RESUMEN

Rolling element bearing faults significantly contribute to overall machine failures, which demand different strategies for condition monitoring and failure detection. Recent advancements in machine learning even further expedite the quest to improve accuracy in fault detection for economic purposes by minimizing scheduled maintenance. Challenging tasks, such as the gathering of high quality data to explicitly train an algorithm, still persist and are limited in terms of the availability of historical data. In addition, failure data from measurements are typically valid only for the particular machinery components and their settings. In this study, 3D multi-body simulations of a roller bearing with different faults have been conducted to create a variety of synthetic training data for a deep learning convolutional neural network (CNN) and, hence, to address these challenges. The vibration data from the simulation are superimposed with noise collected from the measurement of a healthy bearing and are subsequently converted into a 2D image via wavelet transformation before being fed into the CNN for training. Measurements of damaged bearings are used to validate the algorithm's performance.

Asunto(s)

RESUMEN

Many upcoming experiments in antimatter research require low-energy antiproton beams. With a kinetic energy in the order of 100 keV, the standard magnetic components to control and focus the beams become less effective. Therefore, electrostatic components are being developed and installed in transfer lines and storage rings. However, there is no equipment available to precisely map and check the electric field generated by these elements. Instead, one has to trust in simulations and, therefore, depend on tight fabrication tolerances. Here we present, for the first time, a noninvasive way to experimentally probe the electrostatic field in a 3D volume with a microsensor. Using the example of an electrostatic quadrupole focusing component, we find excellent agreement between a simulated and real field. Furthermore, it is shown that the spatial resolution of the probe is limited by the electric field curvature which is almost zero for the quadrupole. With a sensor resolution of 61 V/m/sqrt[Hz], the field deviation due to a noncompliance with the tolerances can be resolved. We anticipate that this compact and practical field strength probe will be relevant also for other scientific and technological disciplines such as atmospheric electricity or safeguarding near power infrastructure.

RESUMEN

Accurate knowledge of the spatial magnetic field distribution is necessary when measuring field gradients. Therefore, a MEMS magnetic field gradiometer is reported, consisting of two identical, but independent laterally oscillating masses on a single chip. The sensor is actuated by Lorentz force and read out by modulation of the light flux passing through stationary and moving arrays of the chip. This optical readout decouples the transducer from the electronic components. Both phase and intensity are recorded which reveals information about the uniformity of the magnetic field. The magnetic flux density is measured simultaneously at two points in space and the field gradient is evaluated locally. The sensor was characterised at ambient pressure by performing frequency and magnitude response measurements with coil and various different permanent magnet arrangements, resulting in a responsivity of 35.67 V/T and detection limit of 3.07 µT/ Hz (@ 83 Hz ENBW). The sensor is compact, offers a large dynamic measurement range and can be of low-cost by using conventional MEMS batch fabrication technology.