The student Beñat Urtasun Marco obtained an EXCELLENT CUM LAUDE grade with Industrial Doctorate mention

The student Beñat Urtasun Marco obtained an EXCELLENT CUM LAUDE grade with Industrial Doctorate mention

---

• Thesis title: Contributions to induction thermography inspection and automated view planning

Court:

• Presidency: Beate Oswald-Tranta (Montanuniversität Leoben)
• Vocal: Ángel Javier García Adeva (UPV/EHU)
• Vocal: Alberto Tellaeche Iglesias (Universidad de Deusto)
• Vocal: Ander Muniategui Merino (Lortek)
• Secretary: Alberto Izaguirre Altuna (Mondragon Unibertsitatea)

Abstract:

The growth in the industry of non-destructive inspection techniques and dimensional control presents continuous challenges in which these processes are expected to be automated on a shorter time span and meet increasingly demanding requirements.

In this context, non-destructive inspection techniques such as active thermography and total dimensional control pose a series of challenges that complicate their application in contexts where reduced inspection times are required. Additionally, their automation can be time-consuming and resource-intensive.

Regarding the inspection of surface integrity, induction thermography is a technique that allows localized heating in surface and subsurface defects, which can be detected with an infrared camera. In this sense, defect detection is highly sensitive to the relative position of conventional induction systems, which can significantly impact their automation. Multiple thermographies may be required by altering the position of the system, thereby extending inspection times.

On the other hand, the automation of surface inspection itself may be an aspect that requires experts for the inspection of complex geometries, leading to prolonged automation timelines and the use of many resources.

As this is a dual problem linked to the requirements of the inspection technique and automation itself, a series of issues affecting induction thermography and the procedural automation of dimensional inspection have been addressed in this thesis.

This thesis deals with this problem by making contributions in the field of induction thermography that enhance the technique with systems allowing multi-directional inspection. To achieve this, a new patented inductor design, a multi-directional induction system capable of generating thermographies equivalent to the rotation of the system, and a new method enabling continuous multi-directional scanning of the surface of the pieces to be inspected have been proposed.

The second aspect aimed at improving the automation of inspections involves the development of a system that generates robot inspection trajectories for dimensional inspection with a 3D scanner on arbitrary geometries. This system has several improvements over classical systems, enabling a reduction in robot planning time from hours to just over a second, without any human intervention.

Contributions:

1. A new multi-directional induction system for induction thermography has been presented. This system allows the generation of directional thermographies, exhibiting a rotation pattern, which has not been observed previously, that has led to the development of a novel processing method, demonstrating a higher detection probability compared to separate analysis. This system, in turn, enables an inspection equivalent to rotating the inspection system, thereby reducing the automation process.
2. A novel type of inductor that induces a more uniformly directional magnetic field than conventional alternatives with a fixed magnetic field. This allows for greater coverage and facilitates the use of multidirectional schemes, as the one proposed in the previous chapter.
3. The proposal of a new system for dynamic thermographic inspections, alternating the orientation of the magnetic field simultaneously. A space-time fusion is applied, generating a set of processable directional thermographies with conventional thermographic algorithms. This further reduces inspection times, avoiding any thermographic dead time associated with coverage movement and system rotation.
4. The development of a novel system for planning the 3D robotic inspection trajectory of arbitrary geometries. It incorporates algorithmic improvements that reduce the search space through combinatorial clustering. Additionally, it includes enhancements in dual optimization of robot inspection time and capture pose selection, demonstrating effectiveness in field tests with complex geometries and a 12-axis kinematic chain.