A novel modeling and drilling optimization method with suitable constraints in geological well. (May 2022)

• Record Type: Journal Article
• Authors: Zhou, Yang; Chen, Xin; Wu, Min; Cao, Weihua
• Is Part Of: Control engineering practice. Volume 122(2022)
• Publication Date: 2022-05
• Subjects: Mud pit volume -- Rate of penetration -- Support vector regression -- Drilling optimization -- Fuzzy comprehensive evaluation; Automatic control -- Periodicals; 629.89
• DOI: 10.1016/j.conengprac.2021.105062
• Languages: English
• ISSNs: 0967-0661

Abstract

Abstract: Drilling is of great importance in developing resources. Rate of penetration (ROP) and mud pit volume (MPV), which reflect drilling efficiency and safety, need to be predicted accurately and maintained in appropriate ranges. Due to the complexity and coupling in the drilling process, drilling safety and efficiency are difficult to improve simultaneously and reliable nonlinear prediction models also need to be constructed. Moreover, the values of optimization objectives are needed to manipulated in proper ranges. This paper presents a modeling and optimization method to solve these problems. In the modeling part, a novel method based on hybrid bat algorithm (HBA) and support vector regression (SVR) is developed to build prediction models, one for predicting ROP and the other for MPV. Then, the objectives intervals are obtained by the fuzzy comprehensive evaluation method, which analyzes the influences of multiple variables. Finally, a novel drilling optimization method is developed to improve ROP and reduce MPV fluctuations according to the objectives intervals and nondominated sorting genetic algorithm II (NSGA-II). The simulation results indicate that prediction models constructed by the developed method have superior performance than other modeling methods. Moreover, the developed method can improve ROP by an average of 35.8% and reduce the fluctuations of the MPV by an average of 34.3%, which means the developed method can improve safety and efficiency at the same time. The developed method is also applied in drilling equipment, the running results further illustrate the validity. Highlights: A reliable modeling method is developed to predict ROP and MPV. The suitable objectives intervals are determined by analyzing the influence of multiple variables. A drilling optimization method is developed to improve ROP and suppress MPV fluctuations.