The coupled model can be used to compare generic hull roll with and without the connected tank.
The results are screening estimates; no result represents a real hull.
The implications of the model's stated moment fit are listed in Section 7.
At an 18 s hull roll period, alpha 0.05, and conduit area A_c = 6.8 m²,
the peak roll reduction screening estimates are 46.1 % at 2 % hull damping and 23.3 % at 5 %.
See Figure 5.1.
At 60 s forcing and 18 s hull period with 2 % hull damping, there is 6.0 % more roll at alpha 0.05
and 29.5 % at alpha 0.20, compared with the 25.0 % slow-forcing limit at alpha 0.20, from 1/(1−α).
See Section 5.2.
The moment fit has 4.3 % RMS error for forcing periods of 20 s and longer; it degrades below about 16 s for a reason not established.
See Section 6.
The reference tank's natural period is 14.3 s [I4]; the 18 s sweep starts at 12.6 s.
Every sampled forcing period below 20 s lies outside the model's stated moment-fit band (≥20 s), including parts of the 21 s sweep.
The 46.1 % peak at 18.5 s lies outside it.
See Appendix A.
Figure 1.1 – Peak roll reduction against alpha at an 18 s hull roll period, 2 % and 5 % hull damping, with the alpha 0.05 results marked.
Figure 1.2 – The connected anti-roll tank in a generic hull, three-dimensional view (detail in Figure 3.2).
2 Introduction
The generic hull roll comparison describes the effect of a connected anti-roll tank for engineers considering the simplified coupled model.
The scope is a normalized hull and the public reference tank, with no actual hull mass, inertia or restoring.
These screening estimates exclude sea-state weighting, structural loads and design qualification.
No full-scale roll reduction has been calculated here.
3 Design data
The calculation uses public reference tank inputs (ASSUMED - to be confirmed), with no actual hull geometry.
The input values and their units are listed in Table 3.1.
3.1 Site or field data
Not applicable: no site or field is represented.
3.2 Structure or equipment data
Each leg measures 20.0 m by 6.0 m (120 m² in plan). Fill depth is 5.0 m; centroid offset is 5.0 m.
Conduit area A_c is 6.8 m². The model supplies effective length, fluid properties and calibrated loss parameters.
In the loss-coefficient unit, n is the loss area exponent in Table 3.1.
Calibration transfer means using the model's calibrated loss parameters in this reference-tank and generic-hull calculation: ASSUMED - to be confirmed.
These screening estimates use the public reference tank, which differs from the illustrative tank on the linked study page [I3].
The transverse leg widths, fill depths and centroid offsets in Figure 3.1 use the dimensions in Table 3.1.
Tank dimensions are recorded assumptions; the hull outline supplies no hull dimensions.
Figure 3.1 – Connected-tank transverse section.
Figure 3.2 shows the same recorded tank dimensions in three dimensions.
Figure 3.2 – Connected anti-roll tank, three-dimensional view: the two legs, the fill, the cross-connecting conduit and the generic hull section around them.
3.3 Environmental data
No sea state is specified. Forcing stays fixed within each tank-on and tank-off comparison and is rescaled between damping groups.
3.4 Vessel data
Generic hull periods are 18 s and 21 s; hull damping is 2 % and 5 % of critical, and alpha is 0.02, 0.05, 0.10, 0.15 and 0.20.
All hull values are ASSUMED; no real hull data is used. The full load-case matrix is in Appendix A.
3.5 Criteria and sources
No design acceptance criterion is specified. The arithmetic checks and the model's stated moment-fit band are distinguished in Section 6.
Their sources are [I4] and [R1].
3.6 Load cases
Every combination of the assumed hull periods, damping and alpha values is retained in Table A.1.
The response sweep covers 0.7 to 1.3 times each hull period; the additional long-period comparison uses 60 s forcing [I1].
4 Analysis methodology
The model links hull roll to water transfer between connected tank legs. Tank moments feed back into the hull roll equation.
Alpha is the tank's quasi-static free-surface moment stiffness divided by the hull's roll restoring. Alpha is dimensionless.
The public U-tube model is used; a software version is not declared. Its exact source is identified in [I2].
Kₛ is the quasi-static tank moment per roll angle.
The public model includes vertical-centroid and local free-surface terms in Kₛ.
Kₛ also includes water redistribution between the legs. Alpha therefore follows the complete public moment convention.
(4.1) The normalized bare-hull impedance is D = 1 − (T₀/T)² + 2iζ(T₀/T).
