The company e-distribuzione S.p.A (the main Italian electricity distributor) has started in 2017 the massive installation of the second-generation smart electrical energy meters. These smart meters have several new features. In particular, the communication functions have been improved to enhance the reliability of the communication towards the concentrator: in addition to the traditional main PLC channel, a radio module has been added with a transmission around the frequency of 169 MHz. This paper assesses the characteristics of the emission due to this radio module, quantifying the total radiated power. The analyses were performed not only inside an anechoic room, to obtain repeatable results, but also in situations close to the normal operating conditions.

1. Introduction

The company e-distribuzione S.p.A. (hereinafter referred to as “ED”) is the main Italian electricity distributor, and it carries out the distribution service in about 7.500 municipalities included in the concession issued by the Ministry of Economic Development in 2003. In these municipalities, ED is also the operator of the measurement activity (Italian Legislative Decree 102/2014), and it is therefore responsible for the installation and maintenance of the meters used to collect the electricity measurement data.

ED has launched the installation of an Advanced Metering Infrastructure (AMI) in 2001, with a project known as Telegestore [1]. This currently working AMI consists of an integrated collection of devices, networks, and computer systems that includes 32 million smart meters at customers’ homes, 380 thousand of data concentrators, and some thousands of meters in medium-voltage/low-voltage (MV/LV) substations fully dedicated to energy service applications (Figure 1). The system is devoted not only to measure and collect electricity metering data, on request or on the basis of predefined schedule, but it is also able to remotely execute commercial operations like activations, deactivations, contractual power limit changes, and bad payer management.

The basic components of Telegestore are the smart meters (Figure 2(a)) installed on each supply node (point of delivery, POD) that communicate in narrow-band power line communication (PLC) with a concentrator that forwards data to a central database system adopting the GSM/GPRS protocol.

In this context, at the beginning of last autumn, ED has started the massive installation of the second-generation smart meters (Enel Open Meter 2G, Figure 2(b)) that will replace the first generation meters at each POD. The second-generation meter is the result of a process that reflects recent developments in the market and in the technology field of metering and remote management. The innovative features of the new smart meter include the availability of customer consumption data every 15 minutes and a dedicated PLC channel to interact with the customer. With this granular timely picture of the daily energy use, traders can develop tailor-made offers and customers, becoming increasingly aware of how they use electricity, and can seize opportunities to achieve greater energy savings. Moreover, they represent a step towards the development of a smart grid [25].

Figure 3 reports the improved communication architecture with some more details. The new meters have two PLC channels: the first (chain 1) is devoted to the communication with the concentrator and the second for communication with in-house devices that could access to metering information through an open protocol (meters and more [6]). To enhance the reliability of the communication towards the concentrator, in addition to the main PLC channel, a radio module, with a transmission frequency at 169 MHz, is used as the backup channel. In fact, radio frequency (RF) communication should only be used when PLC fails, that is, in the case of particularly noisy wired communication (less than 5% of total) and obviously in case of disconnection of the power line (e.g., blackout). The RF communication at 169 MHz is widely used in other applications in the wireless networks of sensors and/or meters (water, gas, electricity, and heat meters [7]). In the considered smart meters, it is applied for the first time in Italy for the electricity metering, and the studies summarized in this paper are aimed to assess the characteristics of the emission spectrum due to the 169 MHz radio module, by quantifying the total radiated power [811].

2. Mathematical Background

The total power radiated (TRP) by an electromagnetic source can be calculated by integrating the intensity of the radiation over the entire solid angle of [8]. Considering the notations shown in Figure 4, it resultswhere is the solid angle element equal to .

It is also possible to define the equivalent isotropic radiated power (EIRP), which is a measure of power density that would be irradiated by an isotropic antenna with the same irradiated power:from which

So, a quantification of the total radiated power can be obtained by measuring the EIRP in different positions.

