In this paper, an efficient and effective Gaussian Kriging metamodeling approach is proposed in the framework of Bayesian maximum a posterior. Different prior densities and particularly, a Jeffreys' noninformative density based hierarchical prior is imposed respectively on the regression coefficients in the mean model and the correlation parameters in the covariance matrix of Gaussian Kriging. And the involved parameters are estimated by the expectation maximization and Fisher's scoring algorithms. Compared against several benchmark and recent methods in literature, the proposed approach is shown capable of simultaneous high prediction accuracy and low computational cost.
Optimum segmented blank holder shape and its variable blank holder gaps (VBHGs) are determined by a sequential approximate optimization (SAO) with radial basis function network. In deep drawing, wrinkling and tearing of blank sheet are major defects. The optimum segmented blank holder shape and its VBHGs are determined to avoid these defects. The Forming Limit Diagram (FLD) is employed to evaluate quantitatively the wrinkling and the tearing. In the numerical examples, a square cup deep drawing is handled. The objective is to minimize the thickness deviation after sheet forming. The wrinkling and the tearing are separately evaluated as the constraints. The dimensions of the segmented blank holder shape and the BHGs are taken as the design variables. The optimization result shows that simultaneous optimization of both the segmented blank holder shape and the VBHGs is one of the effective approaches for improving product quality.
Currently real-time control and online quality estimation of the resistance spot welding (RSW) has benefited a lot from monitoring the electrode displacement. Based on these emerging monitoring techniques a new approach is proposed to determine the optimal welding parameters and help to assess the weld quality. Two causal models are built with the offline trained Bayesian Belief Networks (BBN). The first model which is a pattern determination net deals with the optimal control criteria, i.e. an ideal combination of the maximum electrode displacement and electrode travel velocity, to provide more reliable welding process and qualified welds. The second model which is a weld quality assessment net reveals the dependency of the weld quality on the features displayed by the displacement curve, which can be used for overdesigning the safety welds or for online assessing weld quality as the probabilistic forecasting model. The experimental results show that the proposed approach is valid and feasible to determine the controlled parameters and to predict the weld quality in practices.
The major power information of the main transmission system in machine tools (MTSMT) during machining process includes effective output power (i.e. cutting power), input power and power loss from the mechanical transmission system, and the main motor power loss. These information are easy to obtain in the lab but difficult to evaluate in a manufacturing process. To solve this problem, a separation method is proposed here to extract the MTSMT power information during machining process. In this method, the energy flow and the mathematical models of major power information of MTSMT during the machining process are set up first. Based on the mathematical models and the basic data tables obtained from experiments, the above mentioned power information during machining process can be separated just by measuring the real time total input power of the spindle motor. The operation program of this method is also given.
Constant flow valves have been presented in industrial applications or academic studies, which compensate recess pressures of a hydrostatic bearing to resist load fluctuating. The flow rate of constant-flow valves can be constant in spite of the pressure changes in recesses, however the design parameters must be specified. This paper analyzes the dynamic responses of DSI-type constant-flow valves that is designed as double pistons on both ends of a spool with single feedback of working pressure and regulating restriction at inlet. In this study the static analysis presents the specific relationships among design parameters for constant flow rate and the dynamic analyses give the variations around the constant flow rate as the working pressure fluctuates.
To produce ultra-smooth optical surfaces on hard brittle materials and functional crystal materials without surface damage, atmospheric pressure plasma polishing (APPP) is developed. And improvement of surface quality by APPP process is analyzed by theoretical simulation and experiments. Quantum chemistry simulation proves that convex topography is removed faster than concave topography, thus surface roughness can be reduced further to form ultra-smooth surface. Experiment on single crystal silicon wafer also shows the surface roughness decreases from Ra 1.179nm to Ra 0.483nm. AFM (atomic force microscope) measurement results of specified roughness index taken at different moment in machining process accord well with theoretical simulation. Average reduced-modulus measured by nano-mechanical test system rises from 14.65GPa to 36.10GPa which demonstrates the improvement of surface mechanical properties. Further calculation also indicates the average residual stress is reduced by 7.64 GPa after machining which reflects the elimination of former deformation layer. Element analysis by X-ray photoelectron spectroscopy and X-ray diffraction indicates that chemical composition is also purified without new contamination introduced, and the overall status of machined surface goes closer to ideal silicon surface with better performance.
A gradient-based optimization method for designing linkages with velocity targets is described. Two theoretical application cases are established for four-bar linkage. In the first, a constant-velocity module is proposed for a point on the coupler. In the second, the goal is the velocity components. These cases are studied with and without coordination with the input link. The results obtained are compared with another gradient-based approach, and show that the method works efficiently for these types of target.
