Yinghong Peng

An integrated feature selection and cluster analysis techniques for case-based reasoning

Abstract Feature selection and case organization are crucial steps in case-based reasoning (CBR), since the retrieval efficiency and accuracy even the success of the CBR system are heavily dependent on their quality. However, inappropriate feature selection and case selection together with ill-structured case organization may not only present a dilemma in case retrieval, but also greatly increase the case base. To obtain an efficient CBR system, selection of proper features and suitable cases with appropriate case organization are very important. This paper proposes a hybrid CBR system by introducing reduction technique in feature selection and cluster analysis in case organization. In this study, a minimal set of features is selected from the problem domain while redundant ones are reduced through neighborhood rough set algorithm. Once feature selection is finished, the growing hierarchical self-organizing map (GHSOM) is taken as a cluster tool to organize those cases so that the initial case base can be divided into some small subsets with hierarchical structure. New case is led into corresponding subset for case retrieval. Experiments on UCI datasets and a practical case in electromotor product design show the effectiveness of the proposed approach. The results indicate that the research techniques can effectively enhance the performance of the CBR system.

New CBR adaptation method combining with problem-solution relational analysis for mechanical design

A CBR adaptation way using weighted mean and grey relational analysis is present.Problem-solution relational information is analyzed before adapting similar cases.A weighting strategy with relational information is proposed in weighted mean.The priority of proposed CBR adaptation is validated by empirical comparison. Case based reasoning (CBR) methodology is proved to be a promising methodology on determining the parameter values of new mechanical product by adapting previously successful solutions to current problems. Compared with the sophisticated case retrieval technique, the case adaptation under K-nearest neighbour is still a bottleneck problem in CBR researches, which needs to be resolved urgently. According to the characteristics of parametric machinery design (PMD), i.e., less data and many parameters, this paper employs weighted mean (WM) as a basic model, and presents a new CBR adaptation method for PMD by integrating with problem-solution (PS) relational information. In our proposed adaptation method, prior to adapting the similar cases, the grey relational analysis (GRA) is utilized to investigate the PS relational information hidden in K retrieved cases, and the proposed method is called as GRA-WM. Different from classical WM method, the weighting factor of retrieved case for each solution element adaptation is calculated by multiplying similarity matrix (SM) and relational matrix (RM), and the adapted solution values of new mechanical product are subsequently obtained by calculating the weighted average of solution values of K similar cases. A case study on the power transformer design is given to prove the industrial applicability of GRA-WM. Moreover, the empirical comparisons between GRA-WM and other adaptation methods are carried out to validate its superiority. The empirical results indicate that GRA-WM can offer an acceptable adaptation proposal in application of CBR for mechanical design.

Applying case-based reasoning to cold forging process planning

Abstract On the basis of the practical situation of cold forging process planning, the disadvantages of a rule-based solution are discussed, and a case-based reasoning-based cold forging process planning (CFPP) system model is proposed. Several key problems involved are analyzed, among which a feature-based part representation scheme and a two-level retrieval mechanism are introduced to solve the problems of case representation and case retrieval. It is established in this paper that case-based reasoning-based CFPP is a promising technology for both long-term research and the promotion of efficiency for current cold process planning systems.

Dynamic recrystallisation and dynamic precipitation in AA6061 aluminium alloy during hot deformation

Abstract Deformation behaviour of AA6061 alloy was investigated using uniaxial compression tests at temperatures from 400 to 500°C and strain rates from 0·01 to 1 s−1. Stress increases to a peak value, then decreases monotonically until reaching a steady state. The dependence of stress on temperature and strain rate was fitted to a sinh-Arrhenius equation and characterised by the Zener–Hollomon parameter with apparent activation energy of 208·3 kJ mol−1. Grain orientation spread analysis by electron backscattered diffraction indicated dynamic recovery and geometrical dynamic recrystallisation during hot compression. Deformation at a faster strain rate at a given temperature led to finer subgrains, resulting in higher strength. Dynamic precipitation took place concurrently and was strongly dependent on temperature. Precipitation of Q phase was found in the sample deformed at 400°C but none at 500°C. A larger volume fraction of precipitates was observed when samples were compressed at 400°C than at 500°C.

Simulation and fracture prediction for sintered materials in upsetting by FEM

Abstract Experiments dealing with the upsetting of sintered materials have resulted in the determination of a forming limit in terms of the local compressive and tensile strains. As the specimen is compressed, a plot of tensile strain versus compressive strain can be made. Fracture occurs when the strain path intersects the forming limit. Once the fracture locus is defined by an experimental method, it is possible to use the Finite Element Method to determine if some upsetting operation will exhibit a free surface crack during a deformation process. In this study, a finite element program has been developed to analyze the deformation processes of sintered metals in upsetting. At the same time, the effects of lubrication, height to diameter ratio, initial relative density, die shaping and preform shape on the forming limit of PM products are investigated. The locus strains are then calculated and possible defects leading to material failure are checked. The calculated results agree well with the experimental results.

