A Developed Robust Model and Artificial Intelligence Techniques to Predict Drilling Fluid Density and Equivalent...

Area of Science:

• Petroleum Engineering
• Drilling Operations
• Well Control

Background:

• Regulating formation pressure and preventing kicks are critical in deep well drilling, especially in high-pressure, high-temperature environments.
• Equivalent Circulation Density (ECD) control is vital, particularly in narrow pore-fracture pressure windows.
• Current downhole ECD measurement methods are expensive and operationally constrained.

Purpose of the Study:

• To develop novel models for predicting ECD and mud weight (MW) using surface drilling parameters.
• To overcome the limitations of existing downhole ECD measurement techniques.
• To provide a cost-effective and efficient solution for monitoring drilling fluid properties.

Main Methods:

• Developed two novel models: ECD_effc.m for ECD prediction and MW_effc.m for MW prediction.
• Utilized surface drilling parameters: standpipe pressure, rate of penetration, drill string rotation, and mud properties.
• Employed Artificial Neural Network (ANN), Support Vector Machine (SVM), and Decision Tree (DT) for model development and validation.

Main Results:

• ECD was estimated with a correlation coefficient of 0.9947 and an average absolute percentage error of 0.23% using ANN and SVM.
• MW was estimated with a correlation coefficient of 0.9353 and an average absolute percentage error of 1.66% using DT.
• The developed models demonstrated higher accuracy compared to existing artificial intelligence (AI) techniques and aligned well with Pressure-While-Drilling (PWD) tool data.

Conclusions:

• The novel ECD_effc.m and MW_effc.m models accurately predict ECD and MW from surface data.
• These models offer a cost-effective alternative to expensive downhole sensors and software.
• The models can be applied during well design and drilling operations to optimize mud weight and ECD, saving time and resources.