3. Narrowband Measurements in Domestic Access Networks

Narrowband power line communications (NB PLC) technology for frequency between 9 and 500 kHz has been widely used for automatic metering infrastructure (AMI) [1,2,3,4]. It is well known that NB PLC for frequency 9–500 kHz subjects to hostile channel conditions. Power line channel at low frequencies is characterised by time-selective and frequency-selective interferences, low access impedances and frequency-selective attenuation [2,3,4,5,6,7,8]. Due to negligible cable loss in the low frequencies, the power line channel of access network greatly depends on electrical properties of loads at the network and in particular at customer premises. This dependency is a major source of considerable time variation of power line channel, which makes design of a reliable NB PLC system a challenging task. In China, the influence of electrical appliances in the distribution network and at the customer premises on the performance of NB PLC system is even more evident, partly due to non-strict compliance of electrical appliances with the electromagnetic compatibility (EMC) regulation. To improve the reliability of NB PLC and to enhance its capability to support advanced Smart Grid services, there is still a great need for better understanding of power line channel characteristics in the low frequencies bands.

In China, there are three major types of low-voltage (LV) access network: urban network with high-riser buildings, urban/suburban network with low-riser apartment buildings and rural network with scattered houses. The urban LV access network with high-riser buildings usually uses a dedicated underground power cable for each building, and the cable distance is relatively short. This type of network is considered to be less critical for NB PLC. This chapter addresses power line channel of LV access network in urban/suburban area with low-riser apartment buildings. Noise and attenuation are two major power line channel characteristics. To compare different channels, the so-called link quality index (LQI) is proposed, which combines both noise and attenuation and describes the quality of a particular link.

A proper coupling mode is essential for the performance of NB PLC. Both single-phase and three-phase couplings are possible at the distribution transformer. There are pros and cons between these two approaches. Assisted by measurement results, comparison between these two coupling modes is given.

This chapter is organised as follows. In Section 3.2, an approach to determine the access impedance of single-phase and three-phase coupling is described. In Section 3.3, the three types of LV access network and the network site for the channel measurement are described. Channel measurement results are presented and analysed in Section 3.4.Section 3.5 concludes this chapter.

3.2 Impedance Determination

Access impedance depends on coupling mode. While at the customer side phase to neutral coupling (single-phase coupling) is the default mode, at the transformer, both phase to neutral and three phases to neutral couplings (three-phase coupling) are possible. As will be shown, the access impedance of single-phase coupling at low frequency is already very small. A parallel connection of three phases will further lower the access impedance and hence may lead to additional reduction of effective signal level injected into the grid.

In [7], a voltage/current-based approach is described for the determination of impedance at access point. This approach is adopted here. The measurement set-up is shown in Figure 3.1 with the corresponding coupling network.R_sh is a shunt resistor for the current measurement [7].C denotes the coupling capacitor which has a capacitance of 1 μF. Z_p denotes the parasitic impedance of the entire coupling network, which includes the coupling circuit, coupling leads and fuses. For low frequencies, Z_p can be described by an inductance L and a resistor R with L ≈ 6 μH and R ≈ 0.6 Ω for the used coupling network. Depending on whether a single phase or three phases of the coupling network are actually connected to the power grid, impedance of single-phase coupling, denoted by Z_m,A, Z_m,B, Z_m,C, and of three-phase coupling, denoted by Z_m,ABC, at the reference plane M can be measured. The access impedances at the reference plane L can be determined from, Z_m,A, Z_m,B, Z_m,C, Z_m,ABC, with corresponding calibrations which take into account the coupling network itself. Let Z_l,A, Z_l,B, Z_l,C denote the access impedance of single-phase coupling and Z_l,ABC the access impedance of three-phase coupling at the reference plane L [7]. Then Z_l,A, Z_l,B, Z_l,C can be derived by (X = A, B,C) [7]:

FIGURE 3.1
Impedance measurement set-up.

