## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

Contents lists available at ScienceDirect

**Expert Systems with Applications
**

journal homepage: www.elsevier.com/locate/eswa

**A novel analytic method of power quality using extension genetic algorithm and wavelet transform
**

Meng-Hui Wang ⇑, Yi-Feng Tseng

Department of Electrical Engineering, National Chin-Yi University of Technology, Taichung, Taiwan

a r t i c l e

i n f o

a b s t r a c t

The power quality affects the power stability of power company and customers. In order to avoid economic losses caused by the power disturbances, it is necessary to monitor power parameters. This paper aimed at power quality analyses by wavelet transform and proposed a novel algorithm called extension genetic algorithm (EGA). The paper introduced the fundamental theory of wavelet transform, current applications and the theoretical framework of EGA. Then, it described the deﬁnition of power quality problems and the characteristics of power waves. Finally, this paper compared the analysis results of EGA and other methods. As the results of simulation, this paper mentioned of methods has a very high accuracy. It can also provide an application tool on power quality and data classiﬁcation for future researchers. Ó 2011 Elsevier Ltd. All rights reserved.

Keywords: Power quality analysis Wavelet transform Extension theory

1. Introduction The power quality analysis is an important factor in the modern power systems. The electrical engineer must understand a certain statistical data and system when they analyze the electric power quality issues. Various kinds of electric power quality event used long-term monitoring of voltage to calculate the value rms. And then, we observing the values in a certain unit time of value change value to judgment the voltage were lower, rise or the electricity cut off (Chen, 2007). By the peak voltage and frequency changes, determine whether such problems as voltage ﬂicker or harmonic power. In order to accurate analyze of the various electric power quality problems, usually measure many kinds of the electric power characteristics. Therefore, this paper utilized the wavelet transform to deal with the disturbance waveform of electric power quality and expect to reduce the number of characteristics of pick waveform to maintain the accuracy of the identiﬁcation (Galli & Heydt, 1997). In recent years, the research pointed out the wavelet transform had Multi-Resolution Analysis (MIRA) of characteristics and in accordance with different frequencies of analyzing degree to provide different frame width that the transient change will leave obvious signal. Therefore, some scholars used wavelet transform to monitoring the location of the power system transient and classify different power system accident (Santoso, 1996). When the electric power quality perturbation signal continuous wavelet transform because

⇑ Corresponding author. Tel.: +886 4 23954505 2200; fax: +886 4 23924419.

E-mail address: wangmh@ncut.edu.tw (M.-H. Wang). 0957-4174/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2011.04.032

different input signal, it will show the corresponding wavelet coefﬁcient in the low frequency. If the wave of the signal has high frequency ingredient, it will also produce a bigger wavelet coefﬁcient in the part of high frequency. Therefore, according to the different high and low frequency coefﬁcient, it can be used as recognize of electric power quality perturbation wave. The extension of matter-element theory established the failure of the matter-element model can save a lot of modeling space and the advantages of rapid classiﬁcation. Some of diagnosis case has been successfully (Wang & Ho, 2005; Wang, 2002). This method needs to adjust the weight and set up the matter-element model by experience rule, which can achieve the highest accuracy rate. Therefore, if we can reduce the extension method to the experience rule, this method will increase the universality and the practicality. In view of this, this paper puts forth the EGA and wavelet transform to extract the characteristic value and carry on a classiﬁcation and identiﬁcation to the electric power quality of disturbance wave. In this paper, the methods to propose its very innovation and the highest accuracy rate have 97%. So it has a very high practical value to provide a new reference for scholarly research. 2. Synopsis of wavelet theory The wavelet transform utilize wavelet conversion of the zoom and translation two characteristic to sample time-shifting and scaling as show by Eq. (1). W(j, k) is the discrete wavelet transform in x (k), and w(t) is a ﬁnite energy and quickly constringency of timefunction (Wang & Wang, 2007)

