Publications

2026

1. Lifelong Updating of Digital Twin Models for Degraded Systems Using Autoencoder and LSTM

   Yifan Tang, Mostafa Rahmani Dehaghani, and G. Gary Wang

   Applied Soft Computing, 2026

   Abs DOI

   Digital twin (DT) has been applied for monitoring, control, and decision-making of real-world engineering systems, but how to keep a DT valid as the physical system degrades over the lifecycle remains rarely explored. Instead of identifying degradation parameters from expensive aging experiments, this paper proposes a lifelong update method to capture degradation effects directly from system responses via continuously tuning DT models. The core idea is to represent the system degradation as temporal changes of DT model configurations through degradation stages. During the lifelong update process, an autoencoder compresses DT model parameters into latent features, and an LSTM learns the temporal trend of those features to predict latent features for future degradation stages. Based on the predicted latent features, the DT model configuration is reconstructed to predict the system responses affected by degradation. The proposed lifelong update method is evaluated with two real-world engineering datasets, e.g., the battery degradation dataset from Oxford University and the flight engine degradation dataset from NASA. Results demonstrate that the proposed update method successfully captures the degradation effects on system responses during the lifecycle. The one-tailed t-tests confirm that the proposed method statistically outperforms the conventional fine-tuning method in prediction accuracy and robustness at future unseen stages.

2. In-situ Monitoring and Control of Laser-Directed Energy Deposition with Wire–Part 1: Parameter-Signature-Quality Analysis of Duplex Stainless Steel

   Mostafa Rahmani Dehaghani, Kevin Oldknow, Morgan Nilsen, and 5 more authors

   The International Journal of Advanced Manufacturing Technology, 2026

   Abs DOI

   Laser-directed energy deposition with wire (L-DED/W) has gained attention due to its high deposition and utilization rates. However, components fabricated using this method exhibit inhomogeneous microstructure and anisotropic mechanical properties, primarily stemming from the complex thermomechanical phenomena inherent to the process. To indirectly monitor and control final qualities, such as microstructure and hardness—which cannot be directly measured during deposition—this study employs a parameter–signature–quality (PSQ) framework specifically tailored to duplex stainless steel (DSS) 2209. This paper, the first in a two-part series, systematically investigates the correlations within the PSQ framework with microstructure and hardness as the target qualities. A DSS 2209 ring is fabricated using the L-DED/W process with continuous in-situ melt pool monitoring using a coaxial camera. Advanced image processing techniques are developed and applied to extract critical melt pool signatures, including melt pool width, length, and area, from over 100,000 captured images. Post-deposition characterization involves detailed microhardness testing and microstructural analysis across 288 measurement locations, each replicated at least three times, to quantify microstructural features such as grain size, grain shape, and phase content. The results highlight significant correlations within the PSQ framework, emphasizing that melt pool signatures provide stronger and more sensitive indications of microstructural evolution compared to direct process parameters. It is established that hardness predominantly depends on phase composition (austenite content), followed by grain size, whereas laser power—despite its critical role in controlling melt pool width and thus geometry—has minimal influence on hardness and microstructure. Finite element analysis simulations further support experimental observations by analyzing how the cooling rate varies with the bead positioning and travel speed, influencing the grain size and hardness distributions. This study provides foundational understanding essential for implementing real-time monitoring and closed-loop control strategies for microstructure and hardness in DSS components produced by L-DED/W, which is discussed in Part 2 of this series. Ultimately, the insights gained advance the potential for optimized and consistent mechanical performance in metal additive manufacturing.

3. In-situ Monitoring and Control of Laser-Directed Energy Deposition with Wire–Part 2: Geometry and Hardness Modeling and Closed-Loop Control

   Mostafa Rahmani Dehaghani, Kevin Oldknow, Morgan Nilsen, and 5 more authors

   The International Journal of Advanced Manufacturing Technology, 2026

   Abs DOI

   Laser-directed energy deposition with wire (L-DED/W) offers high deposition and material utilization rates but often suffers from geometric inaccuracies and anisotropic mechanical properties due to unstable deposition dynamics and complex thermomechanical phenomena. To enable in situ monitoring and closed-loop control of final qualities such as geometry and hardness—which cannot be directly measured during deposition—this study employs the parameter–signature–quality (PSQ) framework. This work presents the first in situ multi-input multi-output (MIMO) closed-loop control of both deposition geometry and hardness in a directed energy deposition process. Building on the experimental foundation established in Part 1 of this study, a real-time MIMO control strategy is developed for the L-DED/W process. A long-short-term-memory (LSTM) network is used to model the nonlinear and dynamic relationships between process parameters (power and speed) and melt pool signatures, while hardness is classified using a combined set of process parameters and melt pool signatures within the PSQ framework. These models are integrated with a fuzzy logic controller to achieve closed-loop MIMO control, demonstrating effective regulation of melt pool geometry while maximizing the likelihood of achieving high hardness under process uncertainties. The results demonstrate the feasibility of in situ regulation of otherwise unmeasurable final qualities and highlight fuzzy logic control as a flexible and computationally efficient approach for multi-objective control in metal additive manufacturing.

