7 Causes of Errors in Solar Power Generation Calculations

Table of Contents

• Why solar power generation estimates are prone to errors

• Cause 1: The assumed system capacity does not match the actual capacity

• Cause 2: Azimuth and roof tilt estimates are too rough

• Cause 3: Shadowing conditions are underestimated

• Cause 4: Solar irradiance and weather data are over-averaged

• Cause 5: Loss rates for temperature, soiling, and similar factors do not match the site

• Cause 6: Estimates of self-consumption and exported surplus are overly simplistic

• Cause 7: Insufficient accuracy in acquiring on-site conditions

• Practical items to review in order to reduce errors

• Summary

Reasons Why Solar Power Generation Calculations Are Prone to Errors

Calculating solar power generation looks very straightforward at first glance. If you check the system capacity in kW and multiply it by the area's estimated annual generation, you can get a rough outline of the annual kWh. For that reason, in initial assessments you can quickly compare system sizes, and it is convenient as an entry point for proposals and internal sharing. However, in real-world sites it is not uncommon for there to be a gap between this theoretical value and actual results. Moreover, that gap may be only a few percent, but depending on the conditions it can become quite large.

The reason this error occurs is not necessarily that the calculations are wrong. Rather, in many cases the cause lies not in the formulas themselves but in how assumptions are set, how on-site conditions are interpreted, and how corrections are applied. In other words, errors in solar power generation are not simply arithmetic or formula mistakes but arise from the accumulation of discrepancies between the input conditions and reality.

For example, simply assuming the plant capacity is the theoretical maximum can make the subsequent annual power generation appear overly optimistic. If you take a too rough look at azimuth and tilt, something calculated assuming a south-facing orientation may in reality have many east- and west-facing surfaces, causing annual kWh to fall short of expectations. If you check shading patterns only in summer and overlook winter, actual winter figures may be considerably lower than assumed. Moreover, when yearly weather variability, temperature, soiling, and deviations in operating hours combine, the gap from theoretical values widens further.

For practitioners, the important thing is not to eliminate errors entirely. Solar power generation is influenced by both natural conditions and equipment factors, so a certain amount of variation is unavoidable. What matters is understanding where errors originate and knowing which conditions to revisit in order to improve the accuracy of your estimates. In other words, reducing errors is less about "memorizing the formula for power output" and more about "breaking down and analyzing the sources of error."

Moreover, errors in estimated power generation directly affect the evaluation of the system. If the projected annual kWh is overestimated, the perceived self-consumption, surplus power, electricity cost savings, volume of power sold, and payback period will all appear more favorable. Conversely, if the projection is too conservative, you may miss projects that would otherwise be viable. In short, errors in generation estimates are not merely a matter of output; they affect the entire equipment decision-making process.

This article organizes the seven main causes of errors in calculations of solar power generation and provides practical, work-oriented explanations of how each factor affects generation and how they should be reviewed. By the time you finish reading, you should be able to check, in sequence, which factors are responsible for differences between theoretical and actual values rather than dismissing them based on intuition.

Cause 1: Assumptions about equipment capacity do not match reality

The primary cause is that the assumed system capacity deviates from the actual one. In calculating solar power generation, the system capacity in kW is the starting point for everything. Even if this is slightly overstated, it tends to produce inflated figures for annual generation, self-consumption, surplus power, and projected payback. In other words, if the input kW is off, the entire subsequent estimate is likely to be off as well.

A common occurrence in practice is assuming the theoretical maximum number of panels based on the apparent roof or site area. For example, if a roof looks spacious you might think you can fit that many panels and simply assume that number, but doing so makes it easy to overlook the effects of edge clearances, equipment, maintenance access routes, rooflights, upstands, and similar constraints. As a result, the system capacity tends to be overestimated, and the projected energy generation is correspondingly higher than it will actually be.

Also, even with the same number of panels, the system capacity changes depending on the assumption you make about the output per panel. Whether you assume 0.4 kW, 0.42 kW, or 0.45 kW per panel, the difference widens as the number of panels increases to 20, 50, or 100. In other words, assumptions about system capacity are prone to error not only from roof area but also from how panel specifications are assumed.

Treating multiple roof surfaces as a single group can also be a source of error. If a configuration consists of, for example, 10 panels on the south-facing surface and 10 on the west-facing surface, treating them as 20 panels overall under the same conditions will create large discrepancies later during orientation correction. In practice, accuracy is higher if system capacity is recorded not only as a total but also as a breakdown by surface. In other words, system capacity should not be a single number; it is more practical to organize it to show how many kW are on each surface.