If only a part of the scale is used. 1. Summary of extension theory The extension theory was ﬁrst introduced in 1983 by a China scholar Cai. C . 1] to [À1. Deﬁnition of extension set U is universe of discourse and u is a generic element of U. 1. y ¼ K ðuÞ > 0g A0 ¼ fðu. Table 1 Three different sorts of mathematical sets. 7 . and v is characteristic of value. If the original signal is described as having 1000 points.2. v) a multidimensional matter-element. y ¼ K ðuÞ 2 ðÀ1. yÞju 2 U . . . 1999): W ðj.12492 M. This value could calculate by correlation function that can describe the element to be positive ﬁeld. The matter-element in ET can be described as follows (Cai. The wavelet transform can also use a similar carrier signal. The high-frequency signal carries the detailed changes or partial perturbation. Assuming R = (N. & Sung. 5 ð3Þ vn 3. 1] Extension set Contradictory problems Matter-element model Correlation function Extension [À1. so we usually reduce the data to a sampling point as shown by Fig. kÞ ¼ XX j k xðkÞ2Àj=2 wð2Àj n À kÞ ð 1Þ R ¼ ðN .frequency 1000Points 500Points Fig. The signal after ﬁltering is divided into two parts: high-frequency and lowfrequency. W. v Þ ð2Þ The angle of observation is picked from the characteristic value therefore the continuous wavelet functions can carry the wavelet function on all times and scales. cn . Tseng / Expert Systems with Applications 38 (2011) 12491–12496 cD high-frequency 500Points cA low. There are two main points in extension theory that are matter-element model and extension set (Cai. c2 . 1Þg ð4Þ y = K(u) is correlation function of extension set A. The characteristic is c. C = [c1 c2 Á Á Á cn]T a characteristic vector and V = [v1 v2 Á Á Á vn]T. 3. Chung. . This will make the signal data become longer. yÞju 2 U . Y. low-frequency ﬁlter. Deﬁnition of matter-element In the extension theory. and the region is called joint ﬁeld. a great deal of current factor can be reduced without losing accuracy. 2011). then an extension set A in U is deﬁned as a set of ordered pairs (4): A ¼ fðu. negative ﬁeld or zero boundary (Wang. c .-H. there will be 1000 signals of 2 after the high. 1] The extension sets extent the fuzzy set from [0. Wang. The extension set A and the universe of discourse U are deﬁned as follows: A ¼ Aþ [ A0 [ AÀ Aþ ¼ fðu. y ¼ K ðuÞ < 0g ð5Þ ð6Þ ð7Þ ð8Þ . y ¼ K ðuÞ ¼ 0g AÀ ¼ fðu. 3. the obtained scale is huge. 1] (Das. 2006). 1} Fuzzy set Linguistic variables Fuzzy mathematics model Membership function Ambiguity [0. Compared item Objects Model Descriptive function Descriptive property Range of set Cantor set Data variables Mathematics model Transfer function Precision {0. It can solve contradictory problems which cannot solve by classical methods and fuzzy methods. For most signals the low-frequency is usually an approximate original signal. . 1998). v1 3 v2 7 7 7 . However. c. The signal ﬁlters and reduces the sampling points. A signal ﬁlter is based on the wavelet transform concept.-F. the element is R. The interval is called classical ﬁeld.1. yÞju 2 U . c1 . The matter-element model can describe the data that can analyze the quantitative change and the qualitative change. a value vector of then a multidimensional matter-element is deﬁned as 2 6 6 R ¼ ðN . The distance of extension describes the value from the interval in the region. The u belongs U(u e U). Table 1 shows three different sorts of mathematical sets. The extension set is built by matter-element model. and N is the name of element. yÞju 2 U . V Þ ¼ 6 6 4 N .

