Solar PV Sizing Calculator

Solar PV System Design Guide for AS/NZS 4777.2 and 5033

Sizing a grid-connected solar PV system involves balancing the customer's energy consumption, available roof space, inverter capacity, and local network export limits. A correctly sized system maximises self-consumption (using solar power on-site rather than exporting it at a lower feed-in tariff), minimises payback period, and complies with AS/NZS 4777.2:2020 for inverter energy systems and AS/NZS 5033:2021 for the DC array installation. This calculator takes the site's daily energy usage, location, roof orientation, and panel specifications to determine the optimal array size, inverter capacity, expected annual yield, and estimated payback period. It also checks DC string voltage limits and inverter oversizing ratios to flag compliance issues before installation begins.

Key concepts

• Peak sun hours (PSH) and location. PSH is the number of hours per day that solar irradiance averages 1 kW per square metre. It varies significantly across Australia: southern cities like Melbourne and Adelaide average 3.5 to 4.5 hours, while Darwin and Cairns average 5.0 to 6.0 hours. The calculator uses PSH to convert the array's rated capacity (kWp) into expected daily energy output (kWh). Choosing the wrong PSH value is one of the most common sizing errors.
• DC to AC ratio (oversizing). AS/NZS 4777.2:2020 allows the DC array to be up to 1.33 times the inverter's rated AC output. This is called oversizing or clipping. Oversizing improves energy harvest in the morning and evening when irradiance is below peak, at the cost of minor clipping losses during the midday peak. It is standard practice in Australian residential systems because panels rarely reach their Standard Test Condition rating in real-world conditions.
• String voltage and temperature effects. Panel open-circuit voltage (Voc) increases as ambient temperature drops. AS/NZS 5033:2021 requires that the maximum system voltage be calculated at the coldest expected temperature, not at STC (25 degrees Celsius). In southern Australia where winter mornings can reach 0 to 5 degrees Celsius, a string of panels that is within the inverter's MPPT range at 25 degrees may exceed the maximum input voltage at 0 degrees. The calculator applies the panel's temperature coefficient to compute worst-case Voc.
• Export limits and DNSP rules. Most Australian DNSPs cap single-phase export at 5 kW and three-phase export at 10 kW at the point of connection. Systems larger than the export limit can still be installed but must use export limiting (zero export or dynamic export control) configured at the inverter. The calculator includes export limit configuration so you can size a larger array for self-consumption without violating network connection conditions.

Common scenarios

• Sizing a residential system for a family in Adelaide. A household using 25 kWh per day in Adelaide (PSH approximately 4.2) needs a minimum array of about 6 kW to offset their consumption. After applying a 30% oversize allowance for system losses (soiling, temperature derating, inverter efficiency, cable losses), the recommended array is approximately 7.8 kWp. The electrician enters the consumption, location, and roof details, and the calculator determines the panel count, inverter size, and expected payback period based on current consumption and feed-in tariff rates.
• Checking string voltage for a commercial rooftop in Melbourne. A commercial installation uses 15 panels per string, each with a Voc of 41.2 V and a temperature coefficient of -0.27% per degree Celsius. At STC the string Voc is 618 V, but at a winter morning temperature of 2 degrees Celsius the Voc rises to approximately 644 V. The calculator checks this against the inverter's maximum input voltage (typically 600 V for residential or 1000 V for commercial inverters) and flags a compliance issue if the string exceeds the limit, prompting the installer to reduce the string length.
• Evaluating payback for a system with low self-consumption. A homeowner who works during the day and uses most electricity in the evening will have a self-consumption ratio of only 20 to 30% without a battery. Most of the solar generation is exported at 5 to 8 cents per kWh rather than offsetting grid consumption at 30 to 35 cents per kWh. The calculator shows how this low self-consumption ratio extends the payback period compared to a household with daytime loads, helping the customer make an informed decision about system size and whether to add battery storage.

System sizing components

• Daily energy consumption (kWh): All on-site loads to be offset by solar generation.
• Peak sun hours (PSH): Australia typically 3.5 to 6.0 hours per day depending on location and season.
• Roof area and orientation: North-facing is optimal in the southern hemisphere; east and west orientations reduce output by 10 to 20%.
• Roof tilt: Optimal angle is approximately equal to the site latitude (for example, 35 degrees for Melbourne).
• System losses: Soiling 2 to 3%, cabling 2 to 3%, temperature 5 to 7%, inverter 4 to 5%.

