Residential Solar Site Measurements: Page 4 of 7
Residential Solar Site Measurements
Residential Solar Site Measurements
Figure 4 Common aerial image perspectives include ortho-photos (top left), oblique (top right) and street view (bottom). Each perspective can assist sales and design teams and lower project...
Figure 7 Measuring roof azimuth via aerial images can be more accurate than on-site measurements. The latter can be affected by ferrous material in the building’s roof or frame. The Solmetric Roof...
Figure 8 Roof modeling services provide cost-effective reports that contain an array of site specific information such as roof dimensions, azimuth and tilt values, 3D shadow maps and array layout...
Figure 9 On-site shade measurements may need to be repeated at some point if a system is underperforming and shading from new vegetation growth is suspected. Solmetric’s skyline mapping tool (shown...
Inside this Article
Common image perspectives. Online images are available in different forms and resolutions depending on a site’s location. The following are the three most commonly used.
- Orthophotos: Often referred to as ortho images (see Figure 4), orthophotos are projected onto a map to appear vertically overhead from all locations on the map. This imagery is available throughout the US, but image quality varies.
- Oblique images: These images are taken from offvertical angles and from multiple directions, as shown in Figure 4. Currently, Bing Maps offers free oblique images from north, east, south and west perspectives for the entire US.
- Street-view images: As the name implies, street views are photos taken from public streets. These images have high resolution where imagery is available, but the views and visual access to some buildings may be limited. An example street view is shown in Figure 4.
Roof dimensions. An accurate measurement of roof dimensions is key to sizing a PV system and planning the installation. The most important parameters are the length, width, azimuth and tilt of the various roof surfaces. Length and width determine how many rows and columns of modules fit in the available space. Area is calculated from length and width and used to estimate maximum array capacity. Due to the limited resolution of most free aerial imagery, including ortho and oblique images, it is often difficult to resolve the exact locations of roof valleys and ridges. It can also be challenging to resolve vent pipes and utility service penetrations versus debris, discolored shingles or roof features. Due to these limitations, roof dimensions developed from aerial imagery typically have an accuracy of approximately ± 1 foot for a surface that is parallel to the ground.
When using ortho images to determine roof dimensions, measurements must be corrected according to the cosine of the tilt, since the area of a tilted roof is actually larger than it appears in an image taken from directly overhead. The accuracy of the roof dimensions therefore depends on the accuracy of the tilt used in the calculation. Figure 5 shows how error in the tilt creates errors in the area measurement, depending on the roof pitch. This error can dramatically impact system design in situations that have limited roof area, such as where a row of modules just barely fits—or in reality does not fit. Measuring tilt on-site with an inclinometer is more accurate than doing so from aerial images.
To get improved accuracy from aerial imagery, the roof can be analyzed using oblique views. With the right CAD software tools, an operator can measure the roof from multiple angles and create an accurate model. Using multiple images and incorporating calculations for specific roof types enables operators to overconstrain the geometry equations and improve the accuracy of the roof model.
Annual insolation. Tilt and azimuth are factors in determining the annual insolation for a fixed array in a given location, such as Sacramento, California, as illustrated in Figure 6 (below). Note that insolation does not vary significantly with small changes in tilt, so approximate tilt numbers are usually acceptable for initial energy production estimates. The azimuth, sometimes referred to as the heading of the roof, also factors into the insolation and can be measured using online tools such as the Solmetric Roof Azimuth Tool (see Resources) as shown in Figure 7 (below). Measuring roof azimuth via aerial images can often be more accurate and reliable than on-site measurement because nearby ferrous metals in building frames or rooftops can cause interference that results in errors in the compass reading.
Modeling approaches. CAD modeling requires a considerable investment in software tools, training and dedicated personnel. This may be a significant hurdle for many installers, who may instead opt to use a simpler tool in the presales phase (see “Online Tools for Roof Measurement and Layout,” Sidebar, above). Other installers may choose to use the services of an outside firm. Roof-modeling service providers often present analysis and reports with a 1- to 2-day turnaround and a per-building or per-site fee. Companies offering these services include Aerialogics, Bright Harvest Solar, EagleView Technologies, Pictometry and Precigeo (see Resources). Their reports include detailed roof dimensions and angles, as shown in Figure 8. Image resolution limitations make it difficult to identify gutters, vents and other small on-roof features, which remain a challenge to roof-mapping and analysis providers.
In some cases, professional roofing reports include solar insolation analysis across the roof surface. Shade estimates and insolation charts prepared in this way often do a good job of characterizing the effects of adjacent roof surfaces or dormers, provided the modeling is done correctly. Some roof-modeling services attempt to model shade from nearby trees or buildings, although this has proved difficult in practice. Due to limited image resolution and the inability to overconstrain the CAD problem for a tree model, accuracy is poor for trees and other off-roof obstructions such as utility poles. In addition, images may be out of date and may not account for recent developments such as new construction and tree growth. Seasonal variations, such as those presented by deciduous trees, are difficult to model accurately with the available images because tree branches cannot be adequately resolved.