Search results

The smaller the loss value, the better the model is in studying the training data. accuracy is the accuracy value calculated from the training data for each epoch.
The bigger the valueval_accuracy, the better the model is at recognizing new data.
This shows that the model is getting better at minimizing prediction errors in both training data and validation data.
This shows that the model is getting better at increasing the accuracy of predictions on both training data and validation data. val_loss values ​​andval_accuracy tends to stabilize after the 50th epoch.
Of the 110 images, 90 images were used as training data and 20 images were used as test data.

Introduction Base on the atmospheric environmental quality data through more than 20 years in Tianjin, this paper describes and analyzes the characteristics of atmospheric environmental pollution.
Based on the atmospheric environmental quality monitoring data of Tianjin from 1986 to 2010, the annual average concentration change trend of the main atmospheric pollutants is shown in figure 1.
For an environmental quality index Y, this paper analysis a set of variables which is closely related to x1, x2,..., then this paper finds these variables in the data, and uses the method of linear regression to fill the missing value.
After the standardization of the data, the data have been made as a column of mean as zero and variance as a dimensionless value.
Then by using the method of principal component analysis to dimension reduction of variables, and this paper gets several unrelated principal component to represent more than 85% of the original data information.

But any increase in stirrer speed t did not contribute to size reduction in small particles below 2μm.
It was noted that any increase in the mechanical forces did not contribute to size reduction in small particles below 2μm.
The data indicated that gel content increased significantly with the amount of crosslinking agent.
But any increase in stirrer speed did not contribute to size reduction in small particles below 2μm.

Selective Catalytic Reduction (SCR) is a method that can be used in mobile diesel engine aftertreatment systems to reduce harmful NOx emissions.
Introduction One of the major concerns on heavy duty diesel vehicles is the reduction of Nitric Oxides (NOx).
This requires advanced aftertreatment techniques such as Selective Catalytic Reduction (SCR) systems.
Homogeneous Urea-SCR Reactions: The formation of NH3 out of UWS droplets follows several principle steps [1, 2]: i) Evaporation of water from a fine spray of UWS droplets, NH22COl→NH22COs+H2Og (1) ii) Thermolysis of urea into ammonia and iso-cyanic acid, NH22COs→NH3g+HCNOg (2) iii) Hydrolysis of isocyanic acid, HCNOg+H2O→CO2g+ NH3g (3) Munnannur and Liu [3] used single kinetic rate model in conjunction with homogeneous finite rate chemistry calculations to model thermolysis and hydrolysis of urea with reasonable agreement of experimental data reported by Kim et al. [4].

The simulation results show that the improved hybrid algorithm combines the only solution solving feature and noise reduction ability of the MAP method and strong prior knowledge utilized feature and flexibility of the POCS method, effectively ensures the convergence stability of the restoration and maintains the edge details of the restored image, enhances the effect of the super-resolution restoration.
The accurate degradation functions can ensure the accuracy and validity of data, describe the signal transmission characteristics of the imaging system, improve the performance of the restoration algorithm.
From the subjective evaluation criteria, the noise reduction effect is more obvious, the image edge is well maintained and with less ringing effect.
From the processed results, it assumes that the bicubic interpolation algorithm suffers the noise obstructing; the Wiener filter method has a clear subjective visual perception but still gets the same problem with the bicubic algorithm; the POCS algorithm has the issue of noise reduction and shows the ring effect obviously, the MAP algorithm has the better effect on reducing noise but does not keep the edge retention which shows blur on the re1stored high frequency details; the hybrid super-resolution algorithms overcome the problems mentioned above and acquire better results, the algorithm we proposed also is better than the one proposed in the paper[8] on the subjective evaluation criteria as well as the objective evaluation.