$Z_{l, X} = Z_{m, X} - (R + j ω L + \frac{1}{j ω C}) .$ $Z_{l, X} = Z_{m, X} - (R + j ω L + \frac{1}{j ω C}) .$

(3.1)

The access impedance Z_l,ABC for the three-phase coupling is defined as

$Z_{l, A B C} : = Z_{m, A B C} - Z_{c a l i b r a t i o n} .$ $Z_{l, A B C} : = Z_{m, A B C} - Z_{c a l i b r a t i o n} .$

(3.2)

In a first approximation, the calibration impedance Z_calibmtion is determined to be the impedance Z_m by a short cut of all three phases of the coupling network against the neutral. Due to mutual magnetic coupling between the conductors of the coupling network, Z_calibmUon is different from $(R + j ω L + 1 / j ω C) / 3$ $(R + j ω L + 1 / j ω C) / 3$ . Given the access impedance of single-phase coupling Z_l,A, Z_l,B, Z_l,C, a theoretical value of the access impedance of the three-phase coupling, denoted by $Z_{l, A B C}^{'}$ $Z_{l, A B C}^{'}$ , can be obtained from the parallel connection of Z_l,A, Z_l,B, Z_l,C as

$Z_{l, A B C}^{'} = \frac{1}{(1 / Z_{l, A} + 1 / Z_{l, B} + 1 / Z_{l, C})} .$ $Z_{l, A B C}^{'} = \frac{1}{(1 / Z_{l, A} + 1 / Z_{l, B} + 1 / Z_{l, C})} .$

(3.3)

$Z_{l, A B C}^{'}$ $Z_{l, A B C}^{'}$ is a theoretical value as it neglects the effect of mutual magnetic coupling between the parallel phases. The effect of the mutual magnetic coupling leads to an equivalent higher inductance at each phase compared to the inductance of single-phase coupling. A higher inductance at each phase will lead to an overall higher magnitude of the access impedance of three-phase coupling. Hence, the magnitude Z_l,ABC is larger than the magnitude of $Z_{l, A B C}^{'}$ $Z_{l, A B C}^{'}$ . In other words, $| Z_{l, A B C}^{'} |$ $| Z_{l, A B C}^{'} |$ is a lower bound of Z_l,ABC.

3.3 LV Network Topology and Measurement Site

The majority of the LV power distribution network in China has a radiation topology. There are three major network topologies.

3.3.1 High-Riser Residential Network

In this network, one transformer supplies power to 1–5 high-riser buildings, with each building containing 10–30 floors. There is a direct underground power cable to each building with a cable distance <150 m. This network can be found typically in big cities.

3.3.2 Low-Riser Residential Network

In this network, one transformer supplies power to dozens of low-riser buildings with less than eight floors of each building. Typically, one underground power cable feeds one or a few buildings. The power line distance from the transformer to the main switch of buildings is usually <300 m. This network is quite common in urban and suburban areas.

3.3.3 Scattered Residential Network

In this network, one transformer supplies power to dozens to hundreds of scattered households. The feed line from the transformer to households is usually cascaded through overhead lines. The maximum power line distance is in the range between 500 and 2000 m. This network topology can be found in rural areas.

3.3.4 Description of the Measurement Site

Channel measurement was carried out in a typical urban residential area in north China.Figure 3.2 illustrates the network with three buildings #1, #2 and #3 chosen for the measurement. Each building is supplied by a dedicated underground power cable. The distance between the transformer and house access point (HAP) of the three buildings is about 50, 250 and 350 m, respectively. Buildings #1 and #2 have three blocks; each has six floors and a meter panel with 12 single-phase meters. Building #3 has one block of eight floors and a meter panel with 16 single-phase meters. The distance between the meter panel and the HAP is about 5–20 m.

FIGURE 3.2
Measurement site.

3.4 Measurement Results and Discussion

3.4.1 Impedance

Impedance Z_m of single-phase and three-phase coupling was measured and the corresponding access impedance Z_l was computed with calibrations (3.1) and (3.2) (see Figure 3.1).

