Wavelet Analysis Fault Information and Realization in Virtual Instrument

DUAN Jinying，WU Fentao

1．Xi jing University，Xi’an 710123，China;

2．Xi’an Power Electronics Research Institute，Xi’an 710061，China

1.Introduction

Wavelet analysis has the multi-resolution characteristics in time domain，frequency domain，and ability of characterizing local features［1］．It is very suitable for the analyzing fault transient signal in transmission line．Therefore，the wavelet analysis has been widely adopted and studied in fault location of transmission line［1 －3］．

There exist a variety of methods to complete the Wavelet analysis．For example，it can be achieved by using the wavelet analysis function in the MATLAB special wavelet analysis toolbox(Wavelet Toolbox)．It is very convenient to analyze the signal by wavelet since the toolbox has general function，wavelet function，one dimensional wavelet transform function，two-dimensional wavelet transform function，wavelet packet algorithm as well as in signal and image denoising and compression in wavelet analysis．Although MATLAB is powerful in numerical analysis and computation，the user interface is relatively poor，and the data acquisition，network communication and so on are comparatively cumbersome．Therefore，it is very difficult to analyze actual fault signal based on MATLAB．The Lab VIEW(Laboratory Virtual Instrument Engineering Workbench)language launched by the United States of America is a very good object-oriented graphicalprogramming language，whose source is the graphical diagram，to avoid the traditional language structural distress．

LabVIEW is an open development environment．It uses icons instead of text code to create the application and has a large number of communication interfaces with other application programs in VI database．It also has specialized wavelet and filter group design toolbox，and can achieve wavelet transform calculation by using filter decomposition，reconstruction in signal processing toolbox． However， this method requires the user to understand the relevant knowledge，more importantly，this method requires a large number of tedious repetitive work and the program design process is very complex，and its main-tainability is relatively poor．

Since MATLAB is not good at man-machine interface and data acquisition function，which the Lab VIEW is good at，and Lab VIEW could provide interfaces to call other applications，it will be very convenient for the user if the advantages of these two software could be combined．

In this paper，on the basis of actual transmission line model using ATP simulation data，admixture programming with Mat lab and Lab VIEW are studied by calling the MATLAB Script node directly．It mainly calls binary wavelet denoising in the Lab VIEW，makes the most of them．The simulation results show that binary wavelet is effective in power system fault signal denoising and the graphical interface is good．Therefore，it provides a good base to move to the next step for the fault location．

2.Wavelet analysis of fault information

The literature［1］says that the transient signal contains many fault information of the transmission line，for example，current and voltage transient signal contains the polarity information，current and voltage traveling wave amplitude satisfy certain relation．Fault transient signals contain many parts of the peak or mutation，and the noise is not stationary white noise．The analysis of this signal needs signal processing，removes signal noise component，extracts useful signal，and locates accurately［4］．

Wavelet transform，as a new tool of digital signal processing，has some advantages such as good timefrequency localization ability．It describes singular signal from the time domain and frequency domain in detail simultaneously．Fault signal is kind of changeable high-frequency signal．Therefore，the wavelet transform becomes the most effective mathematical tool when transmission line transient fault signal is analyzed．The basic ideas of fault information denoising could be shown in Fig．1 with wavelet［5］．

Fig．1 Diagram of Wavelet denoising

Once the signal is preprocessed，it is decomposed to multi-scale by using dyadic wavelet transform．The wavelet belonging to noise will be removed and the signal wavelet will be enhanced at the same time．

2.1.Dyadic wavelet transform

Assume function φj，k(t)∈L2(R)，(L2(R)is square and integrals space，that is Energy integral signal space)，if there are two constants A and B，and 0＜A＜B＜∞，this satisfies the stability condition this was almost everywhere，that is

φj，k(x)is dyadic wavelet．Eq．1 is stable condition，if A=B，then it is called the most stable condition．The functions sequence { W2jf( k)}k∈Zis called dyadic wavelet transform of f，where

By definition，with the change of j，dyadic wavelet can be divided into positive frequency axis dyadic adjacent band，realize the signal decomposition of the full frequency．It only for scale parameter on the discrimination and for time domain translation parameter keeps continuous．Another important property of dyadic wavelet is shift invariant，which makes it especially suitable for the detection of singularity signal［6 －7］．

2.2.Denoising characteristics

Compared with the power frequency signal，transmission line fault signal of high frequency and energy is very small，and this is very similar to the noise signal．Wavelet transform can realize multiscale singularity detection，which can be used to signal denoising．Although transmission line fault signal and noise have singularities，the performance in different scale is different［8］．

With the increase of the wavelet scale factor，the results of fault signal wavelet transform are the same or get increased，and the results of the noise signal wavelet transform will be greatly reduced and rapid attenuation．Therefore，the different singularity detection results between fault signal and noise multiscale can realize fault signal de-noising．So as to distinguish the useful signal and noise signal［8］．

