Dequn Li

Numerical Filling Simulation of Injection Molding Based on 3D Finite Element Model

The traditional numerical simulation of injection molding is the 2.5D technique based on the Hele-Shaw approximation, which cannot predict the filling of thick or nonuniform-thickness parts and some flow behaviors such as the fountain flow. A three-dimensional (3D) finite element model is presented to perform more accurate simulations of a fully 3D flow. The model employs an equal-order velocity-pressure interpolation method. The relation between velocity and pressure is obtained from the discretized momentum equations in order to derive the pressure equation. A 3D control volume scheme is introduced to track the flow front. The validity of the model has been tested by case studies and experimental verification.

Numerical simulation of temperature history during the picture tube panel‐forming process

Purpose – This paper aims to develop an integrated cooling simulation for the temperature history of the panel during the forming process.Design/methodology/approach – A local one‐dimensional transient analysis in the thickness direction is adopted for the panel part, which employs finite‐difference method. And a three‐dimensional, boundary element method is used for the numerical implementation of the heat transfer analysis in the mold region, which is considered as three‐dimensional conduction. The Renormalization‐Group turbulence model is applied for the jet impinging cooling. The part and mold analyses are coupled so as to match the temperature and heat flux on the glass‐mold interface.Findings – The paper provides mathematical model and numerical strategy adapted to the problem, with experimental verification that shows a good agreement.Practical implications – Cooling design in the forming operation of picture tube panel is of great importance because it significantly affects the part quality associ...

Study on ultrasonic vibration-assisted adhesive bonding of CFRP joints

Abstract As adhesive bonding does not disrupt the continuity of the fiber integral, it has become an important bonding method of CFRP (carbon fiber-reinforced plastics) joints. Regarding the quality problems in the traditional adhesive bonding process, i.e. the low and unstable adhesion strength, a novel ultrasonic vibration-assisted adhesive bonding method for CFRP joints is proposed. In this method, extra force caused by ultrasonic vibration is introduced into the bonding process to reinforce it. The strengthening mechanism is then analyzed. According to our study, it is found that: (1) ultrasonic vibration can improve the adhesion strength and stability by 52 and 66% in the test. (2) The strengthening mechanism is summarized as: high-frequency vibration contact between adhesive and wall is produced under ultrasonic vibration, which forms better contact; adhesive is driven to penetrate into the surface roughness of the bonding area under ultrasonic vibration, which increases the effective bonding area.

Automatic feature constructing from vibration signals for machining state monitoring

Machining state monitoring is an important subject for intelligent manufacturing. Feature construction is accepted to be the most critical procedure for a signal-based monitoring system and has attracted a lot of research interest. The traditional manual constructing way is skill intensive and the performance cannot be guaranteed. This paper presented an automatic feature construction method which can reveal the inherent relationship between the input vibration signals and the output machining states, including idling moving, stable cutting and chatter, using a reasonable and mathematical way. Firstly a large signal set is carefully prepared by a series of machining experiments followed by some necessary preprocessing. And then, a deep belief network is trained on the signal set to automatically construct features using the two step training procedure, namely unsupervised greedily layer-wise pertaining and supervised fine-tuning. The automatically extracted features can exactly reveal the connection between the vibration signal and the machining states. Using the automatic extracted features, even a linear classifier can easily achieve nearly 100% modeling accuracy and wonderful generalization performance, besides good repeatability precision on a large well prepared signal set. For the actual online application, voting strategy is introduced to smooth the predicted states and make the final state identification to ensure the detection reliability by taking consideration of the machining history. Experiments proved the proposed method to be efficient in protecting the workpiece from serious chatter damage.

Highly flexible and stretchable MWCNT/HEPCP nanocomposites with integrated near-IR, temperature and stress sensitivity for electronic skin

There is an urgent demand for flexible multifunctional sensitive electronic devices in several potential applications, including personalized health monitoring, human motion detection, human–machine interfaces, soft robotics, and so on. Although exciting progress has been witnessed in recent decades, the excessive dependence on inorganic sensors and flexible substrates makes it difficult to attain multi-function sensing and excellent flexibility simultaneously, as well as their complicated fabrication, technical barriers and high cost. Herein, we report on a family of flexible multi-functional sensitive nanocomposites consisting of multi-walled carbon nanotubes (MWCNTs) uniformly distributed in a high elastic form-stable phase change polymer (HEPCP) that exhibit dramatic response to infrared light (IR), temperature and tensile stress in air, together with outstanding flexibility and stretchability. Optimum sensitivity to on/off IR, temperature and tensile stress is demonstrated with electrical conductivity ratios of 103.7, 11.8 and 1084.0 times at room temperature, respectively. The excellent performance of the MWCNT/HEPCP nanocomposites was largely attributed to the cyclic and reversible changes of their MWCNTs conductive network owing to the reversible form-stable phase transitions and high elasticity of the HEPCP substrate, which subsequently affected the thickness of the interfacial HEPCP between adjacent conductive MWCNTs, and thereby the electron tunneling efficiency between the MWCNTs. The novel MWCNT/HEPCP nanocomposites open a new window for multi-function sensing of electronic skin.

