Tiza Group

Home / Guides / Studies and design

Motor Reacceleration After a Voltage Sag

What happens when several motors reaccelerate at once as voltage returns, how bus voltage and time are estimated, and how to stagger them. With example.

By Shadai García, CEO and protection engineer at Tiza Group.

external network, Scc = 75 MVA524.16 kV busbarexternal faultMAMBMCTP2727 trips Crecloses at 5.2 sall three together: busbar at 0.776 puA and B reaccelerate together: busbar at 0.839 puC reaccelerates alone: busbar at 0.912 pubusbar voltage (pu)1.00.80510 sminimum 0.80 puall three together: 0.776 puA and B: 0.839C: 0.912A5.2 sB5.2 sC3.9 sone after another: 3 × 3.9 = 11.8 s, fails the 10 sA and B together, then C: 5.2 + 3.9 = 9.1 sChainBusbar and sourceMotor and loadProtection 27Sequence© tizagrp.comexternal network, Scc = 75 MVA524.16 kV busbarexternal faultMAMBMCTP2727 trips Crecloses at 5.2 sall three together: busbar at 0.776 puA and B reaccelerate together: busbar at 0.839 puC reaccelerates alone: busbar at 0.912 pubusbar voltage (pu)1.00.80510 sminimum 0.80 puall three together: 0.776 puA and B: 0.839C: 0.912A5.2 sB5.2 sC3.9 sone after another: 3 × 3.9 = 11.8 s, fails the 10 sA and B together, then C: 5.2 + 3.9 = 9.1 sChainBusbar and sourceMotor and loadProtection 27Sequence© tizagrp.com
Reacceleration of three motors on a 4.16 kV busbar: all three together pull the busbar down to 0.776 pu and fail; A and B together and then C keep it above 0.80 pu in 9.1 s. Location in the system and operating principle. Illustrative: real values come from the nameplate and the manufacturer's manual.
Short answer

After a voltage sag, the motors that stayed connected have lost speed and, when voltage returns, they all draw a current close to the starting current at once. That current drags the bus voltage down, lengthens the reacceleration and can stall motors. The study estimates the voltage and the reacceleration time from the motor and its load, and decides which motors stay connected, which are disconnected by the undervoltage protection and in what order they return.

Short answer

When the voltage of a bus drops because of a fault and returns when it is cleared, the motors that stayed connected have lost speed. As the voltage recovers, they all draw a high current at once, close to the starting current, to get back to speed. If there are many, the sum of those currents through the source impedance depresses the bus voltage, and with less voltage the motors accelerate more slowly: the high current lasts longer and the sag is prolonged. In the worst case, some motors stall or trip on protection.

The study estimates, from the source and motor data, what voltage the bus has during reacceleration, how long it takes and whether the torque is enough. With that result it decides which motors stay connected, which are disconnected by the undervoltage protection and reconnected later, and in what order. It is a different study from a normal start of large motors: here the motors start together, at different speeds and on the network’s schedule.

What happens during and after the sag

During the sag the motor torque falls with the square of the voltage and is lower than the load torque: the motor decelerates. How much depends on the inertia of the set and on the duration of the sag: a high-inertia load, such as a large fan, loses little speed; a pump with low inertia and almost constant torque loses a lot. Meanwhile, large induction motors can contribute current to the fault; see the prolonged contribution of a large induction motor.

When voltage returns, reacceleration starts with a large slip. The current is several times the rated value, and the reactive power is high because the motor works at a low power factor. If at that moment the undervoltage protection (27) has not yet opened, the motor stays connected and reaccelerates; if it opened, the motor stops and is started again. The difference is set by the time delay of the 27; see the undervoltage trip when starting motors. For the effect of the sag on the plant, see voltage sags in industry.

Study data

  • Short-circuit power at the bus and its X/R ratio, with the lowest-power source possible.
  • Per motor: power, starting current and power factor, torque and current curves against speed. If the nameplate is missing, see the data and assumptions of the starting study.
  • Load torque curve against speed and total inertia.
  • Sag profile: depth and duration.
  • Settings of the undervoltage protection and of the contactors or breakers.
  • Minimum bus voltage criterion and the maximum time the process tolerates.

First-pass estimate

Before a dynamic simulation, a static estimate is made. With the motors locked at their starting impedance and the source reduced to its reactance, the bus voltage is a divider: V = Ssc / (Ssc + Sstart), with the powers in the same unit. The torque of each motor is proportional to V². The acceleration time is estimated with the average torque left over, that of the motor minus that of the load. It is a conservative estimate, because the current drops as the motors accelerate and the voltage recovers.

