Transformer and MV Switchgear Specification for UK Solar PV Projects

Transformer and MV Switchgear Specification for UK Solar PV Projects: What Every EPC Contractor Must Know

The United Kingdom added over 1.5 GW of new solar PV capacity in 2023 alone, and with grid connection queues now stretching beyond five years in some DNO regions, getting the transformer and MV switchgear specification for UK solar PV projects right at the very first submission is no longer a luxury but a commercial necessity. A single misspecified impedance voltage value or an incorrectly selected vector group can trigger a full re-engineering cycle, delay grid connection approval by months, and cost an EPC contractor the entire project margin. This technical guide is written for engineers who need precise, field-tested guidance on transformer sizing, MV switchgear selection, protection relay coordination, and commissioning obligations under the current UK regulatory framework.

Why MV Substation Specification Defines Project Viability on UK Solar Sites

Many EPC contractors treat the MV substation as a late-stage procurement exercise, finalising the transformer and switchgear specification only after the inverter and DC string design are locked in. This is a structural mistake. The solar MV substation specification drives the DNO’s technical acceptance of the grid connection offer, influences the Protection and Automation schedule, and determines whether the project passes the initial G99 pre-assessment without a full load flow study.

On a real 5 MW ground-mount project in the East Midlands, an EPC team discovered during the DNO’s technical review that their selected transformer had a vector group of Dyn1 rather than Dyn11. The DNO’s protection relay on the 11kV feeder was calibrated for Dyn11 phase displacement. The consequence was a 14-week delay to reorder and retest the transformer, a cost exceeding £38,000 in direct procurement and redesign fees, and a revised energisation date that pushed the project outside the original ROI calculation window. The lesson is unambiguous: UK solar electrical design must treat the MV substation as the primary engineering constraint, not an afterthought.

Transformer Sizing Methodology for Solar PV Plant Configurations

Calculating the kVA Rating from AC Inverter Output

The starting point for EPC contractor transformer sizing is the total AC output of the inverter block connected to that transformer. For a string inverter array producing 1,000 kW at unity power factor, the apparent power demand is 1,000 kVA. However, UK-connected solar PV systems are rarely operating at unity power factor at all times. G99 compliance requires reactive power capability across a defined range, typically a power factor range of 0.95 leading to 0.95 lagging at the point of connection. This reactive power obligation increases the apparent power demand and must be factored into transformer kVA sizing.

A practical sizing formula used by experienced medium voltage solar EPC engineers is as follows: take the total rated AC power of the inverters in kilowatts, divide by the worst-case power factor (0.95), then apply a harmonic loading derating factor. For solar PV inverters using modern SiC IGBT switching technology, the Total Harmonic Distortion (THD) on the LV winding rarely exceeds 3%, and most reputable manufacturers apply a K-factor derating of K=4 to K=7. In practical terms, this means selecting a transformer rated at least 110% of the calculated apparent power. On a 4 MW inverter block operating at 0.95 PF, the minimum transformer rating would be 4,210 kVA / 0.95 = 4,421 kVA, rounded up to a standard 4,500 kVA or 5,000 kVA unit depending on the manufacturer’s standard range.

Harmonic Loading and Thermal Derating Considerations

The transformer specification solar PV engineer must account for harmonic-induced eddy current losses. Standard oil-filled distribution transformers are designed to IEC 60076, but this standard does not fully address the elevated harmonic environment produced by PWM inverters. The relevant correction is detailed in IEC 60076-1 and the guidance note within CENELEC HD 538. A K-factor transformer or a transformer with an explicitly uprated eddy current loss factor should be specified when the inverter output THD at full load exceeds 5%.