2025

1. Two-Dimensional Temperature Field Prediction with In-Situ Data in Metal Additive Manufacturing using Physics-Informed Neural Networks

   Pouyan Sajadi, Mostafa Rahmani Dehaghani, Yifan Tang, and 1 more author

   Engineering Applications of Artificial Intelligence, 2025

   Abs DOI

   Accurately predicting the temperature field in metal additive manufacturing (AM) processes is critical for preventing overheating, adjusting process parameters, and ensuring process stability. While physics-based computational models offer precision, they are often time-consuming and unsuitable for real-time predictions. Machine learning models, on the other hand, rely heavily on high-quality datasets, which can be costly and difficult to obtain in the metal AM domain. Existing studies on physics-informed neural networks (PINNs) have made progress in integrating physics with machine learning but often lack in-situ data integration, which is essential for capturing real-time thermal dynamics. Additionally, their methodologies are typically heavily dependent on specific process characteristics, limiting their flexibility. Our work addresses these gaps by introducing a PINN-based framework specifically designed for temperature field prediction in metal AM. The framework incorporates in-situ temperature data gathered during the manufacturing process, combining it with physics-informed inputs and a custom loss function. The approach is demonstrated through two case studies. In the first case, using a small set of experimental data, the model achieves an error below 3 % with a mean absolute error (MAE) of 11 °C. In the second case, using simulation data, the model achieves an error below 1 % with an MAE of 7 °C. In addition, the framework shows promising adaptability for different metal AM scenarios with different geometries, deposition patterns, and process parameters.

2. MIMO System Identification and Uncertainty Calibration with A Limited Amount of Data using Transfer Learning

   Mostafa Rahmani Dehaghani, Pouyan Sajadi, Yifan Tang, and 2 more authors

   International Journal of Systems Science, 2025

   Abs DOI

   Multiple-input multiple-output (MIMO) systems are fundamental in numerous advanced engineering applications, from aerospace to telecommunications, where precise system identification is critical for optimal performance. However, the identification of such systems often faces significant hurdles due to data scarcity, with existing approaches typically requiring substantial amounts of data for effective training. Addressing this challenge, this paper introduces a novel transfer learning framework designed specifically for MIMO system identification under conditions of limited data and inherent uncertainties. The proposed framework is applied to two case studies: the first in metal additive manufacturing, specifically the laser-blown powder-directed energy deposition as the source domain and the laser hot wire-directed energy deposition as the target domain, and the second involving a nonlinear case study of a continuous stirred-tank reactor (CSTR) with a temperature-dependent reaction. The results underscore the framework’s effectiveness in capturing the dynamics of the target systems, including the ability to effectively model nonlinear dynamics. Comparative analyses highlight the benefits of employing dimensionless numbers in dynamic system modelling, offering reduced dimensionality, more physical meaning, and increased model accuracy. Overall, the proposed framework presents a promising approach to enhance system identification in MIMO systems with limited data and uncertainties, with potential applications across diverse domains.

3. Digital Twin Modeling and Updating for Metal Additive Manufacturing Processes

   Yifan Tang

   2025
4. RePaint-Enhanced Conditional Diffusion Model for Generating Designs Under Performance Constraints

   Ke Wang, Nguyen Gia Hien Vu, Yifan Tang, and 2 more authors

   In International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Aug 2025

   Abs DOI

   This paper presents a novel framework integrating the RePaint method with the performance-guided denoising diffusion probabilistic model (DDPM) to complete missing or undefined design components based on any given partial structures while satisfying specified performance targets. RePaint allows a pre-trained diffusion model to synthesize designs under the condition of partial designs, offering flexibility and controllability of generative design results. Using a dataset of parametric ship hull designs, a performance-guided diffusion model is pre-trained to generate designs with targeted total resistance coefficients. Experiment results demonstrate that our method generates new designs from random incomplete ship hull designs with performances close to those generated by the pre-trained models. The study further analyzes factors that might influence the new framework’s performance and reveals that the pre-trained model’s design parameter value distribution has a significant effect on the new method’s output designs. This distribution can be leveraged to guide the synthesis of new designs with desirable performance.

5. A Digital Twin Model Updating Method to Capture Lifecycle System Degradation

   Yifan Tang, Mostafa Rahmani Dehaghani, and G. Gary Wang

   In International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Aug 2025

   Abs DOI

   Iterative update methods are important for digital twin (DT) models to learn from new data and maintain the DT performance during application. However, when considering system lifecycle degradations, current update methods fail to enable DT models to capture system responses affected by degradation over time. To alleviate this problem, degradation models of measurable physical parameters are often integrated into DT construction. This operation is costly since identifying the degradation parameters depends on prior knowledge of the system and requires expensive experiments. To overcome above limitations, this paper proposes a lifelong update method for DT models to capture the effects of degradation on system responses without any prior knowledge and expensive offline experiments on the system. In this work, the DT model is constructed as a feedforward neural network with a fixed structure, and the effect of system degradation is assumed to be reflected by the DT parameters obtained at each degradation stage. During the lifelong update process, an extreme learning machine is adopted to capture how the model parameters change in lifecycle for each layer. The constructed extreme learning machines could predict model parameters of the entire DT model at any future degradation stage. The proposed method is evaluated with the battery degradation dataset from Oxford University. The test results demonstrate that the proposed method could capture effects of system degradation on system responses during the lifecycle and outperform the conventional fine-tuning method in terms of prediction accuracy and robustness at future stages.