and Na is the total mounts. so it is not necessary to explain here.-F. . Step 5.. hG1 L . ck g. which transposed the notion of evolution in nature to computers and imitates natural evolution and selection. GU i Step 6. In general. 2. Basically. random events in which the base sequence of the gene is changed. The basic theory of genetic algorithm The best-known evolutionary algorithm (EA) is the genetic algorithm (GA). Step 2. Building the matter-element model from gene.1. Y. . . the tolerance of error rate Er. The extension method can be found in out at the paper by (Wang. the proposed clustering method involves a combination of the extension theory and genetic algorithm. Step 9. Chen. Step 12. v ja À Ra 6 GjL 6 v ja þ Ra v jb À Ra 6 GjL 6 v jb þ Ra Step 4. Positive field c a Classical domain b 1 5. Calculate the correlation function. The genetic algorithm has the ability to search for an optimal solution within a wide space. The genetic algorithm generally includes the following ﬁve parts: (1) Gene coding: Combining all genes into a chromosome of sequence 0 and 1. 5. The selection of the parental chromosomes put into the mating pool. . k zk ¼ ðGk L þ GU Þ " # jk n X jxjnk À zjk j À ðGjk U À GL Þ=2 þ1 K nk ¼ jk jðGjk i¼ 1 U À GL Þ=2j ð16Þ ð17Þ ð18Þ Step 8. . and vb is lower limit. . The extension theory provides a means for distance measurement in the classiﬁcation process. GU i 7 7 j ¼ 1. Input the training of data that is the value of gene. and AÀ is called negative ﬁeld. This section will present a mathematical description of the EGA. The genetic algorithm constitutes a class of search algorithms especially suited to solving complex optimization problems. Step 3. 2 is the graphic relationship of extension. Normalizing the value of correlation function for the matter-element model to be between 1 and À1. :::. Find the gene of lower limit and upper limit value -1 Joint domain Fig. We deﬁne several variables before using the algorithm. Input the next training of data to repeat Step 6 to Step 8. p2. . . A0 is called zero boundary. The proposed extension genetic algorithm In this paper.. 2 6 6 Rj ¼ 6 6 4 1 N. pn} with ith as follows: pji ¼ fc1 . Tseng / Expert Systems with Applications 38 (2011) 12491–12496 12493 K(x) and gets better accuracy without adjusting the weight (Wang. such as combining components in a novel creative way. the genetic algorithm is also suggested for solving problems in creative design. ck g Step 7.. The training stage Zero boundary d Extension field Negative field x 0 The chromosomes propagate next generation of chromosomes to combine the matter-element models in the EGA. (2) Fitness function: It describes the capability of a certain individual gene to reproduce and is usually equal to the proportion of the individual’s genes in all genes of the next generation. . . and the chromosome rate Ra.M. hG2 L . and repeat Step 5 to Step 9. Step 11. Set the epoch. The graphic relationship of extension sets. Step 10. they ﬁnd solution to a problem by maintaining a population of possible solutions according to the ‘‘survival of the ﬁttest’’ principle. Setting Patterns = {p1. Tseng. & Chao. . Fig. Using the proposed EGA can be simply described as follows: Step 1. (4) Crossover: A process in which chromosomes exchange genes through the breakage and reunion of two chromosomes. .-H. GL . Wang. . the major advantage of using the GA is that the optimal solution is obtained globally (Hwang & He. c1 . 2009). . Mutations are rare. . Produce new gene of lower limit and upper limit value with chromosome rate. xj ¼ fc1 . GL .. The EGA is a kind of supervised learning that ﬁnds the best classical domain k is number of characteristic. A+ is called positive ﬁeld. the crossover rate Cr. GL g ð12Þ ð13Þ ð14Þ The amount of gene in a chromosome is calculated by the function 2 Ã k Ã j. (3) Selection mechanism: It is the intentional manipulation by chromosome of the ﬁtness of individuals in a population to produce a desired evolutionary response. In addition to parameter optimization. c2 . the mutation rate mu. . Calculate the ﬁtness function. c2 . 2006). In the patterns. and the genes implement crossover mechanism. 11 12 12 jk chrom ¼ fG11 L . Fitness ¼ Nr Na ð19Þ Nr is the right amounts. 5 k cn . GU i 3 ð15Þ 7 2 7 c2 . (5) Mutation: A change in a gene resulting in new or rearranged hereditary determinants. m . . and j is the type of pattern. The genes make up the chromosome. hGk L . . The chromosome rate is produced with random generator. Input the next matter-element model. GL . va is the upper limit. v jb i ð9Þ ð10Þ ð11Þ 4. 2004). 2. v ja ¼ minðcjkn Þ v jb ¼ minðcjkn Þ v j ¼ hv ja . i is the total number of genes.

