Data communication through distribution networks for smart grid applications
2015; Institution of Engineering and Technology; Volume: 9; Issue: 6 Linguagem: Inglês
10.1049/iet-smt.2014.0215
ISSN1751-8830
AutoresYousef El Haj, Lutfi Albasha, Ayman H. El‐Hag, Hasan Mir,
Tópico(s)Advancements in PLL and VCO Technologies
ResumoIET Science, Measurement & TechnologyVolume 9, Issue 6 p. 774-781 Research ArticlesFree Access Data communication through distribution networks for smart grid applications Yousef El Haj, Yousef El Haj Electrical Engineering Department, College of Engineering, American University of Sharjah (AUS), Sharjah, UAESearch for more papers by this authorLutfi Albasha, Corresponding Author Lutfi Albasha [email protected] Electrical Engineering Department, College of Engineering, American University of Sharjah (AUS), Sharjah, UAESearch for more papers by this authorAyman El-Hag, Ayman El-Hag Electrical Engineering Department, College of Engineering, American University of Sharjah (AUS), Sharjah, UAESearch for more papers by this authorHasan Mir, Hasan Mir Electrical Engineering Department, College of Engineering, American University of Sharjah (AUS), Sharjah, UAESearch for more papers by this author Yousef El Haj, Yousef El Haj Electrical Engineering Department, College of Engineering, American University of Sharjah (AUS), Sharjah, UAESearch for more papers by this authorLutfi Albasha, Corresponding Author Lutfi Albasha [email protected] Electrical Engineering Department, College of Engineering, American University of Sharjah (AUS), Sharjah, UAESearch for more papers by this authorAyman El-Hag, Ayman El-Hag Electrical Engineering Department, College of Engineering, American University of Sharjah (AUS), Sharjah, UAESearch for more papers by this authorHasan Mir, Hasan Mir Electrical Engineering Department, College of Engineering, American University of Sharjah (AUS), Sharjah, UAESearch for more papers by this author First published: 01 September 2015 https://doi.org/10.1049/iet-smt.2014.0215Citations: 6AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Abstract In this study, a novel solution is introduced to deal with excessive signal attenuation in distribution transformers for power line communication in smart grids. The results presented were not achievable before proposing this solution because of the filtering action of transformer. The solution is based on exploiting a unique frequency resonance spotted in transformer windings at 490 kHz with an amplifying gain (∼2 V/V) and a bandwidth about 40 kHz. It was successfully investigated to transmit a communication signal through 15 kVA distribution transformer. The 40 kHz bandwidth and the associated amplifying gain allowed fast and efficient transmission of data, which was not feasible before. The error of a coded binary phase shift keying signal propagating through a distribution transformer was measured in the laboratory. This simple and effective solution avoids costly hardware, such as bypassing circuits, and does not require sophisticated communication modulation schemes such as orthogonal frequency division multiplexing. A full model was developed for system simulations. The work evaluated factors that may influence the resonance frequency such as distribution cables and transformer ageing. All the simulated results were verified experimentally and met the expectations of the presented solution. 1 Introduction European Technology Platform Smart Grid defines the smart grid as 'an electricity network that can intelligently integrate the actions of all users connected to it – generators, consumers and those that do both – in order to efficiently deliver sustainable, economic and secure electricity supplies' [1]. Evolving the traditional electrical grid into a smart grid is an attractive possibility because it will lead to integrating renewable energy sources into the grid, will help the customers to rationalise their power consumption and it will lead to a reduction in the peak energy demand. Furthermore, the advantages of smart grid integration are not limited to financial figures but also reflect on the environment positively [2]. Therefore the evolved grid, smart grid, will be self-monitoring and self-healing with distributed renewable energy generation units in the grid [3]. Converting the traditional grid into a smart grid requires an established two-way communication between the utilities and the customers. For a smart grid, several communication techniques have been proposed such as global system for mobile communications, power line communication (PLC) and digital subscriber line. PLC technology is the preferred solution by utilities, since the infrastructure is available, is independent of third parties, and it is considered to be a secure medium to carry information [4]. Moreover, PLC