Database structure for storing and processing SHM data for SHM analyses and SHM algorithm research

Abstract

Damage identification in aircraft structures is a complex task. Especially in structural components made out of fiber composites and fiber metal laminates the traditional non-destructive methods for the detection of damage are time consuming and expensive. Structural health monitoring (SHM) can potentially reduce maintenance time and cost, but also be an enabler for condition-based maintenance as well as be used as an information source for the digital twin. The guided wave-based SHM System uses a network of transducers spread over the monitored structure. Additional information from other sensing systems may be also recorded. The data is locally acquired on the aircraft and it needs to be pre-processed, stored and accessible for analyzation. A meaningful implementation of SHM requires the integration of the SHM workflow into the aircraft processes as well as machine learning tools to analyze the acquired data in an adequate time frame. The current work focuses on a sever based application including a database which is storing the data and collating other sensing systems to it, as well as data preprocessing steps. Via the application programming interface, the results, with neglectable data volume, can be handed over for local processing or other web-based tools. Latter enables easy accessibility to demonstration and analyzation of the examined structure in different conditions. Furthermore, this infrastructure provides a good base for machine learning algorithm research (supervised classification and neural networks) in order to gain knowledge out of the additional sensing systems data. This data-management and -processing infrastructure is a necessary step towards the ultimate goal of in-time SHM.