Fault monitoring method and system applied to digital power distribution network

Abstract

The invention relates to the technical field of power distribution network fault monitoring, in particular to a fault monitoring method and system applied to a digital power distribution network. The method comprises the following steps: firstly, dividing each fault time period from a to-be-filtered time period; according to the difference condition of the fault responsivity of all the monitoring points in each fault time period, obtaining the noise measurement of the digital power distribution network in the to-be-filtered time period; adjusting the smoothing coefficient according to the noise metric, and obtaining the adjusted smoothing coefficient of the digital power distribution network in the to-be-filtered time period; performing filtering processing on the data set in the to-be-filtered time period by using the adjusted smoothing coefficient to obtain a filtered data set; and performing fault detection based on the filtered data set. According to the method, the adjusted smoothing coefficient of the digital power distribution network in the to-be-filtered time period is reasonably set, so that the noise reduction effect is improved, and accurate fault monitoring on the digital power distribution network is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network fault monitoring, and in particular to a fault monitoring method and system applied to a digital distribution network. Background Art

[0002] The digital distribution network is an important part of the power system. It transmits electricity to terminal appliances or user equipment. In the distribution network, electric energy enters the transformer station from the transmission network. The transformer converts high voltage electricity into low voltage electricity, and then transmits it to the user equipment through high voltage cables, wires, cable troughs, etc. In order to timely discover potential fault hazards in the distribution network, it is necessary to monitor the operating status of the monitoring points in the digital distribution network. Current and voltage values ​​are important power parameters that reflect the operating status of the distribution network. Through the changes in the power parameters of the monitoring points, abnormal conditions in the distribution network can be discovered in time, such as overload, short circuit, grounding and other faults.

[0003] In digital distribution networks, the power parameters collected by monitoring points are often interfered by noise, resulting in inaccurate data. In order to obtain more accurate data, the collected current and voltage values ​​need to be filtered. The exponentially weighted moving average method is a common data filtering algorithm. The prior art filters the data by the exponentially weighted moving average method. The prior art sets a fixed smoothing coefficient based on empirical values, and does not fully consider the different degrees of noise interference in the distribution network in different time periods, resulting in poor noise reduction effect, making it difficult to ensure accurate fault monitoring of the digital distribution network. Summary of the invention

[0004] In order to solve the problem that the prior art has poor data noise reduction effect and is difficult to ensure accurate fault monitoring of digital distribution networks, the purpose of the present invention is to provide a fault monitoring method and system for digital distribution networks. The technical solutions adopted are as follows:

[0005] A fault monitoring method applied to a digital distribution network, the method comprising:

[0006] Acquire a data set of the digital distribution network in the filtering time period; the data set includes current value data and voltage value data of each monitoring point;

[0007] According to the fluctuation of the current value and voltage value corresponding to the monitoring point at the sampling time with the timing, and the correlation of the current value data and voltage value data corresponding to the preset reference time period of the monitoring point at the sampling time, the fault responsiveness of the monitoring point at the sampling time is obtained; according to the fault responsiveness of all monitoring points at each sampling time in the time period to be filtered, each fault time period is divided from the time period to be filtered; according to the difference of the fault responsiveness of all monitoring points in each fault time period, the noise measurement of the digital distribution network in the time period to be filtered is obtained;

[0008] The smoothing coefficient is adjusted according to the noise measurement to obtain the adjusted smoothing coefficient of the digital distribution network in the filtering time period; the data set in the filtering time period is filtered using the adjusted smoothing coefficient to obtain the filtered data set; and fault detection is performed based on the filtered data set.

