Reliable ultrasonic non-destructive testing and asset management of utility wood poles

点击翻译按钮获取中文摘要

An assessment of the condition of wood poles is a critical step in the development of successful and safe maintenance programmes. Electricity companies consider wood poles to be an asset with high value and high risk of failure. Presently, the condition assessment of wood poles is based on visual inspection and sounding tests using a hammer to detect hollow sections. These inspection tests are subjective; whereas, intrusive tests are expensive and intrusive. Utility owners urgently need a reliable technique to perform cost-effective maintenance and rehabilitation programmes of wood poles. This project is specifically aimed at developing a new methodology for the condition assessment of wood poles. A robust experimental and analytical methodology is proposed; which is based not only on the latest ultrasonic testing methods but also on the latest probabilistic data analysis techniques. Ultrasonic tests are used to measure the residual strength of individual wood poles in-situ; whereas statistical methods are used to assess the reliability of the wood-pole infrastructure as a whole. In ultrasonic testing, changes in medium properties are evaluated from a single point of information out of thousands that are actually recorded. Because the response of ultrasonic transducers is not directly proportional to the actual motion of the medium. A state-of-the-art nano-laser system (recently purchased, RTI grant, 2011) will be used to measure the actual input excitations and medium responses generated by ultrasonic transducers in wood poles. From the actual input and output responses, the internal condition of a wood pole will be evaluated. The proposed project will reduce the costs of the Canadian power generation industry by introducing cost-effective maintenance and rehabilitation operations. The ultrasonic data set generated by an array of ultrasonic transducers attached to a calibrated belt will be analyzed using advanced signal processing techniques to extract the main features from the waveforms; these features will then be used to train a set of statistical classification algorithms. From the statistical classification, an image of the decayed areas in a wood pole will be displayed and the remaining strength evaluated.