Why Ordinary Scales Fail on a Moving Vessel

A scale does not measure mass directly. Its load cell measures the force the load exerts, and the indicator converts that force into kilograms using the gravity value assumed at calibration. On land that assumption holds. On a vessel the deck itself accelerates up and down, tilts and vibrates, so the force on the load cell changes continuously while the mass stays exactly the same. This article explains the physics in plain terms, puts numbers on each effect and shows why waiting or averaging alone is rarely enough.

A scale measures force, not mass

An ordinary electronic scale shows the correct mass only when the platform is stationary and level. The load cell is a strain-gauge sensor that deforms in proportion to the vertical force on it. When the scale is still, that force is simply the weight of the load, and the indicator converts it into kilograms using the gravity value set at calibration.

When the platform accelerates upwards, the load has to be accelerated as well, so it presses on the load cell harder than its weight. When the platform accelerates downwards, it presses less. This is the same effect you feel in a lift: heavier when it accelerates upwards, lighter when it accelerates downwards. HyperPhysics describes the support force in a lift as the weight plus the extra force needed to accelerate the mass. An ordinary scale cannot tell the difference between a heavier fish and a deck that is accelerating upwards. The relative error equals the deck acceleration as a share of normal gravity, which is about 9.8 m/s²: an acceleration of 0.98 m/s² produces a 10% error.

How a vessel moves: heave, pitch and roll

A vessel in waves moves in six degrees of freedom, and three of them create vertical acceleration at the scale. The Delft University of Technology textbook Offshore Hydromechanics defines them relative to the vessel's centre of gravity:

• Heave: translation along the vertical axis, the whole hull rising and falling.
• Pitch: rotation about the transverse axis, bow and stern moving in opposite directions.
• Roll: rotation about the longitudinal axis, port and starboard rising and falling in turn.

Surge, sway and yaw are mainly horizontal. The same textbook shows that the vertical motion at any point on board is built up from heave plus the contributions of pitch and roll, and the rotational contributions grow with the distance from the centre of rotation. A scale mounted far forward or high on a working deck therefore sees more vertical acceleration than one close to the vessel's centre of motion.

Wave-induced motion is roughly periodic, and its peak acceleration depends strongly on how quickly the motion repeats. Doubling the vertical amplitude doubles the acceleration, but halving the period at the same amplitude makes it four times larger. Short, steep seas are therefore more disruptive than long swells of the same height. As a reference point, the Delft textbook's wave-prediction example gives a significant wave height of 1.5 m with an average period of 4.8 s for a 10 m/s wind (Beaufort 5) over a 60 km fetch.

Worked example: how large is the error?

The error on an uncompensated scale is often several percent and can exceed 10% in a moderate sea. The examples below show the peak acceleration for a regular motion of the stated amplitude and period, and what it does to a 10.000 kg load. They are idealised single-motion cases for illustration, not measurements on a specific vessel.

Scenario	Vertical amplitude at the scale	Period	Peak acceleration	Error	10.000 kg load reads between
Gentle heave	0.25 m	6 s	0.27 m/s²	±2.8%	9.72 and 10.28 kg
Moderate heave	0.5 m	6 s	0.55 m/s²	±5.6%	9.44 and 10.56 kg
Short, steep sea	1.0 m	5 s	1.58 m/s²	±16.1%	8.39 and 11.61 kg
Pitch 2°, scale 20 m from pitch axis	0.70 m	6 s	0.77 m/s²	±7.8%	9.22 and 10.78 kg
Roll 5°, scale 5 m off centreline	0.44 m	8 s	0.27 m/s²	±2.7%	9.73 and 10.27 kg

Put the moderate-heave case in context: on a 15 kg scale with a 5 g division, a swing of ±559 g is more than ±110 divisions. The display does not wobble by a few grams; it runs through hundreds of grams every few seconds. These levels are not extreme. Operability criteria quoted in the Delft textbook (NORDFORSK, 1987) set the limit for heavy manual work at an RMS vertical acceleration of about 1.5 m/s², so crews routinely work in conditions where an uncompensated scale is unusable.