[0009] Furthermore, the method for obtaining the fault responsiveness includes:

[0010] For any monitoring point, the current value data and voltage value data of the monitoring point are used as the power parameter data of the monitoring point; according to the average of the upper envelope and the lower envelope of the power parameter data, the baseline data corresponding to the power parameter data is obtained; the baseline data corresponding to the current value data is used as the current baseline data; the baseline data corresponding to the voltage value data is used as the voltage baseline data;

[0011] Obtaining a first fault indicator at the sampling time according to a difference between the sampling time and the baseline data corresponding to the preset surrounding time period;

[0012] According to the difference between the sampling moment and the baseline data corresponding to the previous sampling moment, a second fault indicator at the sampling moment is obtained;

[0013] Obtaining a third fault indicator at the sampling moment according to the absolute value of the Pearson correlation coefficient of the current baseline data and the voltage baseline data corresponding to the preset reference time period of the monitoring point at the sampling moment;

[0014] The first fault indicator, the second fault indicator and the third fault indicator are forwardly integrated to obtain the fault responsiveness of the monitoring point at the sampling time.

[0015] Furthermore, the method for obtaining the first fault indicator includes:

[0016] For any baseline data corresponding to a power parameter data, the mean of all baseline data is calculated within a preset time period around the sampling time to obtain the overall value around the sampling time; the absolute value of the difference between the baseline data at the sampling time and the overall value around the sampling time is calculated to obtain the local difference value of the baseline data at the sampling time;

[0017] The average of the local difference values ​​of the current baseline data and the voltage baseline data at the sampling time is calculated to obtain the first fault indicator at the sampling time.

[0018] Furthermore, the method for obtaining the second fault indicator includes:

[0019] The second fault indicator is obtained according to the second fault indicator acquisition formula, and the second fault indicator acquisition formula includes:

[0020] Among them, Y iis the second fault indicator at sampling time i; ΔU i is the difference value of the voltage baseline data at sampling time i; ΔI i is the difference value of the current baseline data at sampling time i; i-1 is the sampling time before sampling time i; ΔU i-1 is the difference value of the voltage baseline data at sampling time i-1; ΔI i-1 is the differential value of the current baseline data at the sampling time i-1; || is the absolute value symbol; exp() is an exponential function with the natural number e as the base; ε is the preset denominator adjustment value.

[0021] Furthermore, the method for obtaining the noise metric includes:

[0022] The maximum value of the fault response intensity of the monitoring point in the fault time period is used as the fault value to be analyzed of the monitoring point in the fault time period; for any fault time period, all the fault values ​​to be analyzed are sorted in descending order according to the fault values ​​to be analyzed of the monitoring point to obtain the fault monitoring point serial number of the fault value to be analyzed; all the fault values ​​to be analyzed are counted in order according to the order of the fault monitoring point serial numbers to obtain the fault value sequence to be analyzed in the fault time period;

[0023] In the sequence of fault values ​​to be analyzed in all fault time periods, according to the outlier situation of the fault value to be analyzed, the attention weight of the fault value to be analyzed is obtained;

[0024] According to the differences of the fault value sequences to be analyzed in all fault time periods, the time period weight of the fault time period is obtained;

[0025] Forward fusion of the attention weight of the fault value to be analyzed and the time period weight of the fault time period to which it belongs, to obtain the target weight of the fault value to be analyzed;

[0026] Obtaining a target fault value sequence of the digital distribution network according to target weights of all fault values ​​to be analyzed in the fault value sequence to be analyzed in all fault time periods;

[0027] According to the DTW distance between the target fault value sequence of the digital distribution network and the preset standard fault attenuation sequence, the noise measurement of the digital distribution network in the filtering time period is obtained.

[0028] Furthermore, the method for obtaining the attention weight includes:

[0029] Based on a two-dimensional coordinate system, the horizontal axis of the two-dimensional coordinate system is the fault monitoring point number, and the vertical axis of the two-dimensional coordinate system is the fault value to be analyzed. All fault values ​​to be analyzed are mapped into the two-dimensional coordinate system; using the LOF algorithm, the outlier factor of the fault value to be analyzed in the two-dimensional coordinate system is calculated and negative correlation mapping is performed to obtain the attention weight of the fault value to be analyzed.