technology is one of the fastest technologies in conveying the data at low cost and least impact on environment with respect to other technologies [5]. A major challenge when attempting to deploy PLC technology is the transformers in the signal route. A transformer acts like a low-pass filter with cut-off frequency <700 Hz. This attenuates the communication signal because the signal is modulated at higher frequencies in the kilohertz (kHz) and megahertz (MHz) range. This helps to avoid interference with the power signal (50 or 60 Hz) and its harmonics. To overcome signal attenuation at high frequencies, several studies have been conducted. It was suggested to bypass the transformer using an external circuit that allows the signal to propagate around the transformer. This solution requires installing the bypassing circuits at each transformer in the grid. The added installation and operational costs are disadvantages for this approach. Furthermore, the external circuit might reduce the data rate [6, 7]. A radical solution presented as installing electronic smart distribution transformers was proposed. The smart transformer is composed of power electronic devices and microcontroller primarily. Replacing the current distribution transformers with semiconductor smart transformers will achieve the target of having bidirectional communication between utility and customers; nevertheless, the associated cost will be significantly expensive which is a major shortcoming [8, 9]. Another study suggested employing a multicarrier modulation scheme to carry the information, such as orthogonal frequency division multiplexing. The main advantage in this solution is no external by-pass circuitry is required. However, the suggested solution complicates the design of the transmitter and the receiver [10, 11]. It was reported that the transformer filtering can be overcome at certain radio-frequency bands since the transformer acts like a 'window' allowing the communication signal to pass with minimum loss and attenuation at these bands [6]. This concept forms the foundation of this paper with one critical difference. Although in [6], Black investigated to utilise the transformer MHz resonance frequency to transmit their data, in this paper, the kHz will be exploited to transmit the information through the distribution transformer. Transmitting the data in the kHz range has the advantage of avoiding the potential of the signal interference with wireless standards in the MHz and GHz ranges. Reducing the interference will reduce the noise in the system and it will lead to an overall improvement in the system operation. Moreover, the selection of this frequency band is because of the global availability of this band for PLC applications. The transformer is considered as a network of resistors, capacitors and inductors (RLCs). In such a model of the RLC network elements, several resonance frequencies may occur. Primarily, the resonance that occurs at low frequency is because of an interaction between the shunt capacitance and the magnetising inductance of the windings. Resonance also occurs at medium and high frequencies. Beyond the low-frequency range, resonance happens because of an interaction between the air-cored inductances and the series and the shunt capacitors of the windings. Resonance frequency values vary from one transformer to another as the capacitances and the inductances values are function of winding geometry, transformer tap position, internal leads shape and the presence of oil [12, 13]. 2 Methodology Successful transmission of a communication signal through a grid transformer for PLC applications forms the foundation of this paper. The first stage was to measure the resonance frequency of a distribution transformer. To determine the resonance frequency values, the frequency response of the transformer should be obtained. Several methods could be used to obtain the frequency response in this paper, frequency response analysis technique is implemented [14]. It measures the amplitude and phase response of a swept input sinusoid at one terminal of a transformer winding [14]. The captured frequency response was then loaded as a communication channel and simulated in MATLAB. The digitally modulated signals were transmitted through the modelled transformer frequency response (the communication channel) in MATLAB. Finally, the bit error rate (BER) was calculated and compared with the theoretical case, where the channel is impaired only by additive white Gaussian noise (AWGN). The paper involved two experimental setups: frequency response measurement and experimental setup of verifying the simulation results. In the laboratory, the frequency response of the distribution transformer was measured. The experimental setup is composed of an arbitrary function generator (AFG 