Automatic tank gauging (ATG) for fuel tanks: types, working principles and accuracy

An ATG (automatic tank gauge) is a measuring instrument that reads the liquid level in a tank automatically and continuously, usually together with temperature and water level, and converts it into volume using the tank calibration table.

In an Indonesian fuel terminal (TBBM), the ATG sits on the roof of each storage tank and sends its readings to the control room. Operators still dip tanks by hand with a tape, but the ATG gives the live picture between dips and feeds the tank management system that issues tank tickets. This guide covers what an ATG measures, the main gauge types, how accurate they are and how the data reaches the inventory system.

The supporting guides go deeper into each step: tank calibration tables and ullage, volume correction to 15 °C, tank management systems and tank tickets, high-high alarms under API 2350 and integrating ATGs from several brands.

Four measuring principles cover almost every ATG in use. They differ in what touches the product, what else they can measure and where they are normally installed.

Radar gauges send microwaves down to the liquid surface and time the returning echo. Nothing touches the product, so wear is low. Custody-grade models such as the Rosemount 5900S carry OIML R 85 type approvals, and Emerson states a 0.5 mm instrument accuracy for its premium version.

Servo gauges hang a small displacer on a measuring wire. A servo motor keeps the displacer balanced at the surface and counts the wire movement. Because the displacer can be driven down on command, the same gauge can find the water interface. Honeywell lists product level, interface level and temperature among the measurements of its Enraf Servo 954.

Magnetostrictive probes use a float that rides on a rod; the electronics time a pulse along the rod to locate the float. They are typical of underground tanks at fuel stations (SPBU), which is why searches for an SPBU ATG mostly lead to this type.

Hydrostatic systems measure pressure at fixed points on the tank shell and calculate mass and density. Combined with a level gauge, they form the hybrid tank measurement system described in API MPMS Chapter 3.6.

An ATG reports a height, not a volume. Turning that height into a stock figure that two parties can agree on takes a fixed sequence of steps, and every step uses data from outside the gauge.

First the level is looked up in the tank calibration table (strapping table) to give the total observed volume. Free water is subtracted to give the gross observed volume. Temperature and density then set the volume correction factor, which brings the figure to gross standard volume at 15 °C. The guides on tank calibration tables and volume correction factors walk through each step with examples.

API MPMS Chapter 12.1.1 states the principle plainly: to reconcile volumes, different parties must start from the same basic information, meaning the same capacity table, levels and temperatures, whether that information was gathered automatically or by hand. A gauge that reads perfectly but feeds an outdated table still produces the wrong stock.

Floating-roof tanks add one more correction. While the roof floats, it displaces product equal to its own weight, so the table or the TMS applies a roof correction above the level where the roof lands. Readings taken in the zone where the roof is only partly floating are unreliable, and most procedures flag them.

The table is also a legal document. Lampiran I of Permendag No. 24 Tahun 2024 lists vertical cylindrical fixed tanks for petroleum products with a re-verification interval of 6 years, so the TMS must always hold the table from the latest certificate.

Every terminal sees it sooner or later: the ATG shows one figure and the operator's dip tape shows another. Most of the time neither instrument is broken; the two readings were taken under different conditions.

API MPMS Chapter 3.1A sets the practice for manual gauging, including how to verify tape length in the field and how bob weight, temperature and the gauging reference point affect the result. Chapter 3.1B covers accuracy, installation, commissioning, calibration and verification of ATGs. A fair comparison follows both.

• Time. The dip is taken minutes after the logged ATG value, while the tank is still settling after a transfer.
• Reference height. Tank bottoms and roofs move with load and temperature, so the reference height at the dip hatch changes.
• Position. The dip hatch and the ATG nozzle sit at different points of the roof, and a sloping bottom or tilted floating roof reads differently at each.
• Water and sludge. A bottom layer of water or sediment blurs a dip cut and can confuse an interface probe.
• Configuration. A wrong tank offset or a configuration lost after a repair shifts every reading by the same amount.

Record every comparison, including the ones that agree. A long history of comparisons is worth more than one big test, because it shows whether the ATG difference is stable, creeping or jumping after specific work on the tank.

The working rule is to compare like with like: same time, same reference point, tank at rest, and repeated readings instead of a single dip.

The right gauge depends on what the reading will be used for. A gauge for daily inventory and a gauge for custody transfer can look alike on the roof but are specified, approved and maintained differently.

• Service. Custody transfer needs a gauge with metrological type approval and a verification procedure; inventory control can accept a simpler instrument.
• Tank construction. Floating-roof tanks usually need a still pipe for radar, or a servo gauge in a stilling well. Fixed-roof tanks give more freedom of position.
• Product. Heavy or sticky products foul displacers and floats; light products with vapour favour radar antennas designed for condensation.
• Hazardous area. The gauge, its cabling and the junction boxes must carry certification for the zone where they are installed.
• Communication. Confirm the output protocol and how it will reach the existing TMS before ordering, not after delivery.
• Lifecycle. Ask how long the interface modules will be supported. Discontinued interfaces were the root of several faults Citia has handled.

Write these points into the purchase specification and the factory acceptance test. A gauge selected for its accuracy figure alone can still fail at commissioning because of a protocol or mounting mismatch.

Plan maintenance at the same time. Agree which spares stay on site, who may recalibrate the gauge after a repair, and how the tera ulang schedule fits the terminal shutdown plan. These answers decide how long a tank stays without a trusted reading when something fails.

On its own an ATG is a display on a tank roof. Its value to operations comes from the system that collects every gauge, keeps the history and turns levels into stock.

Gauges talk through field interfaces: proprietary buses on older installations, Modbus RTU over RS-485 on many newer ones. An interface unit or gateway polls each gauge and passes level, temperature and status to the server. The INTEGRA TMS page describes this path for Citia's own system: ATG, industrial gateway, SQL database, then tank tickets and stock posting to SAP. Fleets with several gauge brands need extra care, covered in multi-brand ATG integration.

An ATG high alarm warns the operator, but it should not be the only protection against overfill. If the gauge freezes, its alarm freezes with it. API 2350 recommends that the high-high alarm come from a sensor independent of the ATG; see high-high alarms and API 2350.

ATG inventory is also the starting point for loss control. When book stock and measured stock drift apart, the cause can sit in the gauge, the meter or the paperwork, and fuel losses at terminals shows how to separate them.

Every value that leaves the gauge should carry its time stamp and a quality flag. A TMS that keeps showing the last stored level without saying it is old invites a wrong decision during a receipt. Good practice is to grey out stale values and raise a communication alarm, separate from the process alarms.

Most ATG problems Citia meets on site do not start at the sensor. They start in power, wiring and communication, then appear in the control room as frozen or missing values.

At PT Redeco Petrolin Utama in 2024, the assessment found gauges that read wrongly, read nothing or were dead, corrosion and water ingress in outdoor housings, and discontinued IFX-20000 interface modules. Communication paths were restored, readings were compared with stick measurements, and units that could not be repaired were recommended for migration.

At Redeco in 2025, an Endress+Hauser ATG power supply failed after a lightning surge; the blown fuse and the damaged varistor were replaced and the unit was retested. At the Jambi fuel terminal in 2026, the ATG gateway (Advantech ECU-1051) was replaced and surge protectors were added on the RS-485 Modbus and Ethernet lines. The Cikampek tank-monitoring upgrade in 2025 shows the other side: loop checks, I/O tests and functional tests before handover.

The shared lesson is to protect power and communication lines from surges, keep spares for interfaces that are no longer made, and close every repair with an end-to-end test from the gauge to the TMS screen. The ATG integration checklist lists what to prepare before such work.