Figure 3.3 shows an example of access impedances (magnitude and phase) at the transformer. It confirms the observations from other measurements and reports [7,8] that the access impedance is mainly inductive and resistive. The magnitude of the access impedances with single-phase coupling grows from 1 Ω at 30 kHz to 9 Ω at 500 kHz. In this example, the difference of access impedance between the different single-phase coupling modes is small. This is likely due to similar wiring structure and line properties of each phase, and there is no dominant and noticeable load at a particular phase. As expected, the magnitude of the access impedance of three-phase coupling Z_l,ABC is smaller than the corresponding value of single-phase coupling. It is below 1 Ω for frequency below 50 kHz and increases to 6 Ω at 500 kHz. For comparison, the theoretical access impedance of three-phase coupling $| Z_{l, A B C}^{'} |$ $| Z_{l, A B C}^{'} |$ determined from Equation 3.3 is also depicted. As it can be seen, $| Z_{l, A B C}^{'} |$ $| Z_{l, A B C}^{'} |$ is noticeably smaller than $| Z_{l, A B C} |$ $| Z_{l, A B C} |$ . As described in Section 3.2, $Z_{l, A B C}^{'}$ $Z_{l, A B C}^{'}$ corresponds to the ideal case which assumes that there is no mutual magnetic coupling between the phases. Hence, $| Z_{l, A B C}^{'} |$ $| Z_{l, A B C}^{'} |$ is a low bound of $| Z_{l, A B C} |$ $| Z_{l, A B C} |$ .

FIGURE 3.3
Impedance at the transformer: (a) magnitude and (b) phase.

Figure 3.4 shows an example of impedances (magnitude and phase) at the HAP of building #2. Here, the access impedance is largely inductive. The magnitude of the impedance grows steadily from a value below 1 Ω at 30 kHz to 9 Ω at 500 kHz. The difference of the access impedance between different phases is slightly larger than the difference at the transformer. Figure 3.5 shows an example of access impedances at a meter panel of building #2, which is about 5 m apart from the HAP. In general, the access impedance at the meter panel shows a similar behaviour as at the HAP. However, since the meter panel is closer to customer premises, special properties of electrical loads may have more visible impact. In this example, a strong local maximum of the magnitude of the access impedance around frequency 40–70 kHz and at phase B is observed. The magnitude has a peak value of 7 Ω for this low frequency range and the phase of the access impedance is changed by >90°. The load becomes for a short frequency range capacitive. This phenomenon may be caused by a parallel resonance circuit of a load at customer premise.

Figure 3.6 shows a collection of magnitude of access impedances of single-phase coupling measured at different HAPs and different phases. It confirms the trend that the magnitude of the impedances increases from small value in the range of 1–2 Ω at low frequency around 30 kHz to about 8–9 Ω at high frequency around 500 kHz. Local maxima may exist at particular phases depending on particular loads at the network. Impedance value may change with the time due to possible change of grid and/or of electrical loads at the network.

Figure 3.7 gives a comparison of the access impedance (magnitude) measured at phase A–N of the transformer but at two different times. One was measured on 16 September 2012 and the other on 8 November 2012, hence a time gap of nearly 2 months. In this example, the magnitude of the impedance is almost the same after 2 months.

3.4.2 Noise

Noise is a major concern for NB PLC. There are different kinds of noises: background noise, periodic impulsive noises, asynchronous impulsive noise, frequency-selective narrowband interferences, etc. (see Chapter 5) [2–4].Figure 3.8 shows examples of noise waveform at phase A of the transformer. The noise waveform was recorded at two different times, one at night around 0:00 (Figure 3.8a) and one at noon time (Figure 3.8b). The alternate current (AC) mains cycle is also shown as a time reference.Figure 3.8a shows that the noise at night is small and there are no considerable impulsive noises, and the noise level at noon (Figure 3.8b) is significantly higher and excessive periodic impulsive noises can be observed. This may be explained by the increasing usage of electrical appliances at customer premises at noon time.