(b) Fig.5 0 0. random changes in frequency and timing. Find min(Knk). voltage dip. k zk ¼ ðGk L þ GU Þ=2 " # jk n X jxjnk À zjk j À ðGjk U À GL Þ=2 þ1 K nk ¼ jk jðGjk i¼1 U À GL Þ=2j ð22Þ ð23Þ ð24Þ Step 4. Result and analysis This text compliant IEEE STD 1159-1995 (IEEE recommended practice for monitoring electric power quality) to build the information for simulating a power failure. U.1 second (a) Energy 13 12 10 8 6 4 2 0 -2 -4 The ratio of accuracy Accuracy rate (%) Step 13. Tseng / Expert Systems with Applications 38 (2011) 12491–12496 th th th Fig. c1 .06 0.5 1 0. 2. . 5. Until the training is ﬁnished.1. GU i 3 7 7 j ¼ 1.5 0 -0.04 0. Step 14. Input the data that is recognize.. .5 -1 -1.03 0. otherwise go to Step 3. .. If training process is not ﬁnished. Let the next generation of chromosomes to replace the chromosomes. 3 shows the EGA analysis pattern. then the data does not belong to any. 2007.12494 M. . xj ¼ fc1 . 3. The type of parameter adjustment The ratio of training Normal Power failure voltage swell Voltage dip Harmonic voltage Transient voltage 1st 34/40 37/40 36/40 30/40 40/40 37/40 214/ 240 89 2nd 33/40 33/40 32/40 36/40 38/40 36/40 208/ 240 87 3rd 35/40 34/40 33/40 31/40 40/40 39/40 212/ 240 88 4th 35/40 30/40 32/40 35/40 38/40 33/40 203/ 240 85 5th 34/40 37/40 35/40 34/40 40/40 38/40 218/ 240 91 (P. voltage swell. In the training stage 40 samples of every type of power quality problem are 2 6 6 Rj ¼ 6 6 4 1 N. 2003. Step 5. GU i 7 . 2 7 hG2 L . and implement mutation mechanism..2. 1 2 3 4 5 6 7 Er ¼ ð1 À FitnessÞ Â 100% ð20Þ Layer Step 15.02 0. The simulation cycle is set to six cycles. Kwang. 6th and 7th wavelet transform. This paper presents a method that utilizes the wavelet transform to deal with the disturbance waveform in electric power quality. and transient voltage for the power quality problem (Bhattacharyya.09 0. . GU i Step 2. The power harmonic wave and the energy layers with the fundamental frequency removed. k hGk L . . 5th.) 1. c2 . . Until recognizing is ﬁnished. Y. ck g Step 3. power harmonic. Build the matter-element model by optimization solution. . Fig. The EGA needs the parameter characteristics for system training including the max value and min value windows. Calculate the correct rate. 2008). cn . Dash. Testing the ability to training There are numerous problems maintaining power quality. describing 128 points for a cycle.01 0.05 0. hG1 L . Calculate the correlation function. 6. otherwise go to Step 2.-H.. Wang. The recognizing stage Step 1.08 0. 6.07 0. The EGA of analysis pattern. Table 2 The extension with experience rule result.-F. If the Knt is bigger than k. m 5 ð21Þ c2 . . . Power quality maintenance problems include vibration amplitude issues. 4.