2024

1. A Systematic Online Update Method for Reduced-Order-Model-based Digital Twin

   Yifan Tang, Pouyan Sajadi, Mostafa Rahmani Dehaghani, and 1 more author

   Journal of Intelligent Manufacturing, Aug 2024

   Abs DOI

   A digital twin (DT) is a model that mirrors a physical system and is continuously updated with real-time data from the physical system. Recent implementations of reduced-order-model-based DT (DT-ROM) have been applied in aerodynamics and structural health monitoring, where partial differential equations (PDEs) are utilized to update reduced bases and coefficients. However, these methods are not directly applicable when the PDEs of the system are unknown. This paper addresses the online update challenge for DT-ROM in scenarios lacking known PDEs of the system. To tackle the challenge, a systematic online update and application method is proposed. During the online update, the projection residual of online data on the reduced bases determines the necessity of updating reduced bases; the prediction residual of online data obtained by the current DT-ROM is used to decide whether to update the coefficient model. By sequentially evaluating both criteria, the method selectively incorporates essential online data for the online DT model update. During the online application, a criterion defined based on online data is adopted to determine whether the offline DT-ROM or the online one is applied to output final predictions. The capability of the proposed method is tested through three numerical and three engineering problems. Results indicate that the proposed online update method consistently reduces both projection and prediction residuals, thereby progressively enhancing the performance of the online DT-ROM on test data. Meanwhile, the online application method provides a prediction performance better than using offline DT-ROM only. Both demonstrate that the proposed work could be applied to online DT updates where the PDEs of the system are unknown.

2. Selecting Subsets of Source Data for Transfer Learning with Applications in Metal Additive Manufacturing

   Yifan Tang, Mostafa Rahmani Dehaghani, Pouyan Sajadi, and 1 more author

   Journal of Intelligent Manufacturing, Aug 2024

   Abs DOI

   Considering data insufficiency in metal additive manufacturing (AM), transfer learning (TL) has been adopted to extract knowledge from source domains (e.g., completed printings) to improve the modeling performance in target domains (e.g., new printings). Current applications use all accessible source data directly in TL with no regard to the similarity between source and target data. This paper proposes a systematic method to find appropriate subsets of source data based on similarities between the source and limited target datasets. Such similarity is characterized by the spatial and model distance metrics. A Pareto frontier-based source data selection method is developed, where the source data located on the Pareto frontier defined by two similarity distance metrics are selected iteratively. This method is integrated into an instance-based TL method (decision tree regression model) and a model-based TL method (fine-tuned artificial neural network). Both models are then tested on several regression tasks in metal AM. Comparison results demonstrate that (1) the source data selection method is general and supports integration with various TL methods and distance metrics, (2) compared with using all source data, the proposed method can find a subset of source data from the same domain with better TL performance in metal AM regression tasks involving different processes and machines, and (3) when multiple source domains exist, the source data selection method could find the subset from one source domain to obtain comparable or better TL performance than the model constructed using data from all source domains.

3. System Identification and Closed-Loop Control of Laser Hot-Wire Directed Energy Deposition using the Parameter-Signature-Quality Modeling Scheme

   Mostafa Rahmani Dehaghani, Atieh Sahraeidolatkhaneh, Morgan Nilsen, and 4 more authors

   Journal of Manufacturing Processes, Aug 2024

   Abs DOI

   Hot-wire directed energy deposition using a laser beam (DED-LB/w) is a method of metal additive manufacturing (AM) that has benefits of high material utilization and deposition rate, but parts manufactured by DED-LB/w suffer from a substantial heat input and undesired surface finish. Hence, regulating the process parameters and monitoring the process signatures to control the final quality during the deposition is crucial to ensure the quality of the final part. This paper explores the dynamic modeling of the DED-LB/w process and introduces a parameter-signature-quality modeling and control approach to enhance the quality of modeling and control of part qualities that cannot be measured in situ. The study investigates different process parameters that influence the melt pool width (signature) and bead width (quality) in single and multi-layer beads. The proposed modeling approach utilizes a parameter-signature model as F1 and a signature-quality model as F2. Linear and nonlinear modeling approaches are compared to describe a dynamic relationship between process parameters and a process signature, the melt pool width (F1). A fully connected artificial neural network is employed to model and predict the final part quality, i.e., bead width, based on melt pool signatures (F2). Finally, the effectiveness and usefulness of the proposed parameter-signature-quality modeling is tested and verified by integrating the parameter-signature (F1) and signature-quality (F2) models in the closed-loop control of the width of the part. Compared with the control loop with only F1, the proposed method shows clear advantages and bears potential to be applied to control other part qualities that cannot be directly measured or monitored in situ.

