Multi-path model of power line channel based on measurement of trunk network transmission characteristics
Through top-down strategies and measurement data, a multi-path model of power line channels is established, which solves the problem of limited description capability within the frequency range in the prior art, and realizes accurate description of complex frequency responses of power line channels and simplified parameters. , suitable for frequency ranges from 500kHz to 20MHz.
Patent Information
- Application Number
- CN202310518036.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-09
- Publication Date
- 2024-11-12
AI Technical Summary
The existing power line communication model has limited ability to describe the transmission characteristics of the power line network in the frequency range, especially in the frequency range above 150kHz, and it is difficult to accurately describe the transmission characteristics of the power line network, and it is necessary to understand the properties of each component in the network in detail.
Using a top-down strategy, a multi-path model of power line channels is described through a small number of parameters, and a frequency response model is derived using measurement data, covering a frequency range of 500kHz to 20MHz. Taking into account the multi-path propagation of the signal and cable loss, simple parameters are used. Express attenuation factors and weighting factors to establish a generalized multi-path signal propagation model.
The accurate description of the complex frequency response of the power line channel is achieved, covering the frequency range from 500kHz to 20MHz, simplifying the model parameters, improving the practicality and applicability of the model, and effectively describing the transmission characteristics of typical power line networks.
Smart Images

Figure CN118944785A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of power line communication, and in particular to a multipath model of a power line channel based on measured trunk network transmission characteristics. Background Art
[0002] Due to recent developments, the electricity supply system is preparing to migrate from a pure energy distribution network to a multi-purpose medium that provides energy, voice and various data services. In particular, Internet access is currently the focus of various research activities.
[0003] Power line networks differ greatly in topology, construction and physical properties from conventional media such as twisted pair, coaxial cable or fiber optic cable. Therefore, PLC systems have to encounter rather hostile characteristics. For computer simulations aimed at proper system design, models of the main network transmission characteristics are very important.
[0004] Existing models are usually of very limited practical value, since most of them represent a bottom-up approach to describing the behavior of a network consisting of a large number of distributed components. Typically, the matrix contains the properties of these components, either based on scattering parameters or on quadrupole impedance and admittance values. Obviously, detailed knowledge of all components in the network including cables, connectors, connected devices, etc. is required. In practice, it is usually impossible to determine the necessary parameters with sufficient accuracy. The problem with existing models is that they are limited to the frequency range below 150kHz.
[0005] The patent outlines a top-down strategy that treats the communication channel as a black box and describes its transmission characteristics through the frequency response of very few relevant parameters in the frequency range of 500kHz to 20MHz. The structure of the model is based on fundamental physical effects analyzed in a large number of measurements. However, in contrast to previous approaches, the relevant parameters are not derived from component properties but from channel measurements. Summary of the invention
[0006] Purpose of the invention: In order to overcome the deficiencies in the prior art, a multipath model of a power line channel based on measured trunk network transmission characteristics is provided, which is an analytical model that describes the complex transmission function of a typical power line network using only a small number of parameters.
[0007] Technical solution: A multipath model of a power line channel based on measured trunk network transmission characteristics, comprising the following steps:
[0008] (1) Analysis of physical signal propagation effects
[0009] This section studies the impact of communication signals on power line network transmission and derives the basics of frequency response models.
[0010] ① Main network topology
[0011] The house service cable ends at a house connection box and then in-door wiring, which can be terminated by a complex terminal impedance Z H(f) Modeling is performed from the perspective of the access network.
[0012] ②Multipath signal propagation
[0013] The propagation of the signal not only occurs along the line-of-sight path between the transmitter and the receiver, but additional paths must also be considered. The result is a multipath scenario with frequency selective fading. Figure 1
[0016] The attenuation A(f, d) caused by the cable loss increases with the length and frequency. The signal components of the individual paths must be combined by superposition. Therefore, the frequency response from A to C can be expressed as
[0017]
[0018] Where: path delay τ i
[0019]
[0020] where ε r is the dielectric constant, c0 is the speed of light, d i is the length of the cable.