The value of correlation function is 1. Table 4 shows the matter-element model. À2:4913i > > > c4 .5 Error rate 0. Type 4 is voltage swell. 1:343i > : c5 h1:3119. c1 hÀ2:3726. If the value of correlation function is close to À1.04 0. and Fig. 6. That means the data is not belonged to this type. 5(b) shows the layers of energy are over zero. h0:66809. 5(b) shows the layers of voltage dip energy difference without the fundamental frequency. Table 6 is comparing the different sorts of clustering with 240 pieces of testing data. And the EGA has an advantage that is sloughing off the extension of experience rule. When the voltage is normal.2. h1:1198.2. hÀ7:5126. h0:077703. h0:000003. 31:787i > > > > < c2 . 1:5328i > : c5 . In the test stage a sample of every type of acceptable power quality is produced. there are 6 types of fault.03 (a) Energy 30 25 20 15 10 5 0 1 2 3 4 5 6 7 0 100 200 300 400 500 600 700 800 900 1000 Epoch Layer Fig. the wavelet transform can be reasoned to explain the power efﬁciency problem in many cases. and the mutation rate at 0. training times (epoch 1000) 1 2 3 4 5 6 1 2 3 The ratio of accuracy 230/240 228/240 231/240 229/240 231/240 229/240 230/240 228/240 231/240 Accuracy rate (%) 96 95 96 95 96 95 96 95 96 Average accuracy rate (%) 96 8 Intt. Fig. Tseng / Expert Systems with Applications 38 (2011) 12491–12496 12495 (P.05 0. Y. and fuzzy c-means of accuracy rate. h0:0000002. In the text. In Table 5.05 0. The highest accuracy rate of convergence curve. Wang.08 0. The EGA promotes non-progressive convergence. h1:1768. Two hundred and forty pieces of data produce 1000 epochs at a training iteration of 10. Type 3 is power dip. 1:2107i > : c5 . Testing the ability to analysis By the result. when the result is 1. And the result of value is conforming to each type in Table 5.03 0.06 0. hÀ3:0257. À1:1995i > > < c2 . The ﬁrst part is the extension rule of thumb.01 0. 0:54634i R2 ¼ c3 . That is belonged to which type. 1:5236i > : c5 . And the EGA of accuracy rate is 92%. the energy layers are under the zero form 1st to 7th layers. Type 2 is power failure. h0:00144. The training convergence result is recorded. 37:667i > > > > < c2 . The K-means clustering of accuracy rate is 85%. hÀ4:6065. h1:3085. 77:752i > > > > < c2 .02 0.1 second 0. 1:7938i > : c5 . 2002).045 0. Therefore. K-mean. 7:466i randomly produced.-H. The fuzzy c-means clustering of accuracy rate is 70%. 41:274i > > > c4 . This model is obtained after training.07 0. Two hundred and forty pieces of training data are produced in total.055 0. The average accuracy rate is 96% after 10 iterations.M. 5. Table 6 is showing that the EGA of accuracy is batter then the extension.035 0.1. c1 . h3:0461. hÀ1:6369. 8 Har. The training is divided into two parts. The error rate is 0. Fig. The second part is the training EGA. The highest training accuracy rate is 97%. h1:3524.06 0 0.5 -1 -1. c1 . Fig. hÀ0:23328. h0:24499. 4(b) shows the layers of power harmonic energy difference without the fundamental frequency. Table 5 is the value of correlation function that is normalizing between À1 and 1. Type 1 is meaning the problem of harmonic voltage. (b) Fig.09 0. c1 . Table 2 shows the extension with experience rule result. c1 . h1:328. 1:6361i 8 > > Swell. By the testing. the diagnosis is problem of harmonic voltage.075 0. Fig. h0:000003. 1:4996i > > > c4 . The total training data is 240 pieces. hÀ4:3416. 6 shows the highest convergence curve accuracy rate for each epoch. Fig. h2:6703. The voltage dip wave and the energy layers with the fundamental frequency removed. Table 3 shows the training accuracy rate.-F. À0:0588i R4 ¼ c3 . The EGA parameter sets the error rate tolerance to 0. Table 4 The matter-element model. it shows which type is. 5(a) shows the voltage dip wave. h1:3537. hÀ27:01. h0:004125. the 4th layer and the 5th layer of energy is over zero. 4:1231i > > > c4 . Type 5 is the voltage is normal. 4:3545i 8 Osci.07 0. And type 6 is transient voltage (Dugan. 2:9789i > > > > < c2 . c1 . the matter-element model is built by the highest accuracy rate of chromosome in the testing stage. At the result. 6. . U. hÀ2:2624. 39:846i > > > > < c2 . 0:3924i R1 ¼ c3 . 40:614i > > > c4 . 1:5632i 8 Sag .5 1 0. and Fig. 0:24934i R6 ¼ c3 . the crossover rate at 0. 4(a) shows the power harmonic wave. hÀ0:25736. 1:7008i Table 3 The training times of accuracy rate.065 0.08 0.5 0 -0. 4(b) shows when the fault is harmonic voltage.04 0. 0:7138i R3 ¼ c3 . The data of testing are 240 pieces.) 1. hÀ19:05. hÀ3:8565. The number of epochs is 1000. 1:5198i 8 Norm.033. h0:44524. À2:1059i > > > c4 .005. 0:1746i R5 ¼ c3 . 1:6189i > : c5 .

