Modeling the prediction of penetration rate of horizontal directional drilling

In the current practice of horizontal directional drilling (HDD), prediction of the penetration rate (PR) is challenging and relies heavily on practitioners’ experience. Prediction of PR is essential for effective planning and cost management of HDD works. A model that engineers can use to reliably predict PR is not available. This has often led to schedule delays and cost overruns. This study presents a model that quantitatively predicts PR of the penetration process of HDD and accounts for variable ground and operational conditions. On the basis of structured interviews with HDD experts along with the analytic hierarchy process (AHP), the relative importances of the ten factors that significantly affect PR are assessed. A method of scoring PR for each of the significant factors is also presented. The model integrates the obtained scores of PR for the conditions of a given HDD project with the relative weights of the significant factors in a clear methodology to predict PR. Actual records of PR are compared with the corresponding predicted values of PR for seven HDD projects. The comparison shows consistent results and implies a correction relationship with R2 of 0.893 and RMSE of 0.066 m/min. The leave-one-out cross-validation (LOOCV) is performed to examine the model’s robustness. A sensitivity analysis is also made to highlight the impact of variations in the factors’ weights and scoring thresholds. The presented model renders engineers capable of assessing the effects of variations of important factors, such as the type of drilling equipment or type of drilling tools, on the drilling performance. This essentially provides for better drilling performance and reduces uncertainty in scheduling and cost estimation.