310 Sony Tektronix), oscilloscope (Tektronix TDS) and single-phase distribution transformer (15 kVA, 20 k/0.2 kV). Fig. 1 shows the setup that was used to measure the frequency response of the distribution transformer. Fig 1Open in figure viewerPowerPoint Laboratory experimental setup-frequency response measurement The setup that has been used in the laboratory to verify the simulation results was composed of Agilent M9331A, Agilent M9210, 15 kVA (20/0.22 kV) distribution transformer and variable attenuators. The experimental setup is shown in Fig. 2. Fig 2Open in figure viewerPowerPoint Experimental setup to verify the results in the laboratory Agilent M9331A module was used to generate the binary phase shift keying (BPSK) signal. Agilent M9331A is an arbitrary waveform generator (AWG) and it is interfaced with MATLAB. An arbitrary sample of BPSK signal is shown in Fig. 3. Fig 3Open in figure viewerPowerPoint Arbitrary BPSK signal sample Cables act as an attenuator element because of the high-frequency nature of the transmitted signal. Since the distance between the transformer secondary side and the customer load is usually short, the cable attenuation effect is not significant. On the contrary, the cable that connects the substation with the transformer primary is usually long and may cause severe attenuation to the signal. In this test, attenuator elements are used to model the cable attenuation effect from utility substation to the customer transformer. The last used component of the experimental setup is the practical verification of the simulation results is the data acquisition unit. To acquire the data, Agilent M9210A module is used. The module has two channels; hence, the transmitted and the received signals after passing the transformer can be acquired. 3 Frequency response measurement The procedure was performed by injecting a single frequency sinusoidal signal the primary windings (20 kV side) while the output signal at the secondary terminal (220 V) was measured. The frequency of the transmitted sinusoidal signal was swept from 1 to 2000 kHz in steps of 10 kHz, whereas the amplitude was maintained at 2 V (peak-to-peak). The transmitted and the received signal amplitudes (peak–peak) were measured and tabulated. From the measured signal levels, the gain was calculated as ratio of the output to the input voltage peaks. Furthermore, the phase delay was measured between the input and the output signals; therefore the frequency response in terms of gain and phase was measured. Fig. 4 shows the overall frequency response in terms of voltage gain. Fig 4Open in figure viewerPowerPoint Frequency response of distribution transformer (gain against frequency) It is evident from Fig. 4 that the resonance phenomenon appears between 350 and 600 kHz. Zoomed view of the gain response and its corresponding phase is depicted in Fig. 5. It can be seen that for the measured transformer response, there is a resonance that occurs approximately at 490 kHz. Furthermore, it can be noted that the gain is almost 2 V/V at resonance. This gain will amplify the communication signal; therefore it is expected that the BER will be lower at resonance frequency. Moreover, the phase at resonance is linear which means that the system will not encounter phase distortion issues. After completing the frequency response measurements, the results were used in MATLAB as a communication channel, as explained in the next section. Fig 5Open in figure viewerPowerPoint Zoomed section of the frequency response at resonance showing gain and phase responses: zoomed section of the frequency response at resonance showing gain and phase responses 4 Simulation results BPSK modulation scheme was selected to modulate the transmitted message because of its simplicity. The transmitted message was generated in MATLAB as a sequence of random BPSK modulated bits with a carrier frequency that was varied between 440 and 550 kHz in steps of 10 kHz. For each carrier frequency, the simulated bit rate was assumed to be 5 kbps initially and was swept to 15 kbps in steps of 2 kbps. White noise is added to the transmitted signal after being filtered with the obtained frequency response of the transformer. To recover the data and evaluate the system response, an ideal matched filter is used. Finally, the BER is calculated and plotted for a given range of signal-to-noise ratio (SNR) using Monte Carlo simulation. Fig. 6 represents the block diagram of the described procedures and illustrates the major blocks in simulating the communication application. Fig 6Open in figure viewerPowerPoint Functional blocks of testing the communication application using the obtained frequency response Fig. 7 shows a sample of the BER results that were obtained by Monte Carlo simulation where the bit rate is 5 kbps. It can be seen that