An efficient maintenance is a key consideration in systems of railway transport, especially in high-speed trains, in order to avoid accidents with catastrophic consequences. In this sense, having a method that allows for the early detection of defects in critical elements, such as the bogie mechanical components, is a crucial for increasing the availability of rolling stock and reducing maintenance costs. The main contribution of this work is the proposal of a methodology that, based on classical signal processing techniques, provides a set of parameters for the fast identification of the operating state of a critical mechanical system. With this methodology, the vibratory behaviour of a very complex mechanical system is characterised, through variable inputs, which will allow for the detection of possible changes in the mechanical elements. This methodology is applied to a real high-speed train in commercial service, with the aim of studying the vibratory behaviour of the train (specifically, the bogie) before and after a maintenance operation. The results obtained with this methodology demonstrated the usefulness of the new procedure and allowed for the disclosure of reductions between 15% and 45% in the spectral power of selected Intrinsic Mode Functions (IMFs) after the maintenance operation.

A quasi-passive biped (having only one actuator) developed into a Spanish project called "PASIBOT" [1] is presented in this article. We focus on the PASIBOT's topology, kinematics and dynamics, and we describe a program designed for carrying out the corresponding calculations. This code provides for all kinematic and dynamic data, as functions of time, along one step: position, velocity and acceleration of all members, as well as all the forces and torques on each of them, motor torque included. This latter information has helped us to choose the required motor, as this choice depends on some parameters of interest that can be modified in the program, like density or link dimensions. Also, we will be able to get strain-stress data in all links in the course of a step, and then optimize those dimensions. To finish, some results are also presented that confirm the interest of the developed code.

No abstract

No abstract

No abstract