[0021] ③ Cable loss
[0022] The propagating signal is subject to attenuation that increases with length and frequency, using the complex propagation constant
[0023]
[0024] The frequency response H(f) of the matched transmission line can be expressed as
[0025] According to the main cable parameters R', G', C' and L' and the voltage at distance x U (x), as shown below:
[0026]
[0027] The cable parameters C' and L' can be roughly estimated from the geometric dimensions and some material properties. Considering the frequencies in the megahertz range, the resistance per unit length R' is mainly determined by the skin effect and is therefore The conductivity per unit length, G′, is mainly affected by the dissipation coefficient of the dielectric material (usually polyvinyl chloride) and is therefore proportional to f. With typical geometry and material properties, we have R′<<ωL′ and G′<<ωC′ in the frequency range of interest. Therefore, the cable can be considered to have a real-valued characteristic impedance, Z L The complex propagation constant γ can be determined using the simplified expression
[0028]
[0029] Using constants k1, k2, k3 and summing up the material and geometric parameters, the real part of the propagation constant, the attenuation factor α, increases with increasing frequency. However, the exact relationship for a particular cable can be proportional to a mixture of α and f, depending on which of k1 and k2 dominates.
[0030] Based on these derivations and extensive studies of the measured frequency response, an approximate formula for the attenuation factor α is obtained:
[0031] α(f)=a0+a1·f k
[0032] It is possible to describe the attenuation of a typical power line link using only three parameters, which can be easily derived from the measured transfer function. The attenuation of a power line cable can be expressed as
[0033]
[0034] Typically, the parameters a0, a1 and k are derived from the measured transfer function.
[0035] (2) Establishing a channel model
[0036] ① A generalized multipath signal propagation model of transfer function
[0037] Combining multipath propagation with the given frequency and length dependent attenuation, we finally get
[0038]
[0039] The above formula describes the propagation of the signal along the path through the delayed part and the low-pass characteristics, that is, the attenuation increases with the attenuation part. The weighting factor g i summarizes the reflection and transmission factors along the propagation path. Since reflection points may exhibit complex and frequency-dependent values, g i Usually complex and frequency dependent. The signal components of N paths are always added together at the receiving point.
[0040] ②Simplified model
[0041] From a multipath perspective, gi simply describes the weight of path i. Thus, using (2), the final version of the frequency response is given, which has been widely demonstrated in practice. Table 1 explains these parameters.
[0042]
[0043] This is a parametric model that describes the complex frequency response of a typical power line channel and covers all substantial effects on the transfer characteristics in the frequency range from 500kHz to 20MHz, which parameters can be derived from the measured frequency response.
[0044] Table 1 Parameters of transfer function model
[0045] i The number of paths, where the index of the path with the shortest delay is i, = 1 <![CDATA[a0,a1]]> Attenuation Parameters k Exponent of the attenuation factor (typical values are between 0.5 and 1) <![CDATA[g i ]]> The weighting factor for path i is generally complex and can be considered as a combination of the reflection and transmission factors involved. <![CDATA[d i ]]> Path length i <![CDATA[τ i ]]> The delay of path i
[0046] (3) Perform measurement verification
[0047] Verification is done by a Figure 1 The test network was completed with a test network of well-known geometric dimensions. The transmitter is located at position A and the receiver is located at position C. The characteristic impedance of the A and C cables is matched, leaving point D open so that the reflection coefficient r=1. Section (1) is 30 meters long, section (2) is 170 meters long and section (3) is 12 meters long. Sections (1) and (2) consist of straight energy distribution cables of type NAYY150 with a characteristic impedance of approximately 45Ω, and section (3) is an indoor connecting cable of type NAYY35 with a characteristic impedance of approximately 70Ω. The transmitted and received signals are recorded with a digital storage oscilloscope (DSO) and the complex frequency response is calculated off-line on a PC. Since the propagation time of the signal results in larger overall phase values, a more detailed plot can be produced by subtracting the linear phase portion. The resulting "phase detail" plot provides a better insight into the frequency range with phase distortion.
