Search results

The purpose of this study is to establish a novel airfoil icing prediction model using deep learning with geometrical constraints, called geometrical constraints enhancement neural networks, to improve the prediction accuracy compared to the non-geometrical constraints model.

The model is developed with flight velocity, ambient temperature, liquid water content, median volumetric diameter and icing time taken as inputs and icing thickness given as outputs. To enhance the icing prediction accuracy, the model involves geometrical constraints into the loss function. Then the model is trained according to icing samples of 2D NACA0012 airfoil acquired by numerical simulation.

The results show that the involvement of geometrical constraints effectively enhances the prediction accuracy of ice shape, by weakening the appearance of fluctuation features. After training, the airfoil icing prediction model can be used for quickly predicting airfoil icing.

This work involves geometrical constraints in airfoil icing prediction model. The proposed model has reasonable capability in the fast assessment of aircraft icing.

Numerical simulation of the single point incremental forming (SPIF) processes can be very demanding and time consuming due to the constantly changing contact conditions between the tool and the sheet surface, as well as the nonlinear material behaviour combined with non-monotonic strain paths. The purpose of this paper is to propose an adaptive remeshing technique implemented in the in-house implicit finite element code LAGAMINE, to reduce the simulation time. This remeshing technique automatically refines only a portion of the sheet mesh in vicinity of the tool, therefore following the tool motion. As a result, refined meshes are avoided and consequently the total CPU time can be drastically reduced.

SPIF is a dieless manufacturing process in which a sheet is deformed by using a tool with a spherical tip. This dieless feature makes the process appropriate for rapid-prototyping and allows for an innovative possibility to reduce overall costs for small batches, since the process can be performed in a rapid and economic way without expensive tooling. As a consequence, research interest related to SPIF process has been growing over the last years.

In this work, the proposed automatic refinement technique is applied within a reduced enhanced solid-shell framework to further improve numerical efficiency. In this sense, the use of a hexahedral finite element allows the possibility to use general 3D constitutive laws. Additionally, a direct consideration of thickness variations, double-sided contact conditions and evaluation of all components of the stress field are available with solid-shell and not with shell elements. Additionally, validations by means of benchmarks are carried out, with comparisons against experimental results.

It is worth noting that no previous work has been carried out using remeshing strategies combined with hexahedral elements in order to improve the computational efficiency resorting to an implicit scheme, which makes this work innovative. Finally, it has been shown that it is possible to perform accurate and efficient finite element simulations of SPIF process, resorting to implicit analysis and continuum elements. This is definitively a step-forward on the state-of-art in this field.

Limited geometric accuracy is one of the major challenges that hinder the wider application of additive manufacturing (AM). This paper aims to predict in-plane shape deviation for online inspection and compensation to prevent error accumulation and improve shape fidelity in AM.

A sequence-to-sequence model with an attention mechanism (Seq2Seq+Attention) is proposed and implemented to predict subsequent layers or the occluded toolpath deviations after the multiresolution alignment. A shape compensation plan can be performed for the large deviation predicted.

The proposed Seq2Seq+Attention model is able to provide consistent prediction accuracy. The compensation plan proposed based on the predicted deviation can significantly improve the printing fidelity for those layers detected with large deviations.

Based on the experiments conducted on the knee joint samples, the proposed method outperforms the other three machine learning methods for both subsequent layer and occluded toolpath deviation prediction.

This work fills a research gap for predicting in-plane deviation not only for subsequent layers but also for occluded paths due to the missing scanning measurements. It is also combined with the multiresolution alignment and change point detection to determine the necessity of a compensation plan with updated G-code.

Continuous annealing (CA) units usually lack a physical shapemeter; consequently, real-time display and closed-loop control of the strip shape are impossible to achieve.

A shape model for the CA process is established in this study. Specifically, a virtual shapemeter and closed-loop control system based on the advanced parameter acquisition system and information transmission of CA units are developed in C++ programming language. This system realises real-time dynamic shape display, closed-loop control and shape prediction by collecting raw data of steel coils and parameters during CA.

Field test results show that the shape predicted by the virtual shapemeter coincides with the measured shape by over 90 per cent, which fully meets the precision requirement of industrial applications.

Moreover, shape quality is effectively improved without increasing hardware investments.

The purpose of this paper is to reduce the overall computational time of aerodynamic shape optimization that involves accurate high-fidelity simulation models.

The proposed approach is based on the surrogate-based optimization paradigm. In particular, multi-fidelity surrogate models are used in the optimization process in place of the computationally expensive high-fidelity model. The multi-fidelity surrogate is constructed using physics-based low-fidelity models and a proper correction. This work introduces a novel correction methodology – referred to as the adaptive response prediction (ARP). The ARP technique corrects the low-fidelity model response, represented by the airfoil pressure distribution, through suitable horizontal and vertical adjustments.

Numerical investigations show the feasibility of solving real-world problems involving optimization of transonic airfoil shapes and accurate computational fluid dynamics simulation models of such surfaces. The results show that the proposed approach outperforms traditional surrogate-based approaches.

The proposed aerodynamic design optimization algorithm is novel and holistic. In particular, the ARP correction technique is original. The algorithm is useful for fast design of aerodynamic surfaces using high-fidelity simulation data in moderately sized search spaces, which is challenging using conventional methods because of excessive computational costs.