3.Lab VIEW calls Mat Lab to realize wavelet analysis

3.1.ATP fault simulation

ATP Program(The Alternative Transients Program)is a world electromagnetic transient analysis Program(EMTP)，which is widely used，ATPEMTP Program is free for almost every person［9］．

The most dangerous fault of transmission line are all kinds of short circuit．They include single phase earthing，two phases earthing and two phases shorting．The probability of occurrence in a same type short circuit is different，different types of the size of short circuit current is different，which is generally about several times to dozens of times of its current rating．A phase ground fault is simulated and the specific model is shown in Fig．2．

In the process of simulation，ensure invariability of cables’length，change the position of the fault．Fig．3 is a phase ground fault phase current signal．a(chǎn)part from M end and 38 km．

Fig．2 Single-phase earthing fault simulation model

Fig．3 A phase ground fault current waveform

3.2.Lab VIEW calls the MATLAB wavelet analysis

In the process of hybrid programming development virtual instrument，Lab VIEW is used to design GUI，responsible for data acquisition，hardware control，operation control and network communication．And Lab VIEW calls neural network，wavelet analysis algorithm provided by MATLAB［10］．The programming method is calling MATLAB script node in this paper．

MATLAB script node realizes the Lab VIEW and MATLAB script node interactive data by increasing input/output terminal［11］．The block diagram of Lab VIEW called MATLAB is shown in Fig．4．It designs the corresponding man-machine interface of wavelet analysis，which can be easily set up small wave type，layer number，and order number．The original signal is simulated by the ATP．In a real application，it can analyze real-time signal．The results of original signal analysis is shown in Fig．5 and the addition of signal noise after decomposition is shown in Fig．6．

Fig．4 Wavelet decomposition sub vi program

Fig．5 The wavelet decomposition results of original signal in Lab VIEW

Fig．6 Wavelet decomposition after adding noise

Compare the results of Fig．5 and Fig．6，we can see，with wavelet function scale factor increasing，the results of transient signal transformation become bigger，noise transformation results become smaller．Therefore，the method of combining Lab VIEW and MATLAB to realize the wavelet transform can be successfully carried out，and the results are comparatively ideal;On the other hand，wavelet transform is effective used in signal singularity detection problem，and the result is ideal．In the real engineering applications，there exist more complicated signal that needs timely collection，analysis and pro-cessing．In a word，the proposed Lab VIEW with wavelet transform analysis methods provides convenient and effective solutions to these problems．

4.Conclusions

ATP is applied to simulate short circuit fault of transmission line in this paper．Denoised signal is obtained by calling MAT LAB’s wavelet transform in Lab VIEW．Compared the results between noise signal and the wavelet decomposition of original signal，fault information and noise are preliminary identified，the effect and graphical interface is good，which is good for further fault location．

［1］ GE Yaozhong．The Pprinciple and Technology of New Type Relay Protection and Fault Distance［M］．Xian:Xi’an Jiaotong University，2007．

［2］ YANG Leping，LI Haitao，ZHAO Yong．High Level Programming of Lab VIEW［M］．Beijing:TsingHua University Press，2003．

［3］ WEN Yuanfang．High Voltage Technology［M］．Hubei:Huazhong University of Science and Technology Press，2001:135－149．

［4］ KANG Zhongjian，WANG Shenghua．The denoising research of transmission line fault current traveling wave based on the morphological wavelet［J］．Power System Protection and Control，2008，4(19):36 －37．

［5］ YANG Fusheng．Wavelet transform engineering analysis and application［M］．Beijing:science publishing house，1999．

［6］ HU Changhua，LI Guohua．The system analysis and design based on MATLAB6．x-wavelet analysis(second edition)［M］．Xian:Xidian University Publishing House，2004．

［7］ Makming， Pongsak， Bunjongjit． FaultDiagnosis in Transmission Lines Using Wavelet Transform Analysis［C］//Proceedings of the IEEE Power Engineering Society Transmission and Distribution Conference，2002:2246－2250．

［8］ Charfi F．Haddad K A L．Francois B．Power system Fault Monitoring Using Wavelet Transform［C］//IEEE Annual Power Electronics Specialists Conference，2004 IEEE 35th Annual Power PESC04．2004:19－24．

［9］ HAN Lina，YANG Zhijian．The application of electromagnetic transient program EMTP in power system［J］．Guangdong transmission and substation technology，2006(2):19－14．

［10］ Lis Shen．LabVIEW and MATLAB hybrid programming method and its application in diesel engine fault diagnosis［J］．nstrument technology and sensor，2007(1):22－23．

［11］ ZHOU Ping ．The wavelet analysis and application based on Lab VIEW of the realization of［J］．Journal of micro computer information，2007(91):179－180．