Method for determining a state of a power cable in a power supply network by means of broadband powerline communication signals
Patent Information
- Application Number
- EP2024701648
- Authority / Receiving Office
- EP · EP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-01-23
- Filing Date
- 2024-01-22
- Publication Date
- 2025-12-03
AI Technical Summary
Conventional methods for determining the condition of power cables in low-voltage and medium-voltage networks are costly, labor-intensive, and disrupt electricity supply, while existing technologies do not allow for continuous monitoring during operation.
A method using broadband powerline communication signals, where signal-to-noise ratio is measured and analyzed over time to determine the slope coefficient, which indicates the cable's condition, allowing for continuous monitoring without supply interruption, utilizing existing BPL modems and correlating data transmission quality with cable condition.
Enables cost-effective, continuous monitoring of power cable condition, allowing network operators to plan maintenance without disrupting service and improving the accuracy of cable condition assessment.
Smart Images

Figure EP2024051411_02082024_PF_FP
Abstract
Description
[0001] Method for determining the condition of a power cable in a power supply network using broadband powerline communication signals
[0002] The invention relates to a method for determining a state of a power cable in a power supply network, in particular in a low-voltage or medium-voltage network, wherein the power supply network has at least two broadband powerline modems and at least one power cable coupled between the broadband powerline modems for transmitting a data signal.
[0003] Traditionally, the condition of low-voltage power cables is not usually determined, as using conventional methods, such as loss factor measurement, would entail significant effort and expense for grid operators, resulting in supply interruptions for connected customers. Low voltage refers to alternating voltages up to 1000 V and direct voltages up to 1500 V. Low-voltage networks are defined as sections of the power grid that distribute electrical energy to the majority of electrical end users or low-voltage devices. Low-voltage networks are preferably operated with a single-phase voltage of 230 V or a three-phase voltage of 400 V.
[0004] At the medium-voltage level, conventional methods are occasionally used to record the cable condition. However, this method can only provide information about the condition for individual cable sections at the current time.
[0005] The devices or methods known from the prior art do not yet allow the aforementioned disadvantages to be overcome and do not provide a way to continuously determine the technical condition of a power cable in a low-voltage or medium-voltage network using Britband powerline communication during operation. Based on this, the object of the invention is to provide a method with which the technical condition of a power cable in a low-voltage or medium-voltage network can be determined simply and cost-effectively, and without interrupting the power supply.
[0006] This problem is solved by the subject matter of patent claim 1. Preferred developments can be found in the subclaims.
[0007] According to the invention, a method is provided for determining the condition of a power cable in a power supply network, in particular in a low-voltage or medium-voltage network, wherein the power supply network has at least two broadband powerline modems and at least one power cable coupled between the broadband powerline modems for transmitting a data signal, comprising the following method steps:
[0008] 51) Determining a signal-to-noise ratio per transmission frequency of the data signal at the broadband powerline modem for a time step,
[0009] 52) Determining a slope of the signal-to-noise ratio over the transmission frequencies for the time step,
[0010] 53) Repeating steps S1) and S2) for several time steps within a predetermined period to obtain several gradients,
[0011] 54) Interpolating the plurality of gradients for the predetermined period and determining a gradient coefficient, wherein the gradient coefficient indicates the change in the gradients from step S2) over the plurality of time steps,
[0012] 55) Providing a cable length and the cable type of the power cable and normalizing the gradient coefficient over the cable length,
[0013] 56) Determining the state of the power cable depending on a predetermined limit value, wherein the state is specified as a function of the time steps by means of the slope coefficient and an output value.
[0014] A key aspect of the invention is therefore that the technical condition of the power cable can be monitored using broadband powerline modems (BPL modems). Broadband over Powerline (BPL) is a method of power line communication (PLC) that enables digital data transmission over public power distribution lines. BPL uses higher frequencies, a wider frequency range, and different technologies than other forms of power line communication to enable high-speed communication over longer distances. BPL uses frequencies that are part of the radio frequency spectrum allocated for wireless communication services. BPL modems are usually already installed in the power grid, so no modifications are required; the necessary data is continuously retrieved, thus saving costs.
[0015] It has been shown that the quality of data transmission correlates with the technical condition of power cables. Therefore, the signal-to-noise ratio (SNR) is measured across the entire transmission frequency range, particularly from 2 to 28 MHz. The change in SNR across the transmission frequencies can be plotted as a gradient in a graph. This gradient, in conjunction with a baseline value and predetermined or individually defined thresholds, can be used to determine the technical condition.