[0050] Table 2 Test network four-path model parameters
[0051]
[0052] Additional small ripples are visible in the amplitude response due to the non-ideal matching at A and C. In the impulse response, multiple propagation paths are visible. Figure 2 The measured results are compared with the simulation results for the N = 4 path model. The parameter sets used listed in Table II are from the frequency response measurements.
[0053] (4) Application of research model
[0054] ① Attenuation profile
[0055] The attenuation of power line links is mainly affected by multipath fading caused by cable losses and branch point reflections, so each link has its own attenuation profile. However, for links of similar length, similar layout and similar cable types, they have remarkable similarities.
[0056] 1) Short distance link:
[0057] Short-distance links (100-200 meters) with only a few branches (1 to 4) mostly show typical attenuation values starting from a few dB at 500kHz to 40-70dB at 20MHz. Figure 3 An example is shown of a measurement on a 150m link with a 22m branch supplying a small apartment. The amplitude response is fairly smooth with only some weak notches, meaning there is no significant impedance mismatch at the house connection point.
[0058] In addition to the measurement results, Figure 3 Also shown is the amplitude response calculated from the model with path N = 1 and weighting factor g1 = 1 and attenuation parameters k = 0.7, a0 = -2.03·10 -3 m -1 and a1=3.75·10 -7 s / n. This parameter is determined using a least squares fit algorithm. It is clear that the difference between the model and the measured values is small.
[0061] 2) Long distance links:
[0062] A typical 300m long-distance link with PVC-insulated underground cable exhibits higher attenuation, mainly due to the higher losses caused by the length and the many branches (up to 15). In addition, deep notches caused by multipath fading tend to enhance the attenuation. The attenuation typically starts in the 10-30dB range at 500kHz and may reach values in excess of 80dB at 5-8MHz (this is close to the noise floor of the measurement equipment used). Figure 4 The measurement results for a city residential area are shown, with 12 branches supplying mainly small apartments. In addition to the amplitude response, the noise floor of the link is also shown. It is clear that the attenuated signal drops below the noise floor at frequencies above 7MHz. Furthermore, the amplitude response shows several deep notches.
[0063] remove Figure 4 In addition to the measured data in , the simulation results of the attenuation model established according to (10) are also described, with path N = 1 and parameters k = 1, a0 = 6.5·10 -3 m -1 and a1=2.46·10 -9After preprocessing the measured data, the attenuation parameters are estimated using a least squares fitting algorithm so that the frequency range is not affected by deep notches.
[0066] 3) Length profile:
[0067] Since the position and depth of the grooves do not depend on the length of the link but only on the properties of the branches (length and mismatch), it seems reasonable not to take into account the influence of the grooves when specifying a length profile that only includes basic attenuation characteristics. In order to find such a length distribution, about 160 measurements of links using PVC-insulated cables of type NAYY or NYY in the German low-voltage distribution network were evaluated and divided into five length categories, namely 100 m, 150 m, 200 m, 300 m and 380 m. From these categories, the median value of the attenuation versus frequency was calculated and an attenuation model similar to that in Sections 5.1.1 and 5.1.2 was determined. The results are shown in Figure 5 The corresponding parameters are shown in Table 3. Note that we now have median values for attenuation, so it may be very different if a specific link contains notches, other types of cables, or special topologies.
[0070] Table 3: Attenuation parameters corresponding to length distribution
[0071] Distance [m] <![CDATA[g1]]> <![CDATA[a0[m -1 ]]]> <![CDATA[a1[s / m]]]> k 100 1 <![CDATA[9.40·10 -3 ]]> <![CDATA[4.20·10 -7 ]]> 0.7 150 1 <![CDATA[1.09·10 -3 ]]> <![CDATA[3.36·10 -7 ]]> 0.7 200 1 <![CDATA[9.33·10 -3 ]]> <![CDATA[3.24·10 -7 ]]> 0.7 300 1 <![CDATA[8.40·10 -3 ]]> <![CDATA[3.00·10 -9 ]]> 1 380 1 <![CDATA[6.20·10 -3 ]]> <![CDATA[4.00·10 -9 ]]> 1
[0072] ① Detailed model of individual links
[0073] As an example of a frequency response with significant frequency selective fading, Figure 6 Shown are measurements of a 110 m link recorded from 6 row houses, with each branch being approximately 15 m long. In addition to the amplitude response, phase details are plotted against frequency. The very deep notches are a result of multipath propagation caused by equidistant branches of approximately the same length.
