Autotransformer motor starting is governed by current, torque, source impedance, and time. An indicated 9.9% voltage dip is only 0.1 percentage point inside a ±10% steady-state voltage band, and the calculation used a 68% tap even though the motor-starting study used the 75% tap. That mismatch can consume the entire margin before generator dynamics or transition transients are considered.
Starting quantities and acceptance limits
The installation combines two generator sets with approximately 100 MW total capacity on an island network and an 11 kV, 15 MW induction motor. The reported start completes in under 20 seconds on the 75% autotransformer tap. Because the source is islanded, the motor interacts directly with generator electromagnetic response, excitation control, prime-mover loading, and network impedance.
| Quantity or limit | Value | Engineering use | Where to verify |
|---|---|---|---|
| Motor rating | 11 kV, 15 MW | Motor model and load-flow basis | Motor datasheet and certified model |
| Generating capacity | Approximately 100 MW from two sets | Operating scenario and source strength | Generator schedules and power-system model |
| Reported dip | 9.9% | Headline result requiring time-domain confirmation | Vendor calculation report and voltage trace |
| Known voltage tolerance | ±10%, steady state | Not automatically applicable to the initial dip or transfer event | Network design criteria and equipment ride-through data |
| Tap used for dip calculation | 68% | Produces the more favorable source-current result | Calculation input listing |
| Tap used for acceleration study | 75% | Established acceleration in under 20 seconds | Motor-starting study |
| Fault-level case | Minimum system fault level | Defines the weak-source scenario | Fault-level report and model case definition |
The number that matters is not one dip percentage detached from time. Obtain the bus-voltage trace from energization through acceleration and transfer, plus motor current, speed, torque, generator terminal voltage, excitation output, frequency, and real and reactive power. Apply steady-state, transient, protection, and equipment ride-through limits to their respective time intervals.
Starter approaches and deciding criteria
| Approach | Starting behavior | Transfer behavior | Primary advantage | Primary concern in this case |
|---|---|---|---|---|
| Open-transition autotransformer | Reduced-voltage acceleration | Motor is momentarily disconnected during transfer | Simple conceptual sequence | Reconnection can produce a high current and torque transient, light flicker, protection operation, and excitation-system disturbance |
| Closed-transition autotransformer | Reduced-voltage acceleration | Maintains an electrical path during transfer; the autotransformer winding acts as series impedance during the transition | Suppresses the open-transition reconnection event | Requires correct switching topology, interlocking, ratings, and transition analysis at 11 kV |
| Medium-voltage soft starter | Gradually controls applied voltage and current | No autotransformer switching transition | Can ramp current so the AVR responds to increasing load; a power-acceleration mode can also control the added prime-mover load | Must be evaluated for the 11 kV, 15 MW motor, required starting torque, thermal duty, harmonics, bypass arrangement, and system integration |
The confirmed arrangement is closed transition, so retain it as the base design provided the corrected dynamic study passes. A soft starter becomes the stronger alternative when the island system needs controlled current or controlled real-power acceleration rather than fixed-tap starting. Open transition offers no useful benefit here that offsets its reconnection transient.
Tap-dependent current and accelerating torque
For an ideal autotransformer with tap ratio k, motor terminal voltage during starting is approximately kV. At a given slip, motor current is approximately k times the full-voltage locked-rotor current, while source line current is approximately k² times that current. Motor electromagnetic torque is also approximately proportional to k² at the same slip.
The 68% and 75% cases therefore cannot be interchanged:
| Tap | Approximate source-current multiplier | Approximate torque multiplier |
|---|---|---|
k = 0.68 |
0.68² = 0.4624 |
0.4624 |
k = 0.75 |
0.75² = 0.5625 |
0.5625 |
Under the idealized square-law relationship, the 75% tap draws 0.5625 / 0.4624 = 1.216, or approximately 21.6% more source starting current than the 68% tap. If voltage dip scaled linearly with that current, the reported 9.9% result would screen at approximately 9.9% × 1.216 = 12.0%. That 12.0% value is a first-pass calculation, not an acceptance result: actual dip depends on the complex network impedance, motor impedance versus speed, generator dynamics, AVR action, and concurrent loads.
