by how it looks on site can easily lead to inaccuracies. Especially at sites with multiple buildings or many surrounding structures, it's safer not to skip checking for shadows.

Moreover, confusing theoretical values with expected actual power generation is dangerous. If theoretical simulation results are presented as forecast figures without adjustment, you are likely to lose credibility later when losses or performance shortfalls occur. In simulations, it is important to treat theoretical values and practical, operations-oriented estimates separately.

Many of these failures do not arise from the difficulty of the calculations. They occur because people hastily skip over the assumptions. That is why the key to improving accuracy lies not in advanced theory but in keeping the order of checks intact.

How operational staff should proceed without confusion

To prevent practitioners from becoming confused when simulating solar power generation, it is effective to decide in advance which method to use at each stage. First, in the initial assessment, use Method 1's annual coefficient method to grasp the overall scale. At this stage, it is often sufficient to know the annual kWh difference resulting from differences in installed capacity.

Next, once the equipment layout begins to become clear, proceed to Method 2 or Method 3. If the project is better viewed on a monthly basis, Method 2 is appropriate; if there are large differences in the conditions of the installation surfaces, Method 3 is more suitable. At this point, breaking down not only the annual total but also how much power will be generated in which month and on which surface will make later explanations much easier.

At that stage, when preparing proposals or internal approval documents, reflect losses with Method 4 to shift values toward practical figures. At this stage, more conservative numbers that are easier to explain are more valuable than presenting theoretical values as-is. Furthermore, if there are existing performance records or similar cases, applying corrections with Method 5 will bring the forecast closer to site-specific predictions.

In other words, rather than using the most detailed method from the outset, it is important to switch methods according to the stage. With this workflow, you can reach the required level of accuracy without spending excessive effort. For solar power generation simulations, in practice it is more effective to design an approach for using different methods as appropriate than to search for a single “correct” method.

Summary

There are five methods to simulate solar power generation: estimating from installed capacity and an annual coefficient; aggregating using monthly solar irradiance conditions; calculating separately for each installation surface; applying loss coefficients to approximate actual generation; and correcting with measured values or data from similar projects. Each is suited to different situations, and it is not necessary to use all of them at once. The important thing is to choose the appropriate method according to the project stage.

The trick to improving accuracy is not to complicate the formulas, but to raise the quality of input conditions such as system capacity, orientation, shading, and losses. Don’t stop at annual totals — simply breaking them down by month and by surface, and separating theoretical values from practical estimates, will greatly change how usable the simulation is. What is useful in practice is not good-looking numbers, but numbers that can be explained and remain stable.

To make accurate assessments of shadows and placement conditions, it is essential to precisely grasp the on-site spatial relationships. Whether roof-mounted or ground-mounted, if the positions of obstacles or potential installation locations are ambiguous, no matter how refined your formulas are the inputs will be off. If you want to improve simulation accuracy, it is important not to take the understanding of site conditions lightly.

In that respect, for practitioners who need to obtain precise on-site location information, the LRTK feature of iPhone-mounted high-precision GNSS positioning devices is useful. Because it makes it easier to capture candidate equipment locations and the positions of nearby obstructions with high accuracy, it facilitates simulations that reflect shading and layout conditions. Simulating solar power output cannot be completed by desk calculations alone. If you truly want to improve accuracy, establishing a system that can accurately collect on-site conditions makes a significant practical difference.