Presentation Description: Many pretensioned bolt connections are used in wind turbines structures to join different wind tower components. For bolts used to fasten or hold two large structural members together, engineers specify a design tension at which each bolt should be set and at which each bolt should be maintained over time. It is therefore critical to perform periodic audits to ensure that the residual tension of bolts remains sufficient to meet design specifications. The current maintenance methods are labor, time, and cost intensive due to the involvement of hydraulic jacks and torque wrenches. While there have been significant efforts in estimating the residual tension of bolts using Non-Destructive Testing (NDT), NDT methods for predicting bolt tension without baseline measurements or bolt manipulation are virtually nonexistent. The objective of the present work is to predict residual tension of in-situ bolts using NDT methods. These systems and methods eliminate or minimize the inefficiencies and safety hazards of all known methods, as well as provide insight as to the residual tension of bolts used on a wind tower. To this end, an ultrasonic method is used to determine the residual tension of bolts based on the acoustoelastic effect. A post processing software is developed to ensure the consistency and quality of the signals and perform signal analysis. Finally, the regression software tunes a regression model hyperparameters using cross-validation, and reports tension predictions.
Learning Objectives:
An NDT method is developed to estimate the residual tension of an in-situ bolt using the impact of static stress on ultrasound velocity.
The effect of stress on wave speed is material and direction dependent. Using two types of ultrasonic waves, longitudinal and shear waves, removes the need for baseline measurements and bolt manipulation.
Moreover, utilizing Artificial Intelligence (AI) in regression analysis provides the applicability of the product for different bolt sizes and lengths.