Reflections on the 2023 Marrakesh-Safi earthquake
The earthquake was the strongest instrumentally recorded event ever felt in Morocco. Indeed, it appears to be the strongest in Moroccan historyRead more
30-08-2023 | Posted by Joaquín Martí
We are used to thinking of friction as a nuisance, as something that must be overcome for us to achieve what we intend. But this is only one side of the coin, we would be unable to walk, eat with a fork or keep a pencil on the table in a frictionless world. If given a choice, better to manage its drawbacks than to struggle without its advantages.
At times we would like to increase the existing friction, like when we use sticky pads to prevent objects from slipping off smooth surfaces. More often we would like to reduce it, for which we use wheels, bearings, materials like PTFE, or lubricants.
The complexity of the origin of friction (surface roughness, adhesion, deformation, etc.) makes it hard to study it using first principles and it must therefore be determined and investigated by empirical methods. Moreover, any non-trivial problems involving friction are inevitably nonlinear and path dependent. The differences between static and dynamic friction, as well as the possible deterioration of the contact surface during the interaction, contribute further difficulties to the reliable characterisation of friction and to the numerical resolution of the problems that activate it.
Just as an example, we participated recently in the SOCRAT (Seismic simulation of Overhead CRAne on shaking Table), an international exercise jointly organised by OECD-NEA (Nuclear Energy Agency), IRSN (Institut de Radioprotection et de Sûreté Nucléaire), and EDF (Electricité de France). The object was to identify the best modelling practices for crane bridge devices, and the relevant failure criteria, when subjected to seismic motions.
Such cranes involve two sets of wheels running on the corresponding rails. Tests were conducted on a 1:5 scale model placed on a shake table, under a variety of conditions and with different input motions. The problems involved rolling, sliding, wheel-rail separation and impact. Some of the experimental results had been communicated to the participants to allow calibration of parameters, while blind predictions had to be produced for others.
Our results were among the best ones reported, but can only be considered to be moderately successful. And, indeed, friction and its vagaries were mainly responsible for their shortcomings. But it is also true that the objectives of the exercise were rather ambitious, given the complexity of the problem.
When driving a car, friction makes it possible to accelerate, brake and turn. But when the frictional capacity is exceeded and sliding starts to take place, the evolution of the car becomes less predictable, both to the driver and to the analyst solving the numerical problem.