Diagnosis of clamp looseness for hydraulic pipelines using fiber-Bragg-grating-based strain measurement: A feasibility study

Clamp looseness in hydraulic pipelines usually accelerates the wear process of pipelines and results in the fatigue fracture of pipelines. Conventionally, the condition monitoring of pipelines is often difficult and expensive due to the long length of pipes and large number of clamps that, in turn, requires a large capacity sensing network. In recent years, the fiber Bragg grating based sensing technique has begun to attract interests because of its cost-effective and ease of forming the large capacity sensing network. Therefore, this study investigates the feasibility of a new pipeline monitoring system instrumented with fiber Bragg grating sensors, which can measure the dynamic strain along the pipeline. The experiments were performed on a laboratory hydraulic pipeline system, where the impact hammer was used to excite the vibration of pipes. The preliminary selection of significant feature from the measured strain data is made with the use of principal component analysis. Then, the extracted feature with the reduced dimension is analyzed for the diagnosis of pipelines, which utilizes support vector machines to classify the location and severity of clamp looseness. The experimental results show the proposed sensing technique and diagnostic algorithm have a promising potential in developing an economic and reliable monitoring system for online diagnosis of hydraulic pipelines.