4. Physics-Informed Online Learning for Temperature Prediction in Metal AM

   Pouyan Sajadi, Mostafa Rahmani Dehaghani, Yifan Tang, and 1 more author

   Materials, Aug 2024

   Abs DOI

   In metal additive manufacturing (AM), precise temperature field prediction is crucial for process monitoring, automation, control, and optimization. Traditional methods, primarily offline and data-driven, struggle with adapting to real-time changes and new process scenarios, which limits their applicability for effective AM process control. To address these challenges, this paper introduces the first physics-informed (PI) online learning framework specifically designed for temperature prediction in metal AM. Utilizing a physics-informed neural network (PINN), this framework integrates a neural network architecture with physics-informed inputs and loss functions. Pretrained on a known process to establish a baseline, the PINN transitions to an online learning phase, dynamically updating its weights in response to new, unseen data. This adaptation allows the model to continuously refine its predictions in real-time. By integrating physics-informed components, the PINN leverages prior knowledge about the manufacturing processes, enabling rapid adjustments to process parameters, geometries, deposition patterns, and materials. Empirical results confirm the robust performance of this PI online learning framework in accurately predicting temperature fields for unseen processes across various conditions. It notably surpasses traditional data-driven models, especially in critical areas like the Heat Affected Zone (HAZ) and melt pool. The PINN’s use of physical laws and prior knowledge not only provides a significant advantage over conventional models but also ensures more accurate predictions under diverse conditions. Furthermore, our analysis of key hyperparameters—the learning rate and batch size of the online learning phase—highlights their roles in optimizing the learning process and enhancing the framework’s overall effectiveness. This approach demonstrates significant potential to improve the online control and optimization of metal AM processes.

5. An Online Sequential Update Method for Reduced-Order-Model-Based Digital Twin

   Yifan Tang, Pouyan Sajadi, Mostafa Rahmani Dehaghani, and 1 more author

   In International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Aug 2024

   Abs DOI

   Digital twin (DT) refers to any model that reflects a physical system and remains updated with the real-time data from the physical system. Recently, DT constructed with a reduced-order model (DT-ROM) has been applied to aerodynamics and structure health monitoring tasks, whose partial differential equations (PDEs) are used to design the online update formula for reduced bases and coefficients in DT-ROM. Although such online update methods improve the performance of DT-ROM, they are not applicable when the PDEs of a system are unknown. This paper focuses on the online update task where a system is modeled with a DT-ROM, but the system’s PDEs are unknown. To tackle the task, an online sequential update method is proposed. During the update process, the projection residual of online data on the reduced bases is applied to determine whether to update the reduced bases in ROM, and the prediction residual of online data obtained by the offline DT-ROM is adopted to infer whether to update the coefficient models. By checking both criteria sequentially, the online update method selects the necessary online data for the DT update. Three numerical problems and one engineering problem are designed to test the proposed online update method. Testing results demonstrate that (a) the proposed online update method reduces both the projection and prediction residuals gradually, and (b) the performance of the offline DT-ROM model on the testing data is improved gradually during the online update process. Both indicate that the proposed method could be applied to online DT update tasks where the PDEs of the system are unknown.

6. Low-Cost Melt Pool Temperature Prediction Using Visible Light Camera and Machine Learning in Laser Hot-Wire Directed Energy Deposition

   Mostafa Rahmani Dehaghani, Pouyan Sajadi, Yifan Tang, and 1 more author

   In International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Aug 2024

   Abs DOI

   Laser wire directed energy deposition (DED-LB/w) offers notable efficiency in metal additive manufacturing (AM), allowing rapid component fabrication with high material utilization. Despite its advantages, challenges such as anisotropy and uneven mechanical properties arise from unregulated heat application, highlighting the need for precise temperature control during deposition. This study evaluates the use of visible light images to predict melt pool temperature, leveraging Convolutional Neural Networks (CNN), Gaussian Process Regression (GPR), and Artificial Neural Networks (ANN). While CNN is trained directly on images, GPR and ANN utilize extracted features such as melt pool dimensions. The CNN model notably excels, achieving an R-squared value of 0.981, root mean square error of 44.46, and mean absolute percentage error of 2.74%, demonstrating the superior capability of visible light imaging in accurately predicting the melt pool temperature. This success illustrates the considerable potential of integrating predictive models with visible light imaging as a cost-effective alternative to traditional sensory systems. Such integration not only offers a pragmatic solution to the high costs and complexities associated with thermal imaging but also opens new avenues for combining other measurement techniques such as pyrometers to further enhance the prediction accuracy for AM process control. Future efforts will concentrate on implementing these models in real-time metal printing, aiming to enhance microstructure control and advance process automation.