Vibration: small movements, large accelerations

Engine and machinery vibration adds a second, high-frequency disturbance on top of wave motion. Acceleration grows with the square of the frequency, so tiny displacements can produce large accelerations. A deck vibrating with an amplitude of just 0.05 mm at 25 Hz (a machine running at 1,500 rpm) has a peak acceleration of about 1.23 m/s², which is 12.6% of normal gravity.

The saving grace is that vibration is fast. A scale can suppress a 25 Hz disturbance within a fraction of a second, whereas wave motion with a period of 5 to 10 seconds sits in the same time scale as the weighing itself and cannot simply be smoothed out without making the scale very slow. Vibration still matters in three ways:

• It can excite a resonance in a flexible mounting bracket, turning a small disturbance into a large one.
• It adds noise that lengthens the time needed for a stable reading.
• Over months, it loosens fixings and fatigues cables and connectors.

Mounting location and rigidity are covered in Installing a Marine Scale on Deck or in the Fish Hold.

Tilt error

When the platform tilts, a load cell only measures the part of the weight that acts along its own axis, so the scale under-reads. The error is tiny at small angles and grows quickly beyond a few degrees: about 0.015% at 1°, 0.4% at 5° and 3.4% at 15°.

Tilt angle	Under-reading	Error on a 20 kg load
1°	0.015%	3 g
2°	0.061%	12 g
3°	0.137%	27 g
5°	0.381%	76 g
10°	1.52%	304 g
15°	3.41%	682 g

Tilt also introduces side loads. A load cell gives its most reliable output for forces along its measuring axis, while horizontal forces stretch, twist and deform it. Legal metrology reflects this: OIML R 76-1 requires tilt testing for instruments liable to be tilted, and the WELMEC Guide 2 section on weighing instruments installed in ships calls for tilt tests up to 25% (15°) unless the display is blanked at a smaller angle.

Why waiting or averaging is not enough

Averaging reduces random noise, but wave motion is not random noise on the time scale of a weighing. An average only cancels a periodic disturbance cleanly when it spans whole wave periods. With periods of 5 to 10 seconds, that means 10 to 30 seconds per reading to average over several cycles, and wave groups vary in height, so the average still wanders.

There are other problems:

• Operators do not wait. On a processing line a reading that takes 20 seconds is abandoned, and the operator adds extra product "to be safe", which becomes giveaway.
• A basic stability detector on an ordinary scale may never declare the reading stable, or it may lock onto a momentary reading at the top or bottom of a wave.
• Tilt error does not average to zero. It always under-reads, so it biases the mean rather than cancelling out.

This is why marine scales correct for the motion itself instead of relying on averaging alone. What good compensation achieves, and how a buyer can judge it, is explained in Motion Compensation in Marine Scales: What It Does and How to Judge It.

Gravity, temperature and moisture

Motion is the dominant problem, but three slower effects also move the reading on a vessel.

• Gravity. Gravitational acceleration at sea level varies by about 0.5% between the equator and the poles. A scale adjusted in port at one latitude and used on fishing grounds far to the north or south reads differently. See Gravity, Latitude and Why Marine Scales Need Gravity Compensation.
• Temperature. Load cell zero and span drift with temperature. Moving a scale between a chilled fish hold and a sunlit deck is a real temperature step.
• Moisture and build-up. Water, ice, slime and scales accumulating on the platform shift the zero, and a rising zero looks exactly like extra product.

How WPL approaches this

WPL builds scales specifically for these conditions. The M2 Series bench scale and the M3 Series scientific scale are motion compensated; the M5 Series platform scale compensates for motion, tilt and gravity change. Stability behaviour can be tuned per installation with the R10 Configuration Panel, which exposes settings such as marine filtering, stability count and display speed. For the complete picture, start at the Weighing at Sea guide.