Hybrid Model of Machine Learning Method and Empirical Method for Rate of Penetration Prediction Based on Data Similarity

• DOI: 10.3390/app13105870
• Publication Date: 2023-05-11T05:37:24Z
• Authors: Fei Zhou; Honghai Fan; Yuhan Liu; Hongbao Zhang; Rongyi Ji
• Keywords: Robustness; Autoencoder; Rate of penetration

The rate of penetration (ROP) is an important indicator affecting the drilling cost and performance. Accurate prediction ROP has guiding significance for increasing speed reducing costs. Recently, numerous studies have shown that machine learning techniques are effective means to accurately predict ROP. However, in petroleum engineering applications, its robustness generalization cannot be guaranteed. traditional empirical model good ability. Based on quantification data similarity, this paper establishes a hybrid combining method method, which combines high accuracy with overcoming shortcomings any single model. AE-ED (the Euclidean Distance between input reconstructed from autoencoder model) defined measure according similarity each new piece data, chooses corresponding calculate. results show better than model, all evaluation indicators perform better, making it more suitable field.