Treat the PE Naval Architecture and Marine syllabus as a set of classification decisions rather than isolated formulas. For every practice problem, first decide: small-angle or large-angle stability, weight shift or slack tank, intact or damage condition, effective power or delivered power. Then compute. This article teaches the named concepts behind each decision, walks through two scenarios where the plausible first move is the wrong one, and provides a box-barge exercise with expected results and a rubric.
Telling Initial-Stability (GM) Problems from Large-Angle (GZ) Problems
Classify by heel magnitude and available data: if the problem stays within small-angle territory and gives hull geometry, compute GM; if heel exceeds the small-angle range or asks for a righting arm at a stated angle, work from cross-curve or GZ data.
Metacentric height GM is the slope of the GZ curve at the origin, not the curve itself. Build it from the keel upward: KB from draft, BM = I/∇ using the transverse second moment of the waterplane, then KM = KB + BM and GM = KM − KG. Multiplying GM by sin θ approximates GZ only while heel stays small enough that waterline and shape changes remain negligible — the defining assumption of initial stability.
Scenario 1: a question asks for the righting arm of a loaded deck barge heeled 20°, and the hydrostatic table supplies GM. The plausible first move is GZ ≈ GM·sin 20°, a clean one-line answer. The better decision is to read GZ at 20° from the cross curves and correct for the actual vertical center of gravity: GZ_true = GZ_cross − (KG − KG_assumed)·sin θ. Why it matters: beyond small angles the true curve departs from the sine line, and at 20° the gap can be a substantial fraction of the righting arm itself. Committing to GM at that heel misuses a small-angle tool on a large-angle question.
- Cue: heel stated in single degrees under about 5, weights fixed → initial stability, compute KB, BM, KG, GM.
- Cue: heel of 10° or more, or a required GZ or downflooding angle → cross curves plus a KG correction.
- Cue: hydrostatics given as GZ versus displacement curves → always apply the sin θ correction for the difference between your KG and the curve's assumed KG before answering.
| Problem cue | Concept to invoke | What you actually compute |
|---|---|---|
| Small heel, fixed weights | Initial stability via metacentric height | KB + BM − KG, then GM·sin θ if a small-angle arm is needed |
| Heel beyond small-angle range | Large-angle stability via cross curves | Read GZ at the stated angle, correct for actual KG with (KG − KG_assumed)·sin θ |
| Partial heel, upright hull nearly wall-sided | Wall-sided formula | GZ = (GM + ½BM·tan²θ)·sin θ when the sides are vertical near the waterline |
| Angle of maximum righting arm or range asked | Shape of the full GZ curve | Locate the peak or zero crossing on the corrected curve, not from GM |
Handling Slack Tanks: Free Surface Correction versus Weight Shift
A vertical weight move changes KG directly by weight times distance divided by total displacement. A slack tank instead produces a virtual rise of G equal to the free surface moment divided by displacement, and that correction does not shrink with the amount of liquid.
The two mechanisms look alike — both reduce effective GM — but they behave oppositely with fill level. A solid weight raised by distance d shifts the center of gravity upward by w·d/Δ exactly. A slack tank's free surface moment equals the liquid density times i, the second moment of the free-surface plan area about the axis of tilt, so the virtual rise of G is i·ρ/∇. Because i depends on tank plan dimensions, not fill depth, a nearly empty slack tank and a nearly full slack tank of the same plan carry almost the same free surface penalty despite their different liquid weights.
Scenario 2: a stability question gives a service fuel tank roughly half full and asks for the corrected GM after a centerline subdivision is added. The plausible mistake is to reason that half the fuel means half the free surface effect and scale the correction down. The better decision is to compute the virtual rise from i before and after the bulkhead divides the plan area — i for a rectangular surface falls by a factor of four when split about the centerline — and leave the fill fraction out of the correction entirely. Why it matters: treating free surface as proportional to volume understates the penalty of wide, shallow slack tanks, where the correction is geometrically largest, and it misjudges the value of subdividing a tank versus emptying it.
Interlocking Hull Form Coefficients: Cb, Cp, Cm, and Cwp
Block, prismatic, midship, and waterplane coefficients are ratios to their obvious enclosing boxes, and they multiply rather than compete: Cb equals Cp times Cm, so displacement problems decompose into a midship fineness choice plus a longitudinal distribution choice.
Work the algebra in both directions until it is automatic. Displacement volume is L·B·T·Cb; midship area is B·T·Cm; prismatic coefficient Cp is volume divided by (L times midship area), which forces Cb = Cp·Cm. Given any three quantities in that identity you can solve the fourth, and that one-line check catches unit and definition slips before they propagate. The waterplane coefficient Cwp belongs to a different family — it governs waterplane area and inertia, and therefore BM and damping — so never substitute it into the Cb identity.
Practice reading the coefficients as design statements rather than abstract ratios. A high Cm with a low Cp means full sections concentrated midships and fine ends; a low Cm with the same Cb means sections moderately full along most of the length. This becomes practical when a resistance or stability problem asks you to infer form: the same block coefficient supports different waterplane areas, and thus different BM values, depending on how Cwp relates to Cb. Build the habit of estimating waterplane inertia from Cwp through its proportional relationship to L·B³ when a problem gives coefficient data instead of offset tables.
Trim and Draft Problems: Working Through LCB, LCG, and the Trim Moment
A floating body in equilibrium requires LCG to equal LCB longitudinally; any mismatch times displacement is the trimming moment, converted to draft change at the perpendiculars through the moment to change trim one unit.
