A solar plant that exports power raises the voltage at its point of connection, the more so the weaker and more resistive the grid is. If at peak generation the local load is low and the grid voltage is already high, the 59 at the point of connection or at the inverters trips. Other common causes are a poorly configured reactive power control mode, a ground fault on an ungrounded network seen by a phase-to-ground 59, and settings that were not coordinated with the inverter curves.
Symptom
A solar plant connected to the distribution network or to a higher level disconnects on overvoltage. The trip can come from the relay at the point of connection (59) or from the inverters. It tends to concentrate in the hours of peak irradiance, on low-demand days, or to coincide with switching or faults on the utility network. Each outage means energy not delivered and, if it repeats, disputes with the utility.
The most frequent physical cause is the export itself. The current the plant injects flows through the grid impedance and produces a voltage difference between the point of connection and the utility substation. Approximately, that rise is proportional to the grid resistance times the active power plus the reactance times the reactive power, divided by the voltage. On long, resistive medium-voltage networks, active power dominates and the rise can be appreciable. The functions at the point of connection are described in protection of a solar plant and the function itself in overvoltage 59.
Possible causes, from most to least likely
| Cause | Why it happens | What gives it away |
|---|---|---|
| Voltage rise from export | Weak or resistive grid, high generation and low local load | Trips at midday, with voltage that rises with the injected power |
| Grid voltage already high | Utility transformer tap, capacitor banks connected at low demand | High voltage before the plant generates or with the plant disconnected |
| Poorly configured reactive power mode | Unsuitable fixed power factor, setpoint with inverted sign, plant controller in a mode different from the agreed one | The plant injects reactive power when it should absorb |
| Ground fault on an ungrounded network | The healthy phases rise toward the phase-to-phase voltage | Trip coinciding with a ground fault on the utility network, with high 3V0 |
| Settings not coordinated with the inverter | The 59 at the point of connection falls inside the inverter’s support curve or is faster than required | Trip on transient overvoltages that the inverter was supposed to ride through |
| Transient overvoltage | Load rejection, fault clearing, opening of an upstream breaker | Brief spike after a disturbance |
| Incorrect measurement | VT ratio entered wrongly, 110 V secondary treated as 120 V | Relay voltage different from the one measured with an instrument |
| Islanding | The utility network opened and the plant was left feeding a smaller load | Voltage and frequency changing together after an external opening |
What to check first
- Voltage and power trend at the point of connection over several days: whether the voltage rises with the injected active power and drops when the plant is not generating.
- Grid voltage with the plant disconnected, in the morning or at night, to know what value it starts from.
- Reactive power control mode of the plant controller and of the inverters, with its sign and setpoint.
- Trip oscillography: voltages of the three phases, 3V0 and duration, to separate a sustained overvoltage from a transient one or from a ground fault.
- 59 settings at the point of connection and at the inverters against what the protection requirements in force demand and against the voltage ride-through curves.
- Was the overvoltage sustained or transient?
- Transient: Did it coincide with a ground fault on the network, with high 3V0?
- Yes: Ground fault on an ungrounded network seen by a phase-to-ground 59
- No: Overvoltage from switching or fault clearing: check the setting against the inverter curve
- Sustained: Does the voltage rise with the power the plant injects?
- Yes: Voltage rise from export: check the reactive power mode and the grid capacity
- No: Grid voltage already high or incorrect relay measurement
- Transient: Did it coincide with a ground fault on the network, with high 3V0?
How to confirm it
- Load flow calculation of the circuit with the plant at maximum generation and the grid at minimum demand, compared with the recorded voltage. See load flow study.
- Measurement with an analyzer at the point of connection, with the plant at different power and reactive power levels.
- Verification of the relay measurement against an instrument, and of the VT ratio entered.
- Test of the 59 and its time delay with secondary injection. See the solar plant commissioning tests.
How to keep it from happening again
- Interconnection study with the real grid voltage at minimum demand, not only the nominal one.
- Voltage or reactive power control agreed with the utility and tested at commissioning.
- 59 settings coordinated with the inverters’ support curves and with what the protection requirements in force demand.