In one instructive project audit carried out on a 3.4 MW solar site in Lincolnshire, the transformer had been sized exactly to the inverter nameplate kVA with no harmonic derating. After 18 months of operation, the top oil temperature was consistently 8 degrees Celsius above the manufacturer’s predicted curve. The winding hot spot temperature was approaching the 98-degree threshold that triggers accelerated insulation ageing under IEC 60076-7. A transformer replacement was required within four years of commissioning, well inside the design life expectation. Correct harmonic derating at the specification stage would have cost approximately £1,200 more in transformer procurement. The premature replacement cost over £47,000.

UK DNO Technical Requirements for MV Transformer Specification

Impedance Voltage (Uz%) Requirements

The 11kV solar transformer UK specification must satisfy the impedance voltage requirements set by the connecting DNO. In the UK, the four major DNOs (National Grid Electricity Distribution, SP Energy Networks, Scottish and Southern Electricity Networks, and Electricity North West) each publish technical specifications for embedded generation transformers, but there is a common baseline requirement for transformers connecting to 11kV networks: an impedance voltage (Uz%) between 4% and 6%.

Specifying Uz% too low (below 4%) reduces fault current limitation capability, potentially exposing downstream protection equipment and cables to fault levels beyond their rated breaking capacity. Specifying Uz% too high (above 6%) increases the voltage regulation under load, which can trigger automatic voltage regulation (AVR) conflicts on the DNO’s network during periods of high solar generation. The standard DNO preference for a DNO grid connection transformer serving an 11kV-connected solar farm is Uz% = 5%, with a tolerance of +/- 7.5% as defined in IEC 60076-1.

Vector Group Selection: Dyn11 vs Ynd11

Vector group selection is perhaps the most commonly misunderstood element of transformer specification solar PV in the UK market. The overwhelming preference of UK DNOs for embedded generation transformers is Dyn11, which provides a 30-degree phase shift (lagging) between the primary delta winding and the secondary star winding. This configuration offers two critical operational benefits for solar PV applications.

First, the delta primary winding blocks the circulation of zero-sequence currents from the LV side into the 11kV network, which is essential for correct earth fault discrimination on the DNO’s protection relay. Second, the Dyn11 vector group ensures phase displacement compatibility with the majority of UK 11kV network protection relays, which are calibrated for 30-degree lagging transformer connections. The alternative, Ynd11, is occasionally specified for step-up applications where the solar generator is connected to a 33kV bus, but it requires explicit DNO approval and detailed protection coordination study before it is accepted.

On-Load Tap Changer Settings for G99 Compliance

G99 compliance requires the embedded generator to maintain its point of connection voltage within a defined statutory band and to respond to reactive power instructions from the DNO. An on-load tap changer (OLTC) or, on smaller units, an off-circuit tap changer (OCTC) is typically specified with a tapping range of +/- 5% in 2.5% steps. This provides five tap positions, allowing the transformer’s turns ratio to be adjusted to accommodate 11kV network voltage variations across the seasonal export profile.

It is worth noting that many solar EPC teams default to OCTC to reduce procurement cost. However, for solar farms above 500 kW connecting to a rural 11kV network where the no-load voltage can fluctuate by more than 3% during high-generation periods, an OLTC with automatic voltage regulation functionality is strongly recommended. The additional cost of an OLTC on a 1 MVA transformer is approximately £8,000 to £12,000, but it consistently eliminates the curtailment events caused by over-voltage trips on constrained rural feeders.

For further detail on passing the DNO grid connection process, including how transformer specification feeds into the application, see this resource on how to pass grid connection approval.

MV Switchgear Selection Criteria for Solar Substations in UK Conditions

Ring Main Units (RMUs) for Solar Site Applications

MV switchgear solar UK selection begins with understanding the DNO’s connection topology. The majority of UK solar farms below 5 MW connect to the 11kV distribution network via a ring main unit (RMU) arrangement. An RMU typically comprises two load break switches (providing a loop-in, loop-out connection to the DNO’s 11kV ring feeder) and one fuse switch or circuit breaker protecting the generator transformer.

