Intelligent Diagnosis of Bolting System Under Dynamic Impact Available to Purchase

The properties of fasteners as significant parts in varied engineering systems have been widely investigated [1–3], from hard disk drive devices to construction ground control. These problems are complex in nature because every bolting involves different sources of nonlinear and uncertainty characteristics. The interfaces forces such as contact forces, friction forces and bonding are not known in reality. The base forces and deformation could be redistributed non-uniformly in the presence of complex loadings such as shock and impact. Most of the reported studies focused on the design issues, characterization of the linear dynamic properties and energy dissipation of bolting system as well as using bolts with integrated detecting systems or dynamometers. The capability to assess the interface properties of bolting system is important for mechatronics, mechanical, civil and mining engineering [1–4]. Even though some conventional approaches have been used to infer bolting integrity, the diagnosis technology has been lacking. This study is conducted to explore the possibility of the diagnosis bolting interface integrity under impact by using accelerometer signal. The statistical pattern identification such as artificial neural network and support vector machine is used to diagnose the bolting integrity. By integrating the analysis and experimental data, an ANN is established as a nonparametric model to predict the system properties. Numerous numerical and experimental researches have been conducted to characterize the typical bolting system, which enables extraction of varied dynamic features from different mechanisms associated with the failures. This kind of database could be used as feature to characterizing the effect of complex loads on bolting for ANN training. To further illustrate the feature extraction, we investigated system models. Due to its adaptive and nonlinear input–output transformation capabilities, artificial neural network (ANN) has been widely applied in the field of pattern and system recognition. The proposed approach is capable of monitoring the stress/strain history and integrity of bolting interface with the goal of detecting structural damage and defects. The results from simulation, testing and ANN identification demonstrated the high performance of the proposed approach compared with conventional ones. In the following, an approach is proposed to reliably estimate the dynamic properties of a bolt-surrounding solid mass specimen using its impact response signal. The developed approach can be readily extended to the bolting connection in other systems such as mechanical and mechatronics systems.