CLASSIFICATION OF SYSTEMATIC MEASUREMENT ERRORS WITHIN THE FRAMEWORK OF ROBUST DATA RECONCILIATION

LLANOS, CLAUDIA ELIZABETH; SÁNCHEZ, MABEL CRISTINA; MARONNA, RICARDO A.

INDUSTRIAL & ENGINEERING CHEMICAL RESEARCH

AMER CHEMICAL SOC

Año: 2017

0888-5885

A Robust Data Reconciliation strategyprovides unbiased variable estimates in the presence of a moderate quantity of atypicalmeasurements. But estimates get worse if systematic measurement errors that persistin time (e.g., biases, drifts) are undetected and the break down point of therobust strategy is surpassed. The detection and classification of those errorsallow taking corrective actions on the inputs of the Robust Data Reconciliationthat preserve the instrumentation system redundancy while the faulty sensor isrepaired. In this work, a new methodology for variable estimation andsystematic error classification, which is based on the concepts of RobustStatistics, is presented. It has been devised to be part of the real-timeoptimization loop of an industrial plant, therefore it runs for processoperating under steady state conditions. The Robust Measurement Test is definedin this article and used to detect the presence of sporadic and continuoussystematic errors. Also the Robust Linear Regression of the data contained in amoving window is applied to classify the continuous errors as biases or drifts.A performance analysis of the proposed methodology is presented for linear andnon-linear benchmarks. Results highlight the performance of the proposed methodology to detect and classify outliers, biases and drifts for linear and nonlinear benchmarks