[0016] Since the noise power at each BPL modem is unknown and can vary significantly, it is essential to the invention that trends are analyzed. Therefore, the slope of the SNR is determined across the transmission frequencies for several different time steps. Seasonal effects and impulses can preferably be filtered out using low-pass filtering, such as a moving average. The change in the slope over these time steps, i.e., the change in state, can then also be specified as a slope. The change in state is then determined as the slope or change in the slopes of the SNR over the several time steps. This second slope or change in state is described using a slope coefficient. This slope coefficient is determined over the cable length l Ka b ei standardized. The cable type includes, in particular, paper, PVC, or XLPE insulation.
[0017] Using an initial value and the standardized gradient coefficient, a function can be created that indicates the condition of the power cable as a function of time. This is preferably a linear function of the form f (x) = mx + n. Where f is the technical condition, x is the time, m is the gradient coefficient, and n is the initial value. The initial value is understood to be a predetermined and defined value that represents a new cable in perfect technical condition without wear. Preferably, the initial value is determined in advance according to steps S 1) and S 2) for a starting time of an initial power cable. These initial values differ, in particular, depending on the insulation of the power cable.
[0018] The gradient coefficient is interpolated over the entire operating period t in years. The technical condition of the power cable is preferably described by the following linear function:
[0019] State(t) = initial value
[0020] Preferably, the initial values for paper insulation or PVC insulation include the following values:
[0021] Initial valuep api e r = -7.88 Initial valuepvc = -7.99
[0022] The condition of the power cable can then be determined using predetermined or predefined thresholds. Exceeding these thresholds indicates a critical technical condition. These thresholds can be freely selected depending on the network operator's risk affinity.
[0023] Preferably, the predetermined limits depend on the material properties of the insulation of the power cable. In particular, the limits for paper insulation and PVC insulation include the following values: dB
[0024] Limit value p
[0025] H a a nier > —4.94 • 10-3 - , P ler ' MHz m '
[0026] Limit value p Hv Vc L < —9.58 • 10 -3 MH dB By monitoring long-term SNR trends, the cable condition can be continuously monitored and a "live monitoring system" can be established. This provides added value for a network operator, allowing them to monitor the condition of their cable network during ongoing operations without additional effort and plan renewal measures more effectively than before.
[0027] According to a preferred development of the invention, the slope of the signal-to-noise ratios over the transmission frequencies and / or the change in the slope coefficients over the time steps is determined using linear regression. Linear regression (LR for short) is a special case of regression analysis, i.e., a statistical method that attempts to explain an observed dependent variable using one or more independent variables. Linear regression assumes a linear model (LM for short). Therefore, only relationships are considered in which the dependent variable is a linear combination of the regression coefficients (but not necessarily of the independent variables). It has been shown that the power cable exhibits linear aging; therefore, linear regression is suitable for interpolating the values over a longer period of time and determining the slope.
[0028] According to a preferred embodiment of the invention, the predetermined period covers at least three years. The longer the SNR trend tracked across the transmission frequency, the better the significance of the analysis. It has been shown that meaningful results can be achieved with a minimum period of three years.
[0029] According to a preferred development of the invention, the method comprises the following further method steps:
[0030] S3a) Repeating steps S 1) and S2) for several points in time included in the time step, and
[0031] S3b) Filtering the data from step S3b) by determining a median for a defined period of time and low-pass filtering. The data is filtered using medians and low-pass filtering to minimize the influence of noise. To filter out individual noise pulses that occur in the SNR curve, the median of the determined slope coefficients for a defined period of time is first determined for each time step. This ensures that the median is an "information vector" and not a "noise vector." The time step can, in particular, cover one month. Within this month, several measurements were carried out at different times, for example, 30 days. The several time steps, i.e. the months, add up to the predetermined period of time, for example, one, two, or three years.
[0032] According to a preferred development of the invention, the method comprises the following further method steps:
[0033] S3c) comparing the slope of a first time step and the slope of a second time step following the first time step and determining a deviation, S3d) if the deviation is greater than a predetermined threshold, then defining the second time step as a new starting time for the predetermined period.
[0034] In this way, topology changes can be detected and taken into account. Topology changes are detected by comparing the gradients of successive time steps. If the deviation is too large, for example, greater than 0.39 dB,
[0035] ^j-, a topology change has occurred. If a topology change has been detected, the next steps must be performed separately.
[0036] According to a preferred embodiment of the invention, the condition is classified into the following categories: "action required," "no action required," or "monitor cable." The condition can be provided to an evaluation unit or warning unit so that further action can be initiated in the event of a poor condition.
