Search results

It begins with acquiring experiments data.
Usually, the data cannot be directly treated by RS, and corresponding preprocessing should be adopted to improve the quality of data, where discretization is to convert the continuous data to discrete value.
Application in Rotating Arc NGW Obtain raw data To obtain the experiment data, the work piece is machined as showed in Fig. 1.2(b), which imitate the weld of multi-layer and single pass welding.
This will help the welding experts acquire knowledge from the experiment data.
Model reasoning Approximate reasoning is implemented when the RS model is use to predict unseen data.

Then it took advantage of Poyang lake for empirical analysis, and got the function of pollutant reduction cost and function of environmental damage cost by means of multiple regression analysis with the five main data of Poyang lake from 2001 to 2010.
But according to statistics yearbook and other data, Shangrao information is too little to be in operation.So the paper mainly discuss the pollutant reduction between Nanchang and Jiujiang.
Due to some data is not easy to get, we do the following process, The reduction cost of pollutant i is as follows:.
Through the above process,we can get the models of each pollutant with the data of 2010.
And due to limited data samples are few.

Reduction in area is a measure to determine formability on cold heading.
Reduction in area is affected by second phases and inclusions.
The test data of conventional ferrite+pearlite steel [6], tempered martensitic steel [6] and bainitic steel [7] are also plotted in Fig. 3 for the purpose of comparison.
This indicates that reduction in area determines the formability of screw heads.
[6] NIMS data base, Materials Information Technology Station, National Institute for Materials Science, Tsukuba, Japan

Iron reduction rate of Fe, Cu and Ni can be elevated to above 90% by smelting reduction, thus the comprehensive utilization of valuable metals can come true.
In this study, results of the “Double slag” smelting under the condition of smelting reduction have been analyzed from a thermodynamic point of view, and experimental verification of the possibility of “Double slag” smelting under the condition of smelting reduction also has been carried out. 1 Thermodynamic calculation According to the preliminary work the optimum conditions of the “Double slag” smelting reduction can be determined as follows: the slag basicity is 1.1, the adding ratio of JISCO slag is 10%.
Substituting the above data into Eq. 8, the content of manganese in liquid iron can be obtained as follows: . 3 Experimental study on smelting reduction In order to verify the results of thermodynamic calculations, smelting reduction experiment of the “Double slags” was carried out in the electric arc furnace with lining of L3.
Table 4 Tapping output and reduction rate of “Double slag” smelting Test number D1 D2 D3 Tapping amount/ kg 5.6 5.4 5.4 Reduction rate/ ％ 94.9 91.5 91.5 Table 5 Hot metal composition of “Double slag” smelting/ % Test result C Si Mn P S Ni Cu Co D1-T 3.84 0.17 0.21 0.022 0.034 0.85 0.420 0.001 It can be seen from Table 4 and Table 5, in the smelting conditions identified in this process, the iron reduction rate and hot metal chemical composition of double-slag smelting is relatively stable, the iron reduction rates were maintained above 91%, the reduction rate of Ni, Cu in hot metal is above the level of 95%.Compared with the calculational results in Section 2, the content of Si, Ni and S in the experimental results is close to the calculated results, while there are some differences between Mn content and the calculated results, which are mainly attributed to the following reasons, first, the test temperature in the bath is not equal to 1600℃, second, approximate calculation method
Conclusion (1) Test conditions are shown as follows: the weight of Jinchuan slag was 13.5kg, and the weight of JISCO slag was 1.5kg, the composition of the slag has the basicity of CaO : SiO2 = 1.1, the use level of reducing agent H4 was 1.8kg, smelting current strength was 1900A, smelting time was 20min, during the smelting reduction process, reduction rate of iron remained above 91%, the reduction rate of Ni, Cu in hot metal was above 95%; (2) Through thermodynamic calculation, it is found that the content of each element in molten iron of the calculation results is close to that of the experimetal results, thus the purpose of guiding experimental procedure has been achieved; (3) Low alloy iron produced by “Double slag” melting reduction can be smelted into high value-added spring steel.

