Algorithm for Detecting Multiple Partial Blockages in Liquid Pipelines by Using Inverse Transient Analysis

Abstract

SummaryPartial blockages form on the inner wall of the crude-oil pipelines as a result of asphaltene precipitation, scale deposition, and so forth. If not controlled and rehabilitated periodically, these partial blockages can have a serious adverse effect on the efficiency, economy, and safety of the operation of the pipeline. Before each rehabilitation operation, the detection of the local flow-condition deterioration (change in diameter) is necessary for efficiency and economy considerations, especially for long-distance subsea crude-oil pipelines. Most conventional detection techniques require the installment of detecting devices along the pipeline. However, they are economically expensive and even technically impossible for pipelines in operation. The present work focuses on an economically efficient technique that can realize remote nonintrusive measurement (i.e., the pressure-wave technique). The purpose of our research is to develop a method for calibrating multiple irregular partial blockages inside the liquid pipe by using the pressure response in the time domain at certain measuring points along the pipe under the transient state. The method involves the direct problem and the inverse problem. The direct problem is the simulation of the transient flow in the liquid pipe with single or multiple partial blockages. A second-order direct problem solver is developed in the framework of the Godunov-type finite-volume method (FVM). The inverse problem is to determine the partial-blockage distribution by using the pressure response at the measuring point under transient conditions. Our algorithm to solve the inverse problem comprises analytical evaluation and optimization. The analytical evaluation provides a reliable search space for the following optimization procedure, and thus effectively alleviates the local optimum problem. Numerical results demonstrate the efficiency and accuracy of proposed methods for solving the direct and inverse problems.