A Comprehensive Guide to Variable Frequency Series Resonance Withstand Voltage Test Systems (Case Study: 270 kV/108 kV)

Preface: This article compiles information drawn from Wuhan Guodian Zhongxing Electric Power Equipment Co., Ltd., as well as recurring practical issues identified across thousands of user comments. It systematically elucidates these topics in the following sequence: Principles → Apparatus → Wiring → Applications → Practical FAQs → Best Practices. All formulas and typical parameters presented herein allow for numerical substitution and recalculation; readers are encouraged to verify each point by cross-referencing them against the actual physical equipment.

• I. Why Is "Series Resonance" Indispensable in Power Testing?
• II. Principles: What Exactly Is Series Resonance?
• III. Equipment: What Does a Complete 270kV/108kVA System Look Like?
• IV. Wiring: How to Select and Calculate the Three Typical Wiring Configurations?
• V. Applications: What Exactly Do We Test? At What Voltage? For How Long?
• VI. Practical FAQ: Comprehensive Answers to Colleagues' Most Frequent Questions
• VII. Safety & Reliability: 5 Common Pitfalls to Avoid in Field Testing
• VIII. Conclusion: Upholding the Pursuit of Truth in Every Test

For power cables, transformers, GIS (Gas-Insulated Switchgear), switchgear cabinets, motors, and generators—whether during factory acceptance, handover, or preventive maintenance testing—it is essential to subject their insulation to a voltage significantly higher than their rated operating voltage. This serves as a rigorous "stress test" to verify whether the insulation can withstand the applied electrical stress. This type of evaluation is known as an AC Withstand Voltage Test.

However, a challenge arises:

• A 1-kilometer-long 10kV power cable typically possesses a capacitance of approximately 0.25 μF/km; when subjected to a 17.4kV power-frequency withstand test, the resulting capacitive current is around 1.4 A.
• For a section of 110kV cable spanning several kilometers, the capacitive current can reach tens—or even up to a hundred—amperes during a 128kV withstand test.
• If one were to employ a traditional power-frequency test transformer (utilizing direct voltage step-up) for such tasks, the required capacity of that transformer would range from several hundred to several thousand kVA. Such a unit would weigh several tons, rendering it physically impossible to transport to the actual field site for testing.

Consequently, engineers devised an ingenious solution: utilizing LC series resonance for voltage step-up. This method employs a relatively compact variable-frequency power source to establish a series resonant circuit comprising a reactor and the Device Under Test (DUT)—which, by its very nature, acts as a capacitor. During resonance, the voltage is "amplified" by a factor of several tens. In this manner, a testing apparatus weighing merely a few hundred kilograms can generate test voltages reaching several hundred kilovolts, while the power source itself is required to supply only the relatively small current associated with the active power losses within the circuit.

This constitutes the fundamental rationale behind the existence of Variable Frequency Series Resonance (VFSR) testing systems.

A common and simple series resonant circuit—widely encountered in the industry—consists of three components:

They are connected in series (head-to-tail) and powered by a variable-frequency power supply (~U). Note: The specimen under test is, in itself, a capacitor (C). This is a crucial point—and one that also answers a question frequently raised by colleagues: "Does the cable here function as a capacitor?" The answer is: Yes. Since the cable's two conductive layers—the core conductor and the metal shielding—are separated by XLPE insulation, its physical structure is, in fact, that of a cylindrical capacitor.

The opposition that an inductor presents to alternating current is termed "inductive reactance" (XL): XL = 2πfL. The opposition that a capacitor presents to alternating current is termed "capacitive reactance" (XC): XC = 1/(2πfC).

When the frequency f is adjusted to a specific value such that the inductive reactance equals the capacitive reactance:

This f0 represents the resonant frequency. The core mission of a variable-frequency power supply is to continuously sweep through frequencies to locate this specific f0.

At the precise moment resonance occurs, the voltage across the inductor (UL) and the voltage across the capacitor (UC) are equal in magnitude but exactly opposite in direction (exhibiting a phase difference of 180°). Their voltage vectors sum to zero; consequently, the only voltage remaining within the circuit loop is the minuscule voltage drop across the resistor R. The power supply, therefore, needs only to compensate for these losses, requiring virtually no reactive power.

