What are the key takeaways?
• A gauge measures level. The software converts it to volume, tonnes, and stock.
• The chain is level, then volume, then tonnes, then site roll-up. Each link can fail.
• Most errors come from bad inputs, not bad maths.
• A 1% density error on a 1,400 MT tank is about 14 MT of stock that does not exist.
• Good software checks every reading before it reaches a manager.
What does automatic tank gauging software do?
It converts raw gauge signals into numbers the business can use. The gauge measures. The software decides what the measurement means.
Four signals come off every tank:
• Level (mm): the product height, measured by the gauge
• Temperature (°C): used to correct volume to reference conditions
• Density (kg/m³): used to convert volume into mass
• Product ID: which commodity is in the tank
From these, the software calculates volume (m³), quantity (MT) and roll-ups across tanks, farms and sites.
How does a level reading become volume?
The software looks the level up in the tank's strapping table. This is a calibration table that maps liquid height to volume for that specific tank. Tanks are never perfect shapes, so a lookup is more accurate than a formula.
Then it corrects for temperature. Oil expands as it warms. In the deployment below, tanks in a single farm ran from about 25°C to over 53°C. That spread is big enough to matter.
The trap: an outdated strapping table, a level offset, or a temperature tag that has stopped reporting. The result is a volume that looks like a number but is not one.
How does volume become tonnes?
Quantity in MT is volume multiplied by density. Operations and finance use this figure because stock is bought, sold and reconciled by weight.
Here is a real row from a live extract, for a tank rated at 1,500 MT:
Check it: 1,540.41 m³ × 912.5 kg/m³ is about 1,405.6 tonnes. The tank is roughly 94% full. The number is plausible and can be trusted.
The trap: density. A stale or wrongly entered value skews every tonne figure, and nothing looks broken. A 1% error on this one tank is about 14 MT.
How do tank readings become a site stock position?
Roll-up turns single-tank readings into a single number a manager can use. It follows five steps:
1. Read the live quantity (MT) from every tank
2. Group by product or commodity code
3. Sum within each tank farm
4. Sum across tank farms at the site
5. Publish as the daily stock report
At one site in this deployment, the system keeps 26 commodities separate, from crude oils to by-products.
The traps:
• Product ID: if a code is missing or wrong, tonnes land under the wrong commodity or drop out of the total.
• Timing: totals only reconcile if every tank is read at the same moment. Each daily report here uses data captured about five minutes before the 07:00 run, with a fixed, auditable timestamp.
Where does tank gauging data go wrong?
The deployment's own data shows it. Two faults appeared in a single live extract:
• Missing feeds: five tanks returned a "missing column" fault in the same cycle. The gauge or its SCADA tag was not reporting product code, temperature, density or volume.
• An impossible reading: a tank rated at 1,125 MT reported a level of 100,000,000 mm. The computed quantity was about 6,988,972 MT, more than 6,000 times what the tank can hold.
The maths was correct in every row. The input was not. Software that only calculates will publish a bad input as faithfully as a good one.
This is why trust is the real problem. In the 2025 Dun & Bradstreet Manufacturing Pulse Survey, only 36% of manufacturers said they trust their own data.
Here is where the chain tends to break:
What should tank-gauging software check before publishing a number?
Nothing should reach a manager's desk until it passes a few simple checks:
• Completeness: all four signals are present. If one is missing, flag it. Do not calculate around it.
• Range: level sits within the tank's height. Quantity does not exceed rated capacity.
• Plausibility: density and temperature fit the product ID.
• Rate of change: stock cannot move faster than pumps and transfers allow.
• Timestamp: every figure carries a fixed capture time.
This validation layer, between the gauge and the report, is what Ketsol Manufacturing Suite (KMS) is built to provide. In the deployment below, automated range checks and alerts for stopped tags are the next planned step.
What does this look like in a real deployment?
Before KMS, each site ran its own spreadsheet-style printouts. Readings were taken by hand, reconciled by hand and rarely compared across sites.
Today, 11 SCADA systems are connected to a single industrial Ethernet network. Two SCADA platforms write to SQL Server databases. KMS connectors bring them into one hub, with data logged every minute.
Three reports run on top:
• Tank inventory: refreshed every 2 minutes for operators and engineers
• Daily stock roll-up: generated at 07:00 for managers
• Daily production report: a rolling 24-hour window, 07:00 to 07:00
The result is a single data model across all sites. Managers get a time-stamped stock position each morning. Engineers see live level, temperature and density. Operators work in one familiar web client.
What is the bottom line?
The maths behind tank stock is simple. The hard part is trusting the inputs. Automatic tank gauging software earns its place by checking every reading before it becomes a tonne figure.
If you are mapping this chain for your own tank farm, see how Ketsol Manufacturing Suite (KMS) handles data collection, validation and reporting.
Frequently asked questions
What is automatic tank gauging software?
It is software that takes readings from tank gauges, such as level, temperature, density and product ID, and converts them into volume, mass and stock positions. It also combines tanks across farms and sites into one view.
What is the difference between volume and quantity in MT?
Volume (m³) is the amount of space the product occupies. Quantity (MT) is its mass: volume multiplied by density. Finance and operations use MT because stock is traded by weight.
Why does temperature matter in tank gauging?
Product expands as it warms, so volume and density both change. Without a correction, two tanks with the same mass can show different volumes.
How often should tank gauging data refresh?
It depends on who uses it. In this deployment, operators see a view that refreshes every 2 minutes. Managers get a fixed stock snapshot at 07:00.
Can tank gauging software combine data from different SCADA systems?
Yes. Here, two SCADA platforms write to SQL databases. Connectors bring all databases into a single hub with a single data model.
How do you catch a faulty tank gauge reading?
Check each reading before publishing. Confirm every signal is present. Confirm level and quantity sit within the tank's limits. Confirm density and temperature fit the product. Flag any reading that fails.