FIGURE 3.4
Impedance at the HAP of building #2: (a) magnitude and (b) phase.

Figure 3.9 shows a noise waveform recorded at the HAP of building #1 at evening 19:03. It also shows a periodic behaviour related to the AC mains cycle and the noise level is at a minimum at AC zero-crossing points. However, in this example, the noise level in general and the peak of impulsive noises in particular are considerably smaller compared to the value measured at the transformer at the noon time.

FIGURE 3.5
Impedance at a meter panel of building #2: (a) magnitude and (b) phase.

Calculation of autocorrelation of noise waveform over 1 s (50 AC mains cycle) shows that the noise has a wide-sense cyclostationary behaviour as reported in [9,10]. The period corresponds to the half AC mains cycle, that is, 10 ms.Figure 3.10 shows the corresponding standard deviation of noise variance over a period of 10 ms based on recorded noise waveform over 1 s. In contrast to the noise at the night, there are significant variations of noise variance at noon. Two maxima can be observed: one is about 5 ms and the other is about 7.5 ms from the AC zero-crossing points, which coincide with the periodic impulsive noises visible at Figure 3.8b. The difference between the maximum and the minimum of the noise variance is nearly 17 dB (20 × log₁₀(0.7 V/0.1 V)). The noise variance at HAP also shows a maximum at 7.8 ms from the AC zero crossing. But it has a much smaller magnitude compared with the maximum value at the transformer.

FIGURE 3.6
Collection of the magnitude of access impedance measured at different HAPs and different phases, and with single-phase coupling.

FIGURE 3.7
Impedance (magnitude) measured at the same place (phase A–N of transformer) but at two different times, one on 16 September 2012 and the other on 8 November 2012.

FIGURE 3.8
Examples of noise waveform at transformer recorded at night 0:20 (a) and at noon 12:00 (b).

The noise power in the frequency domain is captured with a spectrum analyser with logarithmic video signal averaging and with root mean square (RMS) detector. The noise resolution bandwidth is 10 kHz.

FIGURE 3.9
Example of noise waveform at HAP of building #1, recorded at evening 19:03.

Figure 3.11 shows an example of noise power spectrum at the transformer with the single-phase coupling. The noise power decreases with the frequency. The largest decrease is observed for the phase A, where the noise power is reduced from 80 dBμV at 50 kHz to 33 dBμV at 500 kHz. This is a reduction of >45 dB. At the phases B and C, the reduction of the noise power from 50 to 500 kHz is about 30 dB. The difference of noise power spectrum between phase A and the other two phases B and C is between 10 and 15 dB. The shape of the noise power spectrum at different phases is quite similar indicating a coupling effect among phases.

FIGURE 3.10
Standard deviation of noise variance over half power frequency cycle.

FIGURE 3.11
Noise power spectra at the transformer.

Figure 3.12 shows a collection of noise power spectrums captured at different HAPs and at different phases and with different coupling modes. It provides an impression of variability of noise. It confirms that the noise at low frequencies is significantly stronger than the noise at higher frequencies. The most noise powers are within 50–80 dBμV between 50 and 100 kHz and within 30–60 dBμV between 400 and 500 kHz. The differences can be as large as 30–40 dB. NB PLC operating at low frequency range such at CENELEC A band may face considerably stronger noise.

FIGURE 3.12
Collection of noise power spectrum at different HAPs, different phases and with different coupling modes.

3.4.3 Attenuation

Attenuation is measured between the transformer and HAP of the three buildings. A sweep generator is used. The transmitter is able to inject high signal level for impedances with a magnitude below 1 Ω. The attenuation includes the coupling at transmitter and receiver.