To date, elliptical gear has been commonly used in automobile, automatic machinery, pumps, flow meters and printing presses for its particular non-uniform rotation. However, the dynamic characteristics of elliptical gears have not been clarified yet. In this study, The calculation as well as the experiment of two elliptical gears, which are a single elliptical gear and a double elliptical gear, is carried out to analyze the dynamic characteristics of elliptical gears. General factors including the torque, the rotation speed and the tooth root stress of the test gears are investigated. According to the analysis conducted in this study, the dynamic input torque variation of elliptical gear becomes larger along with the increase of operating gear rotation speed and the experimental one increases much faster than the calculated one over the Critical Rotation Speed of Tooth Separation (CRSTS) of elliptical gear. The experimental input rotation speed varies according to the variation of input torque, leading to the difference between the experimental output rotation speed and the desired one. The calculation results of the CRSTS of elliptical gears are almost equal to the experimental ones. The dynamic load variation ratios of elliptical gear at different angular position as well as their changing trends with operating gear rotation speed are quite different from each other. And the experimental dynamic load variation ratios of elliptical gear show difference from the calculated ones because of tooth separation and tooth impact. The agreement of the calculation and experimental results proves the validity of this study.
Elliptical gear is a typical non-circular gear, which transmits a variable-ratio rotation and power simultaneously. Due to the non-uniform rotation, the vibration and noise of elliptical gears demonstrate particular characteristics which should be paid attention to in practical application. In this paper, two elliptical gears, which are a single elliptical gear and a double elliptical gear, have been investigated to analyze the vibration and noise characteristics of elliptical gears. The corresponding circular gears for comparison are also investigated. General factors including the torque, the rotation speed, the gear vibration acceleration and the gear noise of the four test gears are measured by running test. The root mean square of the Circumferential Vibration Acceleration (CVA) and the sound pressure level of the noise of elliptical gears are obtained from the measured results and compared with those of circular gears to clarify the vibration and noise characteristics of elliptical gears. Furthermore, the frequency analysis of the CVA of elliptical gears is conducted by Fast Fourier Transform Algorithm (FFT) and compared with that of circular gears. The main vibration component of elliptical gear is uncovered according to the obtained frequency spectra. In addition, the Critical Rotation Speeds of Tooth Separation (CRSTS) of elliptical gear is obtained and its relation with load torque is unveiled.
The features of the support structures of the sun gear, carrier and ring gear that are the components of a planetary gear set result in large alignment error in the form of eccentricity or inclination. Such error causes the relative tooth flank error to vary periodically at each pinion position. The side band phenomenon modulated by the rotation of the rotating elements appears around the mesh order components of transmission error. With the conventional method of calculating transmission error, it is not possible to evaluate the side band phenomenon because of the inability to treat the periodical variation in the relative tooth flank error. Therefore, we have developed a new method of simulating the transmission error of planetary gear sets taking into account such periodical variation of the eccentricity and inclination direction of each element. The predicted values are then input into the calculation, making it possible to evaluate the side band phenomenon as well. Periodical variation in the relative tooth flank error due to alignment error returns to the original state when the relative rotation between the rotating elements has completed one revolution. Accordingly, by repeating the calculation until the relative rotation between the sun gear and the carrier or between the carrier and the ring gear has completed one revolution, the new method makes it possible to analyze transmission error taking into account the periodical characteristic of the shape of the tooth flank error due to alignment error. This new calculation method was applied to analyze the transmission error of an actual transmission, and the results showed that the side band phenomenon was also evaluated with relatively good accuracy.
This paper investigates a quantum neural network and discusses its application in control systems. A learning-type neural network-based controller that uses a multi-layer quantum neural network having qubit neurons as its information processing unit is proposed. Three learning algorithms; a back-propagation algorithm, a conjugate gradient algorithm and a real-coded genetic algorithm, are investigated to supervise the training of the multi-layer quantum neural network. To evaluate the learning performance and the capability of the quantum neural network-based controller, we conducted computational experiments for controlling a nonlinear discrete-time plant and a nonholonomic system - in this study a two-wheeled robot. The results of computational experiments confirm both the feasibility and the effectiveness of the quantum neural network-based controller and that the real-coded genetic algorithm is suitable for the learning method of the quantum neural network-based controller.
This paper addresses a pedestrian safety design of front structure of Heavy Goods Vehicle (HGV) by two concepts; firstly by equipping a lower bumper stiffener structure under the front bumper and secondly by putting an airbag in front of the HGV front panel. In this study, HGV-pedestrian collision accident was simulated by the crash analysis solver MADYMO environment, where the HGV model with the speed of 20 km/h was collided with an adult male and with an adult female pedestrian, respectively. The bumper and lower bumper stiffener were varied their positions, while the airbag was adjusted the vent hole size and the position of airbag in front of front panel vertically. The pedestrian injuries that can be sustained during the simulation impact were limited at the critical body parts of head, chest, upper leg; an injury criteria of Head Injury Criterion (HIC), Thorax Cumulative 3ms Acceleration (C3ms) and peak loads of femur, respectively. Because of various parameters and constraints of initial conditions and injury thresholds, a multi-objective optimization design problem considered these main injury criterion is solved in order to achieve the best solution for this study. The results of optimized design parameters for each cases and conditions were obtained and the possibilities of the proposed concept were discussed.