the best performance occurs at the resonance frequency. Fig 7Open in figure viewerPowerPoint Sample of BER results for bit rate 5 kbps As previously discussed, at resonance the frequency response of the tested transformer indicated a gain of 1.796 V/V which will result in increasing the SNR and decreasing the BER. In communication standards, it is required to have BER 2 dB [15]. Sending the modulated message at resonance frequency has achieved superior results compared with all other frequencies. Moreover, the BER curve that is obtained at resonance is better than the theoretical AWGN channel. This is because of the signal amplification, and hence increased SNR at the resonance frequency. The simulation results in Fig. 7 can be visualised by sending a test modulated text message instead of random bits. The text message that is sent and received through the transformer is shown in Fig. 8. The message will be sent seven times and at the receiver side the BER will be computed and the number of errors will be displayed after averaging the received multiple copies of the original message. Fig 8Open in figure viewerPowerPoint Text message that will be sent through the transformer By performing the same previous simulation but sending the message at resonance frequency at 5 kbps, the message is received with BER 2.76 × 10−4 and zero errors because of averaging technique at the receiver side as shown in Fig. 9. Repeating the same simulation, while modulating the message at 400 kHz, resulted in higher BER as depicted in Fig. 10. Fig 9Open in figure viewerPowerPoint Received message in MATLAB workspace along with BER and number of errors in the message modulated at 490 kHz Fig 10Open in figure viewerPowerPoint Received message in MATLAB workspace along with BER and number of errors in the message modulated at 400 kHz The final point to analyse is the effect of varying the bit rate on the BER. It can be seen in Fig. 11 that as the bit rate increases, the BER degrades. Consequently, the best performance of BER was obtained at minimum bit rate (5 kbps). This observation is because of the bandwidth of the resonance frequency. If the data rate is low then it will fit better in the resonance bandwidth while high data rate will require more bandwidth. As data rate spectrum exceeds the resonance bandwidth, the communication signal will be outside the bandwidth and it will suffer from severe attenuation causing degradation in the system performance. Fig 11Open in figure viewerPowerPoint BER curves for different rates for carrier frequency 490 kHz Nevertheless, it is depicted in Fig. 11 that the curves are comparable with each other and they are far better than the theoretical curve. This conclusion allows the system to transmit data at high rate which was not achieved before. Therefore transmitting communication messages at resonance frequency will enhance the system performance in terms of BER. In addition, high data rates can be targeted and implemented if the signal is transmitted at resonance. 5 Factors that influence resonance frequency 5.1 Transformer ageing effect One of the major factors that may affect the resonance frequency is the ageing of the transformer winding. Ageing issue influences several components and parameters in the transformer such as the windings orientation and the elements' dielectric properties. In this paper, the effect of the ageing on the transformer resonance frequency was investigated. To accelerate the ageing of the transformer, the transformer was subjected to several short circuits. This resulted in the flow of high current that might have caused some internal damage in the transformers windings. The tested transformer is dry type transformer; therefore any internal damage that might have occurred inside the transformer is not recoverable. The frequency response of the distribution transformer was measured again following the same procedure highlighted in sweep frequency response analysis (SFRA) measurement. The SFRA test result verifies that there is a shift in the transformer frequency response as shown in Fig. 12. The resonance frequency shifted from 490 to 470 kHz. The bandwidth of the resonance in the transformer stayed almost the same about 40 kHz. Therefore it may be important that the frequency response of the transformers should be investigated regularly to ensure that there is no shift in the resonance frequency. Fig 12Open in figure viewerPowerPoint New and old frequency responses of the transformer 5.2 Distribution cable length effect PLC network method is based on using the power cables and transformers to carry the communication signal from destination to source node. Hence, power cables are vital elements and their behaviour towards the transmitted signals should be evaluated. The frequency responses of the cable at different