1
Detection and Classification of Power Quality EventsBased on WT and ANN for Smart Grids
Saeed Alshahrani1, Maysam Abbod2, Basem Alamri3
123College of Engineering, Design and Physical Sciences, Brunel University London, UK
, ,
Abstract—In this paper, A powerful signal processing method wavelet transform is presented to detect power quality disturbances among one of the Artificial intelligence techniques which is Artificial neural networks as a classification system. As a result of the increased applications of non-linear load, it becomes important to find accurate detecting method. Wavelet Transform represents an efficient signal processing algorithm for power quality problems especially at non-stationary situations. These events are generated and filtered by wavelet as well as extract their features at different frequencies. Thereafter, a training process is done using ANN to classify power quality events.
Index Terms-- Power Quality, events, Feature Extraction, Wavelet Transform, Classification, Artificial Neural Networks.
I. Introduction
T
HEelectrical power devices have increased rapidly which caused a risen upto non-linear load in the electrical power systems including smart grids where is has more in terms of semiconductors devices.These events caused by non-linear loads changes the power supply its rated pure sinusoidal voltage and current. An effective method of detecting power quality events such as (sag, swell, harmonics, interruption, flicker and transient) is dependent on their accurate measurements in time which is still a challenge for researchers and engineers [1].
Fourier Transform (FT) is used in the definitions of power quantities mentioned in IEEE standard 1459-2010[2]. With deep investigation to FT algorithm, it is found that FT is capable to handle stationary events in the power lines but, notanappropriate method for non-stationary situations such as sag and transient[3]. Moreover, FT is unable to trace fluctuated eventswithlimited time[4].Therefore, it is essential to develop analternative method for measuring these quantities and detecting power quality events.
Wavelet transform (WT) represent one of the powerful signal processing algorithms that is used in the field. It gives the power signal both representations in time and frequency domain. WT has been involved in medical researches originally and it expanded to inter power quality field [6]. WT has two categories primarily called continuous wavelet transform (CWT) and discrete wavelet transform (DWT) [7]. Then, it has another derivative called STW which used for power quality events in [8]. Another wavelet transform algorithm foundation is wavelet packet Transform (WPT) which isimplemented to detect power quality indices as in [9].
For smart understanding, it is important to classify data collected from various power quality events to ensure accuracy and measure the proposed performance as described in [10]. Artificial intelligence techniques are used for classifications systems which proven its ability for learning, prediction, training and decision making as defined in [11]. One of the powerful AIs techniques is artificial neural network ANN which is capable in the field of data analysis,pattern matching and classification systems [12]. In power quality analysis, it is used to localise events associated power lines as in [13]. An automatic fault detector is proposed in [14]. It has helped in harmonic identification method presented in [15]. Authors have proposed a fault detection method in transmissions lines using ANN as in [16]. Furthermore, it is been implemented with S-transform for a classification system as described in [17].
II. Wavelet Transform
Detecting, identifying and localization of power quality events needs to be based on a sufficient algorithm especially non-stationary events. Wavelet transform has proved its efficacy with non-stationary events in signal analysis [18]. It can represent distorted signals in time and frequency domain as defined in [19]. Authors in [20] have investigated wavelet transform mathematically as follows:
wherei and jdescribes the integer values and represent wavelet expansion functions. Also are the two coefficients of discrete wavelet transform (DWT) of. These coefficients have the formula:
wherecan be calculated from the mother wavelet through the following scaling:
whereirepresent the scaling parameter in wavelet and j for the translation one. It is necessary to satisfy conditions related to mother of wavelet. Multiresolutionis one of these conditions where the different of two scale equation is given as:
wherecan satisfy the conditions to let wavelet function have a unique value and is the scaling function where it has a relation with mother of wavelet as bellow:
where h in (4) and g in (5) arethe filters of wavelet: high pass filter and low pass filter respectively. From equations mentioned earlier, the j level wavelet value is then calculated as:
where , , are the coefficients at scale j+1 and they are calculated under the availability condition of scale j as follow:
whereis the approximation coefficient and is the detailed one at scale j+1 and they are given.
