Choosing a level measurement that will still work in a year

Level measurement is the area where the wrong choice is most likely to work during commissioning and fail six months later, when the product changes or the weather does.

Guided wave radar

A probe goes into the vessel and a microwave pulse travels down it, reflecting from the surface. Time of flight gives distance.

Largely unaffected by vapour, foam of moderate density, temperature, pressure or dust, because the pulse is guided rather than travelling through the atmosphere.

The limits: it is a contacting measurement, so the probe has to be compatible with the product. It needs a low dielectric product to be checked against the sensitivity, and heavy build-up on the probe eventually confuses it.

If somebody asks for a default recommendation and there is nothing unusual about the application, this is usually it.

Non contact radar

Same physics without the probe: an antenna at the top of the vessel and free space propagation.

Nothing touches the product, so corrosives, sanitary applications and anything with agitation are comfortable. Modern 80 GHz devices have narrow beams and cope with internals better than older 26 GHz ones.

The limits: heavy foam absorbs the return, condensation on the antenna degrades it, and the beam must have a clear path, which vessel internals and agitators may not allow.

Ultrasonic

A pulse of sound, reflected from the surface, timed.

Cheap, non contacting, and easy to install, which is why it is everywhere.

It fails in three common conditions, and all three appear during filling: foam absorbs the pulse, vapour changes the speed of sound so the distance is calculated wrongly, and a turbulent surface scatters the echo. Dust does the same in solids.

Right for a water tank in a benign environment. Wrong for anything that foams, steams or blows dust, and it will appear to work when it is commissioned on an empty vessel.

Hydrostatic

A pressure transmitter at the bottom. Pressure equals density times gravity times height.

Simple, robust, no line of sight required, unaffected by foam or vapour, and cheap.

The catch is in the equation: the reading depends on density. Warm the product and the level appears to fall. Change concentration and the reading moves with no change in level.

For water at a stable temperature it is excellent. For anything whose density varies, it needs compensation, which means measuring density too.

Also note that a sealed vessel needs a differential measurement, because the vapour space pressure adds to the reading.

Capacitance

The probe and the vessel wall form a capacitor whose value changes with the level of product between them.

Robust, no moving parts, works on solids and liquids.

The measurement depends on the product's dielectric constant, so a change in composition or moisture content changes the reading. Coating on the probe is the classic failure: a conductive film effectively extends the probe and the vessel reads full permanently.

Load cells

Weigh the vessel and its contents. Not a level measurement at all, which is precisely the appeal.

Immune to foam, vapour, dust, dielectric, density and every internal fitting, because none of them affect mass.

The costs are mechanical rather than electrical: the vessel must be mounted so it is genuinely free to be weighed, and every rigid pipe, conduit and support that bypasses the cells is an error term. Getting that right on a retrofit is often the reason it is rejected.

The questions that decide it

• What is the product, and does its density, dielectric or composition change?
• Does it foam, steam or produce dust?
• Can anything touch it?
• Is there agitation, or are there internals in the beam path?
• Do you want level, or do you actually want mass?

The last one is worth asking directly. Recipes are usually in kilograms, and a level converted to mass through an assumed density is two measurements pretending to be one.