Search results

A parameter reduction method based on fuzzy rough sets was proposed.
Fundamentals on Fuzzy Rough Sets A structural data used for classification learning can be written as an information system, denoted by , where U is a nonempty and finite set of samples {x1, x2 ,..., xn}, called a universe, A is a set of attributes (also called features, inputs or variables) {a1, a2, ... , am} to characterize the samples.
The objective of rough set based attribute reduction is to find a subset of attributes which has the same discriminating power as the original data and has not any redundant attribute.
Fig.2 Dependency Function There are 6 parameters in the reduction set.
A forward greedy search algorithm is proposed for attribute reduction.

It aims to create an opportunity to develop own patent, alongside cost reduction on the benchmarked systems.
This indirectly reflects the cost reduction outcome on the material consumption related to the bracket design.
The component was fabricated as a prototype and the data of weight reduction was measured.
Meanwhile, the optimization method using TRIZ achieved 75.3% of weight reduction against the existing design.
In the same financial year (2011/2012), the organization really shows significant results in cost reduction performance.

The Laser Polishing tests results have been used as the input data for a design of experiments (DoE), therefore the optimum operation parameters for the process as well as its degree of influence in the melted surface have been defined.
Maximum roughness reduction parameters with CO2 laser.
Maximum experimental roughness reduction with HPDL laser.
Therefore, for the same material and similar roughness reduction rates, the influence of the laser type is relevant.
Conclusions The laser polishing process, in tool steel DIN 1.2379, allows roughness reductions over the 80 percent.

Microbial inactivation data were used to obtain, which after the identification of the critical parameters to reduce the content of spoilage microorganisms in orange juice.To provide a base of scientific knowledge and support the innovation.
Statistical analysis Statistical analysis was carried out using data analysis functions in Microsoft Excel and Significant differences between the results were calculated by analysis of variance (ANOVA).
Effect of the number of pulses on E. coli reduction As t the number of pulses from 4 to 32, the E. coli reduction went up a little at the same imput voltage (Fig. 3).
Effect of the total soluble solids on E. coli reduction Fig.5 indicates that at any total soluble solids, the E. coli reduction decrease a little with the the imput voltage of 4kV,the number of pulses of 64 and the juice thickness of 2mm.
At the total soluble solids of 40ºBrix, the E. coli reduction decrease by approximately 0.56log.

The average error of data from this system is within 5%.
The regularities distribution of the pulling loads are obtained from the experiments data, which can give the theoretical supports to the safety operations about the earth anchor piles.
All the signals can be collected and amplified by charge amplifier and transformed and stored by the data acquisition system and computer.
The output ports of the sensor connect with the data acquisition instrument and the computer by data line.
Finally the real-time collecting and storage of the test data can be realized.

Compression pre-cracking methods [3] were used to initiate cracks at crack-starter notches with constant-amplitude and load-reduction methods used to generate data from threshold to fracture.
Figure 1(a) shows the ΔK-rate test data, and Figure 1(b) shows the ΔKeff-rate data using a plane-strain constraint factor (α) of 1.8.
The term q = 2 was selected to fit high-rate data.
Figure 3(b) shows the test data (symbols) and the predicted results.
The symbols show test data and the curves show life predictions.

A Study on Reduction of Friction in Impact Compressive Test Based on The Split Hopkinson Pressure Bar Method by Using a Hollow Specimen Nobuhiko Kii1,a, Takeshi Iwamoto2,b, Alexis Rusinek3,c and Tomasz Jankowiak4,d 1Graduate School of Engineering, Hiroshima University, 1-4-1 Kagamiyama, Higashi-Hiroshima, Hiroshima, 739-8527 Japan 2Institute of Engineering , Hiroshima University, 1-4-1 Kagamiyama, Higashi-Hiroshima, Hiroshima, 739-8527 Japan 3National Engineering School of Metz, 1 route d’Ars Lquenexy, CS 65820 57078 Metz Cedex 3, France 4Institute of Structural Engineering, Poznan University of Technology, Piotrowo 5, 60-965 Poznan, Poland am124931@hiroshima-u.ac.jp, biwamoto@mec.hiroshima-u.ac.jp, crusinek@enim.fr, dtomasz.jankowiak@put.poznan.pl Keywords: Split Hopkinson pressure bar, Impact compression, Hollow specimen, Reduction of friction coefficient, Johnson-Cook model Abstract.
In addition, reduction of friction effect is attempted based on the experimental and computational studies using ring specimen made of steel and aluminum alloy [6].
Table 2 Material parameters identified from the experimental data until the plastic strain of 0.15 A [MPa] B [MPa] n C m 83.5 28.03 0.34 0.012 0.68 Table 3 Material parameters identified from the experimental data until the plastic strain of 0.3 A [MPa] B [MPa] n C m 83.5 13.85 0.0463 0.012 0.68 5.
That is, it can be considered that friction effect is decreased by a reduction of the contact area.

The value of GIIc was extracted by a new data reduction scheme avoiding the crack length measurement, named Compliance-Based Beam Method (CBBM).
In this work, the interlaminar GIIc of a CFRP laminate is obtained by a new data reduction scheme not requiring the crack length measurement (CBBM).
Moreover, it accounts for the FPZ effects, which does not occur when a real crack length is used in the data reduction scheme.
Ex=1.09E+05 MPa xy=0.342 Gxy=4315 MPa Ey=8819 MPa xz=0.342 Gxz=4315 MPa Ez=8819 MPa yz=0.380 Gyz=3200 MPa Data reduction schemes The classical data reduction schemes to obtain GIIc are usually based on compliance calibration or beam theory.
A new data reduction scheme based on the crack equivalent concept was employed to obtain the interlaminar fracture toughness of a carbon-epoxy laminate in pure mode II.

The data of each material in the roof has been defined based on the thermal conductivity.
The differences between SolidWorks simulation results and experimental data are less than 1%, and the SolidWorks simulation results are agreeable with the experimental data.
Fig. 10: Comparison between experiment and SolidWorks result without insulation for rooftop The comparison data with Glasswool insulation between the experiment and SolidWorks data is shown in Fig. 11.
Fig. 14 presents the result data for without insulation and with insulation applied.
With that, the comparison of the data can be more accurate.

And the outputs are mechanical properties namely ultimate strength, yield strength, elongation, reduction of area, plane strain fracture toughness and microstructure concerned parameters such as β phase fraction, β phase grain size, substructure length and thickness.
The ANN is trained with experimental data and achieves a very good performance, which has already been applied to the optimization of processing for forging of aero-parts.
The inputs are working temperatures, deformation extent, deformation rate and heat treatment conditions, and the outputs are ultimate strength, yield strength, elongation, reduction of area, plane strain fracture toughness, βphase fraction, βphase grain size, substructure length and thickness.
So the experimental temperatures are 923℃, 938℃, 953℃, 960℃, 968℃, 973℃ chosen according to beta and near beta forging theory [10], the deformation extents are 15%,30%, 40%, 50%, the deformation rates are 1.4mm/s, 1.7mm/s and the corresponding heat treatment conditions (as shown in Table 1) are transformed to numerical data because BP network can only deal with numerical ones.
Through comparing the network outputs with the experimental data (Table 3), it can see that the BP network displays very high adaptability and precision, the predicted outputs with the testing samples have relatively low error and most of the generalized outputs with the validating samples are within the 15% error, only a few have high error values.