Suction-anchored tsunami buoy A3: anchor and mooring design basis
Ø4.0 × 8.0 m steel suction caisson and inverse-catenary mooring in 3 850 m of water, Andaman Sea · addendum for the Mw 7.1 event of 22 Sep 2027
Project
Tsunami buoy A3 · suction anchor · TEWS-A3
Investigation
1 · Anchor and mooring of buoy A3
Report
v2 · draft · 22 Sep 2027
Template
TEWS design basis report · r1
Prepared with
Elidyne · records from Runs 1, 2, 3, 4, 5, 6, 7, 8
Prepared by · reviewers
K. Venkat · A. Rao pending
Status: Draft · not released · Event addendum to the design basis. v1, released at G4 on 14 Jan 2027, remains the current release.
AI-assisted engineering output. Results need review by the responsible geotechnical and offshore engineer before use; this is not a certified design.
1SummaryDrafted by Elidyne · review pending
On 22 Sep 2027 the Mw 7.1 Andaman Sea earthquake, 212 km from A3, was detected by BPR-3 in event mode, with a tsunami arrival of at the buoy, and the anchor held: the measured peak tension was and the post-event factor of safety is .With the Ø4.0 × 8.0 m caisson, adopted at G3 on branch v3, the governing case is LC3, the Mw 9 megathrust followed by the 100-year cyclone on clay softened by , at a factor of safety of ; the Ø3.5 × 7.0 m caisson reached in the same case, below the owner’s criterion of 2.0.With the Ø3.5 × 7.0 m caisson of plan v2, LC3 governs at a factor of safety of , below the owner’s criterion of 2.0.The mooring does not depend on the caisson size: the peak line tension is , and the caisson tilt after LC3 falls from to is is , inside the 2.0° alarm band.This is an engineering design basis, not a certified design: the turbidity velocity is an assumption (T-6), the clay update proposed after the event waits for a gate, and the open items in Section 9 are carried as stated.No Run with anchor checks is in scope. Add Run 3 or Run 5 to report results; the limitations in Section 9 still apply.
This edited paragraph cites Runs 3, 5, 7 and 8, and one is now outside the chosen scope. Restore the scope, or ask Elidyne to re-assemble it.
Finding T-2
Ø4.0 × 8.0 m caisson: FoS 2.75 after the Mw 9 then the 100-year cyclone; Ø3.5 × 7.0 m reached 1.74, below 2.0. Accepted by A. Rao at G3 on 2 Nov 2026 from Runs 3 and 5; steel 18 t against 14 t [10].
Finding T-4
The 22 Sep 2027 event was detected and the anchor held: 96 kN measured against 90 ± 12 kN predicted, tilt 0.2° to 0.5°. Proposed from Runs 7 and 8; decided at G4 with this report [19].
2Site and station
Station A3 sits in the Andaman Sea, 212 km west of the Sunda trench, in 3 850 m of water. The seabed is soft pelagic clay, su = 3 + 1.4 z kPa from core GC-7, sensitivity 3 [7]. A 2.6 m discus buoy (4.2 t) carries the satellite and acoustic links; bottom pressure recorder BPR-3 reads to 1 mm, every 15 min normally and every 15 s in event mode. The anchor is a steel suction caisson, Ø4.0 × 8.0 m and 18 t, with the padeye 5.2 m below the mudline [8]. The station model has 146 elements [18].
Input
Version · record
Status
Core GC-7, vane su
v1 [7]
Currentmean for capacity, lower bound for installation (G1)
Caisson drawing TEWS-A3-S-02
v1 [8]
CurrentØ4.0 × 8.0 m, padeye 5.2 m
Polyester certificate
v1 [8]
As provided · MBL 1 450 kN
Cyclone hindcast, 40 years · seismic catalogue
v1 [9]
Currentbasis of Study 2
Deployment log
Mar 2027 [20]
Currentpenetration 7.95 m at 36 kPa
Mooring, top to bottom, as an inverse catenary: chain 120 m, polyester 4 300 m, buoyancy 3.2 t, chain 60 m to the padeye.
AI-assisted engineering output · not certifiedPage 1 of 7
The buoy has to survive the events it exists to detect, in any order. Study 2 set three hazards and five load cases; G2 accepted them on 21 Oct 2026 with the turbidity velocity recorded as an assumption [17].