Figure 3.13 shows an example of attenuation between the transformer and the HAPs of buildings #1, #2 and #3. Three-phase coupling is used at the transformer. The receiver uses single-phase coupling between phase B and neutral N. The attenuation varies between 30 and 70 dB. It is well known that unlike for broadband power line (BPL), for NB PLC at LV access network, there is not necessarily a strong correlation between the attenuation and frequencies, nor between the attenuation and distances. This can be observed in Figure 3.13. Among the three links, the longest link with 350 m has the smallest attenuation between 150 and 250 kHz, while the shortest link with 50 m only has the smallest attenuation for frequency above 425 kHz.

Three-phase coupling at the transformer is widely used, as this coupling may have better chance to cover most PLC nodes at customers directly. However, three-phase coupling has two effects. First, the parallel connection of three phases will result in smaller access impedance, which makes an efficient injection of signal difficult, and second, the signal will further split into the three phases. These together may lead to a remarkable reduction of the signal level at the receiver compared with the case when the signal is only injected into one and the same phase as the receiver.

FIGURE 3.13
Attenuation for the three downlinks.

FIGURE 3.14
Attenuation from the transformer to the HAP of building #2 with single-phase and three-phase coupling at the transformer.

As an example, Figure 3.14 compares attenuation from the transformer to the HAP of building #2 between single-and three-phase coupling at the transformer. For the singlephase coupling, the same phase is used at both sides. In this example, the attenuation with the single-phase coupling is considerably lower than the attenuation with the three-phase coupling. The difference of attenuation value at 150 kHz is nearly 30 dB. Obviously, single-phase coupling is a preferred coupling mode for PLC nodes at the same phase. There are different ways to connect PLC nodes at different phases. An expensive one is to use three transmitters to inject the same signal into the three phases simultaneously. Another possibility is via crosstalk. Many papers, for example, [7,8] report considerable crosstalk between different phases in the considered frequency band. Due to this crosstalk, a PLC node may still be able to directly receive a signal with a sufficient level which is sent at a different phase. It has also been observed that there is significant crosstalk between different phases at the same meter panel. So, PLC nodes coupled at the same phase as the transmitter at the transformer may act as repeater for PLC nodes which are coupled at different phases.

Figure 3.15 shows a collection of attenuation, which are measured between the transformer and different HAPs and with different coupling modes. It provides an impression of variability of attenuations. The attenuation has a large spread which is between 28 and 70 dB. The most attenuation values are in the range 35–60 dB. For the frequency above 250 kHz, the attenuation increases slightly with the frequency.

3.4.4 Link Quality Index

To evaluate power line channel quality and to consider the impact of two important channel characteristics, namely, the noise and the attenuation, the so-called link quality index LQI(f,t) is introduced. LQI(f,t) is defined as the sum of the noise and the attenuation (loss) as follows:

FIGURE 3.15
Collection of attenuations between the transformer and different HAPs.

$LQI (f, t) : = Noise (f, t) + loss (f, t),$ $LQI (f, t) : = Noise (f, t) + loss (f, t),$

(3.4)

where

Loss(f,t) is given as a positive attenuation in dB

Noise(f,t) is the noise power given, for example, in dBμV or dBm measured at frequency f and time t and for a given equivalent noise bandwidth

Hence, LQI(f,t) has the same dimension as Noise(f,t). LQI(f,t) of a link is equivalent to the signal power to be sent at the transmitter to get an equivalent 0 dB signal-to-noise ratio (SNR) at the receiver. Given LQI(f,t) of a link and a signal power at the transmitter Tx(f,t) in dBμV or dBm, the corresponding SNR at the receiver of the link is simply given by

$S N R (f, t) = Tx (f, t) - LQI (f, t) .$ $S N R (f, t) = Tx (f, t) - LQI (f, t) .$

(3.5)

Obviously, the smaller LQI(f,t) is, the less signal power is needed for a required SNR.