[0037] According to a preferred development of the invention, the method comprises the following further method steps: S4a) Carrying out steps S1) to S4) for an incoming data signal of a broadband powerline modem, and
[0038] S4b) Determining a total slope coefficient, wherein the total slope coefficient results from the mean value of the slope coefficient of the incoming data signals.
[0039] In this way, both the forward and return connections of a cable route, or the incoming data signals at the broadband powerline modems, can be described and taken into account when determining the slope coefficient. The average of the two slope coefficients of the forward and return connections is calculated and used as the total slope coefficient m. Kabei the cable route.
[0040] To determine this mean value, the method according to a preferred development of the invention comprises the following further method steps:
[0041] S4b') if one of the slope coefficients of one of the incoming data signals is negative and the power cable is covered with paper insulation, then overwriting the respective slope coefficient with the value zero to determine the mean value of the slope coefficients, or
[0042] S4b") if one of the slope coefficients of one of the incoming data signals is positive and the power cable is sheathed with PVC insulation, then overwriting the respective slope coefficient with the value zero to determine the average value of the slope coefficients.
[0043] To determine the mean value of the slope coefficients of the forward and return connections, it is assessed whether the slope coefficients are positive or negative. Since an improvement in the condition of cables is unrealistic and this only indicates an excessive influence of noise, two rules are defined: For paper-insulated cables, the slope coefficient must not be negative. If this is the case, it is set to zero. For PVC-insulated cables, the slope coefficient must not be positive. If this is the case, it is set to zero. The mean value of the two slope coefficients of the forward and return connections is then calculated and used as the overall slope coefficient m Kabei the cable route. The invention is explained in more detail below using a preferred embodiment with reference to the drawings.
[0044] The drawings show
[0045] Fig. 1 schematically shows a method for determining a condition of a power cable according to a preferred embodiment of the invention, and
[0046] Fig. 2 schematically shows a graphical representation for determining the change in state for a period of time.
[0047] Figure 1 schematically shows a method for determining the condition of a power cable in a power supply network, particularly in a low-voltage or medium-voltage network. The power supply network comprises at least two broadband powerline modems and at least one power cable coupled to the broadband powerline modems for transmitting a data signal.
[0048] In a first step S1, an SNR is recorded on the BPL modem for each transmission frequency of the BPL modem's frequency spectrum for one time step. This SNR is plotted against the transmission frequency, and a slope is determined from it (S2).
[0049] The aforementioned steps are repeated for several time steps of a predetermined period S3. To filter the data, the SNR is recorded for each time step at several different points in time S3a, after which the median is determined and filtered using low-pass filtering S3b. A gradient is therefore determined for several points in time, and from this the median is determined, which is used as the gradient for the time step for the subsequent steps. The gradients of the time steps, which each represent the median of the gradients from several points in time, are then interpolated for the predetermined period. The change in the gradient over time is then specified using a gradient coefficient S4. In parallel, the topology check takes place. Topology changes can be detected by comparing the gradients of two consecutive time steps S3c.If the deviation exceeds a predetermined threshold, a topology change has occurred. In this case, the subsequent steps must be performed separately (S3d).
[0050] The aforementioned steps are performed for both the forward and return connections, i.e., for both the incoming data signals on two interconnected broadband powerline modems (S4a). The slope coefficients of the forward path and the slope coefficient of the return path are averaged (S4b), thus providing a total slope coefficient for the subsequent process steps.
[0051] When determining the mean value, two rules are established: either S4b' if one of the slope coefficients of the incoming data signal or the outgoing data signal is negative and the power cable is sheathed with paper insulation, then the respective slope coefficient is overwritten with the value zero, or S4b" if one of the slope coefficients of the incoming data signal or the outgoing data signal is positive and the power cable is sheathed with PVC insulation, then the respective slope coefficient is overwritten with the value zero. In this way, incorrect values that indicate an unrealistic condition improvement are not taken into account when calculating the mean value.
[0052] The provided slope coefficient is then standardized over the cable length S5 and combined with an output value to form a function of the condition of the power cable S6. It has been shown that the condition of the power cable influences the transfer function and thus the SNR. In paper-insulated cables, a deterioration in condition results in a smaller slope of the attenuation constant over the frequency range under consideration. This in turn means that the transfer function also has a smaller slope and, as a result, the SNR in the upper and lower frequency ranges converges. In PVC-insulated cables, the opposite behavior is observed, with the slope of the attenuation constant increasing as the condition deteriorates over the frequency range under consideration. As the condition of PVC-insulated cables deteriorates, the SNR shows an increasingly larger difference between the upper and lower frequency ranges.If the damping constant is determined for trend analysis at several points in time and then plotted over time, the diagram in Figure 2 appears. It shows that the aging or deterioration of the condition decreases linearly. Therefore, a linear regression is performed on the values to interpolate the values over a longer period and determine a gradient. The gradient is referred to as the gradient coefficient m. Kabei specified and the cable length l Kabei standardized so that, together with an initial value representing a new cable in perfect technical condition without wear, the condition of the power cable can be determined using the following formula:
[0053] State(t) = initial value
[0054] The invention underlying this patent application was developed in a project funded by the BMBF under the funding reference 03 SF0568 A (“Sensors in the Network 2.0”).
