NavArch Handbook
Chapter 1 / Section 1.9
The digits of a number that carry information about its precision. 0.70 m has two significant figures (precision ±0.005 m); 0.7 m has one (±0.05 m). In stability work "GM = 0.7 m" and "GM = 0.70 m" are different statements. Trailing zeros after a decimal point are significant; leading zeros are not.
Round half away from zero (rule A of ISO 80000-1: 2.45 → 2.5) is the common engineering convention; round half to even (rule B: 2.45 → 2.4, 2.55 → 2.6) avoids bias in long sums and is used by most software (IEEE 754 default). State which convention is used when results sit on a criterion boundary — a GM of 0.3449 m rounded to 0.34 fails a 0.35 m criterion; rounded to 0.345 it is ambiguous. Never round in the direction that helps compliance: compare the unrounded value with the criterion and report the margin.
See the table. The principle: report to the precision at which the quantity can be measured or built, and one more figure where the number enters a criterion.
Inputs (with source and units) → formulas (with references) → results → comparison with allowables (unity check or margin) → conclusion. Every table with results should carry the load case, the units in the header and the source of the allowable. The pass/fail statement must quote both the computed value and the limit.
Result unit -
Report UC to two decimals (e.g. 0.87); the margin is (1 − UC) × 100 %.
Result unit %
| Quantity | Report to | Example | Reason |
|---|---|---|---|
| Ship dimensions L, B, D | 0.01 m | L_BP = 100.00 m | Drawing precision |
| Draughts, trim | 0.01 m (1 cm) | T_F = 5.55 m | Draught marks read to 1 cm |
| Displacement | 1 t (0.1 t below 1000 t) | Δ = 6314 t | TPC ≈ 12 t/cm → 1 t ≈ 1 mm |
| Volumes | 1 m³ (0.1 m³ for tanks) | ∇ = 6160 m³ | — |
| KB, KM, KG, GM | 0.01 m (0.001 m in intermediate steps) | GM = 0.70 m | GM criteria in cm |
| GZ | 0.001 m | GZ₃₀ = 0.485 m | Enters area integrals |
| Areas under GZ curve | 0.0001 m·rad | 0.1166 m·rad | Criteria at 0.055 / 0.090 / 0.030 |
| Angles (heel, trim, φf) | 0.1° | φ₀ = 9.6° | — |
| Form coefficients | 0.001 | C_B = 0.700 | — |
| TPC, MCT1cm | 0.01 t/cm; 0.1 t·m/cm | 12.05 t/cm; 59.8 t·m/cm | — |
| Plate thickness, scantlings | 0.1 mm (requirement); 0.5 mm steps (supply) | t_req = 9.3 mm → 9.5 mm (or 10 mm) | Rolled plate availability |
| Section modulus, inertia | 1 cm³; 1 cm⁴ | Z = 307 cm³ | Rule formulas in cm units |
| Stress | 1 MPa | σ = 99 MPa | Material scatter ≫ 1 MPa |
| Forces, moments (structural) | 0.1 kN; 1 kN·m | F = 274.7 kN; M = 2044 kN·m | — |
| Unity checks | 0.01 | UC = 0.50 | — |
| Mass items in loading table | 0.1 t with 0.01 m centres | 12.5 t at Kg 7.20 m | Sum precision |
Worked example
Problem
A fishing vessel's computed fluid GM is 0.3471 m against a required 0.35 m. The hydrostatic KM is known to ±0.01 m. How should the result be reported?
Solution
Report 'GM_f = 0.347 m < 0.35 m required — does not comply; shortfall 0.003 m (0.9 %)'. Rounding to 0.35 m would disguise a failure. The ±0.01 m uncertainty of KM means the true GM may lie between 0.337 and 0.357 m; a result inside the uncertainty band of the criterion is a failure for approval purposes and must be resolved by design change (e.g. 5 t of low ballast lowers KG by about 0.03 m on a 450 t vessel), not by rounding.
Shows what precision is physically achievable in steel construction.