Solar panel orientation and tilt: optimising for yield

Orientation and tilt decisions are made once, at design time, and then live with a system for its entire operating life, which is what makes them worth getting right rather than defaulting to rules of thumb. The physics is well understood — the sun's path across the sky is fully predictable for any latitude and date — yet the optimum answer depends on what the system owner actually values: maximum annual yield, a flatter production profile across the day, better match to an evening-peaking consumption pattern, or simply the constraints of an existing roof.

This article works through azimuth and tilt optimisation for fixed-tilt arrays, the east-west split-array approach increasingly common on commercial roofs, and the seasonal tradeoffs that come with any single fixed configuration. It includes yield tables so a designer can quickly sanity-check whether a given orientation is a two percent or a twenty percent compromise relative to the theoretical optimum, and references the payback calculator for translating orientation losses into economic terms.

None of this replaces a proper simulation for a specific site — shading, local climate, and roof geometry all matter — but understanding the underlying yield surface lets a designer recognise quickly which compromises are cheap and which are expensive.

The yield surface: azimuth and tilt together

For a given latitude, annual specific yield (kWh/kWp/year) as a function of azimuth and tilt forms a smooth surface with a single broad maximum near true south (in the northern hemisphere) at a tilt roughly equal to the site's latitude, though the practical optimum tilt is usually a few degrees lower than latitude because it biases production toward summer months when irradiance is higher, at the cost of winter performance. The surface is broad, not sharp: moving tilt or azimuth away from the theoretical optimum by a moderate amount costs only a small percentage of annual yield, which is why roofs constrained to a particular pitch or orientation are rarely a serious economic problem.

Azimuth deviation from due south (northern hemisphere, mid-latitude)

Tilt deviation from latitude-optimum

These figures are broadly representative for mid-latitude European sites (roughly 40-55°N); actual values shift with latitude and local climate, particularly the balance of direct versus diffuse irradiance, so a Central European site with more diffuse-dominated skies is somewhat more forgiving of orientation error than a site with a very direct-dominated resource such as southern Spain.

Worked example: quantifying an orientation compromise

Consider a roof at 50°N latitude with a usable pitch facing 40° west of due south, roof pitch fixed at 35° (close to the latitude-optimum tilt for that location, so tilt itself is near-ideal). Reference annual specific yield for an optimally oriented, well-designed system at this location is assumed at 1050 kWh/kWp/year, a reasonable order-of-magnitude figure for Central Europe.

From the azimuth deviation table, a 40° deviation sits between the 30° (97%) and 45° (93%) rows; interpolating gives approximately 94.5%.

Adjusted yield = 1050 kWh/kWp/year × 0.945 = 992 kWh/kWp/year

For a 6 kWp system, that is a difference between 6300 kWh/year (optimal) and 5954 kWh/year (actual orientation) — a loss of 346 kWh/year, or about 5.5%. Run through the payback calculator, this modest yield reduction typically extends payback by a matter of months rather than years, reinforcing that near-optimal orientation is usually good enough; chasing the last few percent by, for example, adding tilt frames on a west-facing roof is rarely justified unless roof space is severely limited and yield density matters more than cost per kWp.

East-west arrays

Where roof area is abundant but a single south-facing aspect is not available — or where the goal is a flatter production profile rather than peak annual yield — east-west arrays laid flat or at a shallow tilt against a duo-pitch roof are an established alternative. Two half-arrays, one facing roughly east and one roughly west, together produce a broader, flatter daily generation curve than a single south-facing array of the same capacity, with lower midday peak and higher morning and afternoon output.

This has two consequences worth designing around. First, total annual yield per kWp is typically 10-20% lower than an equivalent south-facing array at optimum tilt, because both halves are permanently at a non-optimal azimuth. Second, for self-consumption-focused systems, the flatter profile can better match household or commercial load patterns that have morning and evening demand peaks and a midday trough, improving self-consumption ratio even though gross yield is lower — a tradeoff worth evaluating against a specific battery sizing strategy where storage might otherwise be used to shift a peaky south-facing profile.

East-west layouts also allow denser packing on flat commercial roofs since rows lean against each other and inter-row shading is reduced compared with south-facing rows at a steeper tilt needing wider spacing, so kWp per available roof area is often higher for east-west even though kWh per kWp is lower — the two effects can roughly offset for kWh per square metre of roof, which is the metric that often matters most on space-constrained commercial roofs.

Seasonal tradeoffs of tilt choice

A single fixed tilt cannot be optimal for both summer and winter, since the sun's elevation varies by roughly 47 degrees between winter and summer solstice at mid-latitudes. Lower tilts (closer to horizontal) favour summer collection when the sun is high, while higher tilts favour winter collection when the sun is low, at the cost of some summer performance due to increased reflection losses at very oblique summer sun angles on a steep panel.

Systems designed to support heating loads (heat pumps, hot water) or off-grid battery systems that must survive winter with minimal generation sometimes deliberately bias tilt steeper than the annual-yield optimum, accepting a small annual yield penalty in exchange for meaningfully higher winter output when it is needed most, since winter irradiance is the scarce resource in those use cases rather than annual totals.

