Strain-Data-Driven Force Reconstruction Using Pseudo-Inverse Matrix

This chapter presents a comprehensive study that focuses on developing and rigorous assessment of a predictive model tailored for estimating bending loads within structural systems. The model leverages strain measurements as input parameters, processed through the pseudo-inverse matrix technique, resulting in highly accurate load predictions. The investigation explores the model's performance across various scenarios, with particular emphasis on its sensitivity to noise and its impact on prediction accuracy. This chapter highlights the model's sensitivity to point load placement (L-f), demonstrating that shifting the point load away from the boundary condition reduces load sensitivity, while positioning it closer heightens sensitivity, making it more susceptible to noise. Moreover, the research investigates scenarios involving sensor removal and malfunction, showing their substantial impact on load prediction. Additionally, the study extends the model's capabilities to predict temperature changes based on thermal strain inputs, achieving close alignment with actual values. In summary, this predictive model proves robust in estimating bending loads, but its sensitivity to noise, load placement, and the importance of sensor quality are emphasized, along with its versatile application in temperature change prediction.