To address this cause, it is important to set the practically deployable equipment capacity rather than the theoretical maximum installed capacity. Thinking in terms of usable area rather than total area, clarifying the number of panels and the output per panel, and organizing by surface are the first steps to reduce error. If you want to reduce errors in power generation, you should first review how you set the input equipment capacity.

Cause 2: Orientation and roof pitch estimates are coarse

The second cause is that estimates of orientation and roof pitch are crude. Even if the system capacity is correct, if you haven’t looked closely enough at which direction the installation will face and at what angle it will be mounted, the annual kWh estimate can easily deviate from reality. In particular, on multi-faced roofs or installations distributed east–west, this mismatch tends to translate directly into errors in annual energy production.

For example, a surface that is near south-facing and an east- or west-facing surface receive different solar irradiation conditions even for the same kW. Furthermore, if the roof pitch changes, the implications of those incident light conditions change between summer and winter. However, in practice the whole is sometimes handled with a sense of “roughly south-facing” or “roughly east–west.” With such coarse assumptions, differences between individual surfaces become obscured and the projected annual kWh becomes rough.

Also, it's dangerous to underestimate the effect of tilt. Different tilts change how solar radiation is received, and in winter the drop in solar elevation can make that difference more pronounced. In other words, looking only at azimuth and ignoring tilt can also cause errors. In practice, azimuth and tilt are separate parameters, but treating them together in power generation calculations makes the results closer to reality.

Furthermore, caution is needed when you try to handle orientation correction with a single average value. If you apply one orientation correction factor to the entire installation, the strengths of the south-facing side and the weaknesses of the east- and west-facing sides will cancel each other out, making the installation’s internal details harder to discern. Dividing the installation capacity by face and reflecting the orientation and tilt conditions for each will, as a result, improve the accuracy of the estimates.

To reduce this cause, it is important to organize the equipment by surface as much as possible. If you separate which surface will carry how many kW, which direction that surface faces, and what its tilt is, you can produce an estimate much closer to reality than processing the whole at once. Not roughly assessing orientation and tilt is a major point in reducing errors in estimated power generation.

Cause 3 Underestimating shading conditions

The third cause is underestimating shading conditions. In calculations of solar power generation, cases that completely ignore shading are rare, but in many projects shading tends to be handled intuitively as "only a little shading" or "there shouldn't be any large shadows." However, because shading changes meaning depending on the time of day, season, and location, even a small oversight can significantly affect annual energy production.

For example, an east-facing surface that is shaded only in the morning and a west-facing surface that is shaded only in the afternoon will have different impacts on power generation even if the shading itself is the same. Furthermore, in winter the sun’s altitude is lower and shadows tend to stretch longer, so obstacles that were negligible in summer can have a large effect in winter. In other words, it is more practical to consider not just whether there is shading, but when, where, and how much shading occurs.

Also, underestimating partial shading can cause errors. In practice, cases where only part of a row, part of a surface, or part of a time period is shaded are more common than cases where the entire installation is shaded. If such partial shading is handled uniformly with only small corrections, the annual difference can be larger than expected. Conversely, overestimating some shading can reduce the installation’s value too much. In other words, unless shading conditions are examined carefully, errors are likely to occur in either direction.

On factory and warehouse roofs, ventilation equipment, skylights, ducts, handrails, and surrounding buildings are common causes of shading. For detached houses, typical causes include neighboring houses, trees, antennas, and utility poles. For carports, walls, nearby buildings, trees, and significant differences in elevation can have a major effect. In other words, because the causes of shading differ by use, it is important not to treat them with a one-size-fits-all approach.

To address this cause, it is effective to carefully check shadows as on-site conditions and, if possible, to examine seasonal differences as well. At minimum, organizing the strength of shadows by surface and simply being aware of differences such as morning shadow, afternoon shadow, and winter shadow can considerably reduce errors. Not underestimating shadows is an important condition for getting closer to actual power generation.

Cause 4 Solar radiation and weather are averaged excessively

The fourth reason is that solar irradiance and weather are treated too much as averages. When estimating annual power generation, it is common to use an average value expressed as kWh per kW per year. This is convenient for comparing system sizes, but as it stands it may not adequately reflect region-specific weather or seasonal variations. In other words, if you try to represent a site solely by the average value, errors are likely to arise.