[0074] To demonstrate the model's ability to cover the complex frequency response of links with multiple propagation paths, the model agrees very well with the measurements, especially in the frequency range below the first notch (4.5 MHz). There are only minor differences between simulation and measurement up to 15 MHz, and even above 15 MHz there is still fair coverage. To get accurate results, sometimes many frequency bands with individual weighting factors g may be needed. i and a path of length d i (44 in the case presented). On the other hand, there is only one set of decay parameters, namely k = 1, a0 = 0 and a1 = 2.5·10 -9 s / m, is usually sufficient.
[0077] ② Reference model
[0078] For the performance analysis of PLC systems, this simplified model with only a small number of paths will be discussed.
[0079] Figure 7 A simplified model of a 110m link is plotted. This model has been simplified to N = 15, covering the main paths of the impulse response. Due to the reduced number of paths, the modeled and measured frequency responses show some differences above 4MHz, especially at the location of the deep slots. However, the model still covers the impulse response very well. The parameters of the 15-path model are listed in Table 4.
[0082] Table 4 Parameters of the 15 path models
[0083] BRIEF DESCRIPTION OF THE DRAWINGS
[0084] Figure 1 : Multipath signal propagation;
[0085] Figure 2 : Test network with N=4 paths, measurements and simulations; (a) amplitude response, (b) phase details and (c) impulse response;
[0086] Figure 3 : Attenuation curve of a 150-meter power line link without a deep notch;
[0087] Figure 4 : Attenuation and noise distribution of a 330 m power line link with a deep notch;
[0088] Figure 5 : Length distribution of power line link attenuation ignoring the effect of notches;
[0089] Figure 6 : Detailed model with 44 paths. (a) Amplitude response. (b) Phase details. (c) Impulse response;
[0090] Figure 7 : Reference model with 15 paths: (a) amplitude response, (b) phase details and (c) impulse response. DETAILED DESCRIPTION
[0091] (1) Analysis of physical signal propagation effects
[0092] This section studies the impact of communication signals on power line network transmission and derives the basics of frequency response models.
[0093] ① Main network topology
[0094] The house service cable ends at a house connection box and then the in-door wiring, which can be done through a complex
[0095] Terminal impedance Z H(f) Modeling is performed from the perspective of the access network.
[0096] ②Multipath signal propagation
[0097] The propagation of the signal not only occurs along the line-of-sight path between the transmitter and the receiver, but additional paths must also be considered. The result is a multipath scenario with frequency selective fading. Figure 1
[0100] The attenuation A(f, d) caused by the cable loss increases with the length and frequency. The signal components of the individual paths must be combined by superposition. Therefore, the frequency response from A to C can be expressed as
[0101]
[0102] Where: path delay τ i
[0103]
[0104] where ε r is the dielectric constant, c0 is the speed of light, d i is the length of the cable.
[0105] ③ Cable loss
[0106] The propagating signal is subject to attenuation that increases with length and frequency, using the complex propagation constant
[0107]
[0108] The frequency response H(f) of the matched transmission line can be expressed as
[0109] According to the main cable parameters R', G', C' and L' and the voltage at distance x U (x), as shown below:
[0110]
[0111] The cable parameters C' and L' can be roughly estimated from the geometric dimensions and some material properties. Considering the frequencies in the megahertz range, the resistance per unit length R' is mainly determined by the skin effect and is therefore The conductivity per unit length, G′, is mainly affected by the dissipation coefficient of the dielectric material (usually polyvinyl chloride) and is therefore proportional to f. With typical geometry and material properties, we have R′<<ωL′ and G′<<ωC′ in the frequency range of interest. Therefore, the cable can be considered to have a real-valued characteristic impedance, Z L The complex propagation constant γ can be determined using the simplified expression
[0112]
[0113] Using constants k1, k2, k3 and summing up the material and geometric parameters, the real part of the propagation constant, the attenuation factor α, increases with increasing frequency. However, the exact relationship for a particular cable can be proportional to a mixture of α and f, depending on which of k1 and k2 dominates.