This explains the source of the comments suggesting that the "external voltage is 0V"—though it is crucial to note that it is the sum of the inductor voltage and the capacitor voltage that effectively cancels out to 0V externally; the voltage across the Device Under Test (DUT) is certainly not 0V. In reality, the DUT (the capacitor C) is subjected to a very high voltage.

This addresses a common—and often perplexing—question within the industry, frequently raised by colleagues in the comments section (as one peer asked: "I just can't figure out how a 5 kV series reactor setup manages to boost the voltage up to over 100 kV").

The answer lies in the Quality Factor (Q):

At resonance, the relationship between the voltage across the specimen (UC) and the power supply voltage (U) is:

In other words, whatever voltage is output by the power source, the voltage across the test specimen is amplified by a factor of Q.

• For a qualified variable-frequency series resonance system, the Q-factor typically falls within the range of 30 to 80.
• With an input power supply of 5 kV (at the secondary side of the excitation transformer) and a Q-factor of 30, the voltage across the test specimen reaches 150 kV.
• The higher the Q-factor, the less strain is placed on the power source; however, the resonance peak becomes sharper and more difficult to locate. Conversely, if the Q-factor is too low, the voltage step-up will be insufficient.

This operates on the same principle as tuning a radio: a radio works by causing an LC circuit to resonate at a specific station's frequency, thereby "amplifying" that frequency signal—the underlying mechanism is essentially identical.

Many veteran engineers, when conducting power-frequency withstand voltage tests (at 50 Hz), would adjust the inductance—typically by switching taps, shifting the iron core, or altering the air gap. This process was both cumbersome and labor-intensive.

Variable-frequency resonance takes the opposite approach: the inductance and capacitance remain fixed (as the test specimen itself is fixed), and the power supply frequency is adjusted to match the resonance point. Typically, the output range of a variable-frequency source is 30 to 300 Hz. The greater the flexibility in frequency adjustment, the better the system's adaptability to test specimens with varying capacitance values. This explains why the control console interface typically displays specifications such as "Input: 0–400 V, 30–300 Hz."

A complete variable-frequency series resonance test system typically consists of five parts:

3.2 270 kV / 108 kVA Configuration Description

Taking a typical 270 kV / 108 kVA Variable Frequency Series Resonant Test System as an example (parameters are subject to recalculation):
Table of Key Parameters

Recalculation Example: 4 sections * 67.5 kV = 270 kV ✓; 4 sections * 0.4 A = ? — Incorrect! When four sections are connected in series, the current remains constant at 0.4 A; therefore, the total capacity = 270 kV * 0.4 A = 108 kVA ✓.

Many people ask: "Why not simply manufacture a single 270 kV reactor section? Wouldn't that be much simpler?"

There are three primary reasons:

1. Insulation Process Complexity: The higher the voltage, the more challenging the coil insulation, external creepage distance design, and oil-paper/SF6 processing become. For a single section, the manufacturing yield drops drastically once the voltage rating exceeds 100 kV.
2. Transportation Difficulties: A single 270 kV reactor section could exceed 4 meters in height and weigh over 2 tons, making it impossible to transport via standard trucks into urban areas.
3. Configuration Flexibility: By dividing the unit into sections, they can be connected in either series or parallel configurations. This allows a single test system to accommodate a wide variety of test objects—a capability that constitutes the "wiring flexibility" we will discuss later.

This constitutes a common and frequently discussed topic among industry peers. In nine out of ten cases, the inability to "locate the resonance point" stems from an error made at this specific stage.

Series connection boosts voltage; parallel connection boosts current (and capacity). If the test specimen has high capacitance, use a parallel configuration; if the test specimen requires high withstand voltage, use a series configuration.

• Total Voltage: 4 * 67.5 = 270 kV
• Total Current: Same as a single section (0.4 A)
• Total Capacity: 270 * 0.4 = 108 kVA
• Total Inductance: 4L₁ (4 times the inductance of a single section)

• On-site AC withstand testing for 110 kV GIS (Test voltage: 1.6Uₘ * √3 / √3 ≈ 184 kV – 218 kV)
• Commissioning AC withstand testing for 110 kV power transformers (80% of factory test value)
• AC withstand testing for 110 kV instrument transformers, surge arresters, and bushings
• All "high-voltage, low-capacitance" equipment within 35 kV / 66 kV systems

• Total Voltage: 2 * 67.5 = 135 kV
• Total Current: 2 * 0.4 = 0.8 A
• Total Capacity: 135 * 0.8 = 108 kVA (Identical to the full-series configuration!)

