Damage assessment in hyperbolic cooling towers using mode shape curvature and artificial neural networks

Abstract

Hyperbolic cooling towers are large thin shell reinforced concrete structures that are used to remove the heat from wastewater and transfer it to the atmosphere using the process of evaporation. During its long service life, a cooling tower can experience damage due to the large temperature variations, environmental degradation, or random actions such as impacts or earthquakes. Such a damage can develop over time and result in the sudden collapse of the cooling tower. To ensure that a cooling tower operates safely and efficiently at all times, it is important to monitor its structural health. In this context, structural health monitoring based on the vibration characteristics of the structure, has emerged as a useful method to detect and locate damage in structures. Hyperbolic cooling towers, due to their particular shape, exhibit rather complex vibration characteristics that do not suit the traditional vibration-based damage detection techniques. This paper develops and applies a damage assessment method using the absolute changes in mode shape curvature (ACMSC) in conjunction with Artificial Neural Networks (ANNs) to detect, locate, and quantify damage in hyperbolic cooling towers. ANN is a machine learning technique that can predict behavioural patterns using a set of data samples and finds use in the damage quantification process. The proposed method for detecting and locating damage is experimentally validated and demonstrated its capability to accurately detect and locate damage. A feed-forward network having one hidden layer with Bayesian algorithm is used to train the artificial neural network. Damage indices calculated from noise polluted mode shape data are used to train the network. The trained network is then used to successfully assess the unknown damage severities in the cooling tower. The outcomes of this paper will enable early warning of damages in the cooling towers and will help towards their safe operation.