Data reconciliation algorithm

The purpose of data reconciliation is to provide more accurate, reliable and consistent measurement data of flows in the whole plant, than what can normally be achieved by direct readings from flow meters, tank gauges and accounting receipts/shipment.

Risk of distortion to raw data

Raw data is often distorted by errors due to instrumentation inaccuracy and failure. Bad data can lead to debates between process engineers, tank farm or process unit managers; each one being convinced that their flowmeters or gauges are properly calibrated. Furthermore, extra maintenance costs are incurred from having to check more instruments than necessary. These consequences, both time-consuming and costly, can be reduced through a material balance employing a data reconciliation treatment.

AVEVA Production Accounting provides a consistent way to handle these problems and presents one set of reconciled data for yield accounting and performance monitoring.

Data reconciliation working principles

The industry has long accepted the fact that measurements disagree. The process of coming up with assumed quantities involves reducing the apparent error ascribed to each instrument, and even choosing intentionally to trust one measurement more than another.

• Accept that certain degree of measurement does not exist: Production accountants must assume that a certain degree of measurement does exist and set standards as to how much error is acceptable, sometimes on a measurement by measurement basis.

• Make reasonable assumptions: By looking at actual measurements, some of which are redundant, production accountants make a reasonable assumption about what the real quantities were. As a result, the production accountant can see, for each measurement, what the difference is between the measured and assumed quantity.

• Use as many measurements for more realistic results: It makes sense to use as many redundant measurements as possible, because this results effectively in a vote that produces the most realistic assumptions.

Since there are many measurements, some of which are only partially redundant, a software tool to assist with this step is makes the process more efficient. If you, the user, can use your intuition or tools to recognize that certain measurements were definitely outside reasonable bounds, those measurements can be ignored and excluded from the process of distributing errors.

AVEVA Production Accounting uses configuration and mathematics to assist you in doing all of the above.