The 30% Feeder Threshold: What Happens to a Distribution Transformer Under High Solar Penetration

· Trillectric

Ask a distribution engineer how much rooftop solar a feeder can take and you will usually hear the same number: about 30% of the transformer rating. It gets repeated in tender documents, in interconnection guidelines, and in internal notes that nobody has revisited in five years.

It is a useful starting point. It is not a limit, and it is not a standard. Two feeders fed by identical 250 kVA transformers can behave completely differently at the same penetration — one stays comfortable at 60%, the other is already tripping inverters at 25%. The difference is rarely the transformer. It is conductor size, feeder length, where along the line the systems sit, and what the daytime load looks like.

With PM Surya Ghar pushing connections faster than most utilities can survey them, the 30% heuristic is now doing load-bearing work it was never designed for. So it is worth being precise about what actually degrades, and in what order.

Why 30% became the number

The heuristic comes from a reasonable instinct: keep midday reverse flow small enough that the feeder never leaves the operating envelope it was designed for. Design assumptions on an LV distribution feeder run one way — from the transformer outward, voltage dropping along the line. Keep injection modest and none of those assumptions break.

The problem is that penetration expressed as a percentage of transformer rating says nothing about the two variables that actually decide the outcome: impedance between the injection point and the transformer, and coincident daytime load. A long, thin rural feeder with lightly loaded daytime consumers will hit voltage problems at a fraction of the penetration that a short, heavily loaded urban feeder tolerates without complaint.

What actually happens, in order

1. Voltage rises at the far end

This is almost always the first constraint, and it is the one least often measured. Current flowing back toward the transformer through line impedance produces a voltage rise rather than a drop. The further the system is from the transformer and the thinner the conductor, the larger the rise.

Consumers at the tail of the feeder see the highest voltage, and they are usually the ones with no solar at all. Their appliances wear. Their complaints arrive at the consumer grievance desk, not the planning desk, which is one reason the connection is often missed.

2. Phase unbalance appears

Most rooftop systems below 10 kW in India are single-phase. Nobody coordinates which phase they land on — the installer connects to whatever is convenient at the meter board. Within a year or two, one phase on a street can carry substantially more injection than the other two.

Unbalance produces its own set of effects: neutral current where the design expected almost none, differing per-phase voltages, and protection settings that were tuned for a balanced picture. Per-phase measurement is what distinguishes real unbalance from a sensor or wiring problem, and it is exactly what single-number monitoring cannot tell you.

3. Reverse power flow through the transformer

At midday the distribution transformer starts exporting upstream instead of stepping down. Transformers handle this direction fine electrically. What does not handle it gracefully is everything configured around the assumption of one-directional flow: automatic tap changers further upstream, directional protection settings, and loss allocation in the billing and energy-accounting chain.

Worth separating two things people conflate here. A flat-topped midday generation curve is normal — that is the inverter capping its AC output at its rating. Reverse flow through the transformer is a different phenomenon entirely, and only the second one changes how the network behaves.

4. Inverter trips, which look like the problem but are not

Once feeder voltage drifts near or past the statutory upper limit, inverters begin disconnecting on over-voltage protection. This is the stage at which the utility usually finds out something is wrong — because generation has stopped and someone calls.

The trips are a symptom. The inverter is doing exactly what IS 18968:2025(/blog/is-18968-2025-india-smart-inverter-standard) expects of it. The underlying condition arrived weeks or months earlier and nobody saw it, because nobody was looking at the right place.

The measurement gap

Here is the structural reason all four of these go unnoticed: on most Indian distribution networks, visibility stops at the 33/11 kV substation. Everything described above happens downstream of that point, on the LV side, where the only routine measurement is a monthly meter read.

Monthly energy totals cannot show a voltage excursion that lasts two hours a day. They cannot show which phase is carrying the injection. They cannot distinguish a feeder that is comfortable at 45% penetration from one that is in trouble at 20%. You get a number for the month and no information about the shape of any day in it.

This is the same gap we wrote about in the context of monitoring under PM Surya Ghar(/blog/monitoring-the-missing-link-pm-surya-ghar). The subsidy programme succeeded at getting systems installed. It did not come with the measurement layer needed to operate the network they are connected to.

What a distribution engineer can actually do

Measure voltage, not just energy. A voltage profile through the day at the tail of the feeder tells you more about headroom than any penetration percentage.

Insist on per-phase data. Unbalance is invisible in aggregate and expensive to fix late.

Look at midday export shape, not monthly totals. The question is not how much a feeder exported in September. It is what the worst hour of the worst day looked like.

Derive your own threshold. Once you have two or three months of measured data on a feeder, you have an observed limit for that feeder. That number is worth more than an inherited rule of thumb, and it will usually be different — sometimes far higher, occasionally much lower.

Sequence the interventions. Re-phasing and tap adjustment are cheap. Conductor upgrades and transformer replacement are not. Measurement is what tells you which one the feeder needs.

The takeaway

The 30% threshold is a placeholder for a measurement nobody took. Treated as a planning prompt it is harmless. Treated as a hard limit it does two kinds of damage: it blocks connections on feeders with plenty of headroom, and it waves through connections on feeders that are already past it.

You cannot manage a penetration limit you cannot see. Before a utility can decide how much more rooftop solar a feeder will take, it needs to know what the feeder is doing today — hour by hour, phase by phase, at the end of the line where the problems show up first.