[0055] List of reference symbols
[0056] 51 Determining a signal-to-noise ratio per transmission frequency for a time step
[0057] 52 Determine a slope of the signal-to-noise ratio over the transmission frequencies for the time step
[0058] 53 Repeat steps S1) and S2) for several time steps
[0059] S3a Repeating steps S1) and S2) for several points in time included in the time step
[0060] S3b Filtering the data
[0061] S3c Comparing the slope of a first time step and the slope of a second time step following the first time step and determining a deviation
[0062] S3d Defining the second time step as a new start time for the predetermined period
[0063] 54 Interpolating the multiple gradients for the predetermined period and determining a gradient coefficient
[0064] S4a Carrying out steps S1) to S4) for the incoming data signals
[0065] S4b Determining an overall slope coefficient
[0066] S4b', S4b“ Overwriting the respective gradient coefficient with the value zero
[0067] 55 Providing a cable length of the power cable and normalizing the gradient coefficient
[0068] 56 Determining the condition of the power cable depending on a predetermined limit value
Claims
Patent claims 1. A method for determining a condition of a power cable in a power supply network, in particular in a low-voltage or a medium-voltage network, wherein the power supply network has at least two broadband powerline modems and at least one power cable coupled between the broadband powerline modems for transmitting a data signal, comprising the following method steps: 51) Determining a signal-to-noise ratio per transmission frequency of the data signal at the broadband powerline modem for a time step, 52) Determining a slope of the signal-to-noise ratio over the transmission frequencies for the time step, 53) Repeating steps S1) and S2) for several time steps within a predetermined period to obtain several gradients, 54) Interpolating the plurality of gradients for the predetermined period and determining a gradient coefficient, wherein the gradient coefficient indicates the change in the gradients from step S2) over the plurality of time steps, 55) Providing a cable length and the cable type of the power cable and normalizing the gradient coefficient over the cable length, 56) Determining the state of the power cable depending on a predetermined limit value, wherein the state is specified as a function of the time steps by means of the slope coefficient and an output value.
2. The method according to claim 1, wherein the output value was determined in advance after steps S1) and S2) for a starting time of an output power cable.
3. The method according to claim 1 or 2, wherein the slope of the signal-to-noise ratios over the transmission frequencies and / or the change in the slope coefficients over the time steps is determined by means of linear regression.
4. Method according to one of the preceding claims, wherein the predetermined limit value is dependent on material properties of an insulation of the power cable.
5. Method according to one of the preceding claims, wherein the predetermined Period covers at least three years.
6. Method according to one of the preceding claims, with the following further method steps: S3a) Repeating steps S 1) and S2) for several points in time included in the time step, and S3b) Filtering the data from step S3b) by determining a median for a defined period and low-pass filtering.
7. Method according to one of the preceding claims, with the following further method steps: S3c) comparing the slope of a first time step and the slope of a second time step following the first time step and determining a deviation, S3d) if the deviation is greater than a predetermined threshold, then defining the second time step as a new starting time for the predetermined period.
8. Method according to one of the preceding claims, wherein the condition is classified into the following categories: "need action" or "no action required" or "watch cable".
9. Method according to one of the preceding claims, with the following further method steps: S4a) Carrying out steps S1) to S4) for an incoming data signal from a broadband powerline modem, and S4b) Determining a total slope coefficient, wherein the total slope coefficient results from the mean value of the slope coefficient of the incoming data signals.
10. The method according to claim 9, wherein the power cable is covered with a paper insulation or with a PVC insulation, with the following further method steps: S4b') if one of the slope coefficients of one of the incoming data signals is negative and the power cable is covered with paper insulation, then overwriting the respective slope coefficient with the value zero to determine the mean value of the slope coefficients, or S4b") if one of the slope coefficients of one of the incoming data signals is positive and the power cable is covered with PVC insulation, then overwriting the respective slope coefficient with the value zero to determine the mean value of the slope coefficients.