Machining vibration states monitoring based on image representation using convolutional neural networks

Abstract Measured signals are usually fed into filters or signal decomposers to extract useful features to assist making identification in state monitoring or fault diagnosis. But what is routinely ignored is that an experienced expert can realize what is happening just by watching the signals presented on the oscilloscope even without the analyzing report. The vision image input and the experience feedback are the two keys in this identification process by the brain. The experience can be easily quantified, like 1 for “good” and 0 for “bad”, and used for identification model construction, while there has been no attempt to use pictured signal as the model input. For closed-loop control system, it is necessary to acquire signal feedback point by point to adjust the system in real time. But for state monitoring and fault diagnosis, the pattern hiding among the signal points is usually more important, which is exactly one of the special fields of image representation to indicate complex interrelationship. Taking machining state monitoring as example, this paper explore the possibility to use the pictured signals as input to construct identification model without traditional feature engineering based on signal analysis. Convolutional neural networks (CNN) is introduced to connect pictured signals to different vibration states with experience feedback. Results validate the proposed method with excellent modeling performance. Time complexity analysis proves this pictured signal image representation based CNN method to be capable to be real-time. Two dimensional image representation is a powerful way to exhibit and fuse information. With high flexibility, the proposed method may be a promising framework for monitoring or fault diagnosis tasks.

Data driven injection molding process monitoring using sparse auto encoder technique

Injection molding process monitoring is quite essential for the stabilization of product quality. One of the most important things is to identify the character of injection batch process. In this study, sparse auto encoder technique is applied to extract features from the raw trajectories of system pressure and screw position. Subsequently, the process condition is identified by performing a classification on the features, in comparison with the raw trajectories data, and the principal components. The mean reconstruction error and the classification accuracy are selected to evaluate the representation capability of the extracted features. The experimental results show that the sparse auto encoder is an effective method of extracting features from the injection processing batch data, indicating that it is useful in injection molding process monitoring.

Feature learning and process monitoring of injection molding using convolution-deconvolution auto encoders

Abstract Feature learning is a generic and fundamental problem in data-based process monitoring of batch processes, such as injection molding. This paper proposes an automatic feature learning method, considering the following two vital characteristic aspects: the mechanical characteristics within a batch and the inherent characteristics among batches. Unsupervised feature learning is performed using convolution-deconvolution auto encoders, and the learned features are applied as predefined parameters for process monitoring, including process condition identification and fault detection. Experiments are carried out with different process conditions. The results indicate that the proposed method achieves improved model generalization ability under various process conditions, which means it can precisely reveal the variable autocorrelations and cross-correlations among different variables. Meanwhile, the learned features achieve higher classification accuracies and offer more optimal solutions for process monitoring. This method has been proven as an efficient means of feature learning, which should be appropriate for other batch processes.

An efficient preconditioned Krylov subspace method for large-scale finite element equations with MPC using Lagrange multiplier method

Purpose – The purpose of this paper is to propose an efficient iterative method for large-scale finite element equations of bad numerical stability arising from deformation analysis with multi-point constraint using Lagrange multiplier method. Design/methodology/approach – In this paper, taking warpage analysis of polymer injection molding based on surface model as an example, the performance of several popular Krylov subspace methods, including conjugate gradient, BiCGSTAB and generalized minimal residual (GMRES), with diffident Incomplete LU (ILU)-type preconditions is investigated and compared. For controlling memory usage, GMRES(m) is also considered. And the ordering technique, commonly used in the direct method, is introduced into the presented iterative method to improve the preconditioner. Findings – It is found that the proposed preconditioned GMRES method is robust and effective for solving problems considered in this paper, and approximate minimum degree (AMD) ordering is most beneficial for th...

Structure, injection molding process and fracture behavior of composite plastics

Abstract High strength and toughness are of importance for composite plastics used as mechanical structure parts. How to provide general and engineering plastics with high performances is one of the critical challenges in advanced manufacturing. In the past decade, the phase dispersion, interfacial interaction, injection molding simulation and fracture behaviors of composite plastics were investigated systematically in our research group in order to simultaneously reinforce and toughen polymer matrix. Here, we review the progress and advances which have been made in the above items.