Key compliance requirements

• Export limiting: Single-phase typically 5 kW, three-phase 10 kW (AS/NZS 4777.2 Cl. 4.2).
• Inverter sizing: Maximum 1.33 times DC array size.
• DC string voltage: Typically 200 to 600 V residential (AS/NZS 5033 Cl. 4.2).
• Anti-islanding: All grid-connected inverters require automatic disconnection on grid failure.

Performance metrics

• Specific yield: 1,000 to 1,400 kWh/kWp/year typical in Australia.
• Self-consumption: Percentage of generation used on-site. Higher self-consumption delivers better return on investment without batteries.
• Payback period: Estimated using approximately $0.30/kWh consumption tariff and approximately $0.05/kWh feed-in tariff (typical current Australian tariffs).

Parameters

Every field this calculator accepts, what it means, the range it accepts, and how it feeds the result. The four fields under Advanced parameters are optional to change but always used.

Daily energy consumption (kilowatt hours per day, 1 to 200)

Average daily site consumption. Take a recent bill and divide the quarterly kilowatt hours by about 90. A typical Australian household sits between 15 and 25. This is the single figure that sets the array size: the tool sizes the system to generate this amount per day, so it targets full annual offset, not a self consumption optimised system. Below 5 or above 100 raises a warning to check the units.

Peak sun hours (hours per day, 2 to 8)

Daily hours at an irradiance of 1 kilowatt per square metre, as a yearly average for the site. Roughly 3.5 to 4.5 for Hobart, Melbourne and Adelaide, 4.5 to 5.0 for Sydney and Perth, and 5.0 to 6.0 for Brisbane, Cairns and Darwin. Outside 2 to 8 the tool warns that the value is outside the Australian range. Use a yearly average, not a summer figure, or the system will be undersized for winter.

Available roof area (square metres, 5 to 500)

Usable area on the chosen roof plane. It does not constrain the array size: the tool sizes the array from consumption and then compares the area needed against this figure, reporting a fail and a warning if it does not fit. Deduct setbacks, walkways, vents and any shaded zone before entering it.

Roof orientation (degrees, 0 to 360, where 0 is north)

Azimuth of the roof plane measured from north. The performance factor is the cosine of the angle from north, squared, so 0 gives 1.00, 45 gives 0.50, and 60 gives 0.25. Above 90 and below 270 the tool warns that the orientation is unfavourable. Important gotcha: at or very near 90 or 270 degrees the factor collapses to zero and the geometry correction, which divides by that factor, returns a meaningless system size. Keep the entry within roughly 0 to 80 or 280 to 360 degrees, and size due east or west roofs by hand.

Roof tilt (degrees from horizontal, 0 to 90)

Pitch of the roof plane. The tool assumes an optimum of 35 degrees and applies a quadratic penalty for deviation, floored at 0.7. A 20 degree roof scores 0.992, a 10 degree roof 0.977, a flat roof 0.955. Tilt is therefore a much weaker driver than orientation in this model.

Panel wattage (watts per module, 200 to 500)

Nameplate output of one module at standard test conditions. Modern modules are 370 to 440 watts; outside that range the tool raises a warning. It sets the module count and therefore the installed capacity after rounding up to a whole module. Note that the roof area calculation uses a fixed 1.7 square metres per module regardless of what you enter here.

System loss (percent, 5 to 30 in the form, 0 to 50 accepted)

Everything except inverter conversion: soiling, cable loss, temperature derate, mismatch, and any shading allowance. Typical 12 to 18 percent, default 14. Do not include inverter losses here, they are entered separately and the two multiply together.

Inverter efficiency (percent, 90 to 99 in the form)

European or weighted conversion efficiency from the inverter datasheet. Modern units are 95 to 98 percent. Below 92 percent the tool raises a warning. Combined with system loss it gives the overall system efficiency: for 14 percent loss and 96 percent inverter, 0.86 times 0.96 equals 0.8256.