Moreover, since the measurement must be taken in two polarizations, vertical and horizontal, it is possible to rewrite the equation based on the linear division of the terms of power with the polarization:

Finally, to calculate numerically the TRP from a set of measured values of the EIRP, it is necessary to perform a calculation of the integral by means of an approximation made by a discretization so that integral becomes a sum of values:

In equation (5), the sampled values of the EIRP, in N positions along the axis and in M positions along the axis θ, have to be inserted.

To take into account of the role of the transmitting and receiving antenna, it is possible to refer to the transmission equation of Friis which binds the received and the transmitted power:where is the wavelength, and are the radiation efficiencies, and are the effective areas, and and are the directivities of transmitting and receiving antenna, respectively [8]. In reception, the antenna “transforms” the intensity of the incident electric field into voltage at its terminals that is then converted into power by the spectrum analyzer with a reference impedance of 50 Ω. Therefore, to calculate the electric field value, it is necessary to introduce, as correction factor, a parameter that is called antenna correction or antenna factor (AF), and it depends on the frequency.

The definition of the antenna factor iswhere E is the instantaneous intensity of the electric field incident on the antenna (V/m), V is the voltage value measured on the output connector of the same antenna (V), and AF is the antenna factor (m−1).

We can rewrite equation (7) using the logarithmic amplitudes (dB) and introduce also that summarizes all the losses of the measurement chain between the antenna and the instrument (for example, the loss of cables and other connection accessories):

3. Experimental Measurement Setup

3.1. Smart-Meter Configuration and Timing Measurements

The apparatus used in this paper for radiated performance testing was representative of the typical production units of the 2G meters, and it was configured to use the highest level of power in transmission on the RF channel. To speed up the measurement procedures, the firmware of the meter has been modified so that, as the device was supplied, it automatically and repetitively performs a normal transmission of metering data with a periodicity of about 2 s. This is not the normal use of the radio frequency channel, as already mentioned above, since the 2G meters use this transmission mode only occasionally (in case of disconnection) or in conditions of an anomalous operation of PLC. So, the normal transmission periodicity is too long to be compatible with repetitive measurements and has been reduced. No other changes have been made in the operation mode.

A preliminary laboratory characterization was done to measure the duration of the radio frequency emissions of a normal data transmission and to verify its periodicity. To this aim, a high-accuracy and high-frequency oscilloscope (LeCroy HDO4000 bandwidth 500 MHz, resolution 12-bit, sample rate up to 2.5 GHz) and a directive antenna (SmartAlaris DF-A0047 combined loop/log-periodic dipole array, frequency range 20–500 MHz) were used. During the tests, the 2G meter was fed through a normal electrical outlet and placed on a nonmetallic support base (wooden table). At the same height but at a distance of 3 m, in order to verify the far field conditions, the receiving antenna was placed and connected to the oscilloscope. The alignment and the distances were verified by means of a laser distance meter.

Specific software was implemented on the oscilloscope to perform the demodulation and the detection of the beginning and the ending of the 169 MHz transmission and to save the measured time intervals. With an acquisition time interval particularly long (20 s), it is possible to analyze both the duration of transmission and the time that elapses between two consecutive transmissions (Figure 5). The oscilloscope was left in acquisition for an hour to perform a statistical analysis. The average measured periodicity was 2.1 s with a variance of about 19% instead the average duration of transmission was 141 ms with a variance less than 1% (Figure 6).

3.2. Measurement Method for TRP

The measurement of the TRP from wireless devices is a well-established topic that was faced in numerous scientific papers [814] and in international standards [15], some of them even specifically focused on smart meters [16].

The alternative test procedures can significantly reduce the measurement time, but, of course, they lead to different levels of measurement uncertainty in the obtained results. A comprehensive spherical spatial scan (Figure 7) could be particularly complex from the point of view of the implementation.

For the accuracy level aimed in this paper, a simplified procedure was preferred, and, mainly, the indications included in [16] were followed. The meter was located at a height of 1 m above the reference plane of a semianechoic room, on a 360° rotating table on the horizontal plane. Measurements were taken with discrete increments of 10° with receiving antenna configured for the horizontal polarizations; then all the angular configurations are repeated with the antenna configured for the vertical polarizations. The measurements were repeated over three planes for a total number of measurements equal to 216.