7. An Online Sequential Update Method for ROM-based Digital Twin

   Yifan Tang

   Jul 2024

   WCCM 2024 / PANACM 2024, Vancouver, Canada

2023

1. Online Thermal Field Prediction for Metal Additive Manufacturing of Thin Walls

   Yifan Tang, Mostafa Rahmani Dehaghani, Pouyan Sajadi, and 6 more authors

   Journal of Manufacturing Processes, Jul 2023

   Abs DOI

   Various data-driven modeling methods have been developed to predict the thermal field in metal additive manufacturing (AM). The generalization capability of these models has been shown with simulation, but rarely tested with online physical printing. Instead, this paper aims to study a practical issue in metal AM, i.e., how to predict the thermal field of yet-to-print parts online when only a few sensors are available. This work proposes an online thermal field prediction method using mapping and reconstruction, which could be integrated into a metal AM process for online performance control. Based on the similarity of temperature curves (curve segments of a temperature profile of one point), the thermal field mapping applies an artificial neural network to estimate the temperature curves of points on the yet-to-print layer from measured temperatures of certain points on the previously printed layer. With measured/predicted temperature profiles of several points on the same layer, the thermal field reconstruction proposes a reduced order model (ROM) to construct the temperature profiles of all points on the same layer, which could be used to build the temperature field of the entire layer. The training of ROM is performed with an extreme learning machine (ELM) for computational efficiency. Fifteen wire arc AM experiments and nine simulations are designed for thin walls with a fixed length and unidirectional printing of each layer. The test results indicate that the proposed prediction method could construct the thermal field of a yet-to-print layer within 0.1 s on a low-cost desktop computer (Intel Core i7-3770 CPU @ 3.40GHz processor, 24.0 GB RAM). Meanwhile, the method has acceptable generalization capability in most cases from lower layers to higher layers in the same simulation, as well as from one simulation to a new simulation on different AM process parameters. More importantly, after fine-tuning the proposed method with limited experimental data, the relative errors of all predicted temperature profiles on a new experiment are sufficiently small, which demonstrates the applicability and generalization of the proposed thermal field prediction method in online applications for metal AM.

2. Review of Transfer Learning in Modeling Additive Manufacturing Processes

   Yifan Tang, M Rahmani Dehaghani, and G Gary Wang

   Additive Manufacturing, Jul 2023

   Abs DOI

   Modeling plays an important role in the additive manufacturing (AM) process and quality control. In practice, however, only limited data are available for each product due to the relatively high AM cost, which brings challenges in building either a high-quality physics-based or data-based model. Transfer learning (TL) is a new and promising group of approaches where the model of one product (source) may be reused for another product (target) with limited new target data. This paper focuses on reviewing applications of TL in AM modeling to help advance research in this area. First, notations, definitions, and categories of TL methods are introduced along with their application scenarios. Then current applications of TL in AM modeling are summarized along with their limitations. Based on reviewed applications, recommendations are given on how to apply TL for a certain AM problem, from the perspectives of source domain determination, TL method selection, target data generation, and data preprocessing. Finally, future research directions about TL in AM modeling are discussed in the hope to explore more potential of TL in improving the AM model quality with limited data.

3. Iterative Uncertainty Calibration for Modeling Metal Additive Manufacturing Processes Using Statistical Moment-Based Metric

   Mostafa Rahmani Dehaghani, Yifan Tang, and G Gary Wang

   Journal of Mechanical Design, Jul 2023

   Abs DOI

   Metal additive manufacturing (AM) has recently attracted attention due to its potential for batch/mass production of metal parts. This process, however, currently suffers from problems including low productivity, inconsistency in the properties of the printed parts, and defects such as lack of fusion and keyholing. Finite element (FE) modeling cannot accurately model the metal AM process and has a high computational cost. Empirical models based on experiments are time-consuming and expensive. This paper enhances a previously developed framework that takes advantages of both empirical and FE models. The validity and accuracy of the metamodel developed in the earlier framework depend on the initial assumption of parameter uncertainties. This causes a problem when the assumed uncertainties are far from the actual values. The proposed framework introduces an iterative calibration process to overcome this limitation. After comparing several calibration metrics, the second-order statistical moment-based metric (SMM) was chosen as the calibration metric in the improved framework. The framework is then applied to a four-variable porosity modeling problem. The obtained model is more accurate than using other approaches with only ten available experimental data points for calibration and validation.

4. Modeling And Optimization of Height-Related Geometrical Parameters for Thin Wall Structures Manufactured by Metal Additive Manufacturing

   M Rahmani Dehaghani, Yifan Tang, Suraj Panicker, and 3 more authors

   The International Journal of Advanced Manufacturing Technology, Jul 2023

   Abs DOI

   Cold metal transfer wire arc additive manufacturing (CMT-WAAM) has attracted attention in recent years due to its ability to print walls with higher dimensional accuracy than regular WAAM. To print near-net shape parts by CMT-WAAM, there is a need to define a set of height-related geometrical parameters (HGPs) that can capture, quantify, and compare the quality of the height of the produced parts. In the presenting study, a set of HGPs, namely, the average height error, maximum height variation, and average absolute slope are defined and assessed. Fifteen single-track multi-layer walls are printed to check the effect of process parameters on the defined HGPs. It is found that the stability and quality of the print cannot be guaranteed by checking the visual appearance of the single beads and at least five-to-ten-layer walls should be printed. It is also found that the travel speed and the wire feed speed have positive monotonic relationships with average absolute slope and maximum height variation, respectively. Correlations between process parameters and HGPs are modeled and optimized using multi-objective optimization, and a validation test is performed to check the validity of the developed models. Moreover, HGPs of walls printed using unidirectional and bidirectional path strategies are calculated and compared. Defined HGPs are able to quantify, capture, and compare the quality of height of a wall with only three parameters. The HGPs can be used in further studies to report and compare the quality of height of thin wall structures.