the EGA to power quality analysis can be the foundation of further research that might lead to other perspective or other problems. Wang. Kwang. W. Application of extension theory to PD pattern recognition in high-voltage current transformers. H. (2005). H.05 0. 63–64. Consequences of poor power quality – An overview. M.02 Voltage dip 0.17 0. (2007). Extension theory and its application. 20(3). (1999). ROC through its Grant No. Tseng. K. S. J. H. machines and drives (pp. Dugan. C.96 1 À0. B. (2007). IEEE Transactions on Power Delivery. 38(5). Y. In this research. F.45 1 0. Study of partial discharge measurement in power equipment using acoustic technique and wavelet transform. Galli. 6164–6169. K.19 1 Result with proper training. 651–656). IEEE Transactions on Power Delivery. C.89 0. 908–915. Detection of power quality disturbance based on binary wavelet transform.07 À0.. (1996). (2002).-H. 278–285. References Bhattacharyya. (2003). (1998). 673–682. Compared methods K-means clustering Fuzzy cmeans clustering The extension theory The extension genetic algorithm The training of accuracy rate (%) 63 64 The testing of accuracy rate (%) 59 61 91 96 85 92 7. 11(2). Power quality assessment via wavelet transform analysis. IEEE Transaction on Power Delivery. 12(2). R. Conclusion This paper supports a method which can advantage accuracy and conveniently with power quality analysis. W. 1939–1946. (1997). TENCON 2007–2007 IEEE region 10 conference (pp. Wang.12496 Table 5 The value of correlation function. IEEE Transactions on Power Delivery. H. 543–547). & Heydt. J. Using thermal image matterelement to design a circuit board fault diagnosis system. 22(3). (2011). In the future. This theory has proved a 96% accuracy rate . the method of using notion of genetic algorithm’s space in search is implemented to enhance the accuracy and to select a best matter-element. Cai. Expert Systems with Applications. Cai.58 0. Y. This fact has proved that the wavelet transform and EGA is an outstanding system for power quality analysis. If it is not trained.. In The 4th IET conference on power electronics. Power quality analysis using S-transform. Fuzzy logic-based fault-type identiﬁcation in unbalanced radial power distribution system. 1 2 3 4 5 6 Table 6 The comparing the different sorts of clustering. 406–411.. EGA is compared extension theory. A novel extension method for transformer fault diagnosis. The data of number 11 45 93 130 182 211 M. Power Quality event characterization. 924–930.66 À1 1 À0. 22(8).-F. NSC-97-2221-E-167-020. the simulate system showed the accuracy rate that is better than before.14 0. G. Acknowledgement The author would like to acknowledge the ﬁnancial support of the National Science Council of Taiwan.21 À0. Journal of Scientiﬁc Exploration.2 À0. Y. 18(2).78 À0. L.56 À1 À1 voltage swell À0.. (2009). Chen. T.75 Normal À1 À1 À0. W. Electrical power systems quality 2E. M..07 À1 1 0. From the result. M. Santoso. 44(7). Y. Transient power quality problems analyzed using wavelets. 1. & Ho. Wang. (2002). 8269–8276. T. Wang.. C. & Sung.. Dash. T.48 À0. C. Wang. & Chao. In the matterelement model of space. Wang. K. The extension set and incompatibility problem. S. Journal of Scientiﬁc Exploration. Inc.37 À0. the accuracy rate would be 92%. 21(1). McGraw-Hill. A. K-means and c-means. IEEE Transactions on Power Delivery. It uses the wavelet transform to extract characteristic value and to build various types of electric power quality problems with EGA. IEEE Transactions on Power Delivery. (2008). & Wang. H. D. IEEE Transactions on Power Delivery. Tseng / Expert Systems with Applications 38 (2011) 12491–12496 The value of correlation function type Harmonic voltage 1 À0. 36(4). Expert Systems with Applications.88 À0. Das.. G. 81–93. Chen.11 0. 1– 3). C.38 Power failure 0. P. In The 42nd international universities power engineering conference (pp.83 Transient voltage 0. Chung. H. (2007). W. (2006). A novel clustering algorithm based on the extension theory and genetic algorithm. 1575–1580. M.49 0.