Key Point: The total capacity remains unchanged; the voltage is simply halved, while the current is doubled. This explains why some commenters asked, "How are the voltage and current calculated for the 2-series, 2-parallel configuration?"—the answer is simply to perform the straightforward vector addition and subtraction as shown above.

• 35 kV medium-length power cables (300 mm² cross-section, approx. 1–2 km in length)
• 35 kV dry-type transformers and oil-immersed transformers
• 35 kV switchgear assemblies (whole-cabinet withstand voltage testing)

┌── L1 ──┐
├── L2 ──┤
Excitation Transformer ───▶ ──┤        ├──▶ Test Object ───▶ Ground
├── L3 ──┤
└── L4 ┘
All 4 reactor sections connected in parallel

• Total Voltage: 67.5 kV (Single-section voltage)
• Total Current: 4 * 0.4 = 1.6 A
• Total Capacity: 67.5 * 1.6 = 108 kVA
• Total Inductance: L₁ / 4 (Inductance reduced to 1/4)

• 10 kV long-distance power cables (300 mm² cross-section, > 2 km in length)
• 10 kV high-capacity generator stator winding withstand voltage testing
• 10 kV high-voltage motor stator withstand voltage testing
• Distribution transformers and distribution cabinets

It is crucial to remember: regardless of how the connection configuration is altered, the total capacity remains constant at 108 kVA; the voltage and current merely shift between a "high-voltage, low-current" state and a "low-voltage, high-current" state. Once this concept is fully grasped, the wiring process will no longer seem mysterious.

This is a question frequently encountered by colleagues in the field. Here is a practical, engineering-oriented estimation procedure:

Reference Values ​​for Typical Capacitance of 10 kV 3-Core Cross-Linked Cables (Per Phase-to-Ground):

Example: For a 10 kV / 300 mm² cable with a length of 2 km, C ≈ 0.32 * 2 = 0.64 μF.

10 kV Cable Withstand Voltage = 17.4 kV (the rationale for this will be explained shortly); the frequency is calculated at 50 Hz (the actual resonant frequency will deviate slightly):
IC = U * 2πf * C = 17,400 * 2π * 50 * 0.64 * 10⁻⁶ ≈ 3.5 A

For a current requirement of 3.5 A, using a 270 kV/108 kVA unit:
• Full-parallel output yields 1.6 A—which is insufficient.
• In other words, for a long cable of this type, a single 270 kV/108 kVA unit is inadequate; a higher-capacity unit (e.g., 270 kV/216 kVA) is required, or one must switch to a unit featuring a "low-voltage, high-current" design (such as a 108 kV/270 kVA model).

Assuming a single inductor section has an inductance of 537 H, four sections connected in parallel result in a total inductance of L = 537/4 ≈ 134 H.
f0 = 1 / (2π√LC) = 1 / (2π√(134 * 0.64 * 10⁻⁶)) ≈ 17 Hz.

At 17 Hz, the frequency falls below the typical industry-standard lower limit of 30 Hz for variable-frequency power sources; consequently, the resonance point cannot be located—this is precisely the issue that has been repeatedly raised in the comments section.

• Shorten the cable segments for testing (e.g., split a 2 km cable into two 1 km sections for separate trials);
• Switch to a reactor with lower inductance (e.g., a specialized low-inductance, high-current reactor designed for 35 kV systems);
• Select a device with a wider frequency modulation range (e.g., a professional, upgraded variable-frequency power source capable of reaching frequencies as low as 20 Hz—a common standard in the industry).

Conclusion: Reactor selection is not a matter of arbitrary guesswork regarding series or parallel configurations. It requires a systematic approach: first, estimate the capacitance; second, calculate the current; and third, verify the resonance frequency. Only when all three steps have been successfully cleared can the chosen wiring configuration be considered correct.