The measurements were carried out in accordance with the provisions of [17], and the TRP is the integral with respect to the spherical coordinates in agreement with equation (7).

3.3. Considerations on Uncertainty

In the execution of electromagnetic field measurements, there are several factors that must be considered for the calculation of uncertainty. In Table 1, there are a list of the main parameters and their typical characteristics of uncertainty that will be considered in the following measurements. This leads to a typical combined uncertainty values of the order of 11% which corresponds to an expanded uncertainty () of the order of 22%.

3.4. Semianechoic Room and Instrumentation

The characterization activities were carried out at the semianechoic chamber at the Electromagnetic Compatibility Laboratory of the Department of Electrical Engineering and Information “Maurizio Scarano” of the University of Cassino and Southern Lazio (Figure 8) whose dimensions are 9.1 m (length) × 5.74 m (width) × 6.24 m (height).

The semianechoic chamber is a normalized test environment shielded with respect to the external environment: the radiated emissions from the equipment under test are not influenced by the locally occurring background electromagnetic noise. In fact, the irradiated disturbances due to radiophonic, radio-television, and radio-mobile broadcasts can be eliminated or strongly attenuated by the shielding so that they do not disturb the immunity test that is carried. The RF shielding test provided the following values for attenuations:(i)Magnetic field 55 dB at 10 kHz, 90 dB at 100 kHz, and 100 dB at 1 MHz(ii)Electric field > 130 dB from 1 MHz to 1 GHz and 100 dB from 1 GHz to 18 GHz(iii)Planar wave > 130 dB from 1 MHz to 1 GHz and 100 dB from 1 GHz to 18 GHz

Moreover, the chamber walls are covered with ferrite tiles and cones with the ability to absorb electromagnetic waves to eliminate the reflected components and have uniformity in the relief mode. Since the floor is not equipped with fixed absorbent cones, the chamber is semianechoic and not fully anechoic.

Finally, the chamber is certified for tests and measurements at 3 meters, according to [1822].

Outside the semianechoic chamber, there is a shielded control room in which the instrumentation used for the emission and immunity measurements are located (Figure 9).

All the operations, in fact, are piloted from the outside because the operators would become sources of disturbance.

The RF shielding test provided the following values for attenuations: 62 dB magnetic field at 10 kHz, 98 dB at 100 kHz, and 124 dB at 1 MHz; electric field > 131 dB from 1 MHz to 1 GHz and 147 dB from 1 GHz to 18 GHz; and planar wave > 132 dB at 500 MHz, 130 dB at 1 GHz, 102 dB at 10 GHz, and 102 dB at 18 GHz.

The instrumentations used for the measurement are as follows: the broadband antenna, type HL-1000 (Figure 8) that is a combined antenna for EMC-measurement purposes composed by logarithmic-periodical and a biconical antenna (frequency range 20–1000 MHz; nominal impedance 50 Ω), and spectrum analyzer Rohde & Schwarz R&S FSH8-2005 (frequency range 100 kHz to 8 GHz; frequency resolution 1 Hz; absolute level uncertainty at 100 MHz < 0.3 dB) (Figure 10). The configuration used for the spectrum analyzer is reported in Table 2. All the used instruments are within the validity period of the calibration certificate.

The detected environmental conditions were as follows: temperature of 24 C and relative humidity of 48%.

In the illustrated condition at 169 MHz, the estimated uncertainty interval, considering a 95% level of confidence, was ±4.2 dB.

3.5. Measurement Setup and Execution of Tests

For the tests, the 2G meter was placed on a nonmetallic support base (a wooden table). In order to verify the conditions of the far field, the receiving antenna was placed at the same height but at a distance of 3 m and then connected to the spectrum analyzer. Alignment and distances were verified by means of the laser distance meter.

The first evaluations were made with the insertion of ferrite cylinders on the supply line (Figure 11) to filter the electromagnetic emissions due to the power supply cables and then to determine the contribution due only to the 2G meter.