lengths are simulated using MATLAB at the frequency interval of the studied transformer. In the literature, there are several models to simulate distribution cables at high frequency such as statistical model and deterministic model. Statistical model is a logarithmic normal distribution model, and it is driven mainly by the multipath effect that the signal encounters to travel from one node to another. The multipath nature of the channel arises from the presence of several branches and impedances mismatches that cause multiple reflections [16]. In this model, the captured distortion is because of low-pass behaviour of the cable as well as the reception of multiple versions of the original signal caused by reflections in the cable because of any mismatch in the impedance that occurs over the line [17]. On the other hand, deterministic model represents the cable as 2-port network using transmission line theory as shown in (1) (1)where A = D = cosh(γl), B = Zo sinh(γl), C = (1/Zo)sinh(γl) 'l' is cable length and 'γ' is the propagation constant [m−1] = and z = R + jωL [Ω/m] and y = G + jωC [s/m]; furthermore, this model allows calculating the transfer function of the channel in the form of a ratio between the voltage at the receiver side and the source voltage in terms of the impedances and transmission matrices coefficients, as shown in (2) [16] (2)The main drawback of this theory is that the complexity of building a complete model increases with the number of branching to different customers. Therefore this model is not recommended to model indoor cables and lines. However, the scope of this paper involves one path from the distribution substation to the customer transformer. Consequently, deterministic model is used in this paper to characterise the power cables representing the segment from the substation up to customer node in the described system. The modelled distribution cable parameters are listed in Table 1. Table 1. Parameters of simulated distribution cable Parameter Value DC resistance 0.727 Ω/km inductance 0.524 mH/km capacitance 0.14 µF/km Another parameter that should be considered is the resistance because of skin effect. Cables are designed to deliver signals at 50/60 Hz; hence, the skin effect factor is not significant and usually the AC resistance is not provided by the manufacture. Nevertheless, considering the skin effect is vital as the communication signal used in this paper is modulated at frequencies in the kHz range. AC resistance can be modelled as a real parameter that is frequency dependent and it will be added algebraically to the cable impedance parameter. To estimate this factor, Carson's equation for modelling distribution cables is used [18] (3)Using (2) and (3), the cable was modelled in MATLAB with only two variables: cable length and signal frequency. Different cable lengths have been considered and Fig. 13 demonstrates the cable frequency response for cable lengths of 100 and 1000 m. It is evident from Fig. 13 that as the frequency increases, the signal attenuation increases because the frequency is positively proportional to the cable impedance as explained in (3). Moreover, and similar to the effect of frequency over the impedance, the cable length is positively proportional to the impedance; hence, the longer the cable, the more attenuation will exist. For example, Fig. 13 shows that the encountered attenuation for 1000 m cables is higher than the case of 100 m at same frequency. Fig 13Open in figure viewerPowerPoint Frequency response of a 100 m distribution cable b 1000 m distribution cable It is depicted in Fig. 12, as shown before, that the attenuation factor is about 0.2 V/V and 0.005 V/V at the frequency of interest, resonance frequency, for cable lengths of 100 and 1000 m, respectively. This corresponds to an attenuation of −14 and −46 dB, respectively. The effect of such high attenuation on the BER will be investigated next. 6 Experimental implementation The simulations' result indicated that the best BER performance was obtained at resonance frequency. Moreover, the simulation concluded that the distribution cable acts as an attenuator at high frequencies and the attenuation will increase as the cable length increases. To verify the simulated results, BPSK signals were sent through the distribution transformer and the distribution cables were modelled as attenuator resistors in the laboratory. This connection represents a case scenario where a communication signal is sent from utility substation to the customer load. Random bits modulated at resonance and non-resonance frequencies are modulated in BPSK scheme. The modulated message is created in MATLAB and generated by the AWG Agilent M9331A. The AWG unit is cascaded with attenuation elements representing the cable effect and is connected with the high-voltage