5. Comparison of Transfer Learning based Additive Manufacturing Models via A Case Study

   Yifan Tang, M Rahmani Dehaghani, and G Gary Wang

   arXiv preprint arXiv:2305.11181, Jul 2023

   DOI
6. Layer-to-Layer Thermal History Prediction for Thin Walls in Metal Additive Manufacturing

   Yifan Tang, Shahriar Bakrani Balani, Akshay Dhalpe, and 5 more authors

   In International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Aug 2023

   Abs DOI

   Thermal history has a great effect on the part properties in metal additive manufacturing such as tensile strength and hardness. To study and control the thermal behavior of the AM processes, various data-driven thermal history modeling methods have been developed and tested on simulation data. However, their in-situ application scenarios are rarely explored and discussed. This paper aims to provide a layer-to-layer thermal history prediction model, which enables predicting the thermal history of a yet-to-print layer based on the data measured from the printed lower layers. First, the thermal behavior is analyzed to reveal the similarities in temperature curves of two successive layers. Then four input variables are identified, including the deposition rate, the relative height of the layer, the printing time, and the dwell time of one layer. Based on the selected input variables and the temperature output, a fully connected neural network with residual connection is designed to simplify the training process. Five numerical simulations are designed to collect temperature curves (curve segments of a temperature profile) on each layer, and one experimental study on wire arc additive manufacturing is completed to record the temperature curves. Based on the collected data, three cases are proposed to test the modeling framework, such as (a) dividing all simulation data into the training set and validation set, (b) training the model based on four simulation runs and its validation with the last simulation, and (c) training the model with all simulation data and testing on the experimental data. The former two cases show great prediction accuracies with a relative error of less than 5% in most cases, which indicates its potential in online prediction when trained with data from the same input conditions or same systems. While the applicability of the model trained only with simulation data should be explored further in real experiments.

7. Defining and Modeling of Height-Related Geometrical Parameters for Thin Wall Structures Manufactured by Metal Additive Manufacturing

   Mostafa Rahmani Dehaghani, Yifan Tang, Suraj Panicker, and 3 more authors

   In International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Aug 2023

   Abs DOI

   Cold metal transfer wire and arc additive manufacturing (CMT-WAAM) is a type of metal additive manufacturing technology that has attracted attention in recent years due to its ability to print walls with less dimensional inaccuracies than regular WAAM. To print near net shape parts by CMT-WAAM, there is a need to define a set of height-related geometrical parameters (HGPs) that can capture, quantify, and compare the quality of the height of the produced parts. In this study, a set of HGPs, namely, the average height error (AHE), maximum height variation (MHV), and average absolute slope (AAS) are defined and assessed. Fifteen single-track multi-layer walls are printed to check the effect of process parameters on the defined HGPs. It is found that the stability of the print cannot be guaranteed by checking the visual appearance of the single beads and at least five-to-ten-layer walls should be printed to check the quality and stability of the print. It is also found that the travel speed (TS) and wire feed speed (WFS) have positive monotonic relationships with AAS and MHV, respectively. Correlations between process parameters and HGPs are modeled and validation tests are performed to check the validity of the developed models. The defined three HGPs are shown to be able to quantify, capture, and compare the quality of a height of a wall and can be used for other similar applications.

2022

1. Surrogate-Assisted Differential Evolution Using Knowledge-Transfer-Based Sampling for Expensive Optimization Problems

   Teng Long, Nianhui Ye, Renhe Shi, and 2 more authors

   AIAA Journal, Aug 2022

   DOI
2. Review of Transfer Learning in Additive Manufacturing Modeling

   Yifan Tang, M. Rahmani Dehaghani, and G. Gary Wang

   In International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Aug 2022

   Abs DOI

   The process-structure-property modeling of additive manufacturing (AM) products plays an important role in process and quality control. In practice however, only limited data are available for each product due to its expensive material and time-consuming fabricating process, which becomes an obstacle to achieve high quality models. Transfer learning (TL) is a new and promising approach that the model of one product (source) may be reused for another product (target) with limited new data on the target. This paper focuses on reviewing applications of TL in AM modeling in order to help further research in this area. To clarify the specific topic, the problem definition is presented, as well as the differences between TL, multi-fidelity modeling, and multi-task learning. Then current applications of TL in AM modeling are summarized according to different TL approaches. To better understand the performances of different TL approaches, several representative TL-assisted AM modeling methods are reproduced and tested on an open-source dataset. Based on the test results, their effectiveness and limitations are discussed in detail. Finally, future research directions about TL in AM modeling are discussed in hope to explore more potential of TL in boosting the AM model performance.