Variable-frequency series resonance withstand voltage testing is applicable to all power equipment that can be modeled as a capacitive load:

• Power cables (common industry practice; applicable to 10 kV – 500 kV systems)
• Power transformers (10 kV – 750 kV)
• GIS, HGIS, and tank-type circuit breakers
• Instrument transformers (voltage and current transformers)
• Surge arresters and bushings
• Large generators and motors (stator windings to ground)
• Complete switchgear assemblies

Inapplicable objects: Purely resistive or inductive loads, and objects with very low capacitance (which are prone to over-compensation and significant resonance point drift).

Some users frequently raise this question. The rules are as follows:
According to the National Standard GB 50150 and cable type designations:
10 kV cable types are typically designated as 8.7/10 kV or 8.7/15 kV. The value to the left of the slash—8.7—is referred to as U0, which represents the rated phase-to-ground voltage; the value to the right represents the rated line-to-line voltage.
Withstand Voltage Testing for New Installations/Commissioning: Test Voltage = 2U0 = 2 * 8.7 = 17.4 kV, held for 60 minutes (note: this is 1 hour, not 1 minute).
Preventive Withstand Voltage Testing: Test Voltage = 1.6U0 = 1.6 * 8.7 = 13.92 kV; the holding duration is determined by specific operational regulations.

A point of contention in the industry forum—the claim that "on-site testing invariably lasts just one minute; I’ve never seen anyone actually perform the full 60 minutes"—highlights a widespread disparity between actual field practices and established standards. While regulations mandate a 60-minute duration (specifically for 10 kV commissioning tests), many field teams, under pressure to meet tight deadlines, cut corners by limiting the test to merely five minutes—or even less. This constitutes a clear violation of protocol; particularly in the case of newly laid cables, the time saved by skipping proper procedures will inevitably have to be repaid later in the form of future equipment failures.

The fundamental difference between a perfunctory "1-minute" on-site test and a rigorous "60-minute" standardized test lies in this: the 1-minute test can only screen out extremely severe insulation defects, whereas the 60-minute test is required to "force out" potential partial discharge sites, water trees, and defects within the semi-conductive shielding layer. Engineers who are truly committed to technical integrity are well aware of this distinction.

This is a critical point specifically highlighted in red text on the wiring diagrams and must be treated with the utmost seriousness.
Reason: A reactor acts as a large coil; any metal located beneath it (such as steel reinforcement bars or metal grating floors) will generate induced eddy currents, effectively acting as a "short-circuited secondary winding." Consequences:

1. The Q-factor drops sharply, the resonance point becomes indistinct, and tuning becomes extremely difficult.
2. The eddy currents cause the underlying metal to heat up; in severe cases, this can result in the metal flooring being scorched or damaged.
3. The reactor itself generates excessive heat due to power losses, leading to insulation breakdown and, ultimately, the burnout of the coil windings.

Correct Procedure: Use an insulating platform (comprising insulators combined with wooden blocks or epoxy resin boards) to elevate the reactor at least 200 mm above the ground. Additionally, ensure that there are no metal objects present within a 1-meter radius directly beneath the reactor.

1. The equipment's external appearance is intact, free from obvious deformation or oil leakage.
2. The grounding cable is securely connected, and the grounding resistance is ≤ 4Ω.
3. The clearance between the high-voltage leads and surrounding metal structures or walls meets the requirements for air insulation (allow a safety margin of 1 cm per kV; for a 270 kV test, the clearance must be no less than 3 meters).
4. Safety barriers, warning signs, and audible/visual alarms are properly positioned and active; unauthorized personnel have been cleared from the test area.
5. Both ends of the cable under test have been disconnected, left suspended (floating), and properly isolated to prevent accidental contact.
6. The voltage divider is wired correctly, and its transformation ratio settings match those configured on the control console.
7. The tap position on the excitation transformer is correctly set to match the requirements of the current test.
8. The emergency stop button is functional, and the overvoltage protection threshold has been configured (typically set at 110% of the intended test voltage).

• Strictly prohibited: Failing to monitor the ammeter while manually adjusting the voltage.
• Strictly prohibited: Entering the safety enclosure/perimeter while testing is in progress.
• Strictly prohibited: Bringing a two-way radio (walkie-talkie) close to the high-voltage zone (as a colleague once asked, "Did you bring a radio too close?"—Yes; strong electromagnetic interference can cause the control system to malfunction).
• Strictly prohibited: Removing the ground wire before the test has concluded.