Hazard
Basis
Design values
Status
Cyclone · 100-year
40-year hindcast, GPD fit
Hs 9.4 m · Tp 13.8 s · wind 38 m/s · surface current 1.6 m/s
Accepted
Earthquake · Mw 9.0 megathrust
Regional catalogue, source at 180 km; IS 1893 spectrum shape
Seabed PGA 0.28 g · 15 equivalent cycles · 96 s above 0.05 g · su −18 % after shaking
Accepted
Turbidity current
No site record; assumed
2.5 m/s · caisson drag 14 kN · chain 6 kN
Assumption · T-6 open
Load cases
LC1cyclone100-year cyclone on intact clay.Run 5
LC2quakeMw 9.0 shaking alone, with the 1-year sea.Run 5
LC3quake → cycloneThe cyclone arrives before the softened clay recovers. Governs.Runs 3, 5
LC4turbidityTurbidity current 2.5 m/s with the 1-year sea; 3.5 m/s as a sensitivity.Run 5
LC5installationSuction installation on the lower-bound su.Run 4
4Anchor capacityLC3 governs
Two caisson sizes were carried to a decision. Every change is a new checked plan version; the adopted size came from a branch.
Figure 1. Plan lineage, v1 to v4, and the event Runs. The branch leaves v2 and is adopted at G3; v2 is kept, not merged [13].
v112 Oct 10:20Proposed from the question: six studies, gates G1–G4; site and station model, hazard loads.Runs 1–2
v221 Oct 11:14Anchor capacity on Ø3.5 × 7.0 m; installation on the lower-bound su, as G1 asked.Runs 3–4
v32 Nov 09:16Branch Ø4.0 × 8.0 m after LC3 reached FoS 1.74; adopted at G3, 2 Nov 17:40.Run 5
v414 Jan 10:05Design basis and report on the adopted caisson; released at G4, 14 Jan 2027 10:30.Run 6
AI-assisted engineering output · not certifiedPage 2 of 7
Main line and adopted branch side by side. A caisson check passes at FoS ≥ 2.0, the owner’s criterion; a line or limit check passes at a ratio ≤ 1.00.
Capacities at 32° load inclination at the padeye. Not checked: tilt is watched against an operations band, not a design limit. Waits for gate: a proposed input, not yet adopted. Sources are records in Section 11.
Ø4.0 × 8.0 m · Run 5 · adoptedØ3.5 × 7.0 m · Run 3 · main v2Required: FoS 2.0 on 185 kNBelow the requirement
Figure 2. Anchor capacity after the Mw 9, Ø3.5 × 7.0 m vs Ø4.0 × 8.0 m. Capacity against load inclination at the padeye on clay softened by 18 %, with the LC3 peak demand of 185 kN at the 32° design inclination. Data: 2 × 91 points [12].
Intact capacity of the adopted caisson is 620 kN; after the Mw 9 it is 508 kN. A third size, Ø4.5 × 9.0 m (24 t), reached FoS 4.40 intact and 3.61 after the quake and was set aside as over-designed [1].
AI-assisted engineering output · not certifiedPage 3 of 7
Lumped-mass dynamics of the 146-element station model, 20 seeds of the 3-hour storm for LC1 and LC3 [3].
Segment
Length
MBL · kN
Peak tension · kN
Utilisation
Top chain
120 m
1 423
185
0.13
Polyester
4 300 m
1 450
174
0.12
Buoyancy
3.2 t net
—
—
—
Bottom chain to the padeye
60 m
1 423
185
0.13
The buoy heels at most 31° against the 45° limit and does not capsize in any of the 20 seeds. The top-chain fatigue life is 68 years [3].
Installation, LC5: on the lower-bound su, the Ø4.0 × 8.0 m caisson needs 38 kPa of suction against 92 kPa allowable for plug heave, a margin of 2.4. At deployment on 18 Mar 2027 it reached 7.95 m at a measured 36 kPa [6][20].
6Calamity sequencesØ4.0 × 8.0 m
Each load case on the adopted caisson, and the event of 22 Sep 2027 as measured. A case passes at FoS ≥ 2.0 [1].
Case
Demand · kN
Capacity · kN
FoS
Status
LC1 · 100-year cyclone, intact
185
620
3.35
Passes
LC2 · Mw 9.0 alone
124
508
4.10
Passes
LC3 · Mw 9.0, then the cyclone
185
508
2.75
Passes · governstilt 1.1°, inside the 2.0° alarm
LC4 · turbidity current, 3.5 m/s sensitivity
—
—
2.20
Passes · caveatedlateral; velocity assumed (T-6)
LC5 · installation, suction in kPa
38
92
2.4
Passesplug heave, lower-bound su
22 Sep 2027 · Mw 7.1 at 212 km, measured
96
596
3.2
Held · caveatedclay update waits for a gate
The event of 22 Sep 2027 · Runs 7 and 8
Time
Signal
Predicted
Measured
04:12:08
Mw 7.1 Andaman Sea, depth 28 km
—
212 km from A3
04:12:21
Seabed acceleration; BPR-3 to event mode, 15 s
trigger 0.15 g
0.21 g
04:14:02
Mooring top tension, peak
90 ± 12 kN
96 kN
04:14
Caisson tilt
band 1.0° / 2.0°
0.2° → 0.5°
04:31:40
Tsunami arrival at A3
4.2 cm
4.6 cm
04:33:10
Warning to the national warning-centre feed
—
90 s after arrival
04:36
Post-event capacity, anchor twin
—
596 kN · FoS 3.2
AI-assisted engineering output · not certifiedPage 4 of 7
The turbidity current velocity, 2.5 m/s, has no site record and is carried as an assumption (Finding T-6). At 3.5 m/s the lateral factor of safety falls to 2.20. One of the six inputs open at intake remains open.
