Condensator Dominit – Filtering and Resonance Damping Restore Voltage Quality Aboard a Cruise Ship
IPQDF gratefully acknowledges Condensator Dominit GmbH (Brilon, Germany) for making this field data available to the engineering community. The measurements and filter design described below are drawn from the manufacturer’s case study on a cruise ship installation; the analysis, standards context, and closing perspective are the author’s own.[1]
01Operating Context: A Ship Is a Weak, Self-Contained Grid
A cruise ship at sea is its own island grid: no utility tie, no infinite bus to lean on, just onboard generators sized to the vessel’s own load. That single fact changes the electrical character of the network more than anything else about the application. Generator sets have limited capacity and comparatively high internal (source) impedance next to a public transmission-fed industrial supply, which means the ship’s short-circuit power — its ability to hold voltage steady against a disturbing load — is inherently low.[1]
Layered onto that weak source is a load population any shore-based PQ engineer would recognize immediately: variable frequency drives on thrusters and propulsion auxiliaries, LED lighting throughout the vessel, and battery chargers, all switch-mode, all nonlinear. On a stiff industrial grid this load mix is manageable. On a weak generator-fed grid, as this case shows, the same loads interact with the network’s own impedance in ways that do not show up onshore.
02The Problem: Flat-Topping, Distorted Zero-Crossings, and Resonance
Pre-installation monitoring documented three distinct symptoms, not one:[1]
- Flat-topping: for a nominal 690 V phase-to-phase supply, the ideal sinusoidal peak is 975 V. The measured peak was 943 V — a peak deficit of roughly 3.3%, enough to affect DC-link recharging in connected frequency converters.
- Distorted, double zero-crossings: the waveform crossed zero more than once per expected transition, a pattern that can cause protective and control systems relying on zero-crossing detection to misfire.
- Overlapping high-frequency voltage distortion: broadband disturbance superimposed on the fundamental, stressing connected network components.
A scoping caveat worth stating plainly: the vendor’s documentation benchmarks the measured spectrum against IEC 61000-2-4, EMC Class 2 — a standard written for industrial plants, not marine vessels. It is a reasonable reference point for judging the harmonic content itself, but it is not the standard that actually governs a ship’s electrical system; that role belongs to classification-society rules (ABS, DNV, Lloyd’s Register) and marine EMC standards such as IEC 60092-101/60533. Readers should treat the IEC 61000-2-4 comparison here as a benchmarking convenience, not a marine compliance claim.[2]
03Why Weak Generator Grids Are Resonance-Prone
The three symptoms above are not independent faults — they are the signature of a single underlying condition: parallel resonance excited by a weak source. Any capacitive element on the network — and ships carry plenty, from EMC filter capacitors to cable capacitance — forms a parallel LC circuit with the source inductance behind it. The resonant frequency of that circuit falls as source inductance rises, which is exactly what happens when short-circuit power is low. On a stiff industrial bus, the same capacitance typically resonates at a frequency high enough to be inconsequential. On a weak generator bus, that resonant peak can land squarely on one of the characteristic harmonics the VFDs and chargers are already producing — and get amplified rather than absorbed.[3]
This also explains the double zero-crossing directly. Near the fundamental’s true zero-crossing, the instantaneous magnitude of the 50/60 Hz component is small, so even a modest superimposed high-frequency oscillation is large by comparison and can drive the waveform through zero multiple times in quick succession. Near the peak, the same absolute ripple is negligible next to the much larger fundamental. The distortion is therefore most visible exactly where zero-crossing-dependent equipment is watching for it.
The practical consequence for a naval architect or ship’s electrical engineer: a filter sized purely to cancel known harmonic orders is necessary but not sufficient on a network like this. Resonance amplification is a network-configuration problem — it shifts as generators come on- and off-line and as breaker lineups change — and a fixed, harmonic-order-tuned filter alone will not chase it.
04Solution: Tuned Harmonic Filtering Plus Resonance Damping
The installation used two functionally distinct module types, matched to the two distinct problems identified above:
- Two SΦIA® modules, H5 and H7 versions — voltage-controlled harmonic filters with intelligent adaptation, targeting the 5th and 7th harmonics specifically. These orders are the expected signature of the six-pulse rectifier front ends common to VFDs and battery chargers.
