Analytical methods for water in crude oil: progress, challenges, and industrial relevance

Abstract

Water determination in crude oil is essential for custody transfer, corrosion control, refinery operation, emulsion management, and production optimization, but reliable measurement is difficult because water may exist as free, emulsified, or dissolved phases. The reported value is also affected by sampling integrity, crude-oil composition, matrix interference, emulsion stability, and calibration drift in online sensors. This review critically evaluates Dean–Stark distillation, Karl Fischer titration, centrifuge testing, infrared, microwave, capacitance, neutron-based, microfluidic, and sensor-fusion methods for water-in-crude-oil analysis. The methods are compared according to operating principle, sensitivity, uncertainty sources, field applicability, calibration requirements, and regulatory relevance. Dean–Stark distillation remains the most broadly accepted reference approach for crude-oil water determination, particularly for emulsified systems, whereas Karl Fischer titration provides higher sensitivity for low-water samples and separated oil phases. Centrifuge testing is useful for rapid field screening but may underestimate water in stable emulsions. Online analyzers enable real-time monitoring but require crude-specific calibration and periodic laboratory anchoring. The novelty of this review is the integration of classical and emerging methods with sampling-uncertainty analysis, multi-criteria decision analysis, value-of-information-based audit scheduling, and artificial-intelligence-assisted recalibration. In an illustrative simulated recalibration case, root-mean-square error decreased from 0.082 to 0.031 v/v %, bias decreased from − 0.041 to − 0.006 v/v %, and expanded uncertainty decreased from ± 0.12 to ± 0.045 v/v %. These results demonstrate the potential of reference-anchored digital correction, while emphasizing the need for field validation before fiscal or regulatory use. Reliable water determination therefore requires a traceable workflow combining representative sampling, reference laboratory methods, fit-for-purpose field tools, and auditable digital calibration.