For example, even if the annual totals are the same, the makeup of generated power can differ considerably between regions that consistently show high output in spring and autumn, regions that experience large drops during the rainy season or winter, and regions that suffer strong high-temperature losses in summer. If you look only at the annual average, these differences are not visible. Viewing the data by month shows that the patterns vary substantially across spring, summer, autumn, and winter. In other words, relying on annual values alone makes errors caused by seasonal differences hard to detect.

Another common cause is using solar radiation and sunshine duration interchangeably. Sunshine duration can serve as an indicator of how long it was sunny, but it is the amount of solar radiation that more directly correlates with power generation. If you judge solely by sunshine duration, you can easily overlook differences in light intensity and solar altitude even within the same hour. Therefore, if you use sunshine duration, it is important to adopt the mindset of converting it into solar radiation.

Furthermore, if corrections for cloudy or rainy conditions are handled with a single annual coefficient, month-to-month differences become obscured. A cloudy summer afternoon and an all-day overcast in winter have different implications for equipment. In other words, weather corrections should be considered not only by their averages but also in terms of seasonal variations and monthly trends.

To reduce this cause, it is useful to check monthly values as well as annual values at least once, and to understand the meanings of solar radiation data and sunshine duration separately. Averages are convenient, but it is important to use them with the awareness that they can easily be a source of error.

Cause 5 Loss rates for temperature, dirt, and other factors do not match site conditions

The fifth cause is that the loss rates for factors such as temperature and soiling are not appropriate for the site. In estimates of solar power generation, it is common to multiply the theoretical value by a loss coefficient. This is the correct approach, but if that coefficient is casually reduced to a single number, discrepancies with the actual site are likely to arise. The meaning of a loss rate changes depending on what it includes and to what extent.

For example, in summer solar radiation is strong, but high temperatures readily cause power output to decline. In winter, although lower temperatures are advantageous for efficiency, power generation falls if there is snowfall or overcast weather. In other words, the impact of temperature varies seasonally, and using a single annual value tends to introduce large errors in summer and winter.

The same applies to soiling and equipment variability. In locations where dust and dirt are common—such as factories and warehouses—areas prone to salt and wind exposure like coastal regions, or sites with nearby trees that produce a lot of fallen leaves, it may be advisable to assume a slightly higher loss rate than the typical one. Conversely, for projects in stable environments, applying an excessively high loss rate can lead to undervaluing the equipment.

Also, if you reduce the loss rate to a single number, you lose visibility into what is affecting performance and by how much. This is because you won’t be able to tell which of temperature losses, conversion losses, wiring losses, soiling, aging, etc., is the primary cause. Then, when there is a discrepancy with actual results, it becomes difficult to know what to review. In short, loss rates are convenient, but precisely because they are so convenient, they can easily hide the sources of error.

To address this cause, it is necessary, at a minimum, to be aware of what the loss rate consists of. If you clarify what you regard as general losses and what you separately adjust as site-specific conditions, errors are likely to be substantially reduced even when using the same coefficients. It is not always necessary to subdivide the loss rate, but it is important to consider the different meanings separately.

Cause 6: Interpreting self-consumption and surplus too simplistically

The sixth cause is that the way self-consumption and surplus are interpreted is too simplistic. It is closer to an error in equipment value than an error in the generation amount itself, but in practice it becomes a very significant problem. This is because, even if the total generated power is as expected, if estimates of self-consumption or surplus are off, the expected savings, the amount of electricity sold, and the projected payback can change substantially.

For example, even if annual generation is 10,000kWh, how much of that can be used on-site depends on the timing of demand. Factories and offices with high daytime loads tend to have higher self-consumption rates, while homes that are often unoccupied during the day tend to have larger surpluses. In other words, even if total generation is the same, the value of the system can differ significantly depending on the breakdown between self-consumption and surplus.

A common cause of error here is assuming a single fixed self-consumption rate. For example, if you arbitrarily assume 40% or 50%, differences due to day of the week, seasonality, time spent at home, operating hours, and HVAC load tend to be overlooked. In practice, looking at monthly and time-of-day trends, even to a small extent, will help reduce errors.

Also, it is risky not to check whether the extra output from increasing system capacity will be self-consumed or become surplus. Increasing the system size will raise generation, but if much of that increase becomes surplus, the value based on self-consumption may not grow that much. In other words, it is less an error in the amount of generation than an error in how the generated electricity is used that can greatly change the equipment evaluation.