Figure 3.16 shows example of LQI for the three downlinks from the transformer to the HAPs of the three buildings. The noise bandwidth is 10 kHz. Three-phase coupling is used at the transformer. The downlink to the building #2 has a minimum LQI value of 85 dBμV at 375 kHz and a maximum LQI value of 109 dBμV at 300 kHz. The downlink to the building #3 has, overall, the highest LQI value and hence requires, on average, the highest signal power than the other two downlinks to obtain the same SNR. Interesting to note is that this link has a quite large distance, which is 350 m. Nevertheless, the smallest LQI of that link is not measured at low frequencies, but rather at higher frequencies between 350 and 500 kHz.

FIGURE 3.16
LQI for the three downlinks between the transformer and the three buildings (HAPs).

Figure 3.17 shows a collection of LQI from different measurements. The noise bandwidth is 10 kHz. The smallest LQI value is about 80 dBμV and the largest LQI value is about 137 dBμV. Most links have an LQI value between 85 and 115 dBμV. A weak tendency can be observed that the LQI decreases with the frequency.

FIGURE 3.17
Collection of LQI from different measurements.

3.4.5 SNR Examples for NB PLC in CENELEC A Band

It is interesting to know what SNR will be available for the LQI values presented in Figure 3.17 if CENELEC A band signal is employed which uses the signal level defined in EN50065-1 [11].

For this purpose, an orthogonal frequency division multiplex (OFDM) signal between 40 and 90 kHz with a peak-to-RMS ratio of 8 dB is assumed. For wideband signal (>5 kHz), EN50065-1 specifies a limit of the peak signal of 140 dBμV for three-phase coupling measured with an artificial mains network (AMN) which has an input impedance between 5 and 20 Ω [11]. This means an RMS limit of 132 dBμV for the assumed OFDM signal. Since the LQI value in Figure 3.17 includes the coupling network in Figure 3.1, to determine available SNR, the RMS of transmitter signal at the input of the coupling network needs to be derived. Since the magnitude of the impedance of the coupling network for frequency between 40 and 90 kHz is in the range of 0.3–1.3 Ω, which is very small compared to the input impedance of 5–20 Ω as defined in [11], the signal loss due to the coupling network can be neglected. Hence, the OFDM signal has approximately an RMS of 132 dBμV between 40 and 90 kHz at the input of the coupling network. The available SNR for LQI from Figure 3.17 with 10 kHz equivalent noise bandwidth can be calculated as

$SNR (f) = 132 dBμV - {10log}_{10} (50 kHz/10 kHz) - LQI (f) = 125 dBμV - LQI (f) .$ $SNR (f) = 132 dBμV - {10log}_{10} (50 kHz/10 kHz) - LQI (f) = 125 dBμV - LQI (f) .$

(3.6)

The SNR value of Equation 3.6 is depicted in Figure 3.18 which shows that in this case, the most available SNR values are between 10 and 40 dB. It can also be observed that the SNR of a particular link may have a large variation with the frequency. The lowest SNR values are found between 60 and 75 kHz with some SNR values far below 0 dB.

FIGURE 3.18
Available SNR for various links with LQI value in Figure 3.17 and for CENELEC A band OFDM signal between 40 and 90 kHz and with a maximum permitted signal limit according to [11].

Actual PLC systems may not exploit the permitted maximum signal limit.Power line intelligent metering evolution (PRIME) specifies OFDM signal between 42 and 89 kHz with a minimum RMS of transmitter signal of 114 dBμV for the three-phase coupling and for the input impedance of 1 Ω [12]. This impedance value is comparable with the value of the coupling network used for the measurement. Hence, for a first approximation, a 6 dB signal loss due to the coupling network is assumed, which leads to an equivalent RMS of transmitter signal of 120 dBμV (114 dBμV + 6 dB) at the input of the coupling network. In this case, PRIME signal will have ∼12 dB less SNR compared with the earlier described OFDM scheme which transmits at maximum permitted signal level [11]. An overall reduction of SNR values in Figure 3.18 by 12 dB means a considerable deterioration of link reliability for PRIME for the various links with LQI value in Figure 3.18.