Figure 2 shows the process of decompositionfor power quality events using discrete wavelet transform DWT.
Fig. 2. Process of decomposition of power signal based on DWT.
- Detection And Identification
In this section, a pure sine wave which has frequency of 50 Hz is generated among nine power quality events (sag, swell, harmonics, interruption, flicker, high frequency transient, low frequency transient, sag with harmonics and swell with harmonics) which represent common quality events in smart grids. Table Ishows power qualityeventswhich are generated according to their parametric equations [21]. Figure 3 shows the output of generated models.
Table I
Signal modelling Of Power Quality Events
PQEs / Equations Models / ParametersPure Sine / / A=1.0 f=50 Hz
Sag / /
Swell / /
Harmonics / /
Interruption / /
Flicker / /
High frequency transient / /
Low frequency transient / /
Sag with harmonics /
/
Swell with harmonics /
/
Fig. 3. Power quality events in smart grids.
- Feature Extraction
In this section, the strategy of extracting power quality events is to transform time domain of the original distorted signal to its energy form which well be the key factor in the classification system. Each event generated are decomposed to 8 levels of multi-resolution analysis and the results for selected PQEs as shown in Figures 4, 5, 6, 7, 8, and 9.
Fig. 4. Features of pure voltage signal extracted based on DWT.
Fig. 5. Features of voltage sag signal extracted based on DWT.
Fig. 6. Features of voltage swell signal extracted based on DWT.
Fig. 7. Features of Voltage Harmonic signal extracted based on DWT.
Fig. 8. Features of voltage interruption signal extracted based on DWT.
Fig. 9. Features of voltage flicker signal extracted based on DWT.
- Classification System
In this study, artificial neural networks (ANN) is implemented to evaluate the performance of the extraction method based on wavelet transform. ANN classifier technique is able to measure the efficiency by training and testing the data extracted from power quality events where each one of these events has its own neural.
After several experiments on the number of hidden layers, ANN can give the best performance for the classifier. As a network learning, results of features extracted of power events are fed to ANN as inputs. Thereafter, learning process is conducted to solve issues attached to data and then solve other learned results. The principle of ANN is based on multiplying inputs with their weights. Then, results are calculated by mathematical process to find the activation of the neural as outlined in Figure 10. Outputs of this work are then reprocessed by anther mathematical functionto initiate a structure if rules to classify different new inputs.
Fig. 10. The Scheme of ANN.
In this section, three phases of training are conducted for data through ANN. First phase, features data of power events are fed to ANN for essential training according to which pattern recognition of inputs. The Second phase is data validation which is done by measuring the network efficiency. Third phase is testing the accuracy to ensure the best performance of the network.
- Results and Discussions
Various signals of power quality events in smart grids were generated and diagnosed according to their parametrical equations. Data analysis of features extracted has been done based on discrete wavelet transform DWT as explained in IV power quality issues.
Thereafter, an investigation of 10 types of these events in smart grids is done sing 8 multiresolutiondecompensation process with respect to IEEE standard requirements: magnitude and frequency [22].
Trainingof ANN is conducted for distorted signals where 40 variables of 10 types of power problems is recorded (pure and others PQEs) toneural. Than after, a database of 3200 inputs is founded based on extracted features of these events using 8 multiresolutionanalysis.
At classification stage, one, two and three hidden layers from 5 to 50 neurons are used to enhance accuracy and explore the best performance of the classifier.
Results of one hidden layer based are shown in Table II. The accuracy is detailed for each quality signal. Moreover, Tables III and IV showthe accuracy of two and three hidden layers and respectively.