(4.2) The coupled impedance is D − αH/Kₛ. H is tank moment per roll angle; Kₛ is its quasi-static limit.
T₀ is the hull roll period (s), T is the forcing period (s), ζ is the hull damping ratio (1), and i is the imaginary unit.
H and Kₛ use the same moment-per-angle units, so both impedances and alpha are dimensionless [I2].
The hull equation uses normalized restoring of 1, with no units. The model sets forcing to give 5° bare roll at the nominal resonance
(ASSUMED - to be confirmed). Forcing stays fixed within each tank-on and tank-off comparison.
Forcing is rescaled between damping groups to keep the 5 degree bare-roll reference, so the comparison also changes the tank amplitude.
The calculation finds each response peak separately over 0.7 to 1.3 times the hull roll period. Reduction equals 100 % times one minus the coupled peak divided by the bare peak.
The input basis and calculation are recorded in [I1] and [I4].
5 Results
All charts select the existing 18 s hull roll period and conduit area 6.8 m².
The wave-period chart uses 2 % hull damping and alpha 0.05. Alpha is normalized stiffness as defined in Section 4.
The calculated tank natural period is 14.3 s [I4].
The reference geometry is retained rather than tuned to the 18 s hull; the tank and hull natural periods therefore differ.
5.1 Peak roll reduction
Figure 5.1 shows peak reduction against hull damping;
Figure 5.2 shows the dimensionless roll amplification factor against forcing period;
Figure 5.3 shows peak reduction against alpha.
The full sensitivity matrix and damping comparison are in Appendix A.
The amplification factor multiplies the normalized wave moment to give roll in radians.
The bare factor is 1/(2ζ) at the nominal hull resonance; a normalized moment of 2ζ times 5° in radians gives the 5° bare roll used here.
Figure 5.1 – Peak roll reduction against hull damping.
Slow water transfer reduces the effective hull restoring. In the slow-forcing limit, the roll ratio approaches 1/(1−α).
At alpha 0.05, that limit gives 5.3 % more roll.
At 60 s forcing and 18 s hull period with 2 % hull damping, there is 6.0 % more roll at alpha 0.05
and 29.5 % at alpha 0.20, compared with the 25.0 % slow-forcing limit at alpha 0.20, from 1/(1−α).
The slow-forcing limit understates the penalty at 60 s. A near-resonance reduction does not remove this penalty.
The comparison is recorded in [I1].
6 Validation and benchmark
The model's stated moment fit and the mathematical hand checks address different quantities.
The moment fit has 4.3 % RMS error for forcing periods of 20 s and longer; it degrades below about 16 s for a reason not established [R1].
No tank-on and tank-off real-vessel comparison establishes these reductions. No full-scale roll reduction has been calculated here.
Alpha = 0 gives zero reduction, or 0 %.
An undamped box hand check copies the model's own moment formula, so it checks arithmetic and the coupling denominator,
not the physics of the moment model. It also checks both natural frequencies [I5].
For that mathematical check, each leg measures 2 m by 1 m, with 1 m depth and 1 m offset.
The conduit has 1 m² area and 1 m effective length. Fluid density is 1 kg/m³; gravity is 1 m/s².
The hand calculation gives Kₛ = 53/6 N m/rad and H = 359/24 N m/rad at 0.5 rad/s.
At alpha 0.1, the response is 2120/1231 times normalized forcing, with no units [I5].
7 Limitations
The model's stated moment fit does not validate coupled hull reduction.
The reductions are screening estimates for generic hulls, with no real-hull evidence; the missing comparison is described in Section 6.
The reference tank's natural period is 14.3 s [I4]; the 18 s sweep starts at 12.6 s.
Every sampled forcing period below 20 s lies outside the model's stated moment-fit band, including parts of the 21 s sweep; see Appendix A notes.
The damping comparisons rescale forcing between damping groups and therefore change tank amplitude; see Appendix A damping comparison.
Calibration transfer and the reference tank dimensions remain assumed; their basis is in Table 3.1.
These screening estimates exclude sea-state weighting, structural loads and design qualification; the assessed scope is in Section 2.
A near-resonance reduction does not remove the long-period penalty; see Section 5.2.
8 Conclusions and next steps
The peak reduction depends on the assumed hull damping and alpha; the separate peak comparisons are shown in Section 5 and Appendix A.
The near-resonance reduction does not remove the long-period penalty in Section 5.2.
