Title : Development and validation of an in-house Python workflow for pressure transient analysis of oil wells against commercial software (KAPPA Saphir)
Abstract:
Pressure transient analysis (PTA) is a core tool for characterizing reservoir properties, diagnosing well condition and guiding stimulation and IOR decisions. Commercial interpretation packages are effective but costly, and their results can depend on licence availability and individual workflow. This paper presents an in-house, Python-based PTA engine developed to give field teams a transparent, repeatable and fast alternative for build-up test interpretation. The tool runs in an interactive Google Colab environment. It ingests a consolidated workbook of well properties, production history and high-frequency gauge data. It then applies hydrostatic datum correction and variable-density downsampling, reconstructs the multi-rate flow history, and computes superposition (Horner) time and Agarwal equivalent time. Diagnostics use a Bourdet-type derivative on log-log plots. Permeability and skin are estimated with Theil–Sen robust regression on the Horner plot, with bootstrap confidence intervals on permeability. The engine also estimates wellbore storage, average reservoir pressure (Dietz method), productivity index, flow efficiency, radius of investigation and drainage volume, and it screens for boundaries. Homogeneous, dual-porosity (Warren–Root), sealing-fault and radial-composite models are available. The workflow was validated on twelve vertical oil wells (Wells X1 to X12) by comparing its results with KAPPA Saphir analytical and semilog interpretations. The wells cover a wide range of conditions, from low to high permeability and from stimulated to severely damaged. For both permeability and skin, the Python results were in close agreement with Saphir in every well, with differences within the range normally expected between two interpreters using the same software. The Python permeability also fell close to the spread between Saphir’s own analytical and semilog results. These results show that the in-house code reproduces commercial-software interpretations for vertical oil wells while keeping the full calculation path open to audit. Extension of the code to gas wells (real-gas pseudo-pressure and non-Darcy skin) and horizontal wells (linear and bilinear flow regimes) is in progress. The authors see the tool as a way to reduce interpretation turnaround time and support timely stimulation and IOR decisions.