- Coordination with the utility on the regulation of its substation and its capacitor banks.
- Power limitation as a last resort if the grid cannot accept more export at certain hours.
What not to do
- Do not raise the 59 setting above what is required to avoid trips: the plant would be outside what the utility approved.
- Do not change the inverters’ reactive power mode without agreement with the utility.
- Do not assume the inverter failed without looking at the grid voltage with the plant out.
- Do not reconnect by hand without respecting the agreed reconnection conditions.
What the interconnection reference values say
IEEE 1547 is the reference for the interconnection of distributed resources and sets, for each resource category, the conditions in which the inverter must trip on overvoltage. In the 2018 edition the default values documented in regulatory filings are 1.1 pu with 2 s and 1.2 pu with 0.16 s. In the 2003 edition, the times were not adjustable: above 1.2 pu the clearing time was 0.16 s, and between 1.1 and 1.2 pu it was 1 s. The PJM ride-through guide, which distinguishes categories II and III of the 2018 edition, keeps 0.16 s with no ride-through above 1.2 pu for both.
These values are not copied to a specific plant. The grid code and the utility’s interconnection requirements govern each case, and the utility can set its own settings at the point of connection. What the comparison does show is how the protection is structured: two or more voltage levels, each with its own time, so that a mild, sustained overvoltage gives the system more time than a high, brief one.
IEC 60255-127, whose 2010 edition defines the minimum requirements of overvoltage and undervoltage relays (protection function, measurement characteristics and delay characteristics), is the reference for the protective relay at the plant substation. The inverter has its own internal limits, and both levels must be coordinated: the inverter’s, the interconnection protection’s and the one required by the grid.
Points that explain trips that are not a plant fault:
- Voltage rise from active power injection in weak networks. As production rises, the voltage at the point of connection rises, and the limit is reached at midday with high irradiance and low demand.
- Tap changer position of the plant and substation transformers, which sets the starting voltage before the inverter injects.
- Voltage measurement on the side where it is measured: the inverter measures at its low-voltage terminals and the interconnection protection at medium or high voltage, and between them there are voltage drops and transformation ratios that must be considered in the setting.
- Reactive power control of the inverter (fixed power factor or voltage-reactive power control): a setting that absorbs reactive power when the voltage rises reduces the risk of tripping, and an incorrect one makes it worse.
- Load disconnections or another plant’s disconnection that produce a temporary rise on the network.
For the analysis, you need the voltages at the point of connection, the tap position and the power injected at the moment of the trip, compared with the utility’s limits.
When to call us
If your solar plant disconnects on overvoltage, send us the connection relay events and the voltage and power trends: we quote the trip analysis. We review and adjust the relay at the point of connection and the plant logic through remote relay configuration, worldwide, and on site we carry out the testing and commissioning of solar plants in Colombia, or elsewhere on request. Write to us.
Related guides
- Protection of a solar plant
- Overvoltage protection 59
- Protection requirements for distributed generation
- Solar plant commissioning tests
References
- IEEE 1547-2018, Standard for Interconnection and Interoperability of Distributed Energy Resources with Associated Electric Power Systems Interfaces.
- IEC 60255-127, Functional requirements for over/under voltage protection.
Frequently asked questions
Why does the solar plant trip on overvoltage at midday?
Because at that hour generation is at its maximum and the circuit load is usually low. The power the plant injects flows through the grid impedance toward the substation and raises the voltage at the point of connection. If the grid voltage was already near the limit, the additional rise pushes the measurement above the 59 setting.
Can reactive power control prevent overvoltage trips?
It can reduce them. Absorbing reactive power offsets part of the voltage rise produced by active power, especially in networks with appreciable reactance. The control mode, the setpoint and its limits must be agreed with the utility and stay within the inverters' capability curve.
Why did the 59 trip during a ground fault on the utility network?
On an ungrounded or high-impedance grounded medium-voltage network, a ground fault raises the voltage of the healthy phases toward the phase-to-phase voltage. A 59 that measures phase-to-ground sees it as overvoltage even though the phase-to-phase voltage is normal.