[0114] Based on these derivations and extensive studies of the measured frequency response, an approximate formula for the attenuation factor α is obtained:
[0115] α(f)=a0+a1·f k
[0116] It is possible to describe the attenuation of a typical power line link using only three parameters, which can be easily derived from the measured transfer function. The attenuation of a power line cable can be expressed as
[0117]
[0118] Typically, the parameters a0, a1 and k are derived from the measured transfer function.
[0119] (2) Establishing a channel model
[0120] ① A generalized multipath signal propagation model of transfer function
[0121] Combining multipath propagation with the given frequency and length dependent attenuation, we finally get
[0122]
[0123] The above formula describes the propagation of the signal along the path through the delayed part and the low-pass characteristics, that is, the attenuation increases with the attenuation part. The weighting factor g i summarizes the reflection and transmission factors along the propagation path. Since reflection points may exhibit complex and frequency-dependent values, g i Usually complex and frequency dependent. The signal components of N paths are always added together at the receiving point.
[0124] ②Simplified model
[0125] From a multipath perspective, gi simply describes the weight of path i. Thus, using (2), the final version of the frequency response is given, which has been widely demonstrated in practice. Table 1 explains these parameters.
[0126]
[0127] This is a parametric model that describes the complex frequency response of a typical power line channel and covers all substantial effects on the transfer characteristics in the frequency range from 500kHz to 20MHz, which parameters can be derived from the measured frequency response.
[0128] Table 1 Parameters of transfer function model
[0129] i The number of paths, where the index of the path with the shortest delay is i, = 1 <![CDATA[a0,a1]]> Attenuation Parameters k Exponent of the attenuation factor (typical values are between 0.5 and 1) <![CDATA[g i ]]> The weighting factor for path i is generally complex and can be considered as a combination of the reflection and transmission factors involved. <![CDATA[d i ]]> Path length i <![CDATA[τ i ]]> The delay of path i
[0130] (3) Perform measurement verification
[0131] Verification is done by a Figure 1 The test network was completed with a test network of well-known geometric dimensions. The transmitter is located at position A and the receiver is located at position C. The characteristic impedance of the A and C cables is matched, leaving point D open so that the reflection coefficient r=1. Section (1) is 30 meters long, section (2) is 170 meters long and section (3) is 12 meters long. Sections (1) and (2) consist of straight energy distribution cables of type NAYY150 with a characteristic impedance of approximately 45Ω, and section (3) is an indoor connecting cable of type NAYY35 with a characteristic impedance of approximately 70Ω. The transmitted and received signals are recorded with a digital storage oscilloscope (DSO) and the complex frequency response is calculated off-line on a PC. Since the propagation time of the signal results in larger overall phase values, a more detailed plot can be produced by subtracting the linear phase portion. The resulting "phase detail" plot provides a better insight into the frequency range with phase distortion.
[0134] Table 2 Test network four-path model parameters
[0135]
[0136] Additional small ripples are visible in the amplitude response due to the non-ideal matching at A and C. In the impulse response, multiple propagation paths are visible. Figure 2 The measured results are compared with the simulation results for the N = 4 path model. The parameter sets used listed in Table II are from the frequency response measurements.
[0137] (4) Application of research model
[0138] ① Attenuation profile
[0139] The attenuation of power line links is mainly affected by multipath fading caused by cable losses and branch point reflections, so each link has its own attenuation profile. However, for links of similar length, similar layout and similar cable types, they have remarkable similarities.