Development of a baseline methodology The methodology includes two parts which are the power supply part calculation of the CO2 emission reduction and the heating part calculation of the CO2 emission reduction. 1.
According to the data of the project, CO2 emission reduction can be calculated as Table 2.
Table 1 Parameters of IGCC power plant Parameter Unit Value Installed capacity MW 250 Power generation efficiency % 48 Running time h 5000 Power generation efficiency of the baseline % 35 Power used in site rate % 3 Desulfurization efficiency % 99 Dedust efficiency % 100 Bituminous coal calorific value GJ/Kg 0.02717 Thermoelectrical ratio [4] % 35 Table 2 Annual emission reductions (unit :) CO2 emissions Baseline emissions Project emissions Leakage Emission reductions Power part 1,083,429 860,269 0 223,160 Heating part 240,875 0 0 240,875 Total 1,324,304 860,269 0 464,035 Environmental benefit and economic benefit According to the 250 MW IGCC power plant data, it can be calculated that IGCC power plant can save 227,763 tons of coal per year [5], and at the same time, the following pollutants (Table 3) can be reduced annually.
Table 3 Pollutants reduction annually (unit: t) Pollutant Baseline emissions Project emissions Leak Emission reductions SO2 634.1 25.5 0 608.6 NOx 4143.7 2496.1 0 1647.6 Smoke 190.3 0 0 190.3 According to the project, it can be calculated for project CO2 emission reductions (CERs) income under the condition of internal rate of return (IRR), detailed data shows in Table 4.
By using the methodology developed, the IGCC emission reductions are calculated.

With the development of on-line monitoring technology of electric power equipment, and the accumulation of both on-line monitoring data and off-line testing data, the data available to fault diagnosis of power transformer is bound to be massive.
Furthermore, it is unable to preprocess data. 3) In the data preparation stage, rough set can be used to fill the incomplete data and reduce the dimensionality of data, for the purpose of reducing the computing amount of association rule mining.
data processing layer itemsets classification layer attribute reduction layer itemsets mining layer rule acquisition layer fault diagnosis layer initial data data processing original itemsets itemsets selecting decision table with attributes attribute reduction simplest decision table frequent itemsets mining frequent itemsets attribute value reduction rule acquisition fault diagnosis classifier fault diagnosis Fig.1.
Attribute reduction layer The operation object of this layer is original decision table, reducing data dimensionality by using attribute reduction algorithm of rough set theory.
Input test data, and use the constructed classifier to classify data whose class attribute value is unknown.

The bioreactor systems are operated under increasingly stringent conditions through the reduction in the hydraulic residence time.
Biological sulfate reduction represents an option for ARD remediation.
Thereafter, the matrix-attached cells were detached in seven sequential steps comprising vigorous agitation for 2 minutes in reactor feed containing 0.4% (v/v) Tween® 20, demonstrated previously to remove the dominant fraction of the microbial community (data not shown).
The SEM micrographs of the colonised carbon microfibres and polyurethane foam corroborate the cell count data (Fig. 1) which indicates more colonisation of the polyurethane foam compared to the carbon fibres.
Marais, S.T.L Harrison, Biomass retention and recycling to enhance sulphate reduction kinetics.

In a bench scale trial biological sulfate reduction was applied to convert anglesite (PbSO4) to galena (PbS).
Galena precipitates in the bioreactor due to the near-neutral pH at which sulfate reduction is carried out.
For the biological reduction of sulfur components hydrogen gas (H2) is a suitable electron donor.
Table 1 presents the elemental composition of the indirect leaching residue (data provided by Tecnicas Reunidas, Madrid, Spain).
Table 1 Elemental composition of the indirect leaching residue and the sludge after the bench scale experiment (data provided by Tecnicas Reunidas) Element % in indirect leaching residue % in sludge after bench scale experiment Pb 51.7 56.4 Zn 0.03 3.9 Fe 4.9 5 Cu 0.42 S-tot 21.8 18.2 SO4 19.9 0.52 The reactor was started up using a synthetic influent containing sulfate.

In AUSCS, ultrasonic flaw signals acquired in the form of digitized data are first preprocessed; then the informative features are extracted using various digital signal processing and pattern recognition techniques; finally, the set of selected features are taken as the basis of flaw identification by training the proper classifier.
Thus, dimensionality reduction techniques need to be applied to transform the original features into a lower dimensional space which is useful for feature reduction to avoid the redundancy [3].
Given training data X with n samples and c classes, let ni be the number of samples in the ith class ().
Therefore, some schemes need be used for dimensionality reduction.
Finally, the features after reduction were fed to BP networks for flaw classification.

To the shortcomings of neural network in fault diagnosis, such as multiple input dimensions and the huge amount of data, some reductions from data based on rough sets theory are derived and unessential attributes were eliminated, an optimized rough set-neural network intelligent system was established.
Based on the above analysis, input data of BP neural network was processed firstly by using rough set, a diagnostic network was built according to reduction results, more satisfactory was achieved.
First reduction rules were mined from data set through rough set, BP neural network was designed by reduction rules and trained by reductive data set.
However, a considerable amount of data was continuous in practical applications, and therefore the data must be discretization.
The discretized data was reduced by using rough set theory, each sub-neural network input after the reduction was achieved.