How to Choose A Right Transformer for Your Solar System

Choosing the wrong distribution transformer for your solar PV project is a costly mistake—not just in upfront capital, but in operational reliability. Oversize it, and you waste 15–20% of your equipment budget on unused capacity. Undersize it, and you face nuisance tripping, accelerated insulation aging, and even catastrophic failure during peak irradiance hours. This article walks you through a 5-step sizing process that moves logically from inverter-side data to a complete transformer nameplate specification. By the end, you will have a fill-in-the-blank technical checklist ready for your RFQ (Request for Quotation).

Before touching any calculation, you must answer one fundamental question: Is this transformer acting as a step-up unit, an isolation unit, or both?

• Step-up transformer – installed directly downstream of the inverter(s) to raise the low AC voltage (e.g., 540V, 690V, or 800V) to the medium-voltage grid level (e.g., 10kV, 20kV, or 35kV). For this role, sizing is driven by total inverter power and simultaneity, while impedance needs to be high enough (typically 6–8%) to limit short-circuit current feeding back into the inverter's IGBT modules.

• Isolation transformer – used primarily to change the grounding scheme or to decouple the PV plant from utility-side zero-sequence harmonics. Here, the selection priority shifts to the connection group (e.g., Dyn11 vs. Yyn0), because this directly determines how zero-sequence impedance behaves under single-phase-to-ground faults.

Capacity sizing is the single most common source of errors in PV projects. Do not rely on the rule of thumb "multiply inverter nameplate by 1.2." Instead, follow this five-substep data-driven procedure:

1. Sum the apparent power of all inverters connected to this transformer – read the value from inverter datasheets (S_inv_total in kVA).

2. Apply a simultaneity factor – because not all inverters produce peak power at the exact same moment due to azimuth/tilt variations and cloud passage. For ground-mounted utility plants, use 0.90–0.95; for rooftop systems with multiple orientations, use 0.85–0.90.

3. Incorporate the operating power factor (PF) – modern smart inverters can run from 0.8 leading to 1.0, but the transformer's kVA must be rated at the worst-case PF where the inverter delivers its maximum reactive power. In practice, use PF = 0.9 as a conservative baseline.

4. Add a harmonic margin – PV inverters generate current harmonics (especially 3rd, 5th, and 7th orders), which cause additional winding heating. A 5% harmonic overload factor is standard for grid-following inverters; for older or low-quality inverters, consider 8%.

5. Round up to the nearest standard kVA rating available in the market (e.g., 1000, 1250, 1600, 2000, 2500, 3150 kVA).

Final formula:

Example: 5 inverters × 1000 kVA each = 5000 kVA; simultaneity 0.92; PF=0.90; harmonic margin 1.05 → S_tr = (5000×0.92)/0.90 ×1.05 = 5,367 kVA → round up to 5,500 kVA (or 6,000 kVA depending on local standard series).

With capacity fixed, the next decision is electrical matching: you need to bridge the inverter-side AC voltage and the utility point-of-common-coupling (PCC) voltage.

First, lock the fixed voltage ratio. Most PV inverters output 540V, 630V, 690V, or 800V (line-to-line). The grid side is typically 10kV, 20kV, or 35kV. Select the nearest standard off-the-shelf ratio – e.g., 0.69/10 kV or 0.8/20 kV. Avoid custom ratios unless the project scale exceeds 50 MW, because custom designs double the lead time and cost.

Second, decide on tap-changing capability. This is where many engineers overspecify or underspecify:

• Off-circuit (De-energized) tap changer (DETC) – sufficient when the grid voltage at the PCC is stable (within ±5% of nominal) and the transmission line from transformer to PCC is shorter than 3 km. DETC saves approximately 10–15% of the transformer cost.

• On-load tap changer (OLTC) – mandatory under any of these conditions:

• - The PV plant connects to a weak grid (short-circuit ratio < 3) where voltage fluctuates by >7% daily.

• - The cable/overhead line distance exceeds 5 km, causing significant voltage drop under full load.

• - The plant participates in grid voltage support (reactive power dispatch) as required by the local grid code.

For OLTC, specify the tap range as ±4×1.25% or ±2×2.5% – the latter is more common in medium-voltage distribution networks.

Short-circuit impedance (Uk%) is not a fixed number – it is a deliberate design choice that balances three competing objectives: limiting fault current, controlling voltage regulation, and minimizing cost.

To select the right Uk%, you must first obtain the system short-circuit capacity (S_sc) from your utility provider – both the maximum and minimum values in MVA. Then apply this logic:

• When S_sc is high (stiff grid), – a higher Uk% (e.g., 7–8%) is preferred because it reduces the through-fault current that could damage inverter output filters and IGBTs.

• When S_sc is low (weak grid), – a lower Uk% (e.g., 5–6%) helps maintain voltage stability under sudden load changes, but you must verify that the transformer itself can withstand the resulting thermal and dynamic stresses.

