DATA VALIDATION & RECONCILIATION

WHAT IS DVR?

DVR is a technology that combines the provided physical information of the studied system/process (chemical species, reactions, thermodynamics, process configuration, physical phenomena, process parameters, equipment characteristics, etc.) in order to build a mathematical model that will be used to adjust the available measurements so as to satisfy all balances and other constraints of the system.

 

DVR utilises the system’s data redundancy as a source of information to reconcile the measurements. Each measurement is corrected as little as possible in order for the corrected measurements to satisfy all the modeled process constraints.

DVR extracts reliable information from otherwise inconsistent raw measurements in order to produce a single consistent set of data representing the studied process’ most probable state.

Data validation denotes all validation and verification actions before and after the reconciliation step. The validation occurs at three different levels :

 

  • Input data filtering ;
  • Results validation ;
  • Gross error detection & elimination.

VALI is an equation-based advanced data validation and reconciliation software which exploits information redundancy and conservation laws to correct measurements and convert them into accurate and reliable information. Unmeasured values are calculated, but VALI also quantifies the precision of reconciled values. In addition, its sensitivity analysis tool shows the interdependence between measurements. Furthermore, VALI can be used online or offline and is integrated in various control systems.

 

This software offers a wide range of usage: upstream, refinery, petrochemical, chemical plants as well as power plants including nuclear power stations. VALI detects faulty sensors and pinpoints degradation of equipment performances (heat rate, compressor efficiency, etc.).

 

Data validation and reconciliation techniques offer many benefits which include:

 

  • Measurement layout improvement;
  • Number of routine physical and chemical analyses decrease;
  • Reduced frequency of sensor calibration – only faulty sensors need to be calibrated;
  • Online optimization tools work with more accurate information;
  • Process data systematic improvement;
  • Early detection of sensors deviation and degradation of equipment performance;
  • Correct plant balances for production accounting and performance monitoring.

Input data filtering

Results validation

Gross error detection & elimination