A Novel Workflow for Interpretation of Single-Well Tracer Tests

Abstract As offshore oil fields reach maturity, enhanced oil recovery (EOR) becomes crucial for maintaining oil production. However, its widespread adoption is often hindered by high uncertainty, significant costs, and lengthy evaluation periods. Single-well tracer tests (SWTT) present a viable solution to these challenges by offering a fast, accurate, and cost-effective method for evaluating waterflood and EOR efficiency in the field. In this process, a reactive tracer is injected into the formation and produced back after a few days. The separation between the tracer and its byproduct serves as a reliable indicator of residual oil saturation (Sor). SWTT is particularly well-suited for offshore applications due to its simplicity, rapid results, and high accuracy. However, interpreting these results can be challenging due to geological heterogeneity. To address this, the research proposes a simplified, fully automated workflow for interpreting SWTT results by combining numerical modeling with advanced optimization algorithms. A comprehensive sensitivity analysis was conducted to identify and assess potential factors contributing to interpretation errors in numerical simulations. It was found that injection volume and dispersitivity could potentially lead to non-unique solutions. Based on the observations, the proposed workflow streamlines the traditionally multi-step interpretation process into a single step with a modified error function that emphasizes the key features of the tracer curve to address the non-uniqueness problem. Two synthetic cases were first investigated to validate the capability of the new platform and compare the performance of two metaheuristic algorithms. The Differential Evolution algorithm demonstrated superior accuracy and stability compared to the Catastrophic Adaptive Genetic Algorithm due to its aggressive evolution strategy and provided more satisfactory results for both cases. The new workflow has shown significant improvements in accuracy when tested in field cases by addressing the non-uniqueness problem. The estimated Sor values aligned well with well-logging data and do not require any human operation, confirming the reliability and effectiveness of the approach. The fully automatic SWTT interpretation workflow has proven feasible through successful field trials. This advancement offers a more efficient and accurate method for evaluating EOR efficiency in offshore oil fields.