[0030] Furthermore, the method for obtaining the time period weight includes:

[0031] Calculate the mean of all the fault values ​​to be analyzed corresponding to the same fault monitoring point number to obtain the reference fault value of the fault monitoring point number; count all the reference fault values ​​in sequence according to the order of the fault monitoring point numbers to obtain the reference fault value sequence of the fault time period;

[0032] The DTW distance between the fault value sequence to be analyzed and the reference fault value sequence in each fault time period is calculated and negative correlation mapping is performed to obtain the time period weight of each fault time period.

[0033] Furthermore, the method for obtaining the target fault value sequence includes:

[0034] For all the fault values ​​to be analyzed corresponding to the same fault monitoring point number, the target weight is used to perform weighted averaging on all the fault values ​​to be analyzed to obtain the target fault value corresponding to each fault monitoring point number; all the target fault values ​​are counted in sequence according to the order of the fault monitoring point numbers to obtain the target fault value sequence of the digital distribution network.

[0035] Furthermore, the method for obtaining the adjusted smoothing coefficient includes:

[0036] The product of the negative correlation mapping result of the noise metric and the preset adjustment index is calculated to obtain the index to be adjusted; the product of the index to be adjusted and the smoothing coefficient is calculated to obtain the adjusted smoothing coefficient.

[0037] The present invention provides a fault monitoring system applied to a digital distribution network, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, steps of a fault monitoring method applied to a digital distribution network are implemented.

[0038] The present invention has the following beneficial effects:

[0039] In order to reasonably filter the data set in the filtering time period, it is first necessary to analyze the noise performance of the monitoring point in the filtering time period. Considering that the fault signal in the distribution network will attenuate with the increase of the transmission distance, the noise will interfere with the transmission and detection of the fault signal, resulting in abnormal attenuation of the fault signal. First, the fault responsiveness is used to reflect the possibility of the fault at the monitoring point at the sampling time. Then the fault time period is divided from the filtering time period, and the fault time period reflects the corresponding time interval of the fault signal transmission. According to the abnormal attenuation of the fault signal, the degree of noise level can be evaluated to obtain the noise measurement of the digital distribution network in the filtering time period; the larger the noise measurement, the greater the possibility of noise in the digital distribution network in the filtering time period, and the more noise reduction processing is needed. The noise measurement reflects the possibility of noise in the filtering time period, and then the smoothing coefficient is adjusted, and the adjusted smoothing coefficient of the digital distribution network in the filtering time period is reasonably set to improve the noise reduction effect, so as to ensure accurate fault monitoring of the digital distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 A flow chart of a fault monitoring method applied to a digital distribution network provided by an embodiment of the present invention;

[0042] Figure 2 A flow chart of a method for acquiring fault responsiveness provided by an embodiment of the present invention;

[0043] Figure 3 A flow chart of a method for obtaining noise metrics provided by an embodiment of the present invention;

[0044] Figure 4 A structural diagram of a fault monitoring system applied to a digital distribution network provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a fault monitoring method and system for a digital distribution network proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0046] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0047] A specific scheme of a fault monitoring method and system for a digital distribution network provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0048] The present invention provides a method and system for fault monitoring in a digital power distribution network. Figure 1 , which shows a flow chart of a fault monitoring method applied to a digital distribution network provided by an embodiment of the present invention, the method comprising the following steps:

[0049] Step S1: Acquire a data set of the digital distribution network in the time period to be filtered; the data set includes current value data and voltage value data of each monitoring point.

[0050] In order to ensure the subsequent data filtering effect, it is first necessary to obtain the data set of the digital distribution network in the filtering time period to provide data support for subsequent data analysis.

[0051] From the monitoring system of the digital distribution network, obtain the data set of the digital distribution network in the time period to be filtered. Considering that in the distribution network, electric energy is input from the transmission network to the transformer station, the transformer converts high voltage electricity into low voltage electricity, and then transmits it to the user equipment through high voltage cables, wires, cable troughs, etc. Each monitoring point is installed with a current sensor and a voltage sensor to collect the current value and voltage value of the monitoring point. For any monitoring point, the current value corresponding to all sampling moments of the monitoring point in the time period to be filtered is counted to obtain the current value data of the monitoring point; the voltage value corresponding to all sampling moments of the monitoring point in the time period to be filtered is counted to obtain the voltage value data of the monitoring point. The current value data and voltage value data of each monitoring point in the digital distribution network in the time period to be filtered are counted to obtain the data set of the digital distribution network in the time period to be filtered.

