The erosion rate prediction for the elbow in shale gas gathering and transportation system: RSM and GA-BP-ANN modeling

Abstract

Elbow erosion caused by sand mixed in shale gas is a significant issue faced by shale gas gathering and transportation systems, posing a serious threat to the safe and stable production of shale gas. In this paper, the factors affecting elbow erosion in the shale gas gathering and transportation system are identified. The computational fluid dynamics (CFD) two-way coupling Euler-Lagrange method is used to investigate the single-action of various parameters on particle trajectory, erosion morphology, and maximum erosion rate, including pipe inner diameter, pipe R/D ratio, pipe bending angle, pipe bending orientation, solid particle velocity, solid particle size, and solid particle mass flow rate. Subsequently, the response surface methodology (RSM) is employed to quantitatively explore the synergistic interactions between these factors that affect the maximum erosion rate of the elbow, and a surrogate model for simple and rapid evaluation of the maximum erosion rate for elbows at the engineering level is proposed. Finally, the backpropagation artificial neural network (BP-ANN) optimized by genetic algorithm (GA) is used to train the samples obtained by Latin hypercube sampling (LHS) in the multi-factor range. By comparison, the RSM surrogate model and the GA-BP-ANN model reveal a favorable level of agreement with the experimental data, while considering various influential factors. It is important to note that each method possesses distinct advantages, limitations, and specific application environments. This research can provide quantitative references for preventing and controlling erosion within the shale gas gathering and transportation system.