The antenna was configured for vertical polarization, the measurements of the radiated emissions were performed for enough time to stabilize the peak value, and the maximum level was recorded (max hold). The position of the 2G meter was rotated with an increase of 10° counterclockwise, and the measuring operations were repeated.

Position changes and measurements were repeated until a complete revolution was reached by returning the device to its initial position (36 measured values). The antenna was configured for horizontal polarization, and all position changes and measurements were repeated. All the described measurements were repeated over three different planes, [17].

4. Measurement Results

The obtained results in terms of emission diagrams are reported from Figures 1217. The device comprehensively presents a good uniformity of emission. Considering the antenna factor and the cable attenuation, interpolating the measured values, and integrating numerically on the spherical surface (see equation (7)), the obtained total irradiated power was , which corresponds to a value of average power density during transmission . Therefore, it results an intensity of average radiation during transmission and a maximum field strength .

In order to compare the measured emission levels with the limits of exposure to electromagnetic fields prescribed by the standards, the emission intensity should be measured and integrated over a certain time period. Referring to [23, 24], a 6 minutes time interval is here accounted. Moreover, as during normal operation, the time distance between transmissions is much higher than 6 minutes (typically 24 hours), and only one transmission was accounted to occur within this time interval. This implies that a duty cycle of 1/2500 is used for the calculation of the average values, obtaining and . These values are much lower than the action level of for thermal effects (100 kHz to 300 GHz), reported in [23].

The measurements were repeated without the ferrite filters obtaining a remarkable increment in the measured values as reported in Table 3. Nevertheless, also in this case, the values are enormously lower than the action level.

The final activities had the purpose to verify whether the power values emitted by the RF transmission module significantly change during the normal operation of the 2G meter, e.g., during the supply of active or reactive energy.

For these activities, an environment similar to that of normal installation was considered (Measurement Laboratory at Engineering Department of the University of Campania “Luigi Vanvitelli,” Figure 18) [2528].

To the described aim, the 2G meter was fed through the calibrator (Fluke Calibration 6105 A/50 A) to obtain a controlled power supply both in terms of amplitude and of power factor. At the same height but at a distance of 3 m in order to verify the far field conditions, the receiving antenna (SmartAlaris DF-A0047) was placed and connected to the spectrum analyzer RSA5106B (real-time spectrum analyzer, 17 dBm 3rd-order intercept at 2 GHz, ±0.3 dB absolute amplitude accuracy to 3 GHz, and real-time spurious free dynamic range of 80 dB). Alignment and distances have been verified by means of a laser distance meter. All the measurement operations were carried out with ferrites inserted on the power supply line, in order to filter the electromagnetic emissions due to the power supply cables and to measure the effect of the 2G meter only. At first, a reference emission level was measured without current flowing. The current level was increased in 10% steps up to 100% of the nominal value with a power factor equal to 1, thus measuring only the active energy. Deviations in the level of emission were recorded. The operations were repeated with power factor 0.7, so determining the flowing of active and reactive energy at the same time. No significant changes in the emission levels were recorded within accuracy limits.

5. Conclusion

This paper describes the characteristics of the emission at 169 MHz due to the RF module used as a backup channel of the Enel Open Meter 2G, by quantifying the total radiated power. The analyses were performed in an anechoic room, to obtain repeatable results, but also in situations close to normal operating conditions. On the basis of the measurement results, the intensity of the emitted field and the power density associated with the transmission of data are far lower than all the regulatory limits, both in peak values during transmission and even more in average values, being in both cases completely irrelevant in terms of electromagnetic pollution. Even if the coupling of the cables can produce positive interactions, which amplifies the values up to determine the doubling of the emission intensity values, the measured levels remain totally negligible. Moreover, no significant deviation was found in the emission pattern in conditions comparable to normal operating conditions.

Data Availability

The data are available upon request to the authors.

Conflicts of Interest

The authors declare that they have no conflicts of interest.