5 Real-Time and Web-Based Structural Damage Detection Network for Multiple Structures 43 5.3 Data Collection and Visualization on the Network A pertinent feature of the monitoring system is to have easy access to data on the framework. The developed framework is an integrated system consisting of a data collection network with storage and visualization capabilities. The system enables the users to collect, store, and conduct analysis through a user interface with a graphical display of real-time and archived data (suitable for permanent monitoring). For the monitoring framework, the system is built to be compatible with MATLAB Data Acquisition Toolbox [53], making it a good fit for a wide range of data acquisition devices. The system is then able to perform the following: • Collect acceleration data from multiple laboratory structures under ambient conditions. • Analyze the raw acceleration data in real time for structural damage detection and provide information on the location and level of damage. Store the information and publish it on a database. • Enable monitoring of the structural condition of multiple structures in real time on the internet. 5.4 System Components, Database, and Web-Based Application For the real-time web-based system, the major components of the MATLAB-based application are hardware for data collection and processing, signal processing algorithms, online database, and web application (Fig. 5.3). First, the sensor network is deployed on each structure which is operated by multiple data acquisition systems connected to a computer. A high-speed internet connection is required to process and share data on the framework. MATLAB Database Toolbox is also needed to be installed to enable communication with the online MySQL database [54, 55]. In the framework, the network for each structure is responsible for processing the collected acceleration data locally to extract the damage indices, assess vibration levels, and then publish them to the online database in real time. Various damage detection algorithms were used to extract damage indices in real time, working directly on the measured acceleration signals from monitored structures [56]. The damage index represents the probability of damage at various locations on the structures through which the existing structural condition can be assessed. The authors implemented the one-dimensional convolutional neural network (1D-CNN) algorithm [57–61], which processes the raw acceleration signal collected from the laboratory Vibration Data Vibration Data Vibration Raw Signals Damage Indices Vibration Levels MATLAB MATLAB MATLAB PC 2 PC 1 PC N Raw Signals SQl Query Real-time Data Real-time Structural Health Monitoring Damage Indices Vibration Levels Raw Signals Damage Indices Vibration Levels Data Fig. 5.3 Real-time web-based damage detection system
RkJQdWJsaXNoZXIy MTMzNzEzMQ==