[0140] 1) Short distance link:
[0141] Short-distance links (100-200 meters) with only a few branches (1 to 4) mostly show typical attenuation values starting from a few dB at 500kHz to 40-70dB at 20MHz. Figure 3 An example is shown of a measurement on a 150m link with a 22m branch supplying a small apartment. The amplitude response is fairly smooth with only some weak notches, meaning there is no significant impedance mismatch at the house connection point.
[0142] In addition to the measurement results, Figure 3 Also shown is the amplitude response calculated from the model with path N = 1 and weighting factor g1 = 1 and attenuation parameters k = 0.7, a0 = -2.03·10 -3 m -1 and a1=3.75·10 -7 s / n. This parameter is determined using a least squares fit algorithm. It is clear that the difference between the model and the measured values is small.
[0145] 2) Long distance links:
[0146] A typical 300m long-distance link with PVC-insulated underground cable exhibits higher attenuation, mainly due to the higher losses caused by the length and the many branches (up to 15). In addition, deep notches caused by multipath fading tend to enhance the attenuation. The attenuation typically starts in the 10-30dB range at 500kHz and may reach values in excess of 80dB at 5-8MHz (this is close to the noise floor of the measurement equipment used). Figure 4 The measurement results for a city residential area are shown, with 12 branches supplying mainly small apartments. In addition to the amplitude response, the noise floor of the link is also shown. It is clear that the attenuated signal drops below the noise floor at frequencies above 7MHz. Furthermore, the amplitude response shows several deep notches.
[0147] remove Figure 4 In addition to the measured data in , the simulation results of the attenuation model established according to (10) are also described, with path N = 1 and parameters k = 1, a0 = 6.5·10 -3 m -1 and a1=2.46·10 -9After preprocessing the measured data, the attenuation parameters are estimated using a least squares fitting algorithm so that the frequency range is not affected by deep notches.
[0150] 3) Length profile:
[0151] Since the position and depth of the grooves do not depend on the length of the link but only on the properties of the branches (length and mismatch), it seems reasonable not to take into account the influence of the grooves when specifying a length profile that only includes basic attenuation characteristics. In order to find such a length distribution, about 160 measurements of links using PVC-insulated cables of type NAYY or NYY in the German low-voltage distribution network were evaluated and divided into five length categories, namely 100 m, 150 m, 200 m, 300 m and 380 m. From these categories, the median value of the attenuation versus frequency was calculated and an attenuation model similar to that in Sections 5.1.1 and 5.1.2 was determined. The results are shown in Figure 5 The corresponding parameters are shown in Table 3. Note that we now have median values for attenuation, so it may be very different if a specific link contains notches, other types of cables, or special topologies.
[0154] Table 3: Attenuation parameters corresponding to length distribution
[0155] Distance [m] <![CDATA[g1]]> <![CDATA[a0[m -1 ]]]> <![CDATA[a1[s / m]]]> k 100 1 <![CDATA[9.40·10 -3 ]]> <![CDATA[4.20·10 -7 ]]> 0.7 150 1 <![CDATA[1.09·10 -3 ]]> <![CDATA[3.36·10 -7 ]]> 0.7 200 1 <![CDATA[9.33·10 -3 ]]> <![CDATA[3.24·10 -7 ]]> 0.7 300 1 <![CDATA[8.40·10 -3 ]]> <![CDATA[3.00·10 -9 ]]> 1 380 1 <![CDATA[6.20·10 -3 ]]> <![CDATA[4.00·10 -9 ]]> 1
[0156] ② Detailed model of individual links
[0157] As an example of a frequency response with significant frequency selective fading, Figure 6 Shown are measurements of a 110 m link recorded from 6 row houses, with each branch being approximately 15 m long. In addition to the amplitude response, phase details are plotted against frequency. The very deep notches are a result of multipath propagation caused by equidistant branches of approximately the same length.
