Search results

From the measured values - after checking and cleaning the data - we created hourly climate data set for further research.
Temperature and relative humidity data evaluation.
To illustrate our measured dataset from 12 October to 20 May we used data contouring.
For calculating the U-value from measured temperature data and heat flow densities, averaging method was used [4].
Using the measured data, hs surface heat transfer coefficients as well as Rs surface resistances can be obtained.

GM(1,1) model has the characteristic with less data, poor information, simple calculation and high precision.
The linear regression model for long-term prediction has strong anti-interference, but highly dependences on raw data, and requires a lot of data.
More original data, more laws, and higher accuracy.
The first is to do accumulate the original data to generate, that is a new data exponentially law, after get the original series accumulate.
Table 4 shows that the combined average relative error of prediction model is minimum, compared to three models to predict the data.

We set the derivative ofin (6) on to be zero: (7) Multiplying (7) by , we have (8) Thus may be expressed as (9) Due to (7) and (9), we have (10) That is to say (11) Similarly, we get on dataset B through (12) (12) In order to eliminate correlation of discriminant transform between data set C and data set A, and between data set C and data set B. we construct the objective function and constraint: (13) replacerespectively, (4) convert into(13).
In order to eliminate data the correlation of discriminant transform between set A and B data set.
We set , where is discriminant transform through the Local projection preserve criterion in data set A,so we construct following objective function and constraint: (17) Where is training samples of data set B, is a diagonal matrix,,is a similarity matrix, ,it is a Laplacian Matrix.
Similarly, we get through (26): (26) In order to eliminate correlation of discriminant transform between data set C and data set A, and between data set C and data set B, we construct the objective function and constraint: (27) is training samples of dataset C, replace respectively.(28)is converted into (27).
We scaled the intercepted images to 60× 48 on three face data set.

The next step was the collection of training data.
Up to now we collected training data from 57 people.
We structured and checked the audio data. 50 samples were suitable for processing.
The new audio data is ca. 1 minute long.
The method does not depend on the data type or data source.

Reduction of the measurement that is caused by the delayed data can affect the whole performance of harmonic state estimation.
However the standard LMS suffers from data dependent behavior and the random discontinuity of the predictor input when the data is delayed also affect the LMS performance.
Delayed data.
Yes No Data Pre screening Harmonic prediction Data delayed?
The number of the data collected to train the predictor was 1800 sample and the mean square error (MSE) is taken as performance indicator which is taken for all training data.

Its functional principle is based on the reduction of the contact shear stress at the sheet surface in the contact zone with the forming tool by means of locally applying a hydrostatic fluid pressure.
Beside technological advantages like increased accuracy of form and dimensions or improved strength, the HBU also allows the reduction of process steps for certain geometries compared to conventional deep drawing [1, 2].
The whole measuring system using the software Diadem, National Instruments, allows the measuring of process parameters, pressures, flange draw-in with high accuracy as well as the process control and analysis of the measured data with high reproducibility.

To use this soil as a subgrade material in the roadway, this soil needs to meet various engineering standard criteria such as deformation, sulfate reduction, strength, and durability for use as subgrade material.
Each mixture was tested three (3) times to provide more precise data, and the average UCS value was computed and analyzed.
After 3-D swelling test, the mixture treated with 6% lime drastically reduced the DC values at 28-day, but the mixtures treated with lime-and-slags had less reduction in the DC values within the same period.

Experimental data have been collected from the literature.
A good agreement has been obtained between experimental data and calculated phase boundaries.
Several data sets are also shown for comparison and the agreement obtained is good.
Accurate experimental data are needed and ab initio estimation can also be used when data are lacking.
Data Monograph No. 9 [8] R.

The best fit to the data was obtained with the Langmuir isotherm (non-linear) with maximum monolayer adsorption capacity (Qo) at 5% magnetic loading MBM adsorbent is 0.203 mg/g with Langmuir constants KL and aL are 2.055 L/g and 10.122 L/mg respectively.
The pseudo-first-order (non-linear) kinetic model provides the best correlation of the experimental data with the rate of adsorption (k1) with magnetite loading 2% and 5%, respectively are 0.024 min-1 and 0.022 min-1.
The adsorption equilibrium data obtained from the experiment were examined to find the most fitting model between Langmuir, Freundlich, and BET (Brunauer–Emmett–Teller) using non-linear analysis methods using SSE (Sum of Squared estimate of Errors) values.
Figure 4 illustrates the fitting of experimental kinetic data for Fe (II) adsorption on different magnetite loading (2 and 5% w/w) to the theoretical model.
By examining the experimental data, we acquired that the adsorption rate is affected by the initial concentration of adsorbate and magnetite loading in adsorbents.

Optical detection methods, such as cadmium reduction and phenol disulfonic acid spectriphotometric, are commonly applied for nitrate-nitrogen detection by using the wavebands of 400～425nm or 510～550nm with an accuracy of 0.1～2mg/L [4].
Besides, problems like size reduction and system integration still remain unsolved, because these nitrate-selective electrodes commonly require inner filling solution.
A micro controller was selected for data processing.
Average potential data of 10 times ISEs readings could be displayed on the micro controller and sent to the lap-top computer. 6 Miniature pumps and 5 miniature valves were used to achieve the solution delivery.