Table II
Classification Accuracy Of Ann With One Hidden Layer
Neural / PQEs / Accuracy Rate (%)C1 / Normal / 90.786
C2 / Voltage Sag / 90.800
C3 / Voltage Swell / 90.657
C4 / harmonics / 91.157
C5 / interruption / 91.104
C6 / flicker / 90.786
C7 / high frequency transient / 90.714
C8 / low frequency transient / 90.914
C9 / sag with harmonics / 90.801
C10 / swell with harmonics / 90.314
Table III
Classification Accuracy Of Ann With Two Hidden Layers
Neural / PQEs / Accuracy Rate (%)C11 / Normal / 91.271
C12 / Voltage Sag / 91.071
C13 / Voltage Swell / 90.829
C14 / harmonics / 91.114
C15 / interruption / 91.057
C16 / flicker / 90.771
C17 / high frequency transient / 90.414
C18 / low frequency transient / 90.603
C19 / sag with harmonics / 90.657
C20 / swell with harmonics / 90.671
Table IV
Classification Accuracy Of Ann With Three Hidden Layers
Neural / PQEs / Accuracy Rate (%)C21 / Normal / 91.128
C22 / Voltage Sag / 91.614
C23 / Voltage Swell / 91.129
C24 / harmonics / 90.714
C25 / interruption / 90.729
C26 / flicker / 90.742
C27 / high frequency transient / 91.114
C28 / low frequency transient / 91.157
C29 / sag with harmonics / 91.300
C30 / swell with harmonics / 90.729
- Conclusions
This paper presents a WT detection method and a robust ANN classifier for power quality events. In this study, it has been proved that wavelet transform is able to overcome Fourier transform limitation especially with non-stationary events whichstated in this paper. Thereafter, Features of these quality events are extracted to 8 levels and a calculation of energy values are done for each event which iseventually used to build database needed for the classifier.
ANN Classifier is implemented and database of a random 3200 signals are generated. Results of the above trained data are analysed, trained and tested to evaluate the accuracy of the classification system. As a result, the accuracy of theclassifier of power quality events were more than 90% for all the 30 neurons which showeffectiveness of ANN classifier for power quality events.
References
[1] J. Arrillaga, M. H. Bollen and N. R. Watson, "Power quality following deregulation," Proc IEEE, vol. 88, pp. 246-261, 2000.
[2] IEEE Standard Definitions for the Measurement of Electric Power Quantities Under Sinusoidal, Nonsinusoidal, Balanced, or Unbalanced Conditions," IEEE Std 1459-2010, pp. 1-50, 2010.
[3] P. G. Axelberg, I. Gu and M. H. Bollen, "Support vector machine for classification of voltage disturbances," Power Delivery, IEEE Transactions on, vol. 22, pp. 1297-1303, 2007.
[4] J. Driesen, T. Van Craenenbroeck, R. Reekmans and D. Van Dommelen, "Analysing time-varying power system harmonics using wavelet transform," in Instrumentation and Measurement Technology Conference, 1996. IMTC-96. Conference Proceedings. Quality Measurements: The Indispensable Bridge between Theory and Reality., IEEE, 1996, pp. 474-479.
[5] Y. H. Gu and M. H. Bollen, "Time-frequency and time-scale domain analysis of voltage disturbances," Power Delivery, IEEE Transactions on, vol. 15, pp. 1279-1284, 2000.
[6] S. Santoso, E. J. Powers and W. Grady, "Power quality disturbance data compression using wavelet transform methods," Power Delivery, IEEE Transactions on, vol. 12, pp. 1250-1257, 1997.
[7] L. Angrisani, P. Daponte, M. D'apuzzo and A. Testa, "A measurement method based on the wavelet transform for power quality analysis," Power Delivery, IEEE Transactions on, vol. 13, pp. 990-998, 1998.
[8] W. G. Morsi and M. El-Hawary, "A new perspective for the IEEE Standard 1459-2000 via stationary wavelet transform in the presence of nonstationary power quality disturbance," Power Delivery, IEEE Transactions on, vol. 23, pp. 2356-2365, 2008.
