Advanced Materials Research Vols. 452-453

Abstract: Using the triangulation method of laser non-contact measurement, the surface shape of CRT's shadow mask and the Q value was high precision measured. Measuring the mesh surface, the laser beam may be all or part through the mesh. The extraction problem of the effective measurement data was studied. Setting the threshold to filter out the invalid data and the data influenced by the mesh, it retained a small amount of valid data. Using data fitting and interpolation methods to process part of the effective data, then extracting the geometry of the mesh surface profile, the good reproducibility results was gotten. The advanced measuring equipment used in production practice was also studied.

Abstract: The precise edges of forging should be extracted from the image when measuring the dimensions of high temperature forging based on the non-contact image measurement technique. It is difficult to distinguish the forging boundary precisely due to the continuous variation of the air temperature near the forging. The two-dimensional simplified cross-section models, including the squaring, chamfering and rounding models, were established by the symmetrical modeling method. The air-cooling process from 1200°C to 800°C of high temperature long-shaft heavy forging was simulated based on ANSYS software. The variation regularities of temperature with the distance and time were obtained, which provides a theoretical basis to determine the edges of high temperature objects in non-contact measurement.

Abstract: Optimal structural design of stiffened plate structure with multi order modal frequencies objective is discussed. It is a structural reverse design problem for the first several order modal frequencies to reach a set of given value. A new method based on bi-directional evolutionary structural optimization (BESO) and size optimization is presented. To solve the porous and irregular shape problem in the optimized structure, a regular shaped grid-like frame structure optimization model is established. The bars in the frame structure are regarded as basic unites of structure modification. The bars’ sensitivity is discussed, and applied to optimization process based on frequency sensitivity. The structural optimal design process can be described as follows: the frequencies are used as the target, the volume is considered as constraint, adjacent sensitivity redistribution method is adopted to suppress numerical instability. Then size optimization method is adopted to conduct detailed design. Finally this method is applied to a stiffened plate structure. Results show that the proposed approach is feasible to achieve given multi order modal frequencies. The optimized structure consists of regular bars, so a clear structure is obtained.

Abstract: Modeling of a multi- ring grooves Magnetorheological（MR）fluid damper designed by ourselves were respectively carried out based on the Bingham model and the Herschel-Bulkley model. By comparing the simulation results of the two models with the true drop experimental result, it was proved that the Herschel-Bulkley model is in good accordance with the experiment and outperformed the Bingham model under high shear rate and high magnetic field. Furthermore, the parameter n in the Herschel-Bulkley model reflects the densification of the MR fluids, and by identifying and selecting the value of n better models and simulation results can be obtained.

Abstract: Due to electromagnetic coupling and structure restriction of the two rotors, the design of anti-rotary PMSM is very complex. Inner motor and outer motor are designed respectively and then integrated in aspect of electromagnetic coupling and size restriction in this paper. Main size and electromagnetic relationship are designed according to objectives, and key design equation and project procedure are presented as well. Parameters effect on motor performance such as efficiency and torque density are analyzed. The performance of the prototype, well matches the theory analysis, whose validity is verified by FEA results

Abstract: Differential Evolution (DE) is one kind of evolution algorithm, which based on difference of individuals. DE has exhibited good performance on optimization problem. However, when a local optimal solution is reached with classical Differential Evolution, all individuals in the population gather around it, and escaping from these local optima becomes difficult. To avoid premature convergence of DE, we present in this paper a novel variant of DE algorithm, called SSDE, which uses the stratified sampling method to replace the random sampling method. The proposed SSDE algorithm is compared with some variant DE. The numerical results show that our approach is robust, competitive and fast.

Abstract: A novel design of high load capacity multiaxis positioning stages with accuracy in the range of nanometers is presented. For strokes of 2mm with no play and high stiffness a general design principle supporting five Cartesian degrees of freedom has been developed using a new parallel kinematic topology based on Parallelogram arrangements. The five uniform feed drives are improved dual mode mechanism with servomotor and ballscrew as macro-actuator and piezoelectric transducer (PZT) with resolution of 1.2nm as micro-actuator. The performance of the setup and its kinematic properties are described as well as resolution of the five motions and their crosstalk. The setup has been implemented with outstanding characteristics and excellent reliability for alignment of a multigrating mosaic compressor in a PW-class CPA-laser.

Abstract: For the Gaussian mixture learning, the expectation-maximization (EM) algorithm as well as its modified versions are widely used, but there are still two major limitations: (i). the number of components or Gaussians must be known in advance, and (ii). There is no generally accepted method for parameters initialization to prevent the algorithm being trapped in one of the local maxima of the likelihood function. In order to overcome these weaknesses, we proposed a greedy EM algorithm based on a kurtosis and skewness criterion. Specifically, we start with a single component and add one component step by step under the framework of EM algorithm in order to decrease the value of the kurtosis and skewness measure which provides an efficient index to show how well the Gaussian mixture model fits the sample data. In such a way, the number of components can be selected adaptively during the EM learning and the learning parameters can possibly escape from local maxima.

Abstract: A short computer program has been developed, for the step-by-step dynamic analysis of cables of cable-stayed bridges, connected to each other and possibly with the deck of the bridge, by very thin pretensioned cross-ties and subjected to variation of their axial forces due to traffic or to a variable wind force. A simplified SDOF model, approximating the fundamental vibration mode, is adopted for every individual cable. The geometric nonlinearity of the cables is taken into account by their geometric stiffness. Whereas, the material nonlinearities of the cross-ties include compressive loosening, tensile yielding and hysteresis stress-strain loops. Seven numerical experiments are performed. Based on them, it is observed that if two interconnected parallel cables have different dynamic characteristics e.g. different lengths, thus different masses, weights and geometric stiffnesses, too, or if one of them has a small additional mass, then a single pretensioned very thin wire, connecting them to each other and possibly with the deck of the bridge, proves effective in suppressing, by its hysteresis damping, the vibrations of the cables.

Abstract: x-ray real time imaging detection technology is a kind of important way for industrial nondestructive test. On the basis of basic theory on X-ray detection, The influence factors on x-ray real time imaging detection precision is analyzed in this article. Through analysis for the focus of X-ray source and the unintelligibility of geometric image, the relation between the optimal amplification multiple and the imaging quality is presented and the electric collimator to solve the influence on imaging quality from the scattered ray. The experimental result shows that the detection resolution ratio is up to 50PL/cm and the sensitivity is up to 1.4 % to solve the on-line real time detection for pore, inclusion and looseness and verify the application feasibility in the detection of cast aluminum parts for x-ray real time imaging detection technology.