The axle axes are checked during maintenance in order to detect any fatigue cracking defect that may appear due to the stresses of the service. Inaccessible for magneticoscopic control, these axes are inspected by conventional ultrasonic techniques by contact, with the use of translators called "mono-element". "When the part to be checked is hidden or inaccessible, we use ultrasound, a volume method, to search for surface defects by a volume method," says Bastien Richard, head of the END competence center of the SNCF railway test agency (AEF). But, although effective, these methods used for more than 40 years have some disadvantages. For example, this requires as many material checks as translators (a TGV motor axis can use up to 4 or 5 translators who must be checked beforehand). In addition, the A-scan signal obtained is a visual operator reserved for initiates and running gear experts who do not practice ultrasound do not necessarily understand their colleagues. In addition to the long control time and the strong attention of the operator required, the control cannot be recorded. Hence the need, for railway maintenance, to evolve towards innovative controls with the use of element technology. »
The AEF has thus developed the Sumax (for "AXe Multielement Ultrasound Scanner"). The equipment consists of a time element ultrasonic device, a probe, a mechanized probe arm and a trolley equipped with an irrigation plant for coupling. The control principle consists in scanning electronically, in a fixed position of the – only – ultrasonic translator, the entire width of the area to be controlled with a beam of variable angles. The control of the area to be examined is then carried out in a single revolution of the axis. At one point in the circumference of the axis, some 135 ultrasonic shots make it possible to scan the entire area to be controlled. After the 360° circumferential scan and an acquisition every 3 mm, the control mapping is generated from the memorization of more than 30,000 basic ultrasonic signals of type A-scan.
This new technical solution has significant advantages, in particular a significant reduction in the overall operating time (divided at least by two) and an improvement in the quality of the inspection thanks to all the ultrasonic data recorded and the maps generated. Also, the probability of detection of defects is greater thanks to the complete mapping of the area to be controlled. In addition, the performance of the control and the conformity of the examination are now demonstrable a posteriori. And no need for grease- or gel-based ultrasonic coupling consumables. Water is used as a coupling agent which provides an economic and environmental gain.
Control robotization is also an important innovation in the axle maintenance process. If, on the demonstrator, a pole allows the operator to no longer hold the probe, a 5-axis robot will soon be integrated into the device. The objective: to free the hands of the operator who can thus refocus on his core business: the interpretation of ultrasonic signals.

Three questions to Bastien Richard, head of the END competence center of the SNCF railway test agency (AEF).
How does the control of an axle axis take place in maintenance?
The accessible parts of the axis are controlled by a video control. Parts hidden by cushioning elements such as wheels, discs or engine bridges are ultrasonic controlled. This ultrasonic control avoids costly and risky wheel dismantling operations, which can damage the elements concerned.
Why did you choose a multi-element ultrasonic technology for control?
Our challenge is to fully integrate industrial technological advances and modernize current control practices for better productivity and controls. For me, elemental ultrasonic technology – now widespread in weld control – is a three-stage rocket. It is first of all a material part with a probe composed of several elements. Then, the second stage of the rocket is a signal processing part with powerful electronics that will make it possible to control these elements and collect a large amount of data. Finally, the third stage of the rocket is a software layer considered as a step of synthesis and simplification of all the numerous and complex data collected. It is image processing that allows the operator to obtain a visual indicating in a binary way whether the controlled part is compliant or not. It is therefore a technique that offers many advantages.
What are the deadlines for the implementation of this new solution?
Conventional ultrasound that has been in place for more than 40 years is well established and the documentation is dense. This will still take time to switch from one methodology to another. However, a first step should be put in place by the end of 2022. A test series will be carried out before the device is put into production. We will take into account the different feedback to improve the solution. The hardware part is not the most complicated to implement. It is necessary to validate the entire process with a complete performance file for procedures that are rather aimed at ultrasonic operators of level 1 CFCM (railway maintenance).