[0052] It should be noted that, in one embodiment of the present invention, synchronous sampling is performed according to a preset frequency, each sampling is taken as a sampling moment, the preset frequency is 10 seconds / time, and the filtering time period is from the beginning to the end of one hour. It should be noted that, in order to facilitate calculations, all indicator data involved in the calculations in the embodiments of the present invention are subjected to data preprocessing to eliminate the dimension effect. The specific means of removing the dimension effect are technical means well known to those skilled in the art and are not limited here.

[0053] Considering that the distribution network is subject to different degrees of noise interference in different time periods in actual application scenarios, for example, in relatively quiet time periods such as at night or in the early morning, the distribution network is less subject to noise interference; in time periods with severe weather, the distribution network is more subject to noise interference. If the same smoothing coefficient is used to filter the data of different time periods in the distribution network, the noise reduction effect may not be ideal for time periods with greater noise interference; and for time periods with less noise interference, useful signal information may be lost due to excessive filtering. The present invention analyzes the degree of noise performance in the filtering time period, and then adjusts the smoothing coefficient to improve the noise reduction effect, so as to ensure accurate fault monitoring of the digital distribution network.

[0054] Step S2: According to the timing fluctuations of the current value and voltage value corresponding to the monitoring point at the sampling moment, and the correlation between the current value data and the voltage value data corresponding to the preset reference time period of the monitoring point at the sampling moment, the fault responsiveness of the monitoring point at the sampling moment is obtained; according to the fault responsiveness of all monitoring points at each sampling moment in the time period to be filtered, each fault time period is divided from the time period to be filtered; according to the differences in the fault responsiveness of all monitoring points in each fault time period, the noise measurement of the digital distribution network in the time period to be filtered is obtained.

[0055] In order to reasonably filter the data set in the filtering time period, it is first necessary to analyze the noise performance of the monitoring point in the filtering time period. Considering that the fault signal in the distribution network will attenuate with the increase of the transmission distance, the noise will interfere with the transmission and detection of the fault signal, resulting in abnormal attenuation of the fault signal. First, the fault responsiveness is used to reflect the degree of fault performance of the monitoring point at the sampling time. Then, the fault time period is divided from the filtering time period, and the fault time period reflects the corresponding time interval of the fault signal transmission in the digital distribution network. According to the abnormal attenuation of the fault signal, the degree of noise level can be evaluated and the noise measurement of the digital distribution network in the filtering time period can be obtained; the larger the noise measurement, the greater the possibility of noise in the filtering time period, and the more noise reduction processing is needed.

[0056] To analyze the degree of fault performance of the monitoring point at the sampling time, refer to Figure 2, which shows a flow chart of a method for acquiring fault responsiveness in one embodiment of the present invention. Preferably, in one embodiment of the present invention, the method for acquiring fault responsiveness includes:

[0057] Step S201: For any monitoring point, use the current value data and voltage value data of the monitoring point as the power parameter data of the monitoring point; obtain the baseline data corresponding to the power parameter data according to the average of the upper envelope and the lower envelope of the power parameter data; use the baseline data corresponding to the current value data as the current baseline data; use the baseline data corresponding to the voltage value data as the voltage baseline data.

[0058] The power parameter data reflects the normal characteristics of the power parameters at the monitoring point. The baseline data can more easily identify abnormal fluctuations or changes in the power parameters, thereby discovering potential faults in a timely manner.

