Sensor Placement Strategy for Pipeline Condition Assessment Using Inverse Transient Analysis

Abstract

Inverse transient analysis (ITA) has been recognized as a useful technique for pipeline condition assessment, such as leak detection and pipe wall thickness estimation. The effectiveness and accuracy of the inverse analysis are dependent on the sensor placement design; however, previous research on this topic is limited. This paper investigates how the number and location of pressure sensors affects the identifiability of pipeline parameters in the ITA approach. An analytical analysis demonstrates that infinite pipe parameter combinations can produce almost the same pressure responses at specific observation locations, which means that the identifiability of the pipe parameters will be poor if sensors are installed at these locations. Numerical sensitivity studies and multiple ITA case studies are conducted to investigate the relationship between the sensor locations and the parameter identifiability. It is found that at least three sensors are needed, and given the first two sensors are N reaches apart (i.e. N pipe segments in the inverse model), the third sensor should not be placed at nodes that are separated from any of the first two sensors by an integer multiple of N reaches.