[0158] To demonstrate the model's ability to cover the complex frequency response of links with multiple propagation paths, the model agrees very well with the measurements, especially in the frequency range below the first notch (4.5 MHz). There are only minor differences between simulation and measurement up to 15 MHz, and even above 15 MHz there is still fair coverage. To get accurate results, sometimes many frequency bands with individual weighting factors g may be needed. i and the length of the path d i (44 in the case presented). On the other hand, there is only one set of decay parameters, namely k = 1, a0 = 0 and a1 = 2.5·10 -9 s / m, is usually sufficient.
[0161] ③ Reference model
[0162] For the performance analysis of PLC systems, this simplified model with only a small number of paths will be discussed.
[0163] Figure 7 A simplified model of a 110m link is plotted. This model has been simplified to N = 15, covering the main paths of the impulse response. Due to the reduced number of paths, the modeled and measured frequency responses show some differences above 4MHz, especially at the location of the deep slots. However, the model still covers the impulse response very well. The parameters of the 15-path model are listed in Table 4.
[0166] Table 4 Parameters of the 15 path models
[0167]
[0168] The present invention derives a complex frequency response model of a PLC link in the frequency range of 500kHz to 20MHz from the physical effects, i.e., multipath signal propagation and typical cable losses. The measured results on a test network with known parameters are in good agreement with the simulation results. In addition, the applicability of the model to real-world networks is demonstrated.
[0169] The model makes it possible to study different network topologies and investigate their impact on the PLC system performance by simulation. In addition to attenuation curves, detailed models can be built for individual links. Furthermore, reference models of typical channels can be defined for standardized PLC system evaluation and performance comparison. This is an extremely important aspect, for example, for upcoming PLC service providers in order to be able to select the best system for their application from a variety of competitors.
Claims
1. A multipath model for power line channels based on measurements of trunk network transmission characteristics, It is characterized in that The steps include: (1) Analysis of physical signal propagation effects This section studies the impact of communication signals on power line network transmission and derives the basics of frequency response models. ① Main network topology The house service cable ends at a house connection box and then in-door wiring, which can be terminated by a complex terminal impedance Z H(f) Modeling is performed from the perspective of the access network. ②Multipath signal propagation The propagation of the signal not only occurs on the line-of-sight path between the transmitter and receiver, but additional paths must also be considered. The result is a multipath scenario with frequency selective fading. The attenuation A(f, d) caused by the cable loss increases with the length and frequency. The signal components of the individual paths must be combined by superposition. Therefore, the frequency response from A to C can be expressed as Where: path delay τ i where ε r is the dielectric constant, c 0 is the speed of light, d i is the length of the cable. ③ Cable loss The propagating signal is subject to attenuation that increases with length and frequency, using the complex propagation constant The frequency response H(f) of the matched transmission line can be expressed as According to the main cable parameters R', G', C' and L' and the voltage at distance x U (x), as shown below: The complex propagation constant γ can be determined using the simplified expression: Using the constant k 1 , k 2 , k 3 The material and geometric parameters are summarized. The real part of the propagation constant, i.e. the attenuation factor α, increases with increasing frequency. However, the exact relationship between a particular cable can be proportional to the mixture of α and f, depending on k 1 and k 2 Which one dominates. Based on these derivations and extensive studies of the measured frequency response, an approximate formula for the attenuation factor α is obtained: α(f)=a 0 +a 1 ·f k It is possible to describe the attenuation of a typical power line link using only three parameters, which can be easily derived from the measured transfer function. The attenuation of a power line cable can be expressed as: Usually, parameter a 0 , a 1 and k are derived from the measured transfer functions. (2) Establishing a channel model ① A generalized multipath signal propagation model of transfer function Combining multipath propagation with the given frequency and length dependent attenuation, we finally get: ②Simplified model From a multipath perspective, g i This simply describes the weight of path i. The final version of the frequency response is: (3) Perform measurement verification The transmitter is located at position A and the receiver is located at position C. The characteristic impedance of the A and C cables is matched