side of the transformer (20 kV). On the other side of the transformer, low-voltage side (220 V), the signal is measured by Agilent M9210A module. Fig. 14 shows the blocks of the experiment. It is worth to mention that the experimental test was performed after studying the ageing effect of the transformer. Hence the new frequency response characteristic in Fig. 12 is used. Fig 14Open in figure viewerPowerPoint System blocks of the experiment setup The practical test was implemented in two phases. The main criterion of investigation in these two phases is the effect of the resonance frequency while the bit rate is fixed. In the first phase, a random message of bits was modulated at the resonance frequency and in the second phase, the message was sent at non-resonance frequency. The bandwidth of the transmitted data for both phases was selected to fit within the bandwidth of the resonance frequency. Since the frequency response indicated that the bandwidth is 40 kHz, the message bandwidth was set to be 20 kHz. Initially, the data were transmitted directly through the transformer without any attenuation element. After that, an attenuator element was inserted in the path between the AWG and the transformer to simulate the cable effect. The attenuator value was swept from 10 to 52 dB in steps of 10 dBs and the transmitted signal voltage was maintained to be 0.34 Vpeak. The BER was calculated for all cases including no attenuators case. Table 2 summarises the BER results against the attenuator value while the signal was sent at resonance frequency. Table 2. BER results against the attenuator value for message sent at resonance frequency kHz Attenuator element value BER no attenuator (direct connect) 0 10 dB 0 20 dB 0 30 dB 0 40 dB 6.5 × 10−3 52 dB 2.563 × 10−1 As shown in Table 2, the BER was zero up to 40 dB attenuation. This means that the bit reception is achievable with zero errors as long as the cables are attenuating the signals by factor <40 dB which correspond to a cable length <1000 m. If the cables are providing higher attenuation, then repeaters or amplifiers must be placed in the signal route to overcome the signal attenuation in the cables. To verify the effect of the carrier frequency, the same experiment was performed but at 550 kHz (non-resonance frequency). The transmitted signal voltage amplitude was 0.34 Vpeak and the bit rate was maintained to be about 20 kHz. By testing the same attenuators elements and by calculating the BER for each case, a summary of BER against the attenuators elements for signal propagating through the transformer is shown in Table 3. Table 3. BER results against the attenuator value for message sent at 550 kHz Attenuator element value BER no attenuator (direct connect) 0 10 dB 0 20 dB 0 30 dB 3.48 × 10−2 40 dB 2.796 × 10−1 52 dB 4.570 × 10−1 By comparing the results obtained in Table 3 with the ones in Table 2, it can be concluded that the performance at resonance frequency is better than non-resonance. Sending data at resonance resulted in zero BER for up to 40 dB attenuation factor before the transformer, whereas sending data at non-resonance resulted in zero BER for attenuation values <30 dB. It is expected that modulating data at frequencies that are farther than the resonance will result in higher degradation in BER. The obtained results in this section, matches the expectation from the response in Fig. 12. The response in Fig. 12 indicated that sending data at non-resonance frequency will attenuate the signal significantly and vice versa. These observations were verified in the comparison between Tables 2 and 3. 7 Conclusions Transmitting communication signals through power lines has been investigated. The presented solution validated the feasibility of establishing a bidirectional communication link for smart grid applications. The corner stone of the discussed work was to identify the resonance frequency of the distribution transformer of interest. One of the key methods to explore the frequency response of a transformer is SFRA test. SFRA test provided an overview of the expected response of transformer and identified the resonance frequency and the resonance bandwidth. Sending the messages at the transformer resonance frequency (490 kHz) allows the signals to pass through distribution transformer with apparent voltage gain and minimum attenuation. The main criterion to consider while modulating the bits is to ensure that the bit rate is commensurate with the available resonance bandwidth. It has been found for the investigated transformer that the bandwidth is 40 kHz. The characterised frequency response obtained from SFRA test was modelled in MATLAB. Furthermore, Monte Carlo BER simulations were performed to investigate the effect of the resonance frequency and the resonance bandwidth. Monte Carlo BER