To address this cause, it is effective to keep power generation, self-consumption, and surplus as separate figures. Not only using annual values, but breaking them down into monthly and daily figures and overlaying them makes discrepancies in self-consumption much more visible. To improve the accuracy of generation estimates, it is necessary to review them including how the system is used.

Cause 7: Insufficient accuracy in acquiring site conditions

The seventh cause is insufficient precision in acquiring site conditions. Many of the causes we have examined so far ultimately come down to this single point. Equipment capacity, orientation, shading, losses, and self-consumption—all of these items will cause the figures to fluctuate more when site conditions are imprecise. In other words, more than the formula itself, the precision of what is being entered greatly affects the magnitude of the error.

For example, if the position of the roof edge is ambiguous, the effective area will also be ambiguous. If the positions of obstacles are unclear, shadow corrections will be coarse. If the roof surface orientation or slope is uncertain, azimuth corrections will be off. If the elevation differences with surrounding buildings are not visible, the impact of winter shadows is easily underestimated. In other words, no matter how elegant the equations used, if the site conditions are coarse, power generation estimates will be prone to variability.

Also, insufficient site condition data affects both the inputs for equipment capacity and the corrections. This is very troublesome. If the input kW is high and the corrections are lax, errors stack up doubly. Conversely, if site conditions are captured accurately, both the way equipment capacity is set and the way corrections are applied become much more stable. If you want to improve the reliability of estimates in practice, it is often quicker to raise the accuracy of the inputs.

To reduce this cause, it is important to review how on-site conditions are obtained. Rather than relying solely on drawings, if you can accurately confirm how shadows fall, the positions of equipment, the orientation of roof surfaces, and elevation differences, the assumptions for calculations become considerably stronger. In other words, one of the most practical ways to reduce errors in power generation estimates is to capture the site accurately.

Operational aspects to review to reduce errors

Based on the seven causes discussed so far, we can also organize what should be reviewed in practice to reduce errors. The first important thing is to bring the inlet equipment capacity closer to the adopted value rather than the theoretical value. Simply using usable area instead of total area, and the number of actually placeable units instead of the total number of units, will make subsequent calculations considerably more stable.

Next, don’t stop at the annual average; take a look at monthly differences at least once. If you check how things change across spring, summer, autumn and winter, you can significantly reduce errors caused by averaging solar irradiation and losses. Furthermore, instead of treating the self-consumption rate as a fixed value, even looking a little at the time-of-day and seasonal differences in demand will reduce misestimation of the equipment value.

Above all, improve the accuracy of the on-site conditions. If the positional relationships remain ambiguous, shadows, orientations, and areas all become prone to variation. Rather than complicating the calculation formulas, accurately capturing the site conditions often results in higher estimation accuracy. In other words, to reduce errors, reviewing the quality of the inputs is more effective than devising mathematical tricks.

Summary

As causes of errors in calculating solar power generation, seven factors are particularly important: assumptions about installed capacity that differ from reality; coarse estimates of azimuth and roof pitch; underestimation of shading conditions; overly averaged treatment of solar irradiance and weather; loss rates for temperature, soiling, and the like that do not match site conditions; overly simplistic interpretation of self-consumption and surplus; and insufficient accuracy in acquiring on-site conditions. Each of these often overlaps with the others and compounds the error more than when viewed individually.

The important thing is not to dismiss errors as mere discrepancies, but to break them down and consider which assumptions were off. Was the capacity of the incoming equipment overestimated? Were orientation or shading treated too optimistically? Was the assessment of weather or losses too rough? Was the estimate of self-consumption overly simplistic? Once you can isolate and identify the causes, the accuracy of the estimates is likely to improve considerably.

Also, if you truly want to reduce these kinds of errors, it is essential to accurately capture the spatial relationships on site. If roof edges, obstacles, elevation differences, the way shadows fall, and the positions of equipment remain ambiguous, then no matter how neatly you refine the calculations, the final kWh will be prone to variability. In particular, shadowing and area conditions are aspects where the on-site spatial relationships directly affect system capacity and power generation.

In that respect, as a means to accurately grasp on-site positional relationships, the iPhone-mounted GNSS high-precision positioning device LRTK is extremely effective. Because it makes it easier to accurately record the positions of roof edges and obstacles on site, it becomes easier to improve the accuracy of power generation estimates that take into account orientation, shading, and layout conditions. If you want to reduce sources of error in solar power generation calculations and move closer to figures that are truly usable, properly capturing site conditions with measures like LRTK is a significant advantage in practice.