Modern SF6-filled RMUs from manufacturers such as ABB, Schneider Electric, and Lucy Electric are compact, require minimal maintenance, and are rated to IEC 62271-200 for indoor and outdoor installation. The critical specification parameters for a solar RMU in UK conditions are a rated voltage of 12 kV (corresponding to the nominal 11 kV network with allowance for voltage variation), a rated short-time current of 16 kA for 1 second (matching the 11 kV system fault level in most UK urban and semi-rural areas), and an IP54 enclosure rating minimum for outdoor installation in the UK’s humid climate.

Air-Insulated Switchgear vs Gas-Insulated Switchgear

For solar farms above 5 MW where the connection is made at 33 kV, or where a multi-feeder substation arrangement is required, the EPC contractor must choose between air-insulated switchgear (AIS) and gas-insulated switchgear (GIS). This is a decision that carries both technical and environmental implications that are increasingly scrutinised by UK planning authorities.

AIS offers lower capital cost and simpler maintenance access but requires a larger substation footprint, typically 30% to 50% larger than an equivalent GIS installation. In the context of a UK solar planning consent where the substation compound size is constrained by the planning condition or the agricultural land classification, this footprint difference can be decisive. GIS uses SF6 gas as the insulating medium, which delivers a compact, weatherproof installation with a 25-year maintenance interval, but SF6 has a Global Warming Potential (GWP) of 23,500 times that of CO2, and the UK is moving toward restricting its use under the F-Gas Regulation revision. Several UK DNOs now require documentation of SF6 leak testing and a GIS retirement and gas recovery plan as part of the IEC 62271 type test evidence submission.

A pragmatic approach adopted by experienced medium voltage solar EPC contractors on UK 33kV solar substations is to specify vacuum circuit breaker (VCB) AIS panels with stainless steel enclosures and class IP55 ingress protection, sized to fit within the planning-approved compound, using a factory-assembled MV switchroom to reduce on-site installation time and to provide a controlled humidity environment without the environmental liability of SF6.

Protection Relay Coordination Under Engineering Recommendation G99

IDMT Overcurrent and Earth Fault Protection Settings

Engineering Recommendation G99 (ER G99) sets the framework for protection relay coordination on UK embedded generation connections. Every solar MV substation specification must include a protection philosophy document and a preliminary relay setting schedule before the DNO will issue a formal connection offer. The protection elements required as a minimum for an 11kV-connected solar farm are IDMT overcurrent protection (51), definite time overcurrent (50), earth fault protection (51N/64), loss of mains protection (81Lo/Hi, ROCOF), and over/under voltage protection (27/59).

The IDMT overcurrent relay settings must coordinate with the DNO’s upstream protection at the grid supply point. A standard coordination requirement is a grading margin of 0.4 seconds between the generator protection relay and the DNO’s feeder protection. This grading margin accounts for circuit breaker operating time, relay overshoot, and current transformer transient errors. In practice, this means the generator-side relay must operate faster than the DNO’s feeder relay by at least 0.4 seconds at any given fault current level.

On a 2 MW solar site in Yorkshire, a protection coordination study revealed that the default relay settings supplied by the RMU manufacturer (a standard UK distribution overcurrent curve with a time multiplier setting of TMS = 0.3) would actually operate faster than the DNO’s feeder relay under a 2,000 A fault condition, causing the solar farm circuit breaker to open before the DNO’s auto-recloser, a sequence that contravenes G99 Section 8.9. Adjusting the TMS to 0.4 corrected the coordination without any hardware changes, but this simple adjustment required a formal relay setting submission to the DNO and a six-week review period.

Differential Protection for Larger Solar Installations

For solar PV installations above 5 MW, or where the connection transformer is rated above 5 MVA, the DNO’s protection schedule will typically require biased differential protection (87T) on the generator transformer in addition to the overcurrent and earth fault elements. Transformer differential protection uses current transformers on both the primary (11kV or 33kV) and secondary (LV) windings to measure the differential current between the two windings. Under normal conditions, the differential current is zero. Under an internal transformer fault, an unbalanced differential current is detected and the protection relay trips both the HV and LV circuit breakers within 30 to 80 milliseconds.