The experimental data show that hydrogel have high sensitivity towards Vanadium and Selenium, adequate selectivity and reproducibility, and a wide operative linear range of concentrations for the determination of Vanadium (V) and Selenium (Se) in water samples.
Whether the goal is qualitative or quantitative data, Raman analysis can provide key information, easily and quickly [30]. 
Three oxidation peaks and two reduction peaks were identified in the hydrogel voltammogram (figure 7).
This was confirmed by scan rate dependent studies, which showed no shift in peak position but increased as the scan rate was increased (data not shown).
However guided by standard reduction potential tables as well as a detailed stepwise analytical protocol, unambiguous quantification is possible.

The drawing textures of aluminum, copper, gold, silver, and Cu-7.3% Al bronze wires are approximated by major <111>+minor <100>, except silver wire, which can have the <100> texture at extremely high reductions.
Introduction The drawing textures of aluminum (stacking fault energy (SFE): 166 mJ/m2 [1]) [2], copper (SFE: 78 [1]) [3,4], gold (SFE: 45 [1]) [4], silver (SFE: 22 [1]) [5], and Cu-7.3% Al bronze (SFE: <20) [6] wires are approximated by major <111>+minor <100>, except silver wire, which can have the <100> texture at extremely high reductions.
The drawing textures of aluminum [2], copper [3,4], gold [4], silver [5], and Cu-7.3% Al bronze [6] wires are approximated by major <111>+minor <100>, except silver wire, which can have the <100> texture at extremely high reductions.
(left) β-fiber intensity lines of Cu-16%Mn alloys after various rolling reductions [8].
The mobility data of tilt grain boundaries in aluminum [16] show that the average mobility of a <111> tilt boundary is increasingly higher than that of a <100> boundary with increasing temperature.

The results indicate options for reduction of the loses by replacement of old type of transformers and preparation of the MV system for upgrade and change the voltage level from 10 kV to 20 kV.
This paper analyses the reduction of technical losses when the existing 110/10 kV transformers are replaced with 110/20 kV transformers, and the MV grid voltage is changed from 10 to 20kV.
A reduction of network losses would make an important contribution to the EU’s plan to increase energy efficiency in electricity supply.
Total loss of the active power and reactive in the whole power system of Kosovo Reducing the losses in substations “Prishtina 2” and “Prishtina 3” by replacement of the old type of transformers and cables and change the MV voltage level from 10kV to 20kV A study was conducted in order to estimate the option for the reduction of the technical losses when the transformers 110/10 kV are replaced with transformers 110/20 kV, transformers 10/0.4 kV replaced with transformers 20/0.4kV and power distribution cables with section 95 mm2 with new cables with section 150 mm2 in substations “Prishtina 2” and “Prishtina 3”.
References [1] Data from KOSTT –transmission system and market operator; [2] Hadi Saadat, Power System Analysis Third Edition 3rd (third) EDITION.

The measured shrinkage reduction was up to 55.3% at the age of 180d of drying.
The datas in table 4 show that expansion increases with the increasing of dosage of UMEA replacing of cement.
Furthermore, there is a significant reduction with respect to the reference cement pastes in the drying shrinkage at the end of testing age from 0.304% to 0.136%, which the shrinkage reduction is up to 55.3%, when the amount of UMEA is 10%.
The datas in table 5 indicate that, the cement mortars with UMEA have larger restrained expansion and lower restrained shrinkage than the cement mortar without UMEA.
When the amount of expansive agent is 10%, the measured drying shrinkage reduction was up to 55.3% at the age of 180d, the initial cracking time is delayed by 17d and the crack width is only 1/6 of the reference mixture. 3) When the deformation was restrained, the cement mortars with UMEA have larger restrained expansion and lower restrained shrinkage than the cement mortar without UMEA.

The numerical results showed that the simulation data agree well with experimental test data.
Contralateral welding The same side welding Fig.7 Stress of W1 at different time Fig.8 Stress curve of welding Figure 7 indicates that the welding of W3 on the same side has little effect on W1, while the contralateral welding of W2 leads to the longitudinal stress reduction of W1 and residual stress of W1 varies greatly.
In the first welding sequence, welding of W2 leads to the temperature of materials near seam zone of W1 to above 200 ℃, which is equivalent to a post-processing, leading to longitudinal stress reduction of W1.