Electromagnetic design utilising finite element or similar numerical methods is computationally expensive, thus efficient algorithms reducing the number of objective function calls to locate the optimum are sought. The balance between exploration and exploitation may be achieved using a reinforcement learning approach, as demonstrated previously. However, in practical design problems, in addition to finding the global optimum efficiently, information about the robustness of the solution may also be important. In this paper, the aim is to discuss the suitability of different search algorithms and to present their fitness to solve the optimization problem in conjunction with providing enough information on the robustness of the solution.

Two novel strategies enhanced by the surrogate model based weighted expected improvement approach are discussed. The algorithms are tested using a two‐variable test function. The emphasis of these strategies is on accurate approximation of the shape of the objective function to accomplish a robust design.

The two novel strategies aim to pursue the optimal value of weights for exploration and exploitation throughout the iterative process for better prediction of the shape of the objective function.

It is argued that the proposed strategies based on adaptively tuning weights perform better in predicting the shape of the objective function. Good accuracy of predicting the shape of the objective function is crucial for achieving a robust design.

This paper aims to present an integrated optimisation‐modelling computational approach for virtual prototyping that helps design engineers to improve the reliability and performance of electronic components and systems through design optimisation at the early product development stage. The design methodology is used to identify the optimal design of lead‐free (Sn3.9Ag0.6Cu) solder joints in fine‐pitch copper column bumped flip‐chip electronic packages.

The design methodology is generic and comprises numerical techniques for computational modelling (finite element analysis) coupled with numerical methods for statistical analysis and optimisation. In this study, the integrated optimisation‐modelling design strategy is adopted to prototype virtually a fine‐pitch flip‐chip package at the solder interconnect level, so that the thermal fatigue reliability of the lead‐free solder joints is improved and important design rules to minimise the creep in the solder material, exposed to thermal cycling regimes, are formulated. The whole prototyping process is executed in an automated way once the initial design task is formulated and the conditions and the settings for the numerical analysis used to evaluate the flip‐chip package behaviour are specified. Different software modules that incorporate the required numerical techniques are used to identify the solution of the design optimisation problem related to solder joints reliability optimisation.

For fine‐pitch flip‐chip packages with copper column bumped die, it is found that higher solder joint volume and height of the copper column combined with lower copper column radius and solder wetting around copper column have a positive effect on the thermo‐mechanical reliability.

The findings of this research provide design rules for more reliable lead‐free solder joints for copper column bumped flip‐chip packages and help to establish further the technology as one of the viable routes for flip‐chip packaging.

Coil shape quality is the external representation of strip product quality, and it is also a direct reflection of strip production process level. This paper aims to predict the coil shape results in advance based on the real-time data through the designed algorithm.

Aiming at the strip production scale and coil shape application requirements, this paper proposes a strip coil shape defects prediction algorithm based on Siamese semi-supervised denoising auto-encoder (DAE)-convolutional neural networks. The prediction algorithm first reconstructs the information eigenvectors using DAE, then combines the convolutional neural networks and skip connection to further process the eigenvectors and finally compares the eigenvectors with the full connect neural network and predicts the strip coil shape condition.

The performance of the model is further verified by using the coil shape data of a steel mill, and the results show that the overall prediction accuracy, recall rate and F-measure of the model are significantly better than other commonly used classification models, with each index exceeding 88%. In addition, the prediction results of the model for different steel grades strip coil shape are also very stable, and the model has strong generalization ability.

This research provides technical support for the adjustment and optimization of strip coil shape process based on the data-driven level, which helps to improve the production quality and intelligence level of hot strip continuous rolling.

One major problem preventing further application and benefits from additive manufacturing (AM) nowadays is that AM build parts always end up with poor geometrical quality. To help improving geometrical quality for AM, this study aims to propose geometrical deviation identification and prediction method for AM, which could be used for identifying the factors, forms and values of geometrical deviation of AM parts.

This paper applied the skin model-based modal decomposition approach to describe the geometrical deviations of AM and decompose them into different defect modes. On that basis, the approach to propose and extend defect modes was developed. Identification and prediction of the geometrical deviations were then carried out with this method. Finally, a case study with cylinders manufactured by fused deposition modeling was introduced. Two coordinate measuring machine (CMM) machines with different measure methods were used to verify the effectiveness of the methods and modes proposed.

The case study results with two different CMM machines are very close, which shows that the method and modes proposed by this paper are very effective. Also, the results indicate that the main geometrical defects are caused by the shrinkage and machine inaccuracy-induced errors which have not been studied enough.

This work could be used for identifying and predicting the forms and values of AM geometrical deviation, which could help realize the improvement of AM part geometrical quality in design phase more purposefully.

This paper gives a review of the finite element techniques (FE) applied in the area of material processing. The latest trends in metal forming, non‐metal forming, powder metallurgy and composite material processing are briefly discussed. The range of applications of finite elements on these subjects is extremely wide and cannot be presented in a single paper; therefore the aim of the paper is to give FE researchers/users only an encyclopaedic view of the different possibilities that exist today in the various fields mentioned above. An appendix included at the end of the paper presents a bibliography on finite element applications in material processing for 1994‐1996, where 1,370 references are listed. This bibliography is an updating of the paper written by Brannberg and Mackerle which has been published in Engineering Computations, Vol. 11 No. 5, 1994, pp. 413‐55.