What is Data Reconciliation?
Data Reconciliation is a technique allowing to adjust the measured data and to give estimates to unmeasured variables where possible, in such a way that this set of measured as well as estimated data satisfies heat and material balance equations.
This reconciled data can then be used for process monitoring , process analysis and evaluation, process optimization and "what if " studies.
In the process of building a plant model, Data Reconciliation can identify erroneous measurements and locate inaccurate instruments .
In terms of plant instrumentation analysis, the same technique can be a powerful tool for the generation of strategic placement of instruments to produce cost effective designs. In a study, it can be used to generate a snap shot of the plant to determine instrumentation adequacy and/or redundancy .
Data Reconciliation turns real time process data that are subject to random error as well as gross error into consistent and reliable information . Such information is essential for effective plant operation and management. It uses statistically sound techniques to reconcile flow, temperature, and composition measurements in such a way that material, enthalpy and component balances around each unit in a plant are satisfied.
A Data Reconciliation software can be interfaced directly with your plant's distributed control system (DCS) or centralized database, and run in an on-line mode.
With data Reconciliation, we take advantage of the redundant equations, which otherwise are often overlooked by the engineers.
Mathematically, Data Reconciliation is nothing more than minimizing a sum of errors (the difference between each measured data and its reconciled value) weighted by the standard deviation of the measurement, subject to a number of constraints (the balance equations). A solution can be found by introducing Lagrange coefficients and by maximizing the likelyhood criteria.
Chuck Kelley's Answer: Synchronization is really about making sure that you have extracted from the appropriate sources and that all the data is available (i.e., run through the ETL process and loaded). Data reconciliation and database reconciliation I believe are synonyms. Both have to do with the reconciling the data and measures to the source systems.
Data reconciliation is element level checking where each element is valid. This includes matching the source and reflecting an accurate, valid value.
Database reconciliation focuses on the integrity and quality of the entire database or data set. Database reconciliation is a superset of data reconciliation.
We frequently use data reconciliation to address the integrity or accuracy of individual records of data quantities. Database reconciliation is frequently discussed at the completion of a data migration effort where we want to understand if the validity of the entire data set is still intact.