Data Reconciliation

Why Raw Instrument Data Isn't Good Enough Anymore

Every flow meter, level gauge, and analyzer on a plant drifts, has some margin of error, and occasionally goes offline. On its own, that's manageable — but once you try to use that raw data for production, yield, and loss accounting, small individual errors compound into numbers that simply don't balance: feedstock in doesn't match product and losses out, even though physically it has to. Data reconciliation is the process of resolving that gap — taking raw measurements and adjusting them, within their known uncertainty, so they satisfy mass and energy balance constraints, producing a single, statistically consistent, gap-free “best estimate” of what's actually happening in the plant.

Why It Matters Right Now

Traditional accounting methods that simply take meter readings at face value are increasingly out of step with what's required. As emissions taxes and carbon accounting frameworks (including Alberta's TIER program) put real financial weight behind accurate production and emissions numbers, the cost of an unreconciled, inconsistent data set stops being an internal nuisance and starts being a compliance and audit risk. Instrument calibration and quality are central to this: reconciliation doesn't fix a badly calibrated instrument, but it does make miscalibration visible — flagging exactly which measurements are inconsistent with the rest of the balance, which is often the fastest way to find a drifting meter before it causes a bigger problem.

What a Reconciliation Project Involves

Capstone's approach typically starts with a preliminary analysis of your current feedstock and production numbers to see how well they already match up, followed by a review of instrumentation and calibration practices, building the reconciliation model itself, and then ongoing monitoring so the model keeps working as conditions change rather than becoming a one-time exercise. The goal throughout is preserving data integrity — giving your team accuracy and accountability they can defend in an audit, not just a cleaner-looking spreadsheet.

Who This Is For

This work matters most for plant managers and process engineers responsible for production accounting, loss accounting, or emissions/regulatory reporting — particularly anywhere feedstock and production numbers have been hard to match, or where a regulator or internal audit has raised questions about measurement consistency. 

If matching feedstock and production numbers has become a recurring headache, or emissions accounting is under more scrutiny than it used to be, a preliminary analysis is a low-commitment way to see where the gaps actually are before investing in a full solution.