Large Motor Starting: Voltage Dip and Study

Short answer

An induction motor starts by drawing several times its rated current at a low power factor, and that current, flowing through the impedance of the grid, the transformer and the cables, pulls down the voltage across the whole bus. If the dip is large, the motor itself fails to accelerate, contactors drop out, drives shut down and protection can trip. A motor starting study (IEEE 3002.7) calculates that dip with the plant model and defines the starting method and the settings that keep it within what the equipment tolerates.

What happens in the grid when a large motor starts

An induction motor started direct-on-line draws the locked-rotor current while the rotor is stopped: several times its rated current and at a low power factor. The exact value is on the nameplate and the data sheet (on NEMA motors, from the locked-rotor kVA code letter). This current flows through the utility network, the plant transformer and the cables to the motor, and produces a voltage drop in each impedance.

The result is a voltage dip on the bus that feeds the motor and, to a lesser degree, on neighboring buses. It lasts as long as the acceleration: from fractions of a second on a pump to many seconds on a mill, an induced-draft fan or a loaded belt. As the motor gains speed, the current falls and the voltage recovers.

Two things make the problem grow:

The grid is weak compared with the motor. The lower the short-circuit power at the bus compared with the motor starting power, the larger the drop. A plant transformer with high impedance or a long cable makes the result worse.
The plant runs on its generator. A standby generator has a much higher impedance than the grid, and its voltage regulator needs time to respond. A motor that starts without trouble on the grid may not start on the generator set.

Effects of the voltage dip in the plant

The torque of an induction motor varies roughly with the square of the voltage. If the voltage at its terminals falls, the available torque falls much more, and the motor takes longer to accelerate or cannot overcome the load torque. What is seen in the field:

Effect	Why it happens
The motor does not finish starting or takes very long	Torque falls with the square of the voltage and does not exceed the load torque
Contactors dropping out in other circuits	Coil voltage falls below its holding value
Drives and electronic equipment shutting down	DC bus voltage falls below its threshold
Light flicker	Dip repeated with each start
Trip of the motor or incomer protection	Start longer than assumed in the settings, or 27 operation
Rotor heating	Each extra second of start adds heat to a rotor already working at its limit

These symptoms are often mistaken for equipment failures. The drive that shuts down or the contactor that drops out is not damaged: it is reacting to a voltage the study did not foresee.

What a motor starting study calculates

The IEEE reference is IEEE 3002.7, the recommended practice for motor starting studies in industrial and commercial systems. The study uses the same plant model as the load flow study and the short-circuit study, and adds the dynamic model of the motors and their mechanical loads.

Motor starting study

Data. Network, transformers, cables, motor nameplate and load curve
Model. Impedances, motor and load resisting torque
Scenarios. Normal grid, minimum grid and standby generator
Simulation. Voltage, current and speed during acceleration
Criteria. Minimum voltage per bus and starting time
Recommendations. Starting method, taps and protection settings

There are two levels of analysis:

Static: calculates the voltage drop at the instant of starting with the locked-rotor impedance. It is enough to rule out clear cases.
Dynamic: simulates the acceleration over time with the torque-speed curve of the motor and the load. It gives the real starting time, the evolution of voltage and current, and whether the motor reaches rated speed. It is needed for high-inertia loads or weak grids.

The results are compared with what the connected equipment tolerates and with the conditions at the point of connection with the utility. When a start affects other customers, the utility may ask the plant to show that the dip it causes is under control. Each utility sets its own criteria in its connection rules, and we apply the ones that govern your project.

How the voltage dip is reduced

The study does not only say whether there is a problem: it compares solutions. The most common:

Starting method. Star-delta, autotransformer, reactor, soft starter and variable frequency drive all reduce the starting current, each with its own cost in torque. The soft starter and the autotransformer lower the current at the expense of torque; the drive starts with a current near rated and full torque. In medium voltage the options are described in medium-voltage motor starter.
Transformer taps so the bus runs higher before the start, without overshooting at no load.
Start sequence: do not start two large motors at once, and start the largest first when the load is lower.
Compensation: starting capacitors or dynamic compensation in specific cases.
Changing the transformer or the cables when the system is at its limit.

Each option affects protection. A soft starter changes the shape of the current; a drive introduces harmonics (see VFD harmonics); a longer start requires reviewing the safe locked-rotor time.

The link with protection settings

Starting is the moment when the conditions that lead to a trip come together the most. For that reason the starting study feeds the settings:

The motor and incomer 50/51 overcurrent must let the real starting current and time pass, with margin. See overcurrent trip at start or energization.
Bus undervoltage 27 must not act during the dip of a normal start, but must act on a real loss of the source. See undervoltage 27.
The motor locked rotor and prolonged start protection needs the calculated starting time, not an assumed one. See locked rotor 51LR and prolonged start 48.
The 49 thermal model must accept the number of starts the process needs.

We do your starting study

Our team models your plant and simulates the start of your motors in the real operating scenarios (grid, minimum grid and standby generator). We deliver the minimum voltage at each bus, the starting time and the recommendations for starting method and settings. We do the study remotely, worldwide, with your data. See our engineering design and studies service and send us your single-line diagram and the motor data through contact.

References

IEEE 3002.7-2018, IEEE Recommended Practice for Conducting Motor-Starting Studies and Analysis of Industrial and Commercial Power Systems.
NEMA MG 1, Motors and Generators (locked-rotor kVA code letters).
IEEE C37.96, IEEE Guide for AC Motor Protection.

Frequently asked questions

How much voltage dip is acceptable when starting a motor?

There is no single value: it depends on what is connected to the bus. The calculated minimum voltage is compared with what each device tolerates (the torque of the motor being started, the holding voltage of contactors, the range of drives and electronic equipment) and with what the utility requires at the point of connection. The study sets the criterion for each bus.

Why does the breaker trip when a large motor starts?

The usual causes are an instantaneous 50 set below the starting current, a 51 curve with no margin for the real acceleration time, undervoltage 27 acting during the dip, or a CT that saturates with the asymmetrical component. A start longer than expected, caused by low voltage, makes the first three worse.

When is a motor starting study needed?

When the motor is large compared with the short-circuit power of the bus, when a new motor is installed or the transformer is changed, when the plant starts operating on a standby generator, or when there are already symptoms: flickering lights, contactors dropping out, drives shutting down or starts that never finish.