DVR is a technology that combines physical information about the studied system or process, including chemical species, reactions, thermodynamics, process configuration, physical phenomena, operating parameters, and equipment characteristics, to build a mathematical model that adjusts available measurements so they satisfy all system balances and constraints.

Data Validation and Reconciliation (DVR) uses data redundancy within a system as a source of information to reconcile measurements. Each measurement is adjusted as little as possible so that the corrected values satisfy all modeled process constraints. Through this approach, DVR extracts reliable information from otherwise inconsistent raw data and produces a single, coherent dataset representing the most probable state of the studied process.

Data validation encompasses all verification and validation actions performed before and after the reconciliation step. Validation is carried out at three distinct levels: input data filtering, results validation, and gross error detection and elimination.

VALI is an equation-based advanced data validation and reconciliation software that leverages information redundancy and conservation laws to correct measurements and transform them into accurate and reliable information. In addition to reconciling measured values, VALI calculates unmeasured variables and quantifies the precision of reconciled results. Its sensitivity analysis tools reveal interdependencies between measurements, providing deeper insight into process behavior. VALI can be deployed online or offline and integrates with control systems.

The software supports a wide range of applications across upstream operations, refineries, petrochemical and chemical plants, and power generation facilities, including nuclear power stations. VALI detects faulty sensors and identifies equipment performance degradation, such as heat rate losses or reduced compressor efficiency.

Data validation and reconciliation techniques deliver significant benefits, including:

• Measurement layout improvement

• Reduced physical and chemical analyses

• Lower sensor calibration frequency, with calibration focused only on faulty sensors

• Improved accuracy for online optimization tools

• Systematic improvement of process data quality

• Early detection of sensor drift and equipment performance degradation

• Accurate plant balances for production accounting and performance monitoring