Prediction of drilling rate of penetration (ROP) using hybrid support vector regression: A case study on the Shennongjia area, Central China

Rate of penetration (ROP) prediction is crucial for the optimization and control in drilling process due to its vital role in maximizing the drilling efficiency. This paper proposes a novel intelligent model to predict the drilling ROP considering the process characteristics. First, the geological background and the drilling process of the case study are described. Based on the mechanism and frequency spectrum analysis, the strong nonlinearity and different low-frequency and high-frequency data noises between the data variables are detected. After that, the intelligent model is established via three stages. In the first stage, a wavelet filtering method is introduced to reduce these noises in the drilling data. In the next stage, the model inputs are determined by the mutual information method, which significantly decreased the model redundancy. In the last stage, a hybrid bat algorithm is proposed to optimize the hyper-parameters of the support vector regression model. Finally, the proposed model is validated by using the data from a drilling site in the Shennongjia area, Central China. The results demonstrate that the proposed method outperforms eight well-known methods and another three methods without different data preprocessing procedures in prediction accuracy.