Worked example

Example data (assumed values, not measured): 4.16 kV bus with a short-circuit power of 75 MVA. Three identical motors; each draws 7.2 MVA at starting with rated voltage (6 times its 1.2 MVA rating). Average motor torque at rated voltage: 1.0 (unit of the example); load torque: 0.30, constant. At rated voltage the motor accelerates in 3.0 s. Criteria of the example: minimum bus voltage 0.80 pu during reacceleration and a total time to recover the three motors of 10 s at most.

1. All three together. Starting power: 3 × 7.2 = 21.6 MVA. V = 75 / (75 + 21.6) = 0.776 pu. The torque is 0.776² = 0.603, and 0.603 − 0.30 = 0.303 is left over. The time is 3.0 × (1.0 − 0.30) / 0.303 = 6.9 s. The voltage does not comply: 0.776 < 0.80.

2. Two together. 14.4 MVA: V = 75 / 89.4 = 0.839 pu. Torque 0.704; 0.404 left over. Time: 3.0 × 0.70 / 0.404 = 5.2 s. The voltage complies.

3. One alone. 7.2 MVA: V = 75 / 82.2 = 0.912 pu. Torque 0.832; 0.532 left over. Time: 3.0 × 0.70 / 0.532 = 3.9 s.

4. A sequence. First A and B together (5.2 s), then C (3.9 s): 5.2 + 3.9 = 9.1 s in total, with a minimum voltage of 0.839 pu. It meets both criteria. If the three started one after the other, the total would be 3 × 3.9 = 11.8 s, and it would not meet the 10 s of the process.

5. Current. With A and B together the bus current is 6 × 0.839 = 5.0 times the rated current of each motor, for 5.2 s. It is compared with the locked-rotor thermal limit of each motor; the example assumes it is met.

Motors reacceleratingBus voltageTime
All three together0.776 pu6.9 s
Two together0.839 pu5.2 s
One alone0.912 pu3.9 s

The example assumes that C is disconnected for a short time by the undervoltage protection and keeps a similar speed; if it stops completely, it is calculated as a start from standstill, with its own time.

How the motors are staggered

  • Different undervoltage delays. The motors you prefer to drop first are disconnected by the 27 after a short time; those of a critical process, after a longer time or not at all.
  • Delayed reconnection. Each disconnected motor returns with a different delay, in the order of the study, so they do not coincide.
  • Priority by process. The order is set by what the process tolerates, not by motor size.
  • Start limit. Each motor has a maximum number of starts per hour; see the starts per hour of a motor.

Mistakes to avoid

  • Studying reacceleration with the motor model of a start from standstill.
  • Using the maximum short-circuit power of the source: the critical case is the minimum.
  • Ignoring the load torque curve and assuming a constant torque with no support.
  • Setting the 27 without checking which motors it leaves to reaccelerate.
  • Forgetting the current and heating of each motor during reacceleration.

Checklist

  • Minimum short-circuit power of the bus confirmed.
  • Data of each motor and torque curve of each load.
  • Sag profile (depth and duration) defined.
  • Bus voltage and reacceleration time estimated for each group.
  • Minimum voltage criterion and maximum process time set.
  • 27 and reconnection settings coordinated with the chosen order.
  • Thermal limits and starts per hour reviewed.
  • Case confirmed with dynamic simulation.

References

  • Calculations and data: our own, with assumed data for the example. The first-pass method (voltage divider and torque proportional to the square of the voltage) is a simplification; the final study is confirmed with dynamic simulation.

Need this study?

The reacceleration study of your motors, with bus voltage, times, undervoltage settings and the reconnection sequence, is part of our engineering design and studies service. We do it for projects in any country. Write to us.

Frequently asked questions

Why do several motors reaccelerate worse than a single one?

Because each motor draws a high current and all of them flow through the same source impedance, which causes a voltage drop proportional to the total current. With less voltage, motor torque falls with the square of the voltage, the motor takes longer to accelerate and the high current lasts longer. With a few motors the voltage recovers; with too many, the bus does not rise and some motors stall.

What data is needed to study reacceleration?

The short-circuit power at the bus, the data of each motor (power, starting current, torque curve), the torque curve and inertia of the load, the sag profile and the settings of the undervoltage protection. Without the load torque curve, the acceleration time is an estimate with no basis.

Is reacceleration the same as starting?

No. In reacceleration the motor keeps part of its speed, so the current is lower than in a start from standstill and the time is shorter. But if the sag was long and the motor stopped, the case is treated as a start. The study must cover both cases according to the sag duration.

WhatsApp