Image Recognition With a Limited Number of Pixels for Visual Prostheses Design

With the rapid development and crossover among the information science, microelectronics, material science, and biomedical disciplines, the visual prosthesis makes visual reparation possible. Because the number of stimulation electrodes is strictly limited by various complicated factors, it is necessary to determine the minimum visual requirements to achieve useful artificial vision for image recognition. This research has studied how many pixels individual images need to have to be correctly and economically recognized by blind subjects. In order to extract the figure of the image with a limited number of pixels, we have proposed a wavelet-based image processing methods, and six resolutions (8 × 8, 16 × 16, 24 × 24, 32 × 32, 48 × 48, and 64 × 64) are investigated. Psychophysical experiments have been designed to verify our proposed image processing method and to investigate the recognition accuracy with a limited number of pixels. The results show that the recognition accuracy increases with the number of pixels. The recognition accuracy varied with tested images, when a resolution of 24 × 24 was used: six of the eight image objects were recognized with an accuracy of >50%, and the remaining two of the eight image objects were recognized with an accuracy of <50%. Moreover, when the resolution is more than 32 × 32, the increase of the recognition accuracy is no longer obvious. We also have investigated the impact of different perspectives of the same object to the recognition accuracy. The experiment shows that providing multiview image sequences, subjects can receive more visual information to obtain higher recognition accuracy.

Hybrid weighted mean for CBR adaptation in mechanical design by exploring effective, correlative and adaptative values

Hybrid weighted mean method using various knowledge for CBR adaptation is present.We explore effective, correlative and adaptative values from similar case data.We integrate effective, correlative and adaptative values into weighted mean model.Using effective, correlative and adaptative values improve the adaptation accuracy. The implementation of case based reasoning (CBR) adaptation in parametric mechanical design can generate the design solution to unknown design problem by adapting similar solutions from other problems already solved. Classical weighted mean (WM) method is a common statistic adaptation method because of its domain independent and easily to be implemented, but with lower adaptation accuracy. A new hybrid WM (HWM) method for CBR adaptation in mechanical design is proposed in this paper, and its contribution is taking advantage of various implicit knowledge hidden in similar case data to improve the performance of WM. To achieve this goal, multiple similarity analysis (MSA), grey relation analysis (GRA) and inductive adaptability analysis (IAA) are firstly used to systematically explore the effective value (EV) of similar case for new design problem, the correlative value (CV) between problem and solution features, and the adaptative value (AV) of similar cases solution element for new adaptation situation, respectively. Then CV, EV and AV compose the integrated weight value of each solution element of similar case in HWM, and the optimal proportion of EV, CV and AV on the integrated weight is also discussed. Based on the parametric transformer design cases, the comparisons of adaptation performances between HWM and other statistical and intelligent methods were carried out, and the empirical results show that HWM has the better adaptation performance than other comparative methods by comparing the adaptation accuracy.

Defect prediction during conform process by FEM

Abstract On the basis of the rigid viscoplastic FEM theories, a surface defect prediction method for plastic forming process by FEM numerical simulation has been deeply studied in this paper. Some technical treatments and algorithms of defect prediction are proposed. Using this method, the processes of defect initiation and development during CONFORM process are predicted successfully, and some critical technological parameters are obtained. Therefore, it is also an extension for application fields of rigid viscoplastic FEM to predict technology defect.

Incorporating adaptability-related knowledge into support vector machine for case-based design adaptation

Abstract In case-based design systems, the adaptation operation based on similar cases is a difficult and complex step, and the more adaptable cases usually could make larger contribution for adaptation generation than less ones. Under this ideology, this paper addresses a new case adaptation method which uses support vector machine (SVM) incorporating adaptability-related knowledge provided by the retrieved cases, called adaptability-involving SVM (ASVM). The knowledge of adaptability includes the adaptability characteristic of old cases returned by the adaptability analysis and the guideline that the training data from adaptable case should be given higher weight to build SVM model. So the content of this work presented here consists of two parts. The first one is to explore the adaptable property of old cases by utilizing decision tree technology. The second one is to study the construction of ASVM adaptation model in terms of retrieved cases. We first employ the differences between test and retrieved cases to assemble the adaptation pattern data for ASVM model training. Then the higher adaptability coefficients are given to the training data from more adaptable cases than those from less adaptable cases. We adopt ASVM in actual power transformer design to illustrate its feasibility, and carry out comparison researches with different numbers of retrieved cases in the different data sets to validate its superiority, through comparing the adaptation error results with those provided by other classical methods. Empirical results show that ASVM is feasible and validated for case adaptation.

Explicit Simulation of Roll Forming Process with EAS Solid-shell Elements

Solid‐shell elements can be seen as a class of typical double‐surface shell elements with no rational degrees of freedom, which are more suitable for analyzing double‐sided contact problems than conventional shell elements. In this study, an EAS‐based solid‐shell element is implemented into the explicit finite element formulation to simulate roll forming process. A twelve‐parameter enhanced assumed strain (EAS) method is adopted to solve for the locking pathologies. Accuracy of the explicit solid‐shell finite element model is excised through two NUMISHEET benchmark tests. Afterwards, a U‐channel forming is simulated with the present explicit model. Numerical results of longitudinal strains and final geometries are compared with experiment as well as calculated by the commercial software ABAQUS. The solid‐shell element is found more applicable in dealing with roll‐forming process than ABAQUS inherent elements. Potential of the explicit solid‐shell model in analyzing cold roll forming process is confirmed.