Set up every trim problem the same way: compute LCG from weights and their longitudinal positions, read or compute LCB from the hydrostatics at the current draft, and take the longitudinal separation as the trimming lever. The moment to change trim by one centimeter follows from the longitudinal metacentric height and length, and it converts the imbalance into end drafts. Drafts fore and aft then follow from the mean draft plus or minus the trim share measured from the center of flotation, not from midships.
The named concept that separates clean solutions from tangled ones is the center of flotation (LCF), the point about which the ship trims at constant displacement. A common setup error is applying the trim change symmetrically about midships; the correct split uses the LCF position from the hydrostatics table. A second distinction: mean draft at midships and mean draft at the LCF are not interchangeable once trim exists. As a practice exercise, write two versions of the same loading problem — one limited by mean draft at midships, one by draft at the LCF — and observe how the allowable cargo differs between them. That habit forces you to identify the governing reference point before distributing the trim.
The Resistance and Powering Chain: From Hull to Propeller Without Gaps
Trace power in named stages: effective power is total resistance times speed; delivered power is effective power divided by propulsive efficiency components; each stage has its own efficiency factor, and conflating stages is the standard setup error.
Name the stages explicitly when you practice. Effective power PE is the towrope demand of the bare hull. Add appendage and air drag, apply a hull efficiency and a propeller (open-water) efficiency, and you arrive at delivered power PD at the propeller; from there, shaft and gear losses lead to brake power. Each factor — hull efficiency, relative rotative efficiency, open-water efficiency, shaft transmission efficiency — has a definition in terms of thrust and torque identities, and exam-style questions test whether you can slot a given efficiency into the right ratio rather than multiply everything together.
The second named concept is the decomposition of hull resistance into frictional and residual parts, with frictional scaling by Reynolds number and residual components treated as scaling with Froude number. Model-test extrapolation adds a form factor (1+k) to carry part of the viscous pressure drag with the frictional scaling, plus a correlation allowance. The practical discipline: when a problem gives model resistance data and asks for ship power, check which scaling each component uses and whether a form factor has already been applied. Applying model-scale friction unmodified, or double-counting the form factor, produces plausible-looking numbers that answer a different question.
Hull Girder Bending: Combining Still-Water and Wave Loads on Paper
Longitudinal strength problems superimpose still-water bending moment from the loading condition and wave-induced bending moment from a standard wave condition, then compare the resulting stress against the section modulus at the deck or keel.
Structure the calculation by loading, not by formula. First distribute weight and buoyancy along the length; the imbalance of their curves produces the shear force and bending moment curves for the still-water condition. Then superimpose the wave-induced moment for the governing hogging or sagging case. The midship section modulus Z = I/c converts the total moment into deck or keel stress, with I built from longitudinal structural members treated as effective flanges. Hogging puts the deck in tension, sagging puts the keel in tension — stating the fiber and the sign before computing prevents sign errors that silently flip the answer.
Keep two distinctions alive while practicing. Local strength (plating, frames, stiffeners under hydrostatic pressure) and hull girder strength (the ship as a beam) use different load models and different section properties; a question about hatch coaming scantlings is not a hull girder question. And the effective breadth of plating in bending matters: counting every longitudinal member at full effectiveness overstates I. When a paper problem specifies which members are fully effective, follow that specification literally; when it asks you to reason about the midship section qualitatively, identify flange, web, and neutral axis roles before any arithmetic.
A Box-Barge Exercise, Self-Check Rubric, and Preparation Sequence
Use one simple geometry — a rectangular barge — to test whether you can classify stability problems, then run a fixed rotation through the six syllabus areas and grade yourself against a written rubric each cycle.
Exercise: take a box barge, L = 60 m, B = 10 m, T = 3 m, KG = 4.0 m, floating upright in water of unit density. Compute KM = T/2 + B²/(12T) = 1.5 + 2.78 = 4.28 m, so GM = 0.28 m. Now estimate the righting arm at 15° of heel two ways: the small-angle estimate GM·sin 15° gives about 0.072 m, while the wall-sided formula (GM + ½BM·tan²15°)·sin 15° gives about 0.098 m. Expected observation: the two answers differ by roughly a third at a heel many people would casually call moderate, and the gap widens with angle. That numerical gap — not a memorized rule — is what teaches you when GM must be set aside for GZ data.
Run a self-check rubric after each practice cycle: (1) you can derive KB, BM, KM, and GM for a box or simple ship form without notes; (2) you can state, with the i-based reason, why the free surface correction is independent of fill depth; (3) given cross curves, you apply the KG correction with the correct sign; (4) you can set up a trim problem from LCG − LCB through end drafts via the LCF; (5) you can name each efficiency in the PE-to-brake-power chain and place a given value in the right ratio; (6) you can identify hogging versus sagging stress fibers before computing section modulus. Score each item yes or no; a low score is a study signal, not a readiness verdict. For an adaptable sequence, rotate: hull geometry one cycle, then the stability chain, then trim, then the powering chain, then structures, then seakeeping and maneuvering, then a mixed set drawn from all six before repeating with harder conditions. One administrative note: for the current exam format, calculator policy, and scheduling, rely on NCEES directly rather than secondary summaries.
- Rubric item for classification: for any new problem, write one line naming the concept before touching the calculator — if you cannot, the problem is not ready to solve.
- Rubric item for free surface: predict the correction change when a rectangular tank is subdivided at the centerline (a factor of four reduction in i), then verify by computing both cases.
- Sequence adaptation: shorten rotation cycles as rubric items turn to yes, and let the mixed set — not a single topic — become the dominant practice mode in later cycles.
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.