Search results

A database of the loading history of such components and the associated thermal data will enable an estimate to be made of the life expectancy of similar components.
In the case of a real time monitoring system, thermal data is transmitted to a pc via an Ethernet/UDP interface.
Software analysis of thermal data.
Continuous monitoring of thermal data could also lead to an arrangement where warning signals are issued automatically allowing a reduction in the frequency and time of inspections.
Summary Thermal data of components can be obtained by safe, non-contact means using infrared cameras.

PCA is sensitive to the relative scaling of the original variables and has been widely used for multivariate data analysis and dimension reduction.
Often, its operation can be thought of as revealing the internal structure of the data in a way that best explains the variance in the data.
Step ①:Enhance the sensitivity of data.
The AR model established on experimental data thus is different.
The principal component analysis result of experimental operation factors indicates that, with dimensionality reduction, PCA effectively retain the feature information of the original data and remove the distractions.

Therefore, apply the data mining technology to the management of electrical equipment, automatically analyze and classify Power system data, Find out useful information to help managers analyze production can greatly improve the production efficiency, Bring more economic benefit [3, 4] Data Mining Application in Fault Diagnosis Using the data mining technology to solve the fault diagnosis work principle of power equipment can be summarized as the following three steps; the specific process is shown in Fig1.
ID3 Algorithm Improvements Three Factors Affecting the Original ID3 Algorithm The original ID3 algorithm is based on the main storage, When using this algorithm dig a large number of data in the database , the Shown outstanding issue is the efficiency of the algorithm significantly sharply decreased or even not feasible; The original ID3 algorithm need to provide the data in data files form, This not only need more data transformation work before data mining, but also against algorithm and database applications to realize seamless integration, This also affects its practicability to a large extent; Database technology is one of the most advanced and effective technology to manage and control large amounts of data in the database, The original ID3 algorithm does not use the advantage of data operation of database technology, This also influences the application efficiency of a mass of data to a great extent.
Algorithm Improvement Ideas Based on the above consideration, this article thought of using embedded SQLESQL/C#, through the help of the database management system, efficient data queries directly from the database of storing the fault data samples.
Validation and Comparison for the Efficiency of Algorithm In order to verify the implementation efficiency of the algorithm, this paper choose the letter-recognition data sets in the UCI machine learning database, random extract1000,5000,10000,15000,20000 data samples from the data sets, test and verify the improved ID3 algorithm respectively.
Top 10 algorithms in data mining.

This algorithm can directly deal with matrix data like face images.
PCA is one of the most famous techniques for data representation and compression.
The first item in Eq. 1 is similar to FCM, which depicts the data structure of both labeled and unlabeled data.
Experimental Results 4.1 Data Normalization.
Fig. 1 Face normalization Fig.2 Face data set 1.2 Experimental Result.

The results thus justified the inevitability of use of Tween-80 at higher concentrations to maintain surface tension reduction profile of electrolyte solutions.
The processing of resultant XRD intensity data of bulk and nano PbCrO4 product (Figure 3a-d) using in-built software Diffraceva and the corresponding data matching using ICDD data showed clearly that the XRD spectrum of bulk (without surfactant) batch displayed the (210), (102), (211), (221), (122) and (303) peaks corresponding to 2θ of 25.990°, 27.021°, 28.892°, 40.395°, 42.360° and 49.623° respectively, which represented the orthorhombic crystal orientation while, the Tween-80 mediated products (PbCrO4 nanorod) obtained via solution spray process exhibited intense Bragg reflections characteristic of monoclinic polymorph with (200), (012), (-131) and (132) peaks corresponding to 2θ of 25.7°, 29.6°, 40.0°, and 49.3°.
Secondly, for same impinging solution spray process, the micrographs advocated the influence of surfactant concentration on size reduction and stabilization of PbCrO4 nano crystals.
At Tween-80 addition corresponding to 10 g/L, major size reduction to the tune of 543 x 37.8 nm in length and diameter respectively was achieved (Figure 4c).
The FESEM micrographs shows major size reduction to the tune of 543 x 37.8 nm in length and diameter respectively with well separated, elongated nanorods at 10 g/L Tween-80 usages.

The relative influences of the thickness reduction, the roll radii, and the friction factors on the void length at the front and rear of the inclusion, respectively, are systematically examined.
Pan et al. [1] presented experimental data for the cold roll cladding of aluminum-stainless steel, copper-stainless steel and mild steel-stainless steel.
Figure 5 illustrates the influence of the reduction percentage on the relative void lengths at the exit for various friction factor ratios during the rolling of complex sheets containing inclusions.
It is clear that as the thickness reduction percentage increases, the relative void lengths at the front and rear of the inclusion, i.e.
This result is to be expected since a greater thickness reduction requires a higher load force at the roll gap, and this increased load force causes fracture voids around the inclusion.

It is very difficult to get complete data records with all 60 attributes, many records are incomplete and it is adaptable to do attribute reduction for them.
Though the input gas data is expressed in the percent of volume, the input data is still dispersant.
The diagnosing process is described as the following: Step 1: We use the first BAM network to separate normal state from all the data set, and we get normal data set and abnormal data set.
For abnormal data set, do step 2.
For overheating data set, do step3, for discharge data set, do step 4.

The received baseband OFDM signal from receiver The received baseband OFDM signal is given by , (1) where , after removing CP, the received signal is given by ， (2) where is the time-domain data sequence and is additive white Gaussian noise(AWGN).
Since the data of each subblock is the 1/P of the whole OFDM block after subblock tracking, then the Normalized Doppler shift of each subblock is just the 1/P that of the whole OFDM block.When P is appropriate,each subblock turns to be time-invariant.
Therefore, (2) can be further derived as (3) whereis the data of each subblock,;is the time window;，， is the CIR of the subblock period and is middle of the subblock period.
After FFT modulation, formula (3) can be rewritten by (4) where is the modulated signal corresponding to the subblock ,it yields (5) where ,withis the FFT modulated signal with and are the frequency do-main channel transfer function of the subblock period and data signal, respectively.
By (8), we have the signal of the subblock (10) Accordingly, the estimation of data block can be obtained by (11) B.

Based on in-situ observational data, we analyzed energy features of strong ground motion induced by short delay blasting in unsaturated loess field through three aspects of time history, frequency spectrum and attenuation.
For time histories, the data show exploding ground motion has two essential characteristics, i.e. larger peak ground acceleration (PGA) and shorter duration.
Fig. 2 shows the time-history data recorded at the centre position (UD, vertical component; SN and EW, horizontal components).
Time histories of the exploding ground motion For actual duration features of exploding ground motion, we adopted a quantitative criteria of acceleration () to analyze the time-history data as the follow
Using Eq. 2, duration features of recorded data (see Fig. 2) are summarized in Fig. 3 (the percentage representing the ratio of effective duration to duration).

The line data and also for load data for 33-bus system has been gain from [19-23].
Same as other test system, a high resistance and high reactance value will be modified in R and X column in line data to simulate the open line.
Result and Analysis Losses Reduction.
In the other hand, the percentage of power reduction for case 2 and case 3 are 37.6% and 38.2% respectively.
Thus, this EPSO method has lower loss reduction compare with other two methods (PSO and EP).