Predicting the ultimate failure load Of composite hardware using Artificial neural network and Acoustic emission data

Abstract

Proof testing of composite structures is complicated by the fact that newlinemost composite structures do not exhibit the same elastic plastic behaviour newlinefound in metal structures Excluding macroscopic discontinuities as long as newlinethe stress is kept below the yield point there is little plastic deformation and newlinetherefore no noticeable degradation in the structural integrity of metal newlinestructures This phenomenon does not hold true for fiber matrix composites newlinein which the structural integrity begins to degrade as soon as the fibers begin newlineto break The common proof testing loads of 70 80 percent of the expected newlinefracture strength used on a metal design can cause significant fiber failures in newlinecomposite structures thereby degrading its structural integrity In order to newlineavoid this a procedure needs to be adopted that uses reasonably lower proof newlineloading for composites and that would also accurately determine the ultimate newlinestrength of the structure Acoustic Emission AE study is a high sensitivity nondestructive newlinetesting technique for detecting active microscopic events in a material newlineAcoustic emission signals collected from composite hardware during proof newlinetesting were interpreted with an Artificial Neural Network ANN and the newlineultimate failure strength was predicted The experimental research was started newlinewith the failure strength prediction attempt on composite tensile coupons newlineSubsequently the same methodology was followed to a burst pressure newlineprediction on Glass Fiber Reinforced Plastic GFRP pressure vessels newline newline