INTRODUCTION TO DATA RECONCILIATION AND GROSS ERROR DIAGNOSIS Process Data Conditioning Methods

In any modern chemical plant, petrochemical process or refinery, hundreds or even thousands of variables such as flow rates, temperatures, pressures, levels, compositions, etc. are routinely measured and automatically recorded for the purpose of process control, online optimization or process economic evaluation. Modern computers and data acquisition systems facilitate collection and processing of a great volume of data, often sampled with a frequency of the order of minutes or even seconds. The use of computers not only allows data to be obtained at a greater frequency, but has also resulted in the elimination of errors present in manual recording. This in itself has greatly improved the accuracy and validity of process data. However, the increased amount of information can be exploited for further improving the accuracy and consistency of process data through a systematic data checking and treatment.