The higher tap also provides approximately 21.6% more motor torque at the same slip. This explains why the 75% tap may accelerate successfully while the 68% tap produces a better-looking voltage calculation. If reduced-voltage torque does not exceed load torque plus accelerating requirements throughout the speed range, the motor reaches transfer at insufficient speed. Applying full voltage at that point can drive current toward a direct-on-line starting condition and impose a severe torque step.
Generator and network dynamic representation
Minimum fault level is a useful weak-network input, but a single fault-level number cannot reproduce the complete motor-starting response. For a balanced three-phase bus, the magnitude of the equivalent source impedance can be screened from short-circuit apparent power using |Zth| ≈ VLL² / Ssc, with voltage in consistent units. The full model also needs the impedance angle or X/R relationship and the generator dynamic data.
Synchronous, transient, and subtransient reactances represent different time regions. Synchronous reactance belongs to the sustained machine characteristic; it is not the correct sole basis for the transfer event or the early generator-voltage response. For the dip point under review, transient impedance is the relevant reduced generator representation, while subtransient behavior addresses the earliest electrical cycles. A time-domain generator model avoids forcing one reactance onto the entire event.
Labeling the period after 1–2 seconds as purely a synchronous-reactance period is insufficient without checking generator time constants and excitation action. During a start lasting up to 20 seconds, the AVR can increase excitation and partly recover generator terminal voltage. Its response depends on sensing, control tuning, limiters, available excitation, and the system operating point. The AVR may be too slow to counter the brief switching transient yet still materially affect the longer acceleration interval.
Model both generator sets in the actual minimum-fault-level operating configuration. Confirm which sets are online, their pre-start real and reactive loading, bus voltage, excitation state, governor response, transformer impedances, and other connected loads. The study must distinguish an electromagnetic voltage depression from frequency or real-power effects caused by accelerating a 15 MW load.
Closed-transition switching mechanism
Open transition disconnects the motor before reconnecting it to the supply. During the interruption, residual motor voltage continues with its own magnitude, frequency, and phase. Reclosing against an unfavorable phase relationship can produce a current and torque transient larger than the reduced-voltage running values, with possible light flicker, data-system disturbance, or nuisance operation of protective devices.
In the closed-transition autotransformer arrangement described here, opening the transformer star point leaves transformer winding impedance in series with the motor during the changeover. The motor remains electrically related to the supply rather than being allowed to drift freely before reconnection. This is different from a closed-transition star-delta scheme that uses transition resistors and another switching device; those components must not be added to the autotransformer design merely by analogy.
Closed transition reduces the principal open-transition risk, but it does not eliminate switching duty. Switching sequence, overlap, breaker or contactor timing, transient recovery voltage, temporary winding current, motor torque, and protection behavior still require manufacturer confirmation. At 11 kV, equipment ratings and the proposed vacuum switching implementation must be reviewed from the actual single-line diagram and sequence, not inferred from a sketch.
Recommended study and design procedure
- Reconcile the tap inputs. Run voltage, current, torque, and acceleration calculations with the same 75% tap proposed for starting. Retain the 68% case only as a separate alternative and verify that it can accelerate the connected load without stalling.
- Obtain the calculation basis. Request the model input listing, motor equivalent-circuit or validated dynamic model, load torque-versus-speed curve, combined inertia, autotransformer impedance and taps, switching sequence, initial bus voltage, concurrent loading, and minimum-fault-level case definition.
- Model the generators dynamically. Include the applicable generator electromagnetic data, AVR and excitation system, limiters, governor or prime-mover response, and pre-start loading. Use the transient model for the assessed dip and retain subtransient detail where the earliest cycles and protection duties matter.