[0059] In one embodiment of the present invention, it should be noted that obtaining the upper envelope and lower envelope of data is a basic operation in the field of signal processing and analysis, which is well known to those skilled in the art and will not be described in detail here. The mean of the upper envelope and lower envelope of the power parameter data is calculated to obtain the baseline data corresponding to the power parameter data; the baseline data corresponding to the current value data is used as the current baseline data; and the baseline data corresponding to the voltage value data is used as the voltage baseline data.

[0060] Step S202: Obtain a first fault indicator at the sampling moment according to a difference between the sampling moment and the baseline data corresponding to the preset surrounding time period.

[0061] Considering that the current and voltage values ​​in the distribution network are stable during normal operation, but will fluctuate greatly when a fault occurs, the possibility of a fault at the sampling moment is evaluated by comparing the baseline data at the sampling moment with the average level of the baseline data in the preset surrounding time period, thereby obtaining the first fault indicator at the sampling moment.

[0062] In one embodiment of the present invention, for any baseline data corresponding to the power parameter data, the mean of all baseline data is calculated within a preset time period around the sampling moment to obtain the overall value around the sampling moment; the absolute value of the difference between the baseline data at the sampling moment and the overall value around the sampling moment is calculated to obtain the local difference value of the baseline data at the sampling moment;

[0063] The mean of the local difference values ​​of the current baseline data and the voltage baseline data at the sampling moment is calculated to obtain the first fault indicator at the sampling moment. It should be noted that the method for obtaining the first fault indicator at any sampling moment is described here, and the method for obtaining the first fault indicator at other sampling moments is the same. It should be noted that the present invention is described for the time period to be filtered when there are multiple fault signal transmissions in the digital distribution network.

[0064] In one embodiment of the present invention, a method for obtaining a preset surrounding time period of a sampling moment includes: using a sampling moment and a time period corresponding to a preset number of sampling moments before the sampling moment as the preset surrounding time period of the sampling moment. In one embodiment of the present invention, the preset surrounding number is 7, and the implementer can set it according to the implementation scenario.

[0065] For the above steps, the surrounding overall value represents the average level of the baseline data in the time period around the sampling moment; the absolute value of the difference between the baseline data at the sampling moment and the surrounding overall value is calculated to reflect the degree of deviation between the baseline data at the sampling moment and the average level of the baseline data in the surrounding time period, that is, the local difference value of the baseline data at the sampling moment. For the current value data and voltage value data, the local difference values ​​of their corresponding baseline data at the sampling moment are calculated respectively, and then the average of these local difference values ​​is taken to obtain the first fault indicator at the sampling moment. The first fault indicator reflects the possibility of a fault at the sampling moment by integrating the local differences of the current and voltage baseline data.

[0066] Step S203: obtaining a second fault indicator at the sampling moment according to the difference between the baseline data corresponding to the sampling moment and the previous sampling moment.

[0067] Considering that when a fault occurs in the distribution network, there is a certain relationship between the fluctuation of voltage and current corresponding to the fault signal, and the noise is random, the change of current and voltage corresponding to the noise signal is also random; therefore, the possibility of fault at the sampling moment is evaluated by the ratio of the current and voltage changes, and the second fault indicator at the sampling moment is obtained.

[0068] In one embodiment of the present invention, a method for obtaining a second fault indicator includes:

[0069] The second fault indicator is obtained according to the second fault indicator acquisition formula, and the second fault indicator acquisition formula includes:

[0070] Among them, Y i is the second fault indicator at sampling time i; ΔU i is the difference value of the voltage baseline data at sampling time i; ΔI i is the difference value of the current baseline data at sampling time i; i-1 is the sampling time before sampling time i; ΔU i-1 is the difference value of the voltage baseline data at sampling time i-1; ΔI i-1 is the differential value of the current baseline data at the sampling time i-1; || is the absolute value symbol; exp() is an exponential function with the natural number e as the base; ε is the preset denominator adjustment value.