- One RΕSI-Mod (Resonance Elimination System) — not tuned to a harmonic order at all, but designed to damp the resonant amplification described in Section 03, adapting passively as network conditions change.[1]
| Location | Onboard, cruise ship |
|---|---|
| Installation date | 2024 |
| Nominal supply | 690 V phase-to-phase; ideal peak 975 V |
| Dominant disturbances | Flat-topping, distorted/double zero-crossings, high-frequency resonance |
| Root cause | Low short-circuit power (weak generator-fed island grid) interacting with nonlinear VFD/LED/charger loads |
| Mitigation | 2× SΦIA® mod (H5 + H7, tuned) + 1× RΕSI-Mod (broadband resonance damping) |
The design pairs a tuned element for known harmonic orders with a broadband element for everything else — a common pattern in harmonic mitigation generally — but the broadband element here is doing a categorically different job than a typical broadband harmonic filter. It is not mopping up residual harmonic content; it is damping a resonance that would otherwise amplify whatever harmonic energy happens to coincide with it, under network conditions that are themselves not fixed.
05Results: Peak Recovered, Zero-Crossings Cleaned, Resonance Damped
The voltage peak recovered from 943 V to 969 V against an ideal of 975 V — closing most of the pre-installation deficit. Zero-crossings became clean and single-valued, and the high-frequency resonant distortion was eliminated.[1] The manufacturer’s THDV trend recording shows total harmonic distortion of voltage stepping down in two stages as the modules were brought online — first with the SΦIA filters, then further with RΕSI added — from roughly 6% down to approximately 2–2.5%. As with the brewery case, these are values read from a trend chart rather than a tabulated report and should be treated as approximate.
Fig. 1. Measured voltage peak before and after installation, against the theoretical ideal peak for a 690 V phase-to-phase supply.[1]
Fig. 2. Approximate THDV as filter modules were brought online in stages. Values read from the manufacturer’s trend recording; treat as indicative rather than exact.[1]
Fig. 3. Illustrative reconstruction of how a small high-frequency ripple superimposed on the fundamental produces a double zero-crossing near the waveform’s true zero, while remaining negligible near the peak. This is a stylized illustration of the reported mechanism — not the original recorder trace.
Net result: voltage peak restored to within roughly 0.6% of the theoretical ideal (from a 3.3% deficit), clean zero-crossings, and elimination of the resonance-driven high-frequency distortion — achieved by pairing harmonic-order-specific filtering with a separate resonance-damping stage rather than relying on either alone.
06The Power Quality Perspective: What This Case Study Illustrates
This case is really a weak-network problem wearing a marine disguise. A ship’s generator-fed bus is an extreme, self-contained version of the same weak-network condition that shows up onshore wherever a facility is fed from a small, high-impedance source — a remote feeder, an islanded microgrid, a site running on backup generation during an outage. Article 4 of this series works through exactly this contrast on a fixed industrial network, comparing a 6-pulse drive’s harmonic and unbalance performance under weak (100 MVA, X/R = 1) versus strong (500 MVA, X/R = 3) source conditions. The cruise ship sits at the far end of that same spectrum: a source so limited that resonance, not just harmonic magnitude, becomes the dominant concern.
The broader lesson for anyone specifying mitigation on a weak or islanded network: know which problem you’re actually solving. Harmonic-order-tuned filtering answers a static question — which frequencies does this load produce. Resonance damping answers a dynamic one — what does the network do to whatever frequencies are present, under conditions that keep changing. Treating the second question as if it were the first is a common and costly mistake on any weak-source system, marine or onshore.
References
- Condensator Dominit GmbH, “Case Study: Improving Power Quality on Cruise Ships,” Brilon, Germany, 2024. Available: condensator-dominit.de.
- IEC 61000-2-4:2002, “Electromagnetic compatibility (EMC) – Part 2-4: Environment – Compatibility levels in industrial plants for low-frequency conducted disturbances.” Cited by the source document for benchmarking purposes only; not a marine governing standard.
- IEEE Std 519-2022, “IEEE Standard for Harmonic Control in Electric Power Systems” — background on source impedance, resonance, and harmonic interaction.