- Simulate the complete event. Begin before starter energization and continue through reduced-voltage acceleration, closed transition, application of full voltage, recovery, and settling. Use a time resolution capable of displaying the switching transient as well as the multi-second start.
- Check acceleration margin. Plot motor torque and load torque against speed. Confirm positive accelerating torque throughout the intended range and verify that motor speed at transfer prevents a near-direct-on-line current step.
- Apply separate acceptance criteria. Compare the sustained portion with the ±10% steady-state tolerance. Obtain transient voltage limits, equipment ride-through curves, protection thresholds, and permissible generator excursions for the initial and transfer events.
- Run credible sensitivities. Test the weakest allowed source configuration, maximum permitted generator loading, low initial voltage, realistic tap tolerances, load-torque variation, delayed transfer, and the protection settings active during starting.
- Compare a soft starter if margins remain inadequate. Specify the required current ramp, starting torque, acceleration time, motor thermal duty, and any power-acceleration requirement. Compare its simulated bus voltage, generator reactive demand, and prime-mover loading against the corrected closed-transition case.
Verification and acceptance evidence
| Observed symptom | Likely mechanism | Deciding record |
|---|---|---|
| Dip exceeds prediction immediately after energization | Incorrect source impedance, initial operating point, or motor locked-rotor model | High-resolution bus voltage, motor current, and generator terminal records |
| Voltage recovers during acceleration | AVR excitation response | Excitation output, reactive power, field limits, and terminal voltage |
| Large current spike at transfer | Low transfer speed, switching sequence error, or ineffective closed transition | Switch status, speed, phase currents, and bus voltage around transfer |
| Motor fails to reach transfer speed | Insufficient reduced-voltage torque or excessive load torque/inertia | Torque-speed curves, speed trace, and acceleration time |
| Protection operates despite acceptable average dip | Short transient, current element, undervoltage timing, or generator protection interaction | Relay event report, oscillography, element pickup, and trip logic |
Acceptance requires more than the minimum voltage value. Record the tap actually connected, start duration, speed at transition, peak and time-resolved current, minimum bus and generator terminal voltages, recovery time, frequency, generator real and reactive power, excitation response, switching states, and protection margins. Confirm autotransformer and motor thermal duty for the recorded acceleration time and permitted restart sequence using their manufacturer data.
A commissioning test should use synchronized disturbance recording across the motor feeder and generator buses. Compare measured traces against the study rather than comparing only headline minima. Any unexplained deviation in transition timing, current, voltage, or acceleration requires model correction before repeated full-duty starts.
FAQ
Can I accept a 9.9% motor-start voltage dip against a ±10% limit?
Only for the portion explicitly covered by the ±10% steady-state criterion, and only after recalculating with the actual 75% tap. Obtain separate transient and ride-through limits for energization and closed-transition switching.
Does a 68% autotransformer tap draw less source current than a 75% tap?
Yes. Under the ideal square-law relationship, the source-current multipliers are 0.4624 and 0.5625; the 75% case is approximately 21.6% higher.
Can I use minimum fault level alone to calculate the voltage dip?
Use it to define the weak-source case, not as the complete dynamic model. Add generator transient behavior, AVR and excitation controls, pre-start loading, network impedance angle, motor dynamics, and the switching sequence.
Does closed transition remove the motor-start transient?
It removes the deliberate open interval and reduces the out-of-phase reconnection risk. Energization, changeover, full-voltage connection, and electromagnetic torque changes still create transient duties that must appear in the simulation.
Can I proceed if the dynamic model or transient limits are unavailable?
Stop when the tap basis cannot be reconciled, the motor lacks demonstrated accelerating torque, or the transient voltage and switching duties have no acceptance criteria. Request the complete study inputs and time traces from the equipment manufacturers and power-system designer. Escalate unresolved generator, AVR, starter-sequence, or protection questions through the manufacturers’ official engineering support channels before commissioning.