What is statistical inventory reconciliation?
Analytics on tank data becomes a regulated leak method the moment you act on it, and the performance bar is written down.
Statistical inventory reconciliation
SIR
Statistical inventory reconciliation, or SIR, is a leak detection method. It reads a tank's stock records, deliveries and sales over weeks to decide whether the tank is losing product. Under 40 CFR 280.43(h) it must detect a leak rate of 0.2 gallon per hour, or a release of 150 gallons within 30 days.
A tank gives you noisy data. Gauge readings drift, temperature moves the volume, deliveries land mid-day and sales never stop. Any single day's numbers are useless for deciding whether product is escaping.
SIR takes a month of that noise and asks a narrower question. Is there a consistent loss underneath it, large enough to matter? A vendor analyses the data off-site and returns a result: pass, fail, or inconclusive. That third answer is real and it is common.
The numbers are the regulation, not a benchmark
Most software claims about accuracy are marketing. This one is federal text. Tank owners must use a release detection method that meets a published standard, and each method carries its own figure.
That changes what a build may claim. A tool that graphs variance is doing analysis. One offered as the release detection method has to be certified against the standard by a third party, with the evidence on file. Blur those two and an operator stops looking for a method they still need. That is the most dangerous overclaim in fuel software.
State programs run underground tank regulation locally, and they are often stricter than the federal floor. Your implementing agency has the final word on what it accepts.
- 01SIR: detect 0.2 gallon per hour, or 150 gallons within 30 days, under 40 CFR 280.43(h).
- 02Automatic tank gauging: detect 0.2 gallon per hour, tested at least once every 30 days, under 280.43(d).
- 03Tank tightness testing: detect 0.1 gallon per hour from any portion of the tank that routinely contains product, under 280.43(c).
- 04Any other method: 0.2 gallon per hour or 150 gallons in a month, at a probability of detection of 0.95 and a probability of false alarm of 0.05, under 280.43(i).
- RecordDaily stock, sales, deliveries.
- CorrectTemperature and tank chart.
- AnalyseLoss signal against noise.
- ResultPass, fail or inconclusive.
- ActInvestigate, or report a suspected release.
Inconclusive is not a failure of the method. It usually means the input data was not clean enough, which points at the readings rather than the tank.
Related questions
01How is SIR different from automatic tank gauging?
A gauge measures the tank directly and runs its own test in the tank, usually overnight when the site is still. SIR uses the records a site already produces and finds the signal statistically over weeks. Both must detect 0.2 gallon per hour, so the difference is the data they work from and how long they need.
02Why did our SIR result come back inconclusive?
Usually the input data was too noisy to reach a verdict, not that the tank is fine. The usual causes are a wrong tank chart, missing or guessed readings, a delivery logged on the wrong day, and volumes kept on mixed bases. Clean inputs fix more of these than a better analysis does.
03Does SIR replace the equipment tests?
No. Release detection is one duty among several. Equipment tests, walkthrough checks and corrosion protection all run on their own clocks. Assume one method covers the lot and you find the gap during an inspection.
04Can we build SIR into our own platform?
You can build the data layer that feeds it, and that is usually the valuable part. The method itself has to meet the federal standard and be evaluated against it. Treat it as a certified component you integrate, not a feature you write. Check what your state accepts before you scope either.
Related reading
- Wetstock management →The wider practice SIR draws its data from, and the six causes of variance.
- Does OSHA PSM apply to fuel storage? →The standard most fuel vendors cite incorrectly.
- Energy and fuel software development →What we build for fuel distributors and site operators.
- TankAware →Petroleum inspection and site management for Sutherland Excavating Ltd.
