Hybrid sparse convolutional neural networks for predicting manufacturability of visual defects of laser powder bed fusion processes

Zhang, Ying; Zhao, Yaoyao Fiona

doi:10.1016/j.jmsy.2021.07.002

Cited by 21 publications

(4 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the OC classification task, we employed categorical cross-entropy loss [51], a widely used loss function for multiclass classification problems [52]. Let m denote the total number of possible operational conditions: OC = (c 1 , c 2 , ..., c m ) represents the real operational condition; OC = ( ĉ1 , ĉ2 , ..., ĉm ) represents the operational condition classified by our model.…”

Section: Loss Functionsmentioning

confidence: 99%

Robust-MBDL: A Robust Multi-Branch Deep-Learning-Based Model for Remaining Useful Life Prediction of Rotating Machines

Tran,

Vu,

Pham

et al. 2024

Mathematics

View full text Add to dashboard Cite

Predictive maintenance (PdM) is one of the most powerful maintenance techniques based on the estimation of the remaining useful life (RUL) of machines. Accurately estimating the RUL is crucial to ensure the effectiveness of PdM. However, current methods have limitations in fully exploring condition monitoring data, particularly vibration signals, for RUL estimation. To address these challenges, this research presents a novel Robust Multi-Branch Deep Learning (Robust-MBDL) model. Robust-MBDL stands out by leveraging diverse data sources, including raw vibration signals, time–frequency representations, and multiple feature domains. To achieve this, it adopts a specialized three-branch architecture inspired by efficient network designs. The model seamlessly integrates information from these branches using an advanced attention-based Bi-LSTM network. Furthermore, recognizing the importance of data quality, Robust-MBDL incorporates an unsupervised LSTM-Autoencoder for noise reduction in raw vibration data. This comprehensive approach not only overcomes the limitations of existing methods but also leads to superior performance. Experimental evaluations on benchmark datasets such as XJTU-SY and PRONOSTIA showcase Robust-MBDL’s efficacy, particularly in rotating machine health prognostics. These results underscore its potential for real-world applications, heralding a new era in predictive maintenance practices.

show abstract

Section: Loss Functionsmentioning

confidence: 99%

Robust-MBDL: A Robust Multi-Branch Deep-Learning-Based Model for Remaining Useful Life Prediction of Rotating Machines

Tran,

Vu,

Pham

et al. 2024

Mathematics

View full text Add to dashboard Cite

show abstract

“…The manufacturability of the lattice is also an important aspect to consider when analyzing the resulting models [44]. The manufacturability check is run on the resulting STL models using the Preform 3D printing software [45].…”

Section: Performance Analysismentioning

confidence: 99%

Beam-Based Lattice Topology Transition With Function Representation

Letov

Zhao

2022

Journal of Mechanical Design

View full text Add to dashboard Cite

A lattice structure is a porous periodic structure with unit cells organized according to a pattern. Lattice structures are lightweight parts that are commonly produced by additive manufacturing techniques. Lattice structures require their topology defined which effectively defines the connectivity of their unit cell. Many of these topologies are beam-based, i.e. their unit cell is represented by a network of nodes connected with beams. Such lattice structures require a geometric modeling tool capable of generating their solid model. This paper presents a method to support the topology transition for beam-based lattice structures by controlling the geometric parameters of topologies. This control is made possible with the function representation of the geometry. The work also analyzes how suitable different beam-based lattice topologies are to support the transition. A few case studies are carried out to demonstrate the feasibility of the proposed method.

show abstract

“…3D models can also be voxelized into cubes to generate original features. Sparse representations are deployed to reduce the memory needed to process huge voxelized data [63]. Geometrical features such as relative distances extracted from point clouds can be the input features to train ML models [64].…”

Section: D Datamentioning

confidence: 99%

“…As a result, several works focused on predicting geometry deviations in AM printed parts [115][116][117][118][119]. Inspecting visual defects has been a common way to judge product quality and is used in several ML models with tabular [63], graphic [120,121], spectrum [74] data as inputs. Regression-based ML models are a popular choice to determine the exact geometry of products in AM [122][123][124].…”

Section: Macro Levelmentioning

confidence: 99%

A systematic review on data of additive manufacturing for machine learning applications: the data quality, type, preprocessing, and management

et al. 2022

Self Cite

View full text Add to dashboard Cite

Additive manufacturing (AM) techniques are maturing and penetrating every aspect of the industry. With more and more design, process, structure, and property data collected, machine learning (ML) models are found to be useful to analyze the patterns in the data. The quality of datasets and the handling methods are important to the performance of these ML models. This work reviews recent publications on the topic, focusing on the data types along with the data handling methods and the implemented ML algorithms. The examples of ML applications on AM are then categorized based on the lifecycle stages, and research focuses. In terms of data management, the existing public database and data management methods are introduced. Finally, the limitations of the current data processing methods and suggestions are given.

show abstract

Hybrid sparse convolutional neural networks for predicting manufacturability of visual defects of laser powder bed fusion processes

Cited by 21 publications

References 36 publications

Robust-MBDL: A Robust Multi-Branch Deep-Learning-Based Model for Remaining Useful Life Prediction of Rotating Machines

Robust-MBDL: A Robust Multi-Branch Deep-Learning-Based Model for Remaining Useful Life Prediction of Rotating Machines

Beam-Based Lattice Topology Transition With Function Representation

A systematic review on data of additive manufacturing for machine learning applications: the data quality, type, preprocessing, and management

Contact Info

Product

Resources

About