161 research outputs found
Integration of the AI and ML Approaches for Prediction Analysis in Welding: Review
No full textInternational audienceWelding techniques are presents in almost all industries from automotive, aerospace, building, etc. Despite the fact that the technology behind welding remain simple, it is very complicated to characterize and predict the final mechanics responses without considering all the process of the welding. In particular when we refer to the fatigue as the main constraints are located at these singularities areas. Such challenges include poorly controlled welding parameters and weld geometry, which lead to weld quality problems. Nowadays, with the recent involvement of Artificial intelligence (AI) and Machine Learning (ML) into the industry 4.0, it was just matter of time before it is applied to welding problems. The potential applications of ML and AI in welding are vast, as it should bring more efficient and streamlined compared to what is used to be in the early 90’s. In this review a focus will be done on the recent advanced, and reviews previous investigation on AI applications in welding process control and welding robot control. Quality control of such welds are important when we deal with mass production of sensitive products. The main challenges remain the data acquisition to increase the efficiency of the prediction. Testing are sometime coupled with finite elements simulation for training and testing the ML Perspectives and future challenges regarding the integration and the impact of AI in the welding industry in the era of Industry 4.0
Neural network computation for the evaluation of process rendering: application to thermally sprayed coatings
No full textIn this work, neural network computation is attempted to relate alumina and titania phase changes of a coating microstructure with respect to energetic parameters of atmospheric plasma straying (APS) process. Experimental results were analysed using standard fitting routines and neural computation to quantify the effect of arc current, hydrogen ratio and total plasma flow rate. For a large parameter domain, phase changes were 10% for alumina and 8% for titania with a significant control of titania phase
Advanced study of hydrogen storage by substitutional doping of Mn and Ti in Mg2Ni phase
No full textInternational audienceThe substitutional doping of Mn and Ti in Mg2Ni phase has been investigated by first principles densityfunctional theory calculations. The calculation of enthalpy of formation shows that among the four different lattice sitesof Mg(6f), Mg(6i), Ni(3b) and Ni(3d) in Mg2Ni unit cell, the most preferable site of substitution of Mn in Mg2Ni latticehas been confirmed to be Mg(6i) lattice site. The most preferable site of Ti substitution in Mg2Ni lattice is Mg(6i) positionand the stability of Ti-doped Mg2Ni decreases with the increase of substitution quantity of Ti for Mg
Evaluation of the possibility of determining fluctuations that could demonstrate the presence of particles interacting with photons
No full textInternational audienceWe are investigating the improvement of the precision of the means of measurement to determine if it is possible to have sufficient sensitivity to the detection of the effects of elementary particles which would be characteristic of dark matter. A particle has been proposed and is called axion. There would be an interaction between the axions and the photons using the Primakoff effect under strong magnetic field. Radio frequencies from 460 to 810 MHz would be assumed to be suitable for the mass of the axion, if it exists. It is then interesting to focus on the piezoaxionic effect. If the frequency of the axions could match the natural frequency of a normal mode bulk acoustic of a piezoelectric crystal, one would expect the piezoaxionic effect to occur. One could then rely on the piezoelectric effect to observe the variations on the resonant frequency which can be read out electrically using the best piezoelectric materials. Through this example of development and applications in detection, we propose to decrypt this subject and to show how multidisciplinary skills are necessary to hope that small fluctuations can be detectable
Optimal design of a crystalline and integrated resonator coupled with optical fibre
No full textInternational audienceQ-factor in optical resonators is important issue that quality the device and the type of applications. Due to the advantages of optical resonators in terms of reproducibility on chip (that are designed of various topologies and integration with optical devices), it is very important to get a highest Qfactor. To increase this factor from the lower rang [104 - 106] to higher one [108 -1010] we use crystalline resonators. In practice, it is more complicated to couple an optical signal from a tapered fiber to crystalline resonator than from a defined ridge to a resonator designed on a chip. In this work, we will focus on the simulation and optimization of the crystalline resonators under straight wave guide and subject also to technological constraints of manufacturing. The coupling problem at the Nano scale makes our optimizations problem more dynamics in term of design space.© (2018) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only
- …