so that point D remains open and the reflection coefficient r=1. Section (1) is 30 meters long, section (2) is 170 meters long and section (3) is 12 meters long. Sections (1) and (2) consist of a straight energy distribution cable of type NAYY150 with a characteristic impedance of approximately 45Ω, and section (3) is an indoor connecting cable of type NAYY35 with a characteristic impedance of approximately 70Ω. The transmitted and received signals are recorded with a digital storage oscilloscope (DSO) and the complex frequency response is calculated off-line on a PC. Since the propagation time of the signal results in a larger overall phase value, a more detailed diagram can be produced by subtracting the linear phase portion. (4) Application of research model ① Attenuation profile The attenuation of power line links is mainly affected by multipath fading caused by cable losses and branch point reflections, so each link has its own attenuation profile. However, for links of similar length, similar layout and similar cable types, they have remarkable similarities. 1) Short distance link: Short-distance links (100-200 meters) with only a few branches (1 to 4) mostly show typical attenuation values, starting from a few dB at 500kHz to 40-70dB at 20MHz. Figure 3 shows an example of measurements on a 150m link with a 22m branch supplying a small apartment. The amplitude response is fairly smooth with only some weak notches, which means there is no significant impedance mismatch at the house connection point. In addition to the measured results, Figure 3 also shows the calculated amplitude response for path N = 1 with the weighting factor g 1 = 1 and the attenuation parameters are k = 0.7, a 0 =-2.03·10 -3 m -1 and a 1 =3.75·10 -7 s / n. This parameter is determined using a least squares fit algorithm. It is clear that the difference between the model and the measured values is small. 2) Long distance links: A typical 300 m long link with PVC insulated underground cable exhibits higher attenuation, mainly due to the higher losses caused by the length and the many branches (up to 15). In addition, the deep notches caused by multipath fading tend to enhance the attenuation. The attenuation usually starts in the 10-30 dB range at 500 kHz and can reach values of more than 80 dB at 5-8 MHz (this is close to the noise floor of the measurement equipment used). Figure 4 shows the measurement results of an urban residential area with 12 branches mainly supplying small apartments. In addition to the amplitude response, the noise floor of the link is also shown. It is obvious that the attenuated signal drops below the noise floor at frequencies above 7 MHz. In addition, the amplitude response shows several deep notches. In addition to the measured data in Figure 4, the simulation results of the attenuation model are also described, with path N = 1 and parameters k = 1, a 0 =6.5·10 -3 m -1 and a 1 =2.46·10 -9 After preprocessing the measured data, the attenuation parameters are estimated using a least squares fitting algorithm so that the frequency range is not affected by deep notches. 3) Length profile: Since the position and depth of the grooves do not depend on the length of the link but only on the properties of the branches (length and mismatch), it seems reasonable not to consider the influence of the grooves when specifying a length profile that includes only the basic attenuation characteristics. In order to find such a length profile, 160 measurements were performed on NAYY or NYY insulated cables in the German low-voltage distribution network and divided into five length categories, namely 100 m, 150 m, 200 m, 300 m and 380 m. From these categories, the median values of the attenuation versus frequency were calculated and the attenuation model was determined. ② Detailed model of individual links To demonstrate the model's ability to cover the complex frequency response of links with multiple propagation paths, the model agrees very well with the measurements, especially in the frequency range below the first notch (4.5 MHz). There are only minor differences between simulation and measurement up to 15 MHz, and even above 15 MHz there is still fair coverage. To get accurate results, sometimes many frequency bands with individual weighting factors g may be needed. i and a path of length d i (44 in the case presented). On the other hand, there is only one set of decay parameters, namely k = 1, a 0 =0 and a 1 =2.5·10 -9 s / m, is usually sufficient. ③ Reference model For the performance analysis of PLC systems, this simplified model with only a small number of paths will be discussed. Plot a simplified model of a 110m link. This model has been simplified to N=15, covering the main paths of the impulse response. Due to the reduced number of paths, the model and measured frequency responses show some differences above 4MHz, especially at the location of the deep slots. However, the model still covers the impulse response very well.
Citation Information
Cited By
Modulation method and system for reducing bottom noise of transmitter and improving stability and medium
CN119758258A
Communication equipment wire harness path optimization system based on 5G base station distribution data
CN120499772A