The complexity of specifying differential protection for a Dyn11 solar transformer lies in the inherent 30-degree phase shift between the primary and secondary winding currents. Modern numerical differential relays (such as the Siemens SIPROTEC 7UT86 or the GE T60) can compensate for vector group phase shift in software, eliminating the need for interposing current transformers. However, this software compensation must be explicitly configured during commissioning, and the commissioning engineer must produce a CT polarity and secondary injection test record as part of the FAT documentation package submitted to the DNO.

If you are at the electrical design stage of your solar project, this overview of how a PV solar system is designed provides helpful context for integrating MV protection design into the broader system architecture.

Commissioning and Factory Acceptance Testing (FAT) for MV Equipment

IEC 62271 Type Test Evidence Requirements

Every piece of MV switchgear and every transformer specified for a UK solar project must be accompanied by IEC 62271 type test evidence before the DNO will permit energisation of the grid connection. IEC 62271-100 covers high-voltage alternating current circuit breakers, IEC 62271-200 covers prefabricated metal-enclosed switchgear, and IEC 60076 covers power transformers. Type test evidence is not the same as a routine test certificate: it is a document produced by an accredited independent test laboratory (such as KEMA, CESI, or PEHLA) certifying that a representative sample of the equipment model has been subjected to full dielectric, thermal, and short-circuit type testing.

A frequent procurement error made by EPC contractors on UK solar projects is accepting a transformer routine test certificate in place of type test evidence. The routine test, which includes a ratio test, winding resistance measurement, and oil dielectric test, is conducted on every transformer unit at the factory. The type test is conducted once on a representative unit of a given design series. If the manufacturer has modified the winding design, changed the core lamination grade, or altered the tank geometry since the type test was conducted, the existing type test evidence is technically invalid and a new type test is required. Checking the type test evidence against the actual manufactured unit specification is a due diligence step that experienced UK solar electrical design engineers always perform before FAT sign-off.

DNO Witness Test Obligations

For solar projects requiring a G99 Type B or Type C connection (generally above 1 MW and above 10 MW respectively), the DNO has the right to witness specific factory acceptance tests on MV equipment. These witness tests typically include the transformer ratio and vector group test, the HV winding impulse withstand test (on request), the RMU or switchgear secondary injection test of protection relays, and the trip function test demonstrating correct breaker operation on protection relay command.

Coordinating DNO witness attendance at a transformer factory in Germany, Sweden, or Turkey (where the majority of UK solar transformer manufacturing occurs) requires a minimum of four weeks’ notice and a formal FAT plan submission. EPC project schedules that do not include adequate float for FAT coordination routinely experience energisation delays. The DNO witness test obligation is documented in the Grid Connection Agreement and cannot be waived without formal written consent from the DNO’s protection engineer.

Preparing your project documentation accurately before the grid connection submission stage is critical. This solar design preparation checklist outlines the key technical documents required at each project stage, and the AutoCAD solar planning resource covers the draughting standards required for MV single-line diagrams submitted to the DNO.

Common Specification Errors That Cause DNO Rejection in UK Solar Projects

A review of DNO technical review letters across 23 UK solar projects reveals a consistent pattern of specification errors that result in formal rejection or conditional acceptance of the grid connection technical submission. Understanding these errors in advance allows the EPC contractor to eliminate them before submission, reducing the review cycle from an average of 14 weeks to closer to 6 weeks.