IN-01 Clay strength · resolved: su = 3 + 1.4 z kPa from GC-7.
IN-02 Installation strength · resolved: lower bound, at G1.
IN-03 Cyclone basis · resolved: 40-year hindcast, GPD fit.
IN-04 Seismic scenario · resolved: Mw 9.0 at 180 km.
Run 8 proposes a clay update, su −6 % near the caisson tip, from the event. It waits for a gate; the capacities in Sections 4 and 6 use GC-7 unchanged, and the post-event FoS 3.2 is reported with the update applied [14].
Caveat
The arrival forecast from the source inversion, 4.2 cm, was 9 % below the 4.6 cm measured at BPR-3 [15].
Not run
Seabed slope stability around A3 is not assessed: no bathymetry survey is attached. Nothing is inferred.
Assumed · 6
Six engineering assumptions are stated in Section 8 with their effect and the verification each needs.
Caveat
Tilt is reported against the operations band, 1.0° advisory and 2.0° alarm, not against a design limit; the owner has set none.
AI-assisted design statement
Cat.
Basis
In this report
A
User-provided project data
Site, station, caisson and mooring particulars; FoS ≥ 2.0 criterion
10Discussion logIntros drafted by Elidyne · review pending
Selected exchanges from the investigation thread, quoted as recorded and shortened where marked […]. Answers cite the same records as the rest of this report; nothing in this section adds a conclusion.
14 Oct 2026 11:40 · at G1 on v1Why installation uses the lower bound.
M. Krishnan requested changes Use the lower-bound su for installation, not the mean.
Elidyne answered(reviewed by K. Venkat) Installation will use the lower-bound su from GC-7, which gives the higher suction and the tighter plug-heave margin; capacity keeps the mean. Nothing already run changes; the next plan version carries it.
2 Nov 2026 09:15 · from v2The larger caisson.
K. Venkat asked Try Ø4.0 × 8.0 m and re-run the sequences.
Elidyne answered(reviewed by A. Rao) Branch Ø4.0 × 8.0 m from v2, as plan v3. The site, hazard and mooring steps carry over; the size-dependent steps re-run, +2.1 CU. In Run 5, LC3 rises to from .
2 Nov 2026 15:30 · before G3Why the quake then the cyclone governs.
K. Venkat asked Why does the quake followed by the cyclone govern, rather than the cyclone alone?
Elidyne answered(reviewed by A. Rao) The Mw 9 shakes the clay for 15 equivalent cycles and softens it by . If the cyclone arrives before the clay recovers, the same 185 kN demand meets of capacity instead of […].
22 Sep 2027 04:36 · during the eventWhether the anchor held.
K. Venkat asked Did the anchor hold, and does the design basis still stand?
Elidyne answered(reviewed by K. Venkat) Yes. The twin predicted 90 ± 12 kN; the station measured , and tilt moved from 0.2° to 0.5°, inside the band. Run 8 gives a post-event factor of safety of with su −6 % near the tip; that update waits for a gate.
21 Oct 2026 · open Finding, discussion onlyNot a conclusion of this report.
Open Finding T-6(not accepted) The turbidity current velocity, 2.5 m/s, has no site record. At 3.5 m/s the lateral factor of safety is 2.20, still above 2.0; a current record at A3 would close it. Reported here only as an open question.
No turns from the chosen parts have exchanges selected for the log.
11References and provenanceAlways included
Every number in this report resolves to one of these records; full digests are in the exported appendix.
[1]rec 6b2e…91c4 · Run 5 step 9 · caisson capacity, LC1–LC5, Ø4.0 × 8.0 m and Ø4.5 × 9.0 m · geo.caisson.capacity r1 · DNV-RP-E303
[2]rec d480…1f7a · Run 3 step 7 · caisson capacity, LC1–LC3, Ø3.5 × 7.0 m · geo.caisson.capacity r1
[3]rec 3c9a…e06b · Run 5 step 6 · mooring dynamics, 20 seeds, LC1 and LC3 · moor.dynamics r1 · API RP 2SK
[4]rec 91f0…2ad8 · Run 3 step 8 · caisson tilt after LC3, Ø3.5 × 7.0 m
[5]rec 5e37…b412 · Run 5 step 10 · caisson tilt after LC3, Ø4.0 × 8.0 m
[6]rec a0c4…7e59 · Run 4 step 3 · installation suction and plug heave, both sizes, lower-bound su · geo.caisson.install r1
[7]ev 2f61…c03d · core GC-7, vane su · sha256 8a0e…42d1