Ground-mount and tracking considerations

Ground-mounted systems are not constrained by roof pitch and can be set to the calculated optimum tilt directly, and in some cases justify single-axis tracking, which follows the sun's azimuth through the day (typically rotating on a north-south horizontal axis) and can improve annual yield by roughly 15-25% over a fixed-tilt array at the same location, at added mechanical cost and complexity. Dual-axis tracking, which also adjusts elevation, adds a further few percent over single-axis for meaningfully higher cost and is rarely justified outside of concentrating PV or research applications given the diminishing incremental yield.

Tracking economics depend heavily on the proportion of direct versus diffuse irradiance at the site: locations with high direct normal irradiance (clear, dry climates) see the largest tracking benefit, while sites with a large diffuse fraction (cloudier, higher-latitude climates) see a smaller benefit because tracking cannot meaningfully concentrate diffuse skylight the way it can direct beam radiation.

Shading interaction with orientation choice

Orientation decisions cannot be made independently of shading, covered in more depth in a companion article on shading analysis. A theoretically optimal south-facing, latitude-tilt array that is shaded by a chimney for two hours each winter afternoon may underperform a slightly off-optimal but shade-free orientation on the same roof. Orientation optimisation should always be run alongside a shading assessment rather than in isolation, and roof geometry frequently makes the shade-free option the better one even when it deviates meaningfully from due south.

Practical decision workflow

1. Establish latitude and identify the theoretical optimum azimuth (due south/north depending on hemisphere) and tilt (roughly latitude minus 5-10°).
2. Map available roof planes and their actual azimuth and pitch.
3. Run a shading assessment for each candidate plane before comparing yield.
4. Use yield-surface tables or simulation software to estimate relative annual yield for each realistic option.
5. Weigh yield against installation practicality (roof penetrations, ballast weight for tilt frames, wind loading).
6. For systems with strong self-consumption or seasonal load goals, evaluate east-west or seasonally biased tilt against straightforward yield maximisation.
7. Cross-check the selected configuration's estimated output against inverter sizing assumptions used elsewhere in the design.

Standards and references

• IEC 61215 — Design qualification and type approval for terrestrial PV modules, relevant background for module performance figures used in yield modelling.
• IEC 61853 — PV module performance testing and energy rating, covering performance across irradiance and temperature conditions relevant to seasonal yield estimation.
• IEC 60364-7-712 — PV-specific wiring rules that interact with array layout decisions such as string routing across east-west sub-arrays.
• EN 62446 (aligned with IEC 62446) — Documentation and testing requirements referenced when recording as-built orientation and tilt for commissioning.

Frequently asked questions

Is a south-facing roof always better than east-west?

For maximum annual energy yield, yes, a well-oriented south-facing array typically outperforms an equivalent east-west split by 10-20% per installed kWp. However, if the roof area available for east-west exceeds what a single south aspect offers, total energy production (not yield per kWp) can favour east-west, and if self-consumption match matters more than gross yield, the flatter east-west profile may be preferable even at equal capacity. The right answer depends on which metric — yield per kWp, total kWh, or self-consumption ratio — is the actual design goal.

How much does a flat roof (no tilt) cost in lost yield?

A completely flat, unracked array (0° tilt) at mid-latitudes typically produces 8-15% less annual energy than the same array at optimum tilt, mainly because it collects less of the low-angle winter sun and no tilt reduces soiling self-cleaning during rain. Many flat-roof commercial installations still choose low tilt angles of 10-15° specifically to reduce this loss while keeping wind loading, ballast weight, and inter-row spacing requirements manageable.

Does snow accumulation change the tilt calculation?

In snowy climates, steeper tilts (above roughly 35-40°) shed snow faster than shallow tilts, which can meaningfully improve winter production by reducing the number of snow-covered non-generating days. This is a secondary consideration behind the latitude-based optimum but is one of the more common reasons to deliberately choose a tilt on the steeper side of the annual-yield optimum in far-northern or continental climates with regular heavy snowfall.

How precise does azimuth measurement need to be during design?

Roof azimuth should be measured or read from site plans to within about 5 degrees for design purposes; finer precision than that has negligible effect on yield given how broad and flat the yield surface is near the optimum. Compass measurements should account for magnetic declination, which varies by location and can be several degrees, since true south (or true north) is the relevant reference for solar geometry rather than magnetic south.

Can orientation be corrected after installation?

Fixed-tilt racking orientation generally cannot be adjusted after installation without significant rework, since it is set by the roof structure or ground-mount frame geometry. Tilt is sometimes adjustable within a limited range on certain ground-mount or flat-roof ballasted racking systems, allowing seasonal or one-time recalibration, but roof-integrated residential systems are effectively fixed for their service life, which is why orientation modelling at design stage deserves careful attention rather than being treated as a minor detail.

Sources

• NREL — PVWatts calculator and solar resource data
• PVEducation — Solar position and array orientation
• IEA PVPS — Photovoltaic power systems programme reports
• Fraunhofer ISE — Photovoltaics report