Practical reference table for PV applications:

Beyond Uk%, you must also specify the short-circuit withstand current (I_cw) in kA – typically derived from the utility's maximum fault level. For example, if the 10 kV bus has a fault level of 250 MVA, the symmetric short-circuit current is 250 MVA / (√3 × 10 kV) ≈ 14.4 kA. Your transformer must be certified for this value (or higher) for 2 seconds of thermal withstand and full dynamic peak withstand (I_peak = 2.5 × I_sym for medium-voltage equipment).

The final step transforms a generic transformer into a site-specific solution. You need three site survey data points: altitude, maximum ambient temperature, and pollution/salinity level.

Altitude correction: For sites above 1,000 m, the air density decreases, reducing convection cooling. According to IEC 60076-2, the temperature rise limit must be derated by approximately 1% for every 100 m above 1,000 m. In practice, this means you either:

• – order a transformer with a higher insulation class (e.g., F instead of A) to allow more thermal headroom, or

• – simply round up the kVA one standard size (e.g., from 2,000 to 2,500 kVA) to keep the actual load below the derated capacity.

Ambient temperature: If the annual average exceeds 40°C (common in desert PV farms), apply the same derating logic. For projects in Middle Eastern or Australian inland regions, require the manufacturer to perform a thermal simulation at 50°C ambient and provide guaranteed temperature rise test reports.

Cooling method selection:

• Oil-immersed, ONAN (Oil Natural, Air Natural) – the default choice for large ground-mounted plants. It offers the best cost-to-cooling ratio and requires minimal maintenance. Add a radiator bank for enhanced cooling if the site experiences 10 days/year above 45°C.

• Dry-type, ANAF (Air Natural, Air Forced) – mandatory for rooftop installations on commercial buildings, carports, or any indoor location where fire safety codes prohibit mineral oil. Dry-type units use cast resin or VPI (Vacuum Pressure Impregnation) insulation; specify Class F (155°C) or Class H (180°C) to compensate for their poorer heat dissipation compared to oil.

Enclosure protection (IP code):

• Outdoor, dusty/sandy (e.g., desert or plains) → IP54 (dust-protected and splash-proof) minimum.

• Coastal or offshore (salt fog corrosion) → IP65 with C5-M marine-grade coating on the tank and radiators.

• Indoor/substation → IP20 or IP23 (ventilated) is sufficient.

This section addresses the most frequently asked questions that do not fit neatly into the linear steps but are critical for final decision-making.

Q1: Should I choose a transformer with a higher kVA than the calculated value "just to be safe"?

A: No – oversizing by more than one standard step (e.g., 2,500 instead of 2,000) increases no-load losses (iron losses) significantly because the core is larger. Since PV systems operate at zero output during nighttime, those iron losses run continuously for 8–12 hours daily with zero revenue offset, drastically reducing the plant's internal rate of return. Only oversize if your future expansion plan is confirmed within 2 years.

Q2: Is aluminum winding acceptable for solar transformers?

A: Technically yes, but economically you need to run a Total Cost of Ownership (TCO) comparison. Aluminum windings have higher resistance → higher load losses (copper losses). For a 25-year PV plant, the accumulated extra energy loss from aluminum often exceeds the upfront cost saving. We recommend copper windings for any project 2 MW where the capacity factor 18%.

Q3: Do I need a zig-zag earthing transformer alongside the main power transformer?

A: Only if your PV plant uses an ungrounded or high-resistance grounding scheme and the main transformer's Dyn11 connection does not provide a stable neutral point. For most grid-following plants with a solidly grounded medium-voltage network, the Dyn11 transformer itself provides a neutral on the low-voltage side – no extra earthing transformer is needed.

Q4: How much does the transformer's efficiency matter in PV projects?

A: Critically. A 0.5% efficiency difference translates to thousands of MWh lost over 25 years. Demand that all bidders provide the weighted efficiency measured at 50%, 75%, and 100% load (with the weights reflecting your site's irradiance profile). This is more meaningful than the nameplate efficiency at full load only. For example, if your site has 1,800 peak sun hours/year, a 0.5% efficiency gain on a 5 MVA transformer saves approximately 5 × 0.005 × 1,800 = 45 MWh/year – enough to power 15 average US homes annually.

Q5: What is the minimum number of transformer suppliers I should quote?

A: At least three. But more importantly, request not just the commercial offer but also the loss capitalization calculation – where each bidder uses your local electricity tariff to convert their guaranteed losses (no-load + load) into a monetary value over the expected life. This allows you to compare offers on a Lifecycle Cost basis, not just the initial purchase price.

After completing Steps 1 through 5 and reviewing the FAQs, compile all outputs into a single-page Transformer Sizing Checklist – this document serves as the backbone of your RFQ.

Selecting the right distribution transformer for a solar PV system is not an art – it is a structured engineering workflow. The five steps presented here convert vague requirements (inverter power, grid voltage, site location) into concrete, verifiable specifications. The ultimate message is simple: do not oversize, do not guess, and do not ignore environmental corrections.

Before you issue your next transformer enquiry, take 20 minutes to fill out the checklist above. Then send it to at least three reputable manufacturers and demand a weighted efficiency loss calculation and a lifecycle cost comparison. This single discipline will save your project between 5% and 12% of the total electrical balance-of-plant cost over the operational lifetime.