International Research Journal of Engineering and Technology ( I RJET) e-ISSN: 2395 -0056 Volume: 02 Issue: 03 | June-2015 www.irjet.net p-ISSN: 2395-0072 © 2015, IRJET.NET- All Rights Reserved Page 2060 Real-Time Damage Detection in Laminated Composite Beams Using Dynamic Strain Response and Modular Neural Arrays for Aerospace Applications Sanjay Goswami 1, 2 and Partha Bhattacharya 2 1 Assistant Professor, Department of Computer Applications, Narula Institute of Technology, Agarpara, Kolkata-109, India 2 Associate Professor, Department of Civil Engineering, Jadavpur University, Kolkata-32, India ---------------------------------------------------------------------***--------------------------------------------------------------------- Abstract Damage detection in multi-layer laminated “Glass Fiber – Epoxy Resin” composite beams have been studied for Structural Health Monitoring applications. A numerical model of the beam is developed using Finite Element Method (FEM). The FEM model is used to simulate damages in the structure, and also mechanical vibrations, actuated at one end of it. Strain responses corresponding to the vibrations are picked up at a different location and are studied to identify the location and severity of the damage. Since multiple parameters like Location, Layer and Severity of the damage are to be identified a single array of multiple neural network units (termed neural-array) is initially proposed which fell short of producing desir ed identification success rates. To address this short coming, a collaborative group of parameter-specific multiple neural arrays is proposed, which finally produced fairly impressive damage identification results. Keywor ds: Structural Health Monitoring; Glass Fiber – Epoxy Resin Composite; Strain Response; Finite Element Method; Modular Neural Network Array. 1. INTRODUCTION Structural Health Monitoring (SHM), an emerging interdisciplinary field, incorporates techniques from computer science, electronics and electrical engineering to solve structural problems related to monitoring of in- service civil structures such as bridges, buildings, aerospace vehicles, etc. The objective is to detect and identify damages that may occur in these structures while in operation. If some levels of damages are detected early, suitable disaster prevention mechanism can be initiated to save public lives. In that sense, it is a mission critical technology. The field of SHM involves study of several individual structural components, such as beams, plates, shells, etc., and integrating them to provide a holistic damage/ failure detection mechanism. Several approaches have been proposed in the open literature by different researchers, ranging from tap testing to modern day acoustics, x-rays, ultrasound and vibration signal processing based techniques. Among them vibration based techniques have been found to be most effective in addressing aerospace domain problems (Farrar et al [01], Carden et al [02], Raghavan et al [03]). Qiao [04] presented some of the most relevant signal processing techniques suitable for vibration based damage analysis. Staszewsky et al [05] elaborated the utility of signal processing based techniques in detecting damage especially in aerospace structures. Sohn et al [06] and Taha et al [07] emphasized the effectiveness of wavelet transform in delamination detection of composite structures. The problem of structural damage detection is essentially a statistical pattern recognition problem, which is supported by the works of Carden et al [02], Raghavan et al [03] and Fan et al [08]. Modern pattern recognition approaches ranging from genetic algorithms, neural networks and support vector machines, suitable for damage detection problems have been discussed in the studies by Bakhary et al [09] and Liu et al [10]. The concept of neural network arrays in pattern recognition has been discussed in Sharkey [11]. Goswami et al [12] investigated the detection of single damages in a structure made of isotropic material (steel). Only two damage parameters, namely Zone and Extent, were identified. In another related work Goswami et al [13] extended the study of [12] to detect multiple damages in a similar type of structure. In the manufacturing of modern day aircrafts and automobiles, laminated composite structures are being increasingly used as they are light weight and have a very high strength-to-weight ratio. Hence, there is a need to extend further the study of damage identification in laminated structures. The laminated structure consists of