• Incorrect Uz% specification: Specifying Uz% outside the DNO’s accepted range (typically 4% to 6% for 11kV connections) without prior written agreement from the DNO protection engineer.
• Missing or invalid type test evidence: Submitting routine test certificates instead of IEC 62271 or IEC 60076 type test reports from an accredited laboratory.
• Incorrect vector group: Specifying Dyn1 or Dyn5 instead of the DNO-required Dyn11, often as a result of copying a specification from a European project where alternative vector groups are standard.
• Underspecified earth fault protection: Failing to include a restricted earth fault (64) element on the transformer LV winding for transformers above 1 MVA, which is required by most UK DNOs under their protection schedules.
• SF6 switchgear without leak testing documentation: Supplying GIS or SF6-filled RMUs without a zero-leak certificate (F-Gas compliance) from the manufacturing test, which is increasingly required by UK DNOs and planning authorities.
• Tap changer range inadequate for rural feeders: Specifying only a +/- 2.5% tapping range on a transformer connecting to a rural 11kV feeder where the network voltage variation can exceed 4% during low-load, high-generation conditions.

Practical Sizing Example: 4.9 MW Solar Farm Connecting at 11kV

To bring together the principles described above, consider a 4.9 MW ground-mount solar farm in Shropshire connecting to a rural 11kV ring feeder. The inverter configuration is 14 x 350 kW string inverters arranged in two arrays, each array feeding one 2,500 kVA transformer. The DNO is National Grid Electricity Distribution (NGED) Midlands.

Step one is transformer kVA sizing. Each inverter array produces 14 x 175 kW = 2,450 kW per array (two inverters per array string to balance the configuration). At a worst-case power factor of 0.95, the apparent power is 2,450 / 0.95 = 2,579 kVA. Applying a 10% harmonic derating factor gives a minimum rated transformer size of 2,837 kVA. The nearest standard IEC unit size above this is 3,150 kVA, which provides adequate thermal margin and allows the transformer to operate continuously at 90% load without exceeding the 98-degree hot spot threshold.

Step two is specification confirmation. The transformer specification for NGED Midlands is: 3,150 kVA ONAN, 11kV/433V, Dyn11, Uz% = 5%, OCTC +/- 5% in 2.5% steps, K-factor rated K=7, IEC 60076 type tested, oil-filled, ONAN cooling, with a stainless steel conservator tank and bottom filter valve for oil sampling. The RMU specified is a 12 kV, 16 kA, SF6 ring main unit to IEC 62271-200 with two load break switches and one VT-controlled IDMT protection relay feeder panel, IP54 rated for outdoor installation.

This specification, submitted with the correct type test evidence and a completed relay setting schedule based on NGED’s protection data sheet for the relevant feeder, passed the NGED technical review in 7 weeks with a single minor clarification request on the ROCOF relay setting threshold, which was resolved within 48 hours.

Frequently Asked Questions

Conclusion: Getting MV Specification Right the First Time

The specification of transformers and MV switchgear for UK solar PV projects is a discipline that sits at the intersection of electrical engineering, regulatory compliance, and commercial risk management. An EPC contractor who understands the transformer kVA sizing methodology, the DNO’s Dyn11 vector group requirement, the Uz% tolerance window, the RMU versus AIS versus GIS trade-offs, and the G99 protection relay coordination obligations does not simply produce a better technical submission. That contractor wins the programme race, avoids the re-engineering cycles that destroy project margins, and builds the credibility with DNO protection engineers that generates faster review turnaround on future projects.

The most important takeaway from the case studies and technical detail presented in this guide is that every specification decision has a downstream consequence, and those consequences arrive with a price tag attached. The difference between a Dyn11 and a Dyn1 transformer is one number in a specification document, but it represents 14 weeks and £38,000 in one real project. The difference between correct harmonic derating and no derating is £1,200 in procurement versus £47,000 in premature replacement. The details matter enormously in UK solar electrical design, and they matter most at the MV substation level where every component sits on the critical path between inverter output and grid connection.

Start your MV substation specification work early, engage the DNO’s protection engineer before finalising transformer procurement, and treat the FAT and witness test obligations as fixed project milestones rather than optional documentation exercises. That approach, consistently applied, is what separates successful medium voltage solar EPC delivery from the avoidable delays and costs that continue to characterise too many UK solar projects.