Real-Time Prediction Using a Multivariable Grey Prediction Model

Abstract

In recent years, the challenge of managing annular pressure in high-temperature, high-pressure gas wells has become increasingly significant for the global petroleum industry. While extensive research has explored the formation mechanisms and indirect calculation methods of annular pressure, real-time prediction remains underdeveloped. This study addresses this gap by applying grey system theory, specifically grey correlation analysis, to identify key variables affecting annular pressure changes in gas wells. A multivariable grey prediction model was developed to enable early diagnosis and proactive management of abnormal annular pressure. Using 12-hour measured data from a high-pressure gas well in northwest Sichuan, the model demonstrated a maximum prediction error of 0.65%, confirming its effectiveness and providing a novel method for dynamic annular pressure management. The findings underscore the potential of grey system theory in enhancing safety and efficiency in gas well operations.