Export limit (kilowatts, 0 to 30)

The distributor export cap at the point of connection. Commonly 5 kilowatts single phase and 10 kilowatts three phase, but it varies by network and by feeder. Enter 0 to switch the check off entirely. The check compares 0.95 times the selected inverter rating against this figure, so a system that is sized larger than the cap will fail here even though it is perfectly legal with export limiting configured at the inverter.

Phase type (single phase or three phase)

Recorded with the calculation but not used in any formula in this engine. It does not change the inverter selection, the export check, or the yield. Set the export limit yourself to match the supply arrangement.

Assumptions and limits

This is a first pass sizing tool for a single roof plane. Here is exactly what it assumes and where it stops.

What it assumes

• A single roof plane and a flat energy target. The required array is the daily consumption divided by peak sun hours times system efficiency, then divided again by the combined orientation and tilt factor, then rounded up to whole modules. Splitting an array across two orientations, or across two inverters, is not supported.
• Orientation and tilt are indicative curves, not irradiance modelling. Cosine squared for azimuth and a quadratic penalty around a 35 degree optimum for tilt. There is no shading analysis, no horizon profile, no ground reflectance, no seasonal split and no hourly simulation. A site with a real shading obstruction will not be represented at all.
• Compensating for a poor roof by making the array bigger. When the geometry factor drops, the tool divides the required size by it, so a 45 degree roof asks for roughly twice the array. That is a modelling choice, not what a designer would necessarily do. In practice you would usually accept the lower yield rather than double the array and the cost.
• Yield is a flat annual extrapolation. Daily yield is peak sun hours times installed capacity times system efficiency, and annual yield is that times 365. No monthly profile, no temperature variation through the year, no panel degradation over life, no clipping loss when the array exceeds the inverter.
• Self consumption is a crude ratio. It is the daily load divided by the daily generation, capped at 100 percent. It has no knowledge of when during the day the load occurs, so a system sized to just match the load reports a self consumption near 100 percent, which no real site achieves without storage. Treat this figure, and the payback that depends on it, as order of magnitude only.
• Payback uses fixed assumptions. Installed cost of 1.50 dollars per watt, consumption tariff of 30 cents per kilowatt hour and feed in tariff of 5 cents. There is no rebate or small scale technology certificate, no interest, no tariff escalation, no maintenance and no inverter replacement at year 10 to 12.
• Roof area required is a fixed 1.7 square metres per module. It ignores the module dimensions you implied by the wattage, and it ignores setbacks, walkways, penetrations and rail layout. Real usable roof needs are typically higher.
• Inverter selection is a lookup, not an engineering choice. The tool takes the first standard rating from 3, 5, 6, 8, 10, 15, 20, 25 and 30 kilovolt amperes that is at or above the array size and at or below 1.33 times the array size, and if none fits it uses 1.1 times the array rounded up. It does not consider maximum power point tracker count, string current limits, or any actual product range.
• The direct current string voltage check is a placeholder. It estimates the string voltage as 48 times the module count divided by 10, which has no relation to your actual string arrangement, and it fails any result below 200 volts. It is not a temperature corrected open circuit voltage calculation. Do the real calculation at the coldest expected cell temperature using the module temperature coefficient before you fix a string length, and check it against the inverter tracker window and the AS/NZS 5033 maximum system voltage.

What it does not do

• No direct current cable sizing, string fusing, array isolation, switching arrangements or earthing and bonding of the array.
• No alternating current cable sizing and no voltage rise back to the point of supply. Use the voltage rise calculator for that, and expect voltage rise, not array sizing, to be the binding constraint on many jobs.
• No inverter grid protection settings, no volt watt or volt var response curves, no anti islanding verification, and no dynamic or flexible export arrangements.
• No distributor application requirements, no battery storage sizing, and no time of use or demand tariff modelling.

Every result should be checked against the current editions of AS/NZS 4777.1, AS/NZS 4777.2 and AS/NZS 5033, and against the local distributor connection rules, then signed off by the person responsible for the installation. Nothing here is validated or certified, and the underlying figures have not been reviewed by a chartered professional engineer.

Worked examples

Three complete runs with every intermediate figure shown, so you can check the tool against your own arithmetic.