Emissions from brakes have become a public health issue as it is becoming a major contributor to pollution in transport with the electrification of vehicles. Many studies are devoted to noise and particle emissions dealing with the influence of the friction pair materials, the loading conditions, etc. Few of them deal with the physical source mechanisms which are considered a complex problem because it involves complex physical mechanisms in a closed system that cannot be observed. Emissions concern harmful high-frequency squeal noise and particle emissions resulting from material wear. It is proposed here to provide elements of identification of these mechanisms thanks to a multimodal operando instrumentation associated with a discrete follow-up of surfaces. To this end, numerous sensors are introduced, continuous measurements of forces, displacements, accelerations, and temperatures, combined with measurements of nuisances: noise and particle concentration. The surfaces are monitored by observation between each test sequence with optical measurement and profilometry determination. To carry out these measurements, a laboratory device of a pin-on-disc type is used, with a car disc equipped with a pin of material extracted from a brake pad. In addition to facilitating the accessibility of the measurements, the advantage of using a simpler system than a complete brake is to gain access to the understanding of the behavior of the system, particularly in the vibratory domain. The results discussed here are for a material pair consisting of a cast iron and a metal-free automotive friction material. Of particular interest is that the formulation and manufacturing process of this material is well known. Load cycles are performed with varying forces, speeds, and durations to achieve varying temperature levels. The data processing is done in two stages: • A cross-analysis of the data for the construction of tribological monitoring of the surfaces in relation to the load exerted, allowing scenarios to be proposed on the emission events: noise and particles. • A more systematic analysis of the data, using machine learning algorithms to identify the most meaningful measures on emissions and to support or improve the physical scenarios. The results show that the emissions are closely linked to the level of loading applied, which is a known result, but it is shown here that the order of solicitation is also preponderant. Several key emission parameters have been identified and prioritized through multimodal instrumentation and associated data processing. This, together with the surface monitoring, makes it possible to describe the tribological scenario which associates the emissions with the evolution mechanisms of the interface and the materials. This is a key point for the identification of the source mechanisms. This knowledge highlights key points of action for the reduction of emissions.

Dr.-Ing. Mael Thévenot, PostDoctoral Student in Tribology, University of Lille; Prof. Dr. Jean-François Brunel, Teacher-researcher, Université de Lille; Mr. Nathanael Winter, PhD student, Hamburg University of Technology; Ms. Charlotte Geier, PhD student, Hamburg University of Technology; Dr. Merten Stender, Post-doctoral Student, Hamburg University of Technology; Prof. Dr. Norbert Hoffmann, University lecturer and researcher, Hamburg University of Technology; Prof. Dr. Philippe Dufrenoy, University lecturer and researcher, Université de Lille