[0071] In the formula, the differential value of the voltage baseline data reflects the voltage change; the differential value of the current baseline data reflects the current change; The smaller it is, the more likely it is that there is a fault signal. Finally, the second fault index is obtained. The larger the second fault index is, the more the sampling time conforms to the performance of the fault signal. It should be noted that the present invention uses backward difference to calculate the differential value, that is, the data corresponding to the next sampling time minus the data corresponding to the current sampling time. In order to prevent the denominator from being 0, the preset denominator adjustment value in the present invention is set to 0.001, and the implementer can set it according to the implementation scenario.

[0072] Step S204: obtaining a third fault indicator at the sampling moment according to the absolute value of the Pearson correlation coefficient of the current baseline data and the voltage baseline data corresponding to the preset reference time period of the monitoring point at the sampling moment.

[0073] Considering that when a fault occurs in the distribution network, the changes in current and voltage are correlated, while the interference of noise is random, there is no relationship between the changes in voltage and current caused by noise, so as to evaluate the fault possibility at the sampling moment and obtain the third fault indicator at the sampling moment.

[0074] In one embodiment of the present invention, the absolute value of the Pearson correlation coefficient of the current baseline data and the voltage baseline data corresponding to the preset reference time period of the monitoring point at the sampling time is calculated to obtain the third fault indicator at the sampling time. It should be noted that the Pearson correlation coefficient is a technical means well known to those skilled in the art and will not be described in detail here.

[0075] In one embodiment of the present invention, a method for obtaining a preset reference time period of a sampling moment includes: using a time period corresponding to a sampling moment and a preset reference number of sampling moments before the sampling moment as a preset reference time period of the sampling moment. In one embodiment of the present invention, the preset reference number is 9, which can be set by the implementer according to the implementation scenario.

[0076] Step S205: forwardly fuse the first fault indicator, the second fault indicator, and the third fault indicator to obtain the fault responsiveness of the monitoring point at the sampling time.

[0077] Fault responsiveness comprehensively reflects the possibility of fault signals existing at the monitoring point at the sampling time.

[0078] It should be noted that forward fusion is an existing technology well known to those skilled in the art, and forward fusion can adopt simple product, arithmetic mean or other suitable fusion methods. In one embodiment of the present invention, the product of the first fault indicator, the second fault indicator and the third fault indicator is calculated and normalized to obtain the fault responsiveness of the monitoring point at the sampling time. It should be noted that the normalization method adopted is: normalization is performed using the norm normalization function to limit the numerical range to between 0 and 1. Among them, normalization is a technical means well known to those skilled in the art, and the choice of normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0079] In order to analyze the time interval corresponding to the transmission of the fault signal in the digital distribution network, preferably, in one embodiment of the present invention, the method for obtaining the fault time period includes:

[0080] The average of all fault response degrees corresponding to each sampling moment is taken as the overall fault degree of each sampling moment; the sampling moment when the overall fault degree is greater than the preset fault threshold is taken as the fault moment; the time period corresponding to the fault moments with more than the preset number of consecutive faults is taken as the fault time period. In one embodiment of the present invention, the preset fault threshold is 0.59 and the preset number of faults is 6, which can be set by the implementer according to the implementation scenario.

[0081] For the above steps, the higher the overall fault degree, the more serious the fault degree of the distribution network at the sampling moment. The fault moment reflects that the fault degree of the distribution network at the corresponding sampling moment is significant enough to be considered as the presence of a fault signal. Considering that the transmission of the fault signal lasts for a period of time, the time period corresponding to the fault moment with more than the preset number of consecutive faults is taken as the fault time period. This division helps to grasp the actual duration of the fault more accurately.

[0082] Considering that the fault signal in the distribution network will attenuate as the transmission distance increases, noise will interfere with the transmission and detection of the fault signal, resulting in abnormal attenuation of the fault signal. The noise measurement is used to reflect the possibility of noise in the digital distribution network during the filtering period. Please refer to Figure 3 , which shows a flow chart of a method for obtaining a noise metric in one embodiment of the present invention. Preferably, in one embodiment of the present invention, the method for obtaining a noise metric includes:

[0083] Step S211: taking the maximum value of the fault response intensity of the monitoring point in the fault time period as the fault value to be analyzed of the monitoring point in the fault time period; for any fault time period, sorting all the fault values ​​to be analyzed in descending order of the fault values ​​to be analyzed of the monitoring point to obtain the fault monitoring point serial number of the fault value to be analyzed; counting all the fault values ​​to be analyzed in order of the fault monitoring point serial numbers to obtain the sequence of fault values ​​to be analyzed of the fault time period.