Calibration transfer and the bare-roll reference remain ASSUMED.
The missing real-vessel tank-on and tank-off comparison in Section 6 remains the evidence gap for coupled hull reduction.
9 References
The public model basis and the calculation records identify the inputs, results and mathematical checks used in this report.
9.1 Public references
No external publication is used. The moment-fit statement below comes from the model documentation.
[R1] The model's stated moment fit: 4.3 % RMS error for forcing periods of 20 s and longer; degradation below about 16 s has no established cause.
The statement is retained in the model reference [I2]; no independent coupled-hull validation is cited.
[I4] Calculation and figure builder: scripts/capabilities/build_generic_hull_screening.py.
Includes the tank natural-period calculation and chart selections. Reproduction commands and numerical-output conventions are in the builder documentation.
The full matrix retains every calculated hull-period, damping and alpha combination.
Table A.1 gives the peak reduction and the coupled peak forcing period [I1].
Hull roll period (s)
Hull damping (% of critical)
Alpha (1)
Peak reduction (%)
Coupled peak forcing period (s)
18
2
0.02
24.3
18.1
18
2
0.05
46.1
18.5
18
2
0.10
61.4
19.3
18
2
0.15
66.0
20.4
18
2
0.20
67.2
21.4
18
5
0.02
10.5
18.2
18
5
0.05
23.3
18.4
18
5
0.10
37.9
19.1
18
5
0.15
45.6
20.1
18
5
0.20
48.7
21.3
21
2
0.02
20.7
21.2
21
2
0.05
37.8
21.6
21
2
0.10
50.1
22.5
21
2
0.15
55.0
23.4
21
2
0.20
56.9
24.4
21
5
0.02
9.0
21.3
21
5
0.05
19.2
21.6
21
5
0.10
29.8
22.4
21
5
0.15
35.2
23.4
21
5
0.20
37.7
24.5
Table A.1 – Peak reduction sensitivity matrix.
Figure A.1 – Peak roll reduction against alpha, all twenty cases of Table A.1.
In this case (18 s, alpha 0.05), hull damping roughly halves the benefit (ratio 0.505).
The ratio of the 5 % benefit to the 2 % benefit ranges from 0.43 to 0.72 across the matrix.
For alpha 0.02, 0.05, 0.10, 0.15, 0.20 respectively, the ratios are 18 s: 0.43, 0.505, 0.62, 0.69, 0.72; 21 s: 0.43, 0.51, 0.59, 0.64, 0.66.
These comparisons rescale forcing between damping groups and therefore change the tank amplitude [I1].
The wave-period chart shades sampled forcing periods below 20 s, outside the model's stated moment-fit band.
The 18 s sweep starts at 12.6 s; the band also excludes parts of the 21 s sweep [I1], [R1].
All charts select existing calculated samples without interpolation or new cases.
CSV contains each plotted series point; hover or focus points for values and use checkboxes to toggle series.
The static SVG retains axes, units, legend and plotted samples for printing [I4].
Reference tank geometry and calibration transfer are assumed; no actual hull geometry is supplied.
Hull outline and conduit height are symbolic; widths, depths and centroid offsets use the recorded dimensions [I1].
Appendix B Review status
Calculation records, source links and reproduction instructions are identified in [I4].
AI generated the script, calculations, tests and draft text. Engineering checks and acceptance remain pending.
P.E. sign-off remains pending; no engineer signature is asserted.
Appendix C Revision history
Revision
Date
Description
Prepared / checked
Initial
2026-10-24
Generic hull screening estimate added to distinguish model predictions from real-hull evidence.
AI-drafted; engineering check pending
Clarification
2026-10-25
Damping basis, assumptions and validity limits clarified; numeric output stabilized and unpublished comparison removed.
AI-drafted; engineering check pending
Interactive summary
2026-10-25
Input-driven tank section, interactive charts and matching CSV exports added; calculated values retained and limits placed before method.
AI-drafted; engineering check pending
Report layout
2026-10-25
Engineering reporting layout applied to presentation: numbered sections, input table, dimensioned section, figure captions and appendices. Calculated values, claims and conclusions retained.
AI-drafted; engineering check pending
Summary and chart clarification
2026-10-25
Earlier headline clarification recorded: peak roll reduction estimates refer to an 18 s hull roll period, with separate 2 % and 5 % damping cases. Moment-fit provenance and extrapolation, slow-forcing limit, tuning basis, symbols and chart scales clarified; calculation data retained.