- IEEE_1159_Update_Collins__T&D_2008.pdf
- Bollen Et Al-2010-European Transactions on Electrical Power
- Tofudi Com PHuk
- IEC STANDARDS ON POWER QUALITY.pdf
- PQ_Edition_4_EN_141021
- Power Quality
- mazloom-milad.pdf
- Eaton Cutler-Hammer-Metering Devices, Protective Relays & Communications
- Genetic Algorithms
- Power Quality Notes
- Introduction Into Power Quality
- Tb 02600002 e
- VLT 6000 Series Drives for HVAC
- Power Quality Site Survey (Presentation)
- Power Quality
- Dubuc-Deslauriers Interpolation Wavelets Tese
- 7700_Users_Guide
- Evolving Mathematical Functions with Genetic Algorithms - A Team Project on Artificial Intelligence
- WINE-D-16-01154.pdf
- JonesSoule2006_RobustnessinGenerationalvsSteadyStateGA_GECCO06p143
- Allen Bradley Energy Management Application Solution
- Dynamic Route Planning for Car Navigation
- PID Optimal
- Kumar 2014
- CH 3 Quantum Fuzzy
- 1011fyp HK02 FYP1 Proposal
- ejsr_44_4_14
- Design & Simulation of Energy Storage Unified Power Quality Conditioner (EUPQC) for Power Quality Improvement
- ECIS.compressed
- GEH 6273E Generic

- A Joint Adaptive Wavelet Filter and Morphological Signal Processing Method for Weak Mechanical Impulse Extraction
- PQ Event Detection Using Joint 2D-Wavelet Subspaces
- Power Quality Event Classification an Overview and Key Issues
- 01-Dwivedi U D -Shakya Deepti -And Singh S N -2008
- Power Quality Analysis Using Bilinear Time-Frequency Distributions
- an_adaptive_update_lifting_scheme_with_perfect_reconstruction.pdf
- THE PROBLEM OF CONTOURS IN TRANSFORM IMAGE CODING.pdf
- Low-Cost PQ Meter
- Power Quality and EMC in Smart Grid
- IEEE TPD a Self-Organizing Learning Array System for Power 2006
- Power Quality Indicators Estimation Using Robust Newton-Type Algorithm
- Power Quality Aspects of Smart Grids
- PQ Monitoring System for Real-Time Detection and Classification of Disturbances in a Sigle-Phase Power S
- Wavelet Regularity 05 236 by M. Pollicott and H. Weiss June 9, 2005
- Measurement of Current Taken by Fluorescent_pes2006
- A New Method for Measurement of Harmonic Groups in Power Systems Using Wavelet Analysis in the IEC Standard Framework
- A New Reactive, Distortion and Non-Active Power Measurement Method for Nonstationary Waveforms Using Wavelet Packet Transform
- The Lifting Scheme a Custom-Design Construction of Biorthogonal Wavelets
- FWTAA-Ref-6-1
- Power Quality Problem Classification Based on Wavelet Transform and a Rule-Based Method
- A New Morphology Method for Enhancing Power Quality Monitoring System
- The Effect of Time on Ear Biometrics
- A new algorithm for automatic classiﬁcation of power quality events based on wavelet transform and SVM
- A New Approach to Voltage Sag Detection Based on Wavelet Transform
- A Novel Analytic Method of Power Quality Using Extension Genetic Algorithm and Wavelet Transform
- A VXI Power Quality Analyser Implementing a Wavelet Transform-based Measurement Procedure
- A Wavelet-based Approach for Reactive Power Metering in Modern Three-phase Grids Grids
- Algorithm Analysis on Measuring the Begin-End Time of Voltage Sag Based on Biorthogonal Wavelet
- Wavelet representation of voltage ﬂicker

Sign up to vote on this title

UsefulNot usefulClose Dialog## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

Loading