Variational autoencoders have been recently proposed for the problem of process monitoring. While these works show impressive results over classical methods, the proposed monitoring statistics often ignore the inconsistencies in learned lower-dimensional representations and computational limitations in high-dimensional approximations. In this work, we first manifest these issues and then overcome them with a novel statistic formulation that increases out-of-control detection accuracy without compromising computational efficiency. We demonstrate our results on a simulation study with explicit control over latent variations, and a real-life example of image profiles obtained from a hot steel rolling process.
more »
« less
A Weighted Variance Approach for Uncertainty Quantification in High Quality Steel Rolling
This paper proposes a computer vision framework
aimed to segment hot steel sections and contribute to rolling
precision. The steel section dimensions are calculated for the
purposes of automating a high temperature rolling process. A
structured forest algorithm along with the developed steel bar
edge detection and regression algorithms extract the edges of
the high temperature bars in optical videos captured by a
GoPror camera. To quantify the impact of noises that affect
the segmentation process and the final diameter measurements,
a weighted variance is calculated, providing a level of trust in
the measurements. The results show an accuracy which is in line
with the rolling standards, i.e. with a root mean square error
less than 2:5 mm.
more »
« less
- Award ID(s):
- 1903466
- NSF-PAR ID:
- 10161400
- Date Published:
- Journal Name:
- Proceedings of the International Conference on Information Fusion (Fusion 2020)
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract Ultrasonic additive manufacturing (UAM) is a solid state manufacturing process capable of producing near-net-shape metal parts. Recent studies have shown the promise of UAM welding of steels. However, the effect of weld parameters on the weld quality of UAM steel is unclear. A design of experiments study based on a Taguchi L16 design array was conducted to investigate the influence of parameters including baseplate temperature, amplitude, welding speed, and normal force on the interfacial temperature and shear strength of UAM welding of carbon steel 4130. Analysis of variance (ANOVA) and main effects analyses were performed to determine the effect of each parameter. A Pearson correlation test was conducted to find the relationship between interfacial temperature and shear strength. These analyses indicate that a maximum shear strength of 392.8 MPa can be achieved by using a baseplate temperature of 400°F (204.4°C), amplitude of 31.5 μm, welding speed of 40 in/min (16.93 mm/s), and normal force of 6000 N. The Pearson correlation coefficient is calculated as 0.227, which indicates no significant correlation between interfacial temperature and shear strength over the range tested.more » « less
-
more » « less
Recent advancements in the severe plastic deformation process called accumulative roll bonding (ARB) can help to address the long‐standing need for manufacturing lightweight, high‐strength Mg sheet materials. However, the fabrication of Mg alloy‐based laminates via ARB remains a challenge due to the intrinsically poor formability of Mg. Herein, it is shown that Mg‐based composite laminates with refined layers can be fabricated via several room‐temperature ARB cycles with appropriate intermediate annealing and alloy selection. The final laminates made here consist of equal volume fractions of a dilute Mg–Zn–Mn–Ca alloy phase and a pure Nb phase with fine 150 μm layer thicknesses. Deformation texture evolution in both phases within the composite is analyzed via neutron diffraction measurements taken at different stages in the process. The analysis suggests that the annealing step recrystallizes the Mg‐alloy phase. It is also shown that for both phases, the stabilized deformation textures within the composite correspond to the classic stable textures of the individual constituents. Polycrystal texture modeling implies that {10–12} <‐1011> extension twinning developed in the Mg alloy during rolling.
-
Quenching and partitioning (Q&P) processing of third-generation advanced high strength steels generates multiphase microstructures containing metastable retained austenite. Deformation-induced martensitic transformation of retained austenite improves strength and ductility by increasing instantaneous strain hardening rates. This paper explores the influence of martensitic transformation and strain hardening on tensile performance. Tensile tests were performed on steels with nominally similar compositions and microstructures (11.3 to 12.6 vol. pct retained austenite and 16.7 to 23.4 vol. pct ferrite) at 980 and 1180 MPa ultimate tensile strength levels. For each steel, tensile performance was generally consistent along different orientations in the sheet relative to the rolling direction, but a greater amount of austenite transformation occurred during uniform elongation along the rolling direction. Neither the amount of retained austenite prior to straining nor the total amount of retained austenite transformed during straining could be directly correlated to tensile performance. It is proposed that stability of retained austenite, rather than austenite volume fraction, greatly influences strain hardening rate, and thus controls strength and ductility. If true, this suggests that tailoring austenite stability is critical for optimizing the forming response and crash performance of quenched and partitioned grades.more » « less
-
Recent advances pertaining to modeling of grain fragmentation during deformation and recrystallization of polycrystalline metals using viscoplastic self-consistent (VPSC) polycrystal plasticity are combined into a field fluctuations VPSC (FF-VPSC) model. The FF-VPSC model is a higher-order formulation calculating the second moments of lattice rotation rates based on the second moments of stress fields inside grains and resulting intragranular misorientation distributions. The misorientation distributions are used to define a grain fragmentation sub-model for improving predictions of deformation texture evolution and to formulate kinetics sub-models for nucleation as well as to influence the stored energy governing grain growth for the predictions of recrystallization texture evolution. Formation of a copper-like texture in moderately high stacking fault energy (SFE) Cu and a brass-like texture in low SFE brass during rolling to very large strains are successfully predicted using the model. Remarkably, the model also predicts recrystallization textures from the deformation textures of the two metals after adjusting tradeoffs between transition-bands and grain boundary nucleation mechanisms. Additionally, rolling and recrystallization of an interstitial-free steel, tension and recrystallization of AA5182-O, and recrystallization of an additively manufacturing cobalt-based alloy MarM-509 are simulated to predict texture evolution. Through these case studies involving multiple alloys and thermo-mechanical processes we show that, in addition to being predictive with good accuracy, the key advantage of the model lies in its versatility. The FF-VPSC model, simulation results, and insights from the results are presented and discussed in this paper.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.23919/FUSION45008.2020.9190527