3. Iterative Uncertainty Calibration for Modeling Metal Additive Manufacturing Processes Using Statistical Moment-Based Metric

   M. Rahmani Dehaghani, Yifan Tang, and G. Gary Wang

   In International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Aug 2022

   Abs DOI

   Metal additive manufacturing (AM) has recently attracted attention due to its potential for batch/mass production of metal parts. This process, however, currently suffers from problems including low productivity, inconsistency in the properties of the printed parts, and defects such as lack of fusion, keyholing, and un-melted powders. Finite Element (FE) modeling cannot accurately model the metal AM process and has a high computational cost. Empirical models based on experiments are time-consuming and expensive. This paper improves a previously developed framework that takes advantages of both empirical and FE models. The validity and accuracy of the metamodel developed in the previous framework depend on the initial assumption of parameter uncertainties. This causes a problem when the assumed uncertainties are far from the actual values. The proposed framework introduces an iterative calibration process to overcome this limitation. In addition, the u_pooling metric used as the calibration metric in the previous framework is found not as good as the second-order statistical moment-based metric (SMM), after comparing several calibration metrics. The proposed framework is then applied to a four-variable porosity modeling problem. The obtained model is more accurate than using other approaches with only 10 available experimental data points for calibration and validation.

2021

1. Multi-Agent Deep Reinforcement Learning for Solving Large-scale Air Traffic Flow Management Problem: A Time-Step Sequential Decision Approach

   Yifan Tang and Yan Xu

   In 2021 IEEE/AIAA 40th Digital Avionics Systems Conference (DASC), Aug 2021

   DOI

2020

1. Sequential Radial Basis Function-Based Optimization Method Using Virtual Sample Generation

   Yifan Tang, Teng Long, Renhe Shi, and 2 more authors

   Journal of Mechanical Design, May 2020

   Abs DOI

   To further reduce the computational expense of metamodel-based design optimization (MBDO), a novel sequential radial basis function (RBF)-based optimization method using virtual sample generation (SRBF-VSG) is proposed. Different from the conventional MBDO methods with pure expensive samples, SRBF-VSG employs the virtual sample generation mechanism to improve the optimization efficiency. In the proposed method, a least squares support vector machine (LS-SVM) classifier is trained based on expensive real samples considering the objective and constraint violation. The classifier is used to determine virtual points without evaluating any expensive simulations. The virtual samples are then generated by combining these virtual points and their Kriging responses. Expensive real samples and cheap virtual samples are used to refine the objective RBF metamodel for efficient space exploration. Several numerical benchmarks are tested to demonstrate the optimization capability of SRBF-VSG. The comparison results indicate that SRBF-VSG generally outperforms the competitive MBDO methods in terms of global convergence, efficiency, and robustness, which illustrates the effectiveness of virtual sample generation. Finally, SRBF-VSG is applied to an airfoil aerodynamic optimization problem and a small Earth observation satellite multidisciplinary design optimization problem to demonstrate its practicality for solving real-world optimization problems.

2019

1. Multidisciplinary Design Optimization of Long-Range Slender Guided Rockets Considering Aeroelasticity and Subsidiary Loads

   Zhao Wei, Teng Long, Renhe Shi, and 2 more authors

   Aerospace Science and Technology, May 2019

   Abs DOI

   Due to the insufficient rigidity feature and long-range requirement, it is crucial to consider the aeroelasticity and subsidiary loads caused by the Earth rotation and fuel consumption when designing long-range slender guided rockets (LRSGRs). As a typical multidisciplinary design optimization (MDO) problem, the design optimization of LRSGRs confronts two critical challenges, i.e., accurate multidisciplinary modeling and efficient global optimization. To address the challenges, a novel MDO framework including MDO problem definition, multidisciplinary modeling, and metamodel-based optimizer is developed for LRSGR design. The LRSGR MDO problem is formulated to minimize the total mass subject to a number of practical engineering constraints such as bending mode frequencies, miss distance, and fall angle. Several disciplinary models including structure, aerodynamics, propulsion, mass, aeroelasticity, guidance control, and trajectory are established. To enhance the analysis accuracy, structural finite element analysis (FEA), three-channel autopilot, and high-fidelity trajectory models are adopted. In the aeroelasticity model, the unsteady aerodynamic loads are calculated by slender body theory and aerodynamic derivative method. The subsidiary loads including subsidiary Coriolis force, centrifugal inertial force, Coriolis force, and subsidiary Coriolis moment are incorporated in the trajectory model of LRSGRs. Since structural finite element, aeroelasticity, and trajectory models are computationally expensive (about 1.8 hours for one trial of system analysis on a well-equipped workstation), an adaptive radial basis function metamodel-based optimizer is integrated in the framework to solve the LRSGR MDO problem with moderate computational cost. The total mass of the studied LRSGR is decreased by 88 kg (i.e., 14% of the total mass) after optimization, which demonstrates the effectiveness and practicability of the proposed MDO framework for LRSGRs.

