| OCR Text |
Show 165 THE UNIVERSITY OF UTAH RESEARCH POSTERS ON THE HILL 2012 AN AUTOMATED HEALTH MONITORING SYSTEM FOR AERO SPACE STRUCTURES Peter T. Hillyard (V. John Mathews) Department of Electrical and Computer Engineering University of Utah Structural health monitoring systems are used to locate damage in aerospace and in other type of structures. The aeronautical industry invests time and money in structural health monitoring systems to locate damage in aircraft structures to increase their longevity and to ensure flight-worthiness. Composite materials (fiber-reinforced plastics) are becoming a more popular primary load-bearing material in aircraft because of their strength, stiffness, and resilience to corrosion. They are also susceptible to internal damage (for example, delamination) without any visual evidence, making it necessary to frequently monitor composite aircrafts for structural damage. This poster presents a new approach to estimating the location of damage in aerospace structures using an array of piezoelectric sensors that passively monitor the structure for acoustic emission signals generated by damage mechanisms such as delamination, impacts, fatigue, etc. The approach involves estimating the time difference of arrival of the acoustic emission signals at the different sensors and then employing a functional minimization algorithm to locate the source of the acoustic emission. The method will be validated first through simulations in Matlab and then via a physical impact test. Peter Hillyard and Dr. V. John Mathews Electrical and Computer Engineering Student Photo Faculty Photo Plot of simulated error values produced along the surface of the test material. 6 sensors, located on a circle, were used. An independent random amount of noise with a standard deviation of 100 μseconds was added to the actual time delays. Error values were computed every 0.5 cm. This simulation showed only a 1.75 cm error between the estimated and actual location of the source. This simulation varied the radius of the sensor circle and computed the root-mean-square error between the predicted and actual location of the source. 30 independent trials were simulated for each radial distance. An independent random amount of noise with a standard deviation of 100 μseconds was added to the actual time delay values. As expected, the prediction error decreases as the sensor density increases and as the number of sensors increases. Peter Hillyard Dr. V. John Mathews An Automated Health Monitoring System for Aerospace Structures Damage in aircraft material causes structural failure if it is undetected and unrepaired. Damage in aircraft material can occur due to impact, continued strain, delamination of material, and other damage mechanisms. Non-destructive tests (such as traditional ultrasound scans) are one way to locate damage in aerospace structures. Another method to detect damage is the use of sensor networks. Piezoelectric sensors placed on the structure can record acoustic emission signals that are created from any combination of damage mechanisms. The source location is estimated by minimizing the cost function above. The cost function is based on the estimate of the time differences of arrivals of the source signal at the different sensors, the unknown velocities and known sensor locations. 0 (x1,y1) (x6,y6) (x3,y3) (x5,y5) (x4,y4) (x2,y2) E(x0, y0 )= vivj (ti tj )vj (xi x0 )2 +(yi y0 )2 +vi (xj x0 )2 +(yj y0 )2 ( ) j=i+1 N i=1 N1 2 Conclusion This project has shown the feasibility of passively monitoring a structure to estimate the location of a damage event even when the velocity characteristics of the material are unknown. The simulation results showed that this monitoring system is capable of estimating the location of the source best when the radius of the sensor circle is decreased and when the number of sensors on the circle is increased. Simulations showed that with 7 sensors and with an independent random amount of noise with a standard deviation of 100 μseconds added to the actual time delay estimates, an average estimation error was 2.5 cm. Finally this project proposes a new and viable solution for finding damage in aerospace structures and in other practical applications. Abstract Structural health monitoring systems are used to locate damage in different classes of materials. The aeronautical industry invests time and money in structural health monitoring systems to locate damage in aircraft structure to increase the longevity and ensure the flight-worthiness of the aircraft. Several methods currently exist to locate damage in structures made of aluminum which has been the primary load-bearing material used in aircraft structures. Within the last few decades, composites ( ber-reinforced plastics) have been integrated slowly into aircraft structures. Composites are becoming a more popular load-bearing material in aircraft because of their strength, sti ness, and resilience to corro-sion. Locating damage in composites, however, is a relatively new process and is limited by the complex physical properties of the material. While some methods, like passive sensor networks, can locate damage in composites, the algorithms used in this process require prior knowledge of how acoustic emissions (AE) (caused by delamination of the composite, impact, or by some other damage mechanism) propagate in the composite material. This project will improve upon current damage locating procedures that use passive sensor networks by nding damage in isotropic and anisotropic materials without having to characterize the material proper-ties. The approach involves estimating the time di erence of arrival of the AE signals at the di erent sen-sors and then employing a functional minimization algorithm to locate the source of the AE. The method will be validated rst through simulations in Matlab and then via a physical impact test. |
