Prediction of molten pool temperature and processing quality in laser metal deposition based on back propagation neural network algorithm

Gao, Jiali; Wang, Chi; Ye, Hao; Wang, Xu; Zhao, Kai; Ding, Xiaoqing

doi:10.1016/j.optlastec.2022.108363

Cited by 12 publications

(8 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The number of iterations determines the number of training cycles. Too many iterations may lead to overfitting, while too few may result in underfitting [4,5,20,27].…”

Section: Bp Neural Networkmentioning

confidence: 99%

“…Due to the adoption of a single response, the cladding width, in determining the hyperparameters during the selection process, it may lead to poorer predictive performance when predicting the width and dilution rate. This is the reason why SVR and BP cannot fit the mapping relationship between height, dilution rate, and clad quality well [3][4][5][6].…”

Section: Evaluation Of Model Performancementioning

confidence: 99%

“…The process parameters of the laser cladding process have a profound effect on the quality characteristics of the resulting cladded layer (i.e., the geometrical and mechanical properties). The coupling of light, powder, and substrate properties leads to a complex mapping relationship between process parameters and cladding layer [3,4]. In order to optimize laser cladding, many scholars are exploring the influence of the different process parameters on the coating morphology and mechanical [5,6].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Comparative assessment of supervised machine learning algorithms for predicting geometric characteristics of laser cladded inconel 718

Yang,

Geng,

Alfano

et al. 2024

Mater. Res. Express

View full text Add to dashboard Cite

Laser cladding, an innovative surface modification and coating preparation process, has emerged as a research hotspot in material surface modification and green remanufacturing domains. In the laser cladding process, the interaction between laser light, powder particles, and the substrate results in a complicated mapping connection between process parameters and clad layer quality. This work aims to shed light on this mapping using fast evolving machine learning algorithms. A full factorial experimental design was employed to clad Inconel 718 powder on an A286 substrate comprising 64 groups. Analysis of variance, contour plots, and surface plots were used to explore the effects of laser power, powder feeding rate, and scanning speed on the width, height, and dilution rate of the cladding. The performance of the predictive models was evaluated using the index of merit (IM), which includes mean square error (MSE), mean absolute error (MAE), and coefficient of determination (R2). By comparing the performance of the models, it was found that the Extra Trees, Random forest regression, Decision tree regression, and XGBoost algorithms exhibited the highest predictive accuracy. Specifically, the Extra Trees algorithm outperformed other machine learning models in predicting the cladding width, while the RFR algorithm excelled in predicting the associated height. The DTR algorithm demonstrated the best performance in predicting the cladding dilution rate. The R2 values for width, height, and dilution rate were found to be 0.949, 0.954, and 0.912, respectively, for these three models.

show abstract

“…The number of iterations determines the number of training cycles. Too many iterations may lead to overfitting, while too few may result in underfitting [4,5,20,27].…”

Section: Bp Neural Networkmentioning

confidence: 99%

Section: Evaluation Of Model Performancementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Comparative assessment of supervised machine learning algorithms for predicting geometric characteristics of laser cladded inconel 718

Yang,

Geng,

Alfano

et al. 2024

Mater. Res. Express

View full text Add to dashboard Cite

show abstract

“…Random forest, as a type of integration algorithm, is a method that integrates multiple decision trees into a forest to predict the final outcome. Compared with parametric regression methods, this algorithm does not need to test assumptions such as normality and independence of variables, and also does not need to consider co-linearity of multiple variables [6]. As one of the best nonparametric regression models for prediction, it can handle input samples with high-dimensional features without dimensionality reduction.…”

Section: Rf Prediction Modelmentioning

confidence: 99%

“…Wu et al used RF and NN algorithms to predict the residual stresses during arc additive process with arc power, scanning speed, substrate thickness and substrate preheating temperature as input variables [5]. Gao et al predicted the melted pool temperature and cladding layer processing quality of 316L stainless steel and titanium alloy materials using BPNN and RF models [6,7]. In the additive manufacturing of NiTi alloys, Mehrpouya et al used artificial neural network (ANN) nonlinear model to predict the recovery rates and transformation temperatures under the input parameters (laser power, laser scanning speed and pattern filling spacing) for NiTi alloys [8].…”

Section: Introductionmentioning

confidence: 99%

Prediction of single-track deposition quality in additive manufacturing of NiTi shape memory alloy

Wang

Zhao

et al. 2023

2022 Workshop on Electronics Communication Engineering

View full text Add to dashboard Cite

Nickel-titanium (NiTi) shape memory alloy (SMA) has excellent application potential in aerospace because of its shape memory effect and super elasticity, but is still limited by poor machinability and weldability in traditional processing techniques. The coaxial powder feeding laser deposition technology opens a new window in the processing of NiTi SMA components. In this paper, prediction models between the process parameters (laser power, scanning speed and powder feeding rate) and process state parameter (melted pool temperature), deposition quality (track width, track height, microhardness) in NiTi alloy laser metal deposition based on the Back Propagation Neural Networks (BPNN) and Random Forest (RF) algorithms were estab-lished. Thirty single tracks were deposited and measured as training groups. The results show that the average prediction error based on the BPNN model for microhardness, track width, track height and melted pool temperature are 0.37%, 1.88%, 4.45% and 0.91%, respectively, which are better than the RF model. Then, BPNN model was further used to predict deposied quality under the combination of five new process parameters groups. Objective of this study was to provide a guidance for the subsequent optimization of process parameters for the improvement of the deposition accuracy of NiTi alloy parts.

show abstract

Multi-objective optimization of process parameters for laser metal deposition of NiTi shape memory alloy based on neural network and genetic algorithm

Gao,

Wang,

Wang

et al. 2024

Int J Adv Manuf Technol

View full text Add to dashboard Cite

Prediction of molten pool temperature and processing quality in laser metal deposition based on back propagation neural network algorithm

Cited by 12 publications

References 44 publications

Comparative assessment of supervised machine learning algorithms for predicting geometric characteristics of laser cladded inconel 718

Comparative assessment of supervised machine learning algorithms for predicting geometric characteristics of laser cladded inconel 718

Prediction of single-track deposition quality in additive manufacturing of NiTi shape memory alloy

Multi-objective optimization of process parameters for laser metal deposition of NiTi shape memory alloy based on neural network and genetic algorithm

Contact Info

Product

Resources

About