3.5 Conclusion

Narrowband power line channel characteristics in typical underground LV access network in China are evaluated based on measurement results. A three-phase coupling at transformer has very low impedance. This low impedance together with a signal splitting into the three phases may result in considerable signal loss compared to a single-phase coupling. The attenuation has a broad range between 20 and 70 dB. In contrast to BPL, for NB PLC there is not always a clear trend for the attenuation to increase with the frequency and with the distance. On the other side, the noise power decreases with the frequency significantly. Measurements show a reduction of average noise power of 10 kHz noise band-width from a range between 50 and 90 dBμV at 50 kHz to a range between 30 and 60 dBμV at 500 kHz. A considerable difference of noise level between night and day is observable. The noise also exhibits cyclostationary behaviour with a period which corresponds to the half cycle of AC mains. The link quality index (LQI), taking into account both noise and attenuation proved, is a possible measure to compare different channels. An LQI analysis shows that for a given link, there is a strong variation of LQI values along the frequency axis. On average, there is a weak trend in favour of higher frequencies.

Acknowledgements

This work was funded by the project ‘The New Generation Smart PLC Key Technologies Research’ of State Grid Corporation of China (SGCC). The authors wish to thank Prof. Dostert and his assistants for the support in measurement tools.

References

1. Liu, J., Zhao, B., Geng, L. et al., Current situations and future developments of PLC technology in China, Proceedings of IEEE International Symposium Power Line Communications and Its Applications, Beijing, China, 28–30 March 2012.

2. Galli, S., Scagllone, A. and Wang, Z., For the grid and through the grid: The role of power line communications in the smart grids, Proceedings of the IEEE, 99(6), 998–1027, June 2011.

3. Dostert, K., Power Line Communications. Prentice Hall, Upper Saddle River, NJ, 2001.

4. Ferreira, H. C., Lampe, L., Newbury, J. et al., Power Line Communications. John Wiley & Sons, West Sussex, UK, 2010.

5. Sigle, M., Bauer, M., Liu, W. et al., Transmission channel properties of the low voltage grid for narrowband power line communication, Proceedings of IEEE International Symposium Power Line Communications and Its Applications, Udine, Italy, 3–6 April 2011, pp. 289–294.

6. Varadarajan, B., Kim, I., Dabak, A. et al., Empirical measurements of the low-frequency power-line communications channel in Rural North America, Proceedings of IEEE International Symposium Power Line Communications and Its Applications, Udine, Italy, 3–6 April 2011, pp. 463–467.

7. Sigle, M., Liu, W. and Dostert, K., On the impedance of the low-voltage distribution grid at frequencies up to 500 kHz, Proceedings of IEEE International Symposium Power Line Communications and Its Applications, Beijing, China, 28–30 March 2012.

8. Arzberger, M., Dostert, K., Waldeck, T. et al., Fundamental properties of the low voltage power distribution grid, Proceedings of IEEE International Symposium Power Line Communications and Its Applications, Essen, Germany, 2–4 April 1997, pp. 45–50.

9. Katayama, M., Yamazato, T. and Okada, T., A mathematic model of noise in narrowband power line communications systems, IEEE Journal on Selected Areas in Communications, 24(7), 1267–1276, July 2006.

10. Katayama, M., Itou, S., Yamazato, T. et al., Modeling of cyclostationary and frequency dependent power-line channels for communications, Proceedings of IEEE International Symposium Power Line Communications and Its Applications, Ireland, April 2000, pp. 123–127.

11. CENELEC, EN 50065-1 Signaling on low-voltage electrical installations in the frequency range 3 kHz to 148.5 kHz – Part 1: General requirements, frequency bands and electromagnetic disturbances, April 2011.

12. PRIME Alliance, Power Line Intelligent Metering Evolution (PRIME) Specification, v1.3E, 2010.

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.

Table of Contents for 3. Narrowband Measurements in Domestic Access Networks

Create new playlist

Sign In

Sign Up

Table of Contents for
3. Narrowband Measurements in Domestic Access Networks