Beyond Technical Performance: A Hybrid Efficiency Model for HPHT Drilling Fluids

Drilling-fluid selection in high-pressure/high-temperature (HPHT) wells is usually assessed through technical performance and chemical cost, while time overruns and the consequences of fluid losses are treated separately. This study develops a hybrid efficiency framework that combines a Technical Performance Index (TPI), built from mud-weight, loss, time, and hole-stability compliance and weighted with the Analytic Hierarchy Process (AHP), with a Total Real Cost (TRC) that adds an unplanned-time cost proxy and a loss-based environmental cost proxy to chemical cost. The Hybrid Efficiency Index (HEI) is the ratio of the two. The framework was applied to the comparable 121⁄4-in. sections of two anonymized land wells in Northern Kuwait (K-300) and the Eastern Venezuela Basin (V-100), measuring performance against design targets rather than raw operational values. An initial compilation of the inputs favored K-300 by an HEI ratio of 1.65. An audit of every input against the source documents found that the design programs belonged to analogous wells, that the V-100 loss figure combined all loss categories (3,146 bbl, of which 1,153 bbl went to the formation), and that the stability events used did not match the records of the interval. With audited inputs, the TRC was USD 359,485 for K-300 and USD 260,356 for V-100, and the HEI ratio was 0.77, favoring V-100. This result was insensitive to criterion weights, the stability penalty, the mud-weight metric, and the environmental proxy, but reversed (1.23) when the analogous plan was applied without scaling by drilled length; the full audited range was 0.53–1.42. The comparison between the two wells has no competitive intent: it is not a ranking or a competition between countries, regions, operators, or contractors, but a controlled methodological illustration showing that loss accounting and the time reference can outweigh the fluid data themselves, and that traceable data provenance makes such errors detectable. The results are illustrative and do not establish causality.