Silo Inventory Monitoring: Level, Volume or Weight?

Technical guide by Arvin, METRAVON Instruments · Reviewed September 2026

Short answer

Choose the measurement variable that matches the business decision. Use non-contact radar when the primary requirement is continuous level or ullage. Convert level to volume only when the vessel geometry and strapping table are known. Use silo weighing when mass is the required result and the vessel structure can transfer load cleanly to load cells or weigh modules. Guided wave radar is worth evaluating for suitable liquids, interfaces or confined installations, but it is not automatically the best option for dusty bulk-solid silos.

Level, volume and mass are related, but they are not interchangeable. A stable level reading can still produce a poor inventory estimate when the material forms peaks, funnels or wall buildup, or when bulk density changes with moisture, compaction or batch.

Which method fits the requirement?

Selection factors that change the answer

For radar, confirm the material, dielectric behavior, measuring range, vessel diameter, roof geometry, filling point, internal beams, ladders, buildup, dust, process temperature and pressure, nozzle dimensions and hazardous-area requirements. The instrument should have a clear measurement path and should not look directly into the filling stream.

For weighing, calculate the maximum gross load rather than the product net weight alone. Include vessel tare weight, attachments, eccentric loading, impact, wind or seismic loads where applicable, and an engineering safety margin. Check whether inlet and outlet piping, dust-extraction ducts, cable trays, access platforms and mechanical stops can bypass the load cells.

If the value will be used for commercial settlement or regulated metrology, a general inventory system may not be sufficient. The approved instrument class, calibration method, legal requirements and operating conditions must be reviewed for that specific use.

Recommended engineering workflow

1. Define the result: level, ullage, volume, estimated mass or measured mass.
2. Set an accuracy expectation: state the required uncertainty and the operating range over which it applies.
3. Survey the vessel: record dimensions, cone geometry, supports, nozzles, obstructions, filling and discharge locations.
4. Characterize the material: record particle size, dust, moisture, buildup tendency, angle of repose and bulk-density range.
5. Freeze interfaces: confirm power, 4–20 mA or switching outputs, RS485/Modbus requirements, cable routes and system ownership.
6. Define acceptance: agree on reference points, loading conditions, allowable deviation and records before ordering.

Installation and commissioning checks

• Mount radar away from the filling stream, wall, roof ribs and internal structures; record unavoidable false echoes during commissioning.
• Use a segmented strapping table for conical or irregular vessels instead of multiplying height by a single area.
• For silo weighing, level and align support points, control restraints, and use flexible connections where process design permits.
• Verify empty, low, mid and high conditions where practicable, and compare readings during filling, discharge and rest.
• Record raw distance or load, calculated inventory, signal quality, alarms and the independent reference used for acceptance.
• Test loss of power, sensor fault, out-of-range values and communication failure so the control system can distinguish bad data from real process change.
• Repeat reference checks after filling and discharge cycles to expose hysteresis, retained material and mechanical binding.

What should be supplied with an enquiry?

Provide the material name and condition, silo height and diameter, roof and cone drawings, measuring range, filling and discharge arrangement, expected surface profile, bulk-density range, temperature and pressure, dust level, mounting opening, internal obstructions, required output or protocol, hazardous-area classification, inventory calculation method and available site photographs.

Frequently asked questions

Can a level reading be converted directly into tonnes?

Only as an estimate based on a verified vessel-volume relationship and a representative bulk density. If density or surface shape varies, define an uncertainty range and retain the original level value for traceability.

Does heavy dust rule out radar?

No. Non-contact radar is often suitable for dusty bulk solids, but the final choice still depends on range, antenna buildup, mounting position, filling impact and the strength of unwanted echoes. Dynamic testing is more useful than an empty-vessel check alone.

When are multiple measurement points justified?

Consider multiple points for wide silos, multiple filling points, asymmetric discharge or inventory targets that cannot tolerate the uncertainty of a single local surface measurement.

Why can the instrument be stable while inventory remains wrong?

The common causes are an incorrect strapping table, unrepresentative density, wall buildup or an assumed surface model. Audit the raw measurement, conversion algorithm and independent stock check separately.

How to establish a defensible inventory uncertainty

Do not assign one accuracy figure to the complete inventory chain without separating its contributors. For a radar-based result, uncertainty starts with the distance measurement but usually becomes dominated by the surface model, vessel geometry and bulk density. A narrow instrument specification therefore does not guarantee a narrow error in tonnes. Establish an error budget for the operating range that matters: identify uncertainty in distance, reference height, strapping table, retained material, surface shape and density, then test the combined result against independent stock movements. Report an expected range rather than false precision when the inputs vary.

A practical check uses reconciled transfers over several normal cycles. Record opening inventory, verified receipts, verified consumption and closing inventory without changing calculation constants during the trial. Compare the balance with the indicated change and investigate bias separately from random variation. A repeatable bias may point to geometry, zero reference or density. Scatter that changes with fill pattern more often indicates an unrepresentative measurement point or surface model. This method is more useful than calibrating at one convenient level.

Worked decision example for a bulk-solid silo

Consider a tall silo filled pneumatically from one side and discharged through a central cone. Operations need high-level protection, purchasing needs weekly inventory, and production needs confirmation that enough material remains for the next batch. These are three requirements, not one. Use an independent high-high switch for the protective action. Use radar for level trend and remaining headroom. Convert the radar result to estimated volume through the verified geometry, and apply a density range for planning. If financial reconciliation requires a tighter mass result, evaluate load cells or an external weighing reference rather than forcing the level calculation to meet an unsuitable target.

The installation review should verify whether the radar sees the filling peak, the discharge funnel or a representative zone between them. Measurements taken immediately after filling may differ from those taken after settlement even though no material has left. Agree when inventory snapshots are valid, for example after a defined settling time and with conveying stopped. Store raw distance together with calculated volume and mass so future corrections to geometry or density do not destroy traceability.

Ownership, records and lifecycle control

Assign responsibility for the instrument, vessel dimensions, conversion table and bulk-density values. Maintenance may own the sensor, engineering the strapping table, and operations the density selection; without named ownership, a seemingly small change can invalidate the result. Protect calculation constants with access control and revision history. Record why each value changed, who approved it and from which physical measurement or laboratory result it came.

Review the inventory model after structural work, liner replacement, a new filling route, a product change or persistent reconciliation drift. Do not automatically retune the radar when the raw distance remains credible. First determine whether the error is in sensing, geometry, density or the timing of the comparison. This disciplined separation shortens troubleshooting and prevents one correction from hiding another problem.

Data presentation for operators and purchasing

Present the raw level or ullage, calculated inventory, data quality and time of the last valid update together. A single large tonnage number can mislead users when the sensor is in fault, the silo is being filled or the selected density is outdated. Show the density assumption and, where useful, a high and low estimate. Operations may need time to empty or batches remaining, while purchasing needs stock reconciliation; calculate these outputs from the same governed source data rather than maintaining separate spreadsheets.

Set reorder alerts from usable stock, expected consumption, delivery lead time and a reserve, not from an arbitrary percentage. Exclude inaccessible heel and account for planned production. Review alert performance after demand or supplier lead time changes. This connects the measurement system to an actionable decision while keeping its physical uncertainty visible.

Archive each approved report with its source readings and assumptions.

Engineering note: This guide supports preliminary selection. Final performance, certification and supply scope must be confirmed against the selected model, application data, drawings and written technical agreement.