[9] W. G. Morsi and M. El-Hawary, "Wavelet packet transform-based power quality indices for balanced and unbalanced three-phase systems under stationary or nonstationary operating conditions," Power Delivery, IEEE Transactions on, vol. 24, pp. 2300-2310, 2009.
[10] M. Zhang, K. Li and Y. Hu, "A real-time classification method of power quality disturbances," Electr. Power Syst. Res., vol. 81, pp. 660-666, 2011.
[11] W. A. Ibrahim and M. M. Morcos, "Artificial intelligence and advanced mathematical tools for power quality applications: a survey," Power Delivery, IEEE Transactions on, vol. 17, pp. 668-673, 2002.
[12] M. K. Saini and R. Kapoor, "Classification of power quality events–a review," International Journal of Electrical Power & Energy Systems, vol. 43, pp. 11-19, 2012.
[13] A. K. Ghosh and D. L. Lubkeman, "The classification of power system disturbance waveforms using a neural network approach," Power Delivery, IEEE Transactions on, vol. 10, pp. 109-115, 1995.
[14] C. Dilokratanatrakool, P. Na Ayudhya, T. Chayavanich and C. Prapanavarat, "Automatic detection-localization of fault point on waveform and classification of power quality disturbance waveshape fault using wavelet and neural network," in Robotics, Intelligent Systems and Signal Processing, 2003. Proceedings. 2003 IEEE International Conference on, 2003, pp. 142-147.
[15] D. Srinivasan, W. Ng and A. Liew, "Neural-network-based signature recognition for harmonic source identification," Power Delivery, IEEE Transactions on, vol. 21, pp. 398-405, 2006.
[16] K. Silva, B. Souza and N. Brito, "Fault detection and classification in transmission lines based on wavelet transform and ANN," Power Delivery, IEEE Transactions on, vol. 21, pp. 2058-2063, 2006.
[17] N. Huang, D. Xu, X. Liu and L. Lin, "Power quality disturbances classification based on S-transform and probabilistic neural network," Neurocomputing, vol. 98, pp. 12-23, 2012.
[18] G. Heydt and A. Galli, "Transient power quality problems analyzed using wavelets," Power Delivery, IEEE Transactions on, vol. 12, pp. 908-915, 1997.
[19] I. Daubechies, Ten Lectures on Wavelets. SIAM, 1992.
[20] M. Uyar, S. Yildirim and M. T. Gencoglu, "An effective wavelet-based feature extraction method for classification of power quality disturbance signals," Electr. Power Syst. Res., vol. 78, pp. 1747-1755, 2008.
[21] A. Gaouda, M. Salama, M. Sultan and A. Chikhani, "Power quality detection and classification using wavelet-multiresolution signal decomposition," IEEE Trans. Power Del., vol. 14, pp. 1469-1476, 1999.
[22] IEEE Standards Association, "1159-2009. IEEE Recommended Practice for Monitoring Electric Power Quality Industrial and Commercial Applications," 2009.
SaeedAlshahraniPhD research student in Electrical and Electronic Engineering, Brunel University, UK, since 2014. MSc in Automation and Control Engineering, Newcastle University, UK 2009.BSc from King Saud University, 2003.
Dr Maysam F. Abbod received the BSc degree in electrical engineering from the Baghdad University of Technology, Baghdad, Iraq, in 1987, and the PhD degree in Control Engineering from the University of Sheffield, Sheffield, UK in 1992. He is currently a reader in Intelligent Systems at the College of Engineering, Design, and Physical Sciences, Brunel University London, Uxbridge, London UK. His main research interests are intelligent systems for modelling, control and optimization. Dr Abbod is a member of IET and a UK Chartered Engineer.
BasemAlamriBSc in Electrical Engineering from KFUPM, Dhahran, Saudi Arabia. MSc in Sustainable Electrical Power, Brunel University, London. UK. PhD researcher at Brunel Institute of Power Systems, London, UK