In a simply manner, data reconciliation is a mathematic treatment with propose of a better quality of the data in a process. Industrial processes typically have a large number of measured variables, which presents some degree of random errors and, less frequently, gross errors. In this text, in order to simplify the notation and terminology we classify all instrument and process errors in these two categories. Any significant systematic bias is included in the gross error category. Data reconciliation allows the measurements to be adjusted (“reconciled”) to satisfy process restrictions (mass and energy balances) and improve measurements quality. The results obtained by data reconciliation can also provide benefits in custody transfer issues. Custody transfer is the responsibility transfer during the storage and transportation of a measured refined product volume. Any loss or gain resulting in a non-trustful measurement is considered as the transportation company responsibility. Therefore, the work objective is to propose a data reconciliation methodology, in a process involving diesel oil custody transfer in a Transpetro’s terminal (Terminal of Sao Caetano do Sul), in order to evaluate and correct possible inconsistencies, besides to know a single measure that represents better the measurement system. In this study we will use data from static measurement in tanks, dynamic measurement in turbine and ultra-sonic device. A database will be obtained in two basic steps: process modeling and data reconciliation to consolidate the mass balance. The reconciled value shows us that there is a bias in the ultra-sonic meter and the turbine meter measurement is more reliable, as expected.

This content is only available via PDF.
You do not currently have access to this content.