Value Engineering

Attoh-Okine, Nii

doi:10.1002/047134608x.w3358

Cited by 1 publication

(1 citation statement)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[6][7] A training mechanism called back-propagation training algorithm method is so far one of the most popular. [8] In this work, the batch gradient decent with momentum algorithm [9] was used as the training function. It has a strong mathematical foundation based on gradient descent learning.…”

Section: Training Proceduresmentioning

confidence: 99%

Plasma Spray Process On‐Line Control by Artificial Intelligence Methodology

Kanta¹,

Montavon²,

Planche³

et al. 2007

Adv Eng Mater

View full text Add to dashboard Cite

Atmospheric Plasma Spraying (APS) is a complex process in which a powdered feedstock material of particle size distribution ranging usually from 10 to 80 lm is injected into a thermal plasma plume where heat and momentum transfers melt and accelerate them towards the surface of a substrate where they impact, spread and solidify by conductive transfer to form individual near-circular lamellae of a few tens to hundreds micrometers in diameter and a few micrometers thick. From the stacking of these lamellae results the coating. [1] Several physicochemical mechanisms lead to the formation of a coating: [2] -momentum transfers from the plasma jet to the particles and during particle flattening; -mass transfers when particle partial vaporization occurs in the plasma plume and at the particle impact when splashing can take possibly place;-heat transfers from the plasma jet to the particles, from the particles to the ambient atmosphere prior impact, from the particle to the substrate or previously deposited layers during coating built-up mechanisms.Thermal spray processes are special because the coating in-service functional properties derive mostly from its structure. The properties are hence indirectly linked to the selection of the operating parameters. These operating parameters and variables are intimately interrelated via complex -nonlinear -relationships, Figure 1.Understanding and controlling these correlations is mandatory for robust quality control of the process. Thermal spray coatings are more often being prescribed at the initial stages of the design process to become fundamental elements of the engineering system. This paper intends to develop a model-based estimation and control for regulating the particle average velocity and temperature. In a first time, an estimation scheme based on artificial neural network is proposed to predict the in-flight average particle characteristics (i.e., velocity and temperature) as a function of the process power parameters (i.e., arc

show abstract

Section: Training Proceduresmentioning

confidence: 99%