2. Efficient Aero-Structure Coupled Wing Optimization Using Decomposition and Adaptive Metamodeling Techniques

   Teng Long, Yufei Wu, Zhu Wang, and 3 more authors

   Aerospace Science and Technology, May 2019

   Abs DOI

   This paper proposes an efficient decomposition-based optimization framework using adaptive metamodeling for expensive aero-structure coupled wing optimization problems. First, high-fidelity aero-structure coupled analysis method is presented through spatial interpolations of distributed aerodynamic loads and structural deformation. The proposed optimization framework decomposes the original complex coupled optimization problem into 2-D airfoil optimization (i.e., Stage-I) and 3-D wing optimization (i.e., Stage-II) to alleviate the expensive computational costs. In Stage-II, a two-level optimization strategy is tailored to further decompose the 3-D wing optimization into system-level and subsystem-level for dimension reduction. An adaptive response surface method using intelligent space exploration strategy is used to perform optimization tasks involving CFD simulations (i.e., 2-D airfoil optimization and system-level optimization in Stage-II), while sequential quadratic programming is employed to optimize the massive subsystem structure sizing variables. The effectiveness of two-level optimization strategy is validated on a numerical testing problem and a simplified wing optimization case. Finally, the developed models and proposed methods are successfully applied to aero-structure coupled optimization of a high aspect ratio wing. The optimization results demonstrate the effectiveness of the developed aero-structural analysis models and efficiency of the proposed optimization methods.

3. Filter-Based Adaptive Kriging Method for Black-Box Optimization Problems with Expensive Objective and Constraints

   Renhe Shi, Li Liu, Teng Long, and 2 more authors

   Computer Methods in Applied Mechanics and Engineering, May 2019

   Abs DOI

   To reduce the computational cost of solving engineering design optimization problems with both expensive objective and constraints, a novel filter-based adaptive Kriging method notated as FLT-AKM is proposed in this paper. In FLT-AKM, a probability of constrained improvement (PCI) criterion is developed based on the notion of filter to sequentially generate new samples for updating Kriging metamodels of objective and constraints. At each iteration, an infill sample point is allocated at the position where the PCI is maximized to achieve potential improvement in optimality and feasibility. And the Kriging metamodels are consecutively updated by the newly-added infill sample points, which leads the FLT-AKM search to rapidly converge to the global optimum. The performance of the proposed FLT-AKM method is tested on a number of numerical benchmark problems via comparing with several widely-used metamodel-based constrained optimization methods. The comparison results indicate that FLT-AKM generally outperforms the competitors in terms of global convergence and efficiency performance. Finally, FLT-AKM is successfully applied to an all-electric GEO satellite MDO problem. The optimization results show that FLT-AKM is able to find a better feasible design with fewer computational budgets compared with our previous study, which demonstrates the effectiveness and practicality of the proposed FLT-AKM method for solving real-world expensive black-box engineering design optimization problems.

4. Filter-Based Sequential Radial Basis Function Method for Spacecraft Multidisciplinary Design Optimization

   Renhe Shi, Li Liu, Teng Long, and 2 more authors

   AIAA Journal, May 2019

   Abs DOI

   Spacecraft system design is practically a complex multidisciplinary design optimization (MDO) problem. Because of the application of high-fidelity simulation models, the massive computational cost of spacecraft MDO problems becomes a bottleneck and challenging problem in engineering practices. To address the issue, this paper proposes a novel filter-based sequential radial basis function (FSRBF) method for effectively and efficiently solving spacecraft MDO problems. In FSRBF, to handle expensive constraints, a filter is constructed, augmented, and refined based on the concept of Pareto nondomination, which is then combined with a support vector machine (SVM) to construct the filter-based region of interest (FROI) for sequentially bias sampling. During the optimization process, the expensive multidisciplinary analysis process is approximated by RBF metamodels to reduce the computational cost. The RBF metamodels are gradually updated via consecutively sampling within the FROI, which leads the search to rapidly reach the feasible optimum. A number of numerical benchmark problems are used to demonstrate the desirable performance of the proposed FSRBF compared with several alternative methods. In the end, FSRBF is applied to the design of an all-electric GEO satellite and a small Earth observation satellite to illustrate its capability for real-world spacecraft MDO problems. The results show that FSRBF can successfully obtain feasible solutions to improve the design quality of satellite systems. Moreover, the required computational cost of FSRBF is much lower than that of competitive methods, which illustrates the effectiveness and practicality of the proposed FSRBF for solving spacecraft MDO problems.

5. Aero-structure Coupled Optimization for High Aspect Ratio Wings Using Multi-model Fusion Method

   Yifan Tang, Jing Sun, Teng Long, and 2 more authors

   In 2019 IEEE 9th Annual International Conference on CYBER Technology in Automation, Control, and Intelligent Systems (CYBER), May 2019

   DOI

2018

1. Dual-Sampling based Co-Kriging Method for Design Optimization Problems with Multi-Fidelity Models

   Renhe Shi, Li Liu, Teng Long, and 2 more authors

   In 2018 Multidisciplinary Analysis and Optimization Conference, May 2018

   DOI
2. Small-Sample Learning Enhanced Sequential Radial Basis Function for Expensive Aerospace System Design Optimization

   Yifan Tang

   Nov 2018

   Asian Joint Symposium on Aerospace Engineering, Gyeongju, Korea