The Lab: hydrostatics, stability, motion and the sea
The Lab computes the hydrostatics and stability of the hull in a ship scene, and records the ship's motion as a file you can analyze. It is for study and education. It is not a certified design or approval tool, it is not a stability assessment, a class or flag-state approval, or a substitute for a qualified naval architect or structural engineer, and its results may not be relied on for the design, operation or safety of any real vessel, structure or system (see the Terms of Service). The Lab never produces stamped, certified or approved results.
**Open it**
Build a ship scene at /worlds, then choose **Lab** in the toolbar at the top left of the viewer. The Lab is laid out for a desktop screen.
**What it computes**
At the draft the scene states: displacement, volume, the center of buoyancy (KB and LCB), the waterplane area and center of flotation, the metacentric radii (BMt, BMl), the transverse metacenter KMt, and the block, waterplane and midship coefficients. It also computes the cross curve KN from 0 to 60 degrees of heel at constant displacement.
**KG: the number the hull cannot tell you**
GM and the righting arm GZ depend on the ship's center of gravity above the keel (KG), which comes from how the ship is loaded, not from its shape. The Lab never assumes one. Until you enter a KG it shows GM for a range of KG values (GM = KMt - KG) and the KN curve; GZ for any KG is KN - KG x sin(heel). Enter a KG, and the source you have for it, to see GM, the GZ curve and the general intact criteria as indicators. Each criterion shows the computed value beside its threshold, cited to the IMO 2008 Intact Stability Code, with "at or above" or "below": an indicator, not a pass or a compliance finding.
**Tiers: VISUAL and REFERENCE METHOD**
Every result is labeled. A ship's own hull in Worlds is drawn from its length, beam, draft and an estimate of its form, not from its lines plan, so its results are **VISUAL** however carefully they are computed, and the Lab lists why. The benchmark hulls (a box barge, the Wigley hull, a floating cylinder) are exact shapes: with a KG you enter, their results are **REFERENCE METHOD**: a published method, checked against known answers, applied to inputs that are facts. You can compare them with the textbook formulas. Reference method describes how a number was computed; it is not a professional review of anything.
**Exports**
**Download CSV** and **Download JSON** save the hydrostatic table and the curve. Every file starts with its tier, the method and engine version, each input with where it came from, the assumptions, and the units of every column. A CSV's header lines start with #, which most tools (pandas, R, MATLAB) can skip.
**Motion recording**
For a ship afloat, **Record** captures heave, pitch, roll and the wave height at the bow ten times a second while the scene runs; **CSV** or **JSON** saves it with the sea state, wind, speed and heading in the header. The file also adds the encounter frequency at the peak of the sea's spectrum for each row's speed and heading. This is the real-time motion the scene draws, labeled VISUAL: plausible, but not a validated seakeeping result. Nothing is uploaded; the file is made in your browser.
**The sea**
The waves the hull floats on, the waves the scene draws and the waves in the Lab's files are one sea. It is drawn from a wave spectrum (JONSWAP), realized from a seed, so the same scene always has the same sea.
When a scene is set at a real place for now or the next few days, and a licensed wave forecast covers that spot (today US coastal waters from the National Weather Service, and Northwestern Europe from MET Norway), the sea takes its height, period, direction and wind from that forecast. In US coastal waters the National Weather Service forecast also gives the visibility, which sets the scene's haze. Where the forecast gives a swell and a wind sea separately, both are drawn, each from its own direction. Turtini fetches the forecast on its servers; your browser never contacts the agency. Every figure is labeled with the agency and the time it issued the forecast, in the Inspector (tap the water) and in the Lab's files. A sea state you change by hand replaces the forecast.
A historical date, a place no forecast covers, or a forecast that cannot be reached leaves the scene's stated sea state in charge: its height comes from the sea state and its period from the wind, and it is labeled "inferred from sea state". There is no historical wave record wired yet, so a past date never shows a forecast. The **Sea** section shows the spectrum, its significant wave height and peak period, and the height this particular sea actually carries. Every motion file names them in its header with the seed, so a run can be reproduced.
**Wave probes CSV** or **JSON** saves two minutes of the water height, four times a second, at five buoys: the middle of the scene and four points 100 m up-wave, down-wave and to each side. The buoys are moored, so the series is the sea as a buoy would measure it, not as a moving ship meets it (that is the bow column of the motion recording). The sea is a model input, not a measured or forecast sea, and it is labeled VISUAL.