1. Slowly reduce the voltage until it reaches zero.
2. Switch off the variable-frequency power supply.
3. Use an insulated discharge rod (equipped with a series discharge resistor) to perform a staged discharge via the R-G (Resistor-Ground) path: first make contact with the high-resistance terminal, then make contact with the direct-ground terminal.
4. Ensure a minimum discharge duration of 3 minutes (for long cables, a discharge time of 5–10 minutes is required).
5. Only remove the test cables after the specimen has been directly grounded.

Special Reminder: A client in Guangdong once mentioned, "I spent half a day training a customer, trying to demonstrate the instantaneous voltage fluctuations during a test—and ended up burning out my oscilloscope!"—Large capacitive test specimens can retain residual charges amounting to hundreds of joules. Touching them directly with an oscilloscope probe will inevitably result in the probe being destroyed; you must discharge the specimen using a discharge rod before taking any measurements.

Some engineers, in their haste to achieve results, continue to ramp up the excitation voltage even before the resonance point has been successfully identified. This is extremely dangerous:

• At this stage, the circuit is in a "detuned" state; the current may be very high, yet the voltage fails to rise.
• The excitation transformer and reactor are subjected to severe overloading.
• In severe cases, the reactor coils may burn out, or the excitation transformer may begin to smoke.

The Correct Procedure: Perform a low-voltage frequency sweep to locate the resonance point → Lock the frequency → Then raise the voltage to the required test level. The "Automatic Resonance Point Search" function found in equipment from manufacturers such as Wuhan Guodian Zhongxing is designed specifically for this purpose—simply press the "Auto Test" button, and the control console will automatically perform the frequency sweep to find the resonance point, then proceed to ramp up and hold the test voltage. For manual testing, the operator must manually increase the voltage step-by-step while closely monitoring the current fluctuation curve.

The original motivation for writing this educational article stemmed from observing several recurring questions in the comments section:

• "I can't find the resonance point; I’ve manually adjusted the settings dozens of times, but I still can't locate it."
• "On-site testing usually lasts only 1 minute; I’ve never seen anyone perform a 60-minute test."
• "How do you calculate series and parallel configurations? I’ve been in the field for a long time, but I still can't figure it out."
• "How does 5 kV get stepped up to over 100 kV? No one has ever clearly explained the mechanism."

Underlying these questions is a widespread phenomenon within the industry: many practitioners know only how to press buttons, without understanding the physics and principles operating behind them. When equipment fails, they don't know why it failed; when a test concludes, they don't understand why the result was a pass or a fail; and when standards mandate a 60-minute test duration, they perform only 5 minutes on-site. This "good enough" attitude—while perhaps allowing one to muddle through with systems under 35 kV—becomes a critical liability in vital power grids operating at 110 kV or 220 kV. In such high-voltage environments, the time "saved" through shortcuts will inevitably have to be repaid later—often at the cost of a catastrophic cable explosion or a widespread power outage.

Therefore, through this educational series, we hope to convey two key messages:

1. Clarify the Principles: From RLC series circuits to Q-factor amplification, and from series/parallel combinations to proper wiring selection—all the relevant formulas are provided within this text. You can plug in your own values ​​to verify the calculations; there is no need to rely solely on rote memorization or mnemonics.
2. Clarify the Standards: A 60-minute test duration means exactly 60 minutes; voltage dividers must be connected; ground wires must not be omitted; and testing equipment must not be placed directly on metal flooring. These requirements represent hard-won lessons—paid for by the accidents and sacrifices of generations of power grid engineers—rather than arbitrary rules conjured up on a whim by equipment manufacturers.

The field of power system testing is one where the maxim "better safe than sorry" reigns supreme. Our objective is not merely to "pass the test," but to "uncover potential hidden hazards."

The 270 kV / 108 kVA variable-frequency series resonance test system in your hands is far more than just an assembly of inductors, capacitors, copper, and iron. It serves as the final quality checkpoint before a piece of equipment is commissioned into service—and, indeed, acts as the ultimate guardian safeguarding the integrity and reliability of the power grid.

We hope this article encourages you to pause and think for an extra 30 seconds the next time you are on-site, just before you press that "Start" button.

Referenced Standard: GB 50150-2016, Standard for Handover Test of Electric Equipment in Electrical Installation Engineering. Compiled from field testing practices and manufacturer technical documentation.