simulations and the test text messages verified the work because the best performance was obtained at centre frequency of the resonance bandwidth. Additionally, the work investigated the factors that have the potential to influence transformer resonance frequency such as the distribution cables and the ageing phenomenon of the transformer. Distribution cable was modelled in MATLAB and the frequency responses of different lengths of the modelled cable were simulated. The results indicated that the longer the cable, the more attenuation will be encountered. Moreover, the simulation highlighted the effect of the frequency and it concluded there is a significant attenuation at high frequency because of the influence of the skin effect. The simulation results were compared with experimental work in the laboratory. All the laboratory work matched the simulation results and derived equations. Ageing factor is an important topic in distribution transformers. The ageing phenomenon plays a vital role in determining the resonance frequency of the studied transformer. Ageing influences the stray and physical impedance and admittance of transformer insulation. It was shown in this paper that transformer ageing results in a drift in the resonance frequency. This observation was verified experimentally. 8 References 1' Global Smart Grid Federation': (2012 May 1). [Online]. Available at http://www.globalsmartgridfederation.org/smartgriddef.html 2Tuna, G., Gungor, V.C., Gulez, K.: 'Wireless sensor networks for smart grid applications: a case study on link reliability and node lifetime evaluations in power distribution systems', Int. J. Distrib. Sensor Netw., 2013, 2013, pp. 11, http://doi/org/10.1155/2013/796248, Article ID 796248 3Yigit, M., Gungor, V., Tuna, G., Rangoussi, M., Fadel, E.: 'Power line communication technologies for smart grid applications: a review of advances and challenges', Comput. Netw., 2014, 70, pp. 366– 383 (https://doi/org/10.1016/j.comnet.2014.06.005) 4Haidine, A., Adebisi, B., Treytl, A., Pille, H., Honary, B., Portnoy, A.: ' High-speed narrowband PLC in smart grid landscape – state-of-the-art'. 2011 IEEE Int. Symp. on Power Line Communications and its Applications (ISPLC), 2011 5Farhangi, H.: 'The path of the smart grid', Power Energy Mag., 2010, 8, (1), pp. 18– 28 (https://doi/org/10.1109/MPE.2009.934876) 6Black, W.: ' Data transmission through distribution transformers without bypass components'. 2010 IEEE Int. Symp. on Power Line Communications and its Applications (ISPLC), 2010 7Krishnan, V.: ' Transformer bypass circuit'. 2005 Int. Symp. on Power Line Communications and its Applications, 2005 8Han, B.-M., Choi, N.-S., Lee, J.-Y.: 'New bidirectional intelligent semiconductor transformer for smart grid application', IEEE Trans. Power Electron., 2014, 29, (8), pp. 4058– 4066 (https://doi/org/10.1109/TPEL.2013.2284009) 9Quevedo, J.O., Giacomini, J.C., Beltrame, R.C. et. al.,: ' Smart distribution transformer applied to smart grids'. 2013 Brazilian Power Electronics Conf. (COBEP), 2013, pp. 1046– 1053 10Razazian, K., Umari, M., Kamalizad, A., Loginov, V., Navid, M.: ' G3 plc specification for powerline communication: overview, system simulation and field trial results'. 2010 IEEE Int. Symp. on Power Line Communications and its Applications (ISPLC), 2010 11Razazian, K., Yazdani, J.: ' Utilizing beyond CENELEC standards for smart grid technology'. 2011 Second IEEE PES Int. Conf. and Exhibition on Innovative Smart Grid Technologies (ISGT Europe), 2011 12Sandeep, K., Mukesh, K., Sushil, C.: ' Diagnosis of power transformer through sweep frequency response analysis and comparison methods'. XXXII National Systems Conf., 2008 13Kennedy, G.M., McGrail, J.A., Lapworth, A.J.: 'Transformer sweep frequency response analysis (SFRA)', Energize, 2007, pp. 28– 33 14Secue, E.M.J.R.: 'Sweep frequency response analysis (SFRA) for the assessment of winding displacements and deformation in power transformers', Electr. Power Syst. Res., 2008, 78, (6), pp. 1119– 1128 (https://doi/org/10.1016/j.epsr.2007.08.005) 15' ETSI Standards': (2008 April 1). [Online]. Available at http://www.etsi.org/deliver/etsi_ts/125100_125199/125111/07.01.00_60/ts_125111v070100p.pd 16Galli, S., Banwell, T.C.: 'Deterministic frequency domain model for the indoor power line transfer function', IEEE J. Sel. Areas Commun., 2006, 24, (7), pp. 1304– 1316 (https://doi/org/10.1109/JSAC.2006.874428) 17Galli, S., Scaglione, A., Wang, Z.: 'For the grid and through the grid: the role of power line communications in the smart grid', Proc. IEEE, 2011, 99, (6), pp. 998– 1027 (https://doi/org/10.1109/JPROC.2011.2109670) 18Kersting, W.H.: ' Distribution system modeling and analysis' (CRC Press, FL, 2002) Citing Literature Volume9, Issue6September 2015Pages 774-781 FiguresReferencesRelatedInformation
Referência(s)