[0084] In order to analyze the propagation of fault signals, it is considered that in digital distribution networks, fault signals will propagate from the fault source and gradually attenuate as the propagation distance increases. The monitoring point can capture the attenuation of the fault signal at different positions and distances. The fault value to be analyzed is the maximum value of the fault response intensity of the monitoring point during the fault time period. The maximum value not only reflects the severity of the fault, but also implies the attenuation of the fault signal at the monitoring point. The fault value to be analyzed of a certain monitoring point is small, reflecting that the monitoring point is far away from the fault source and the fault signal is greatly attenuated during the propagation process. According to the order of the fault monitoring point serial number, all the fault values ​​to be analyzed are counted in turn, and a fault sequence to be analyzed that can reflect the attenuation of the fault signal can be obtained.

[0085] Step S212: in the sequence of fault values ​​to be analyzed in all fault time periods, obtaining the attention weight of the fault value to be analyzed according to the outlier situation of the fault value to be analyzed.

[0086] In order to eliminate the influence of noise and obtain a sequence that reflects the attenuation of fault signals in the digital distribution network, it is considered that the fault signal will attenuate with the transmission distance and has a lower degree of outliers. Affected by the randomness of the noise signal, the degree of outliers of the noise signal is higher. According to the outliers of the fault value to be analyzed, the attention weight of the fault value to be analyzed is obtained. The larger the attention weight, the more the fault value to be analyzed conforms to the performance characteristics of the fault signal.

[0087] In one embodiment of the present invention, a method for obtaining attention weight includes:

[0088] Based on a two-dimensional coordinate system, the horizontal axis of the two-dimensional coordinate system is the fault monitoring point number, and the vertical axis of the two-dimensional coordinate system is the fault value to be analyzed. All fault values ​​to be analyzed are mapped into the two-dimensional coordinate system; the LOF algorithm (Local Outlier Factor) is used to calculate the outlier factor of the fault value to be analyzed in the two-dimensional coordinate system and perform negative correlation mapping to obtain the attention weight of the fault value to be analyzed. It should be noted that the use of the LOF algorithm to obtain the outlier factor is a prior art well known to those skilled in the art and will not be elaborated here.

[0089] For the above steps, in order to visualize all the fault values ​​to be analyzed in the coordinate system, a two-dimensional coordinate system is constructed. The fault value of each monitoring point is mapped to the corresponding position of the two-dimensional coordinate system to form a series of data points. The data points can intuitively show the changing trend of the fault signal at different monitoring points. The outlier degree of each data point in the two-dimensional coordinate system is calculated. The LOF algorithm is a density-based outlier detection method, which evaluates the outlier degree of a data point by comparing the local density of a data point with other points in its neighborhood. The outlier factor of the fault value to be analyzed in the two-dimensional coordinate system is calculated by the LOF algorithm. After obtaining the outlier factor of each data point, a negative correlation mapping is performed, that is, the smaller the outlier factor, the greater the attention weight assigned. This is because the attenuation characteristics of the fault signal during the transmission process cause it to show a more consistent change trend in space, while the noise signal shows a higher outlier due to its randomness. Therefore, through negative correlation mapping, those data points that are more in line with the characteristics of the fault signal can obtain a higher attention weight.

[0090] Step S213: obtaining the time period weight of the fault time period according to the differences of the to-be-analyzed fault value sequences of all fault time periods.

[0091] Considering that if the fault value sequence in a certain time period is significantly different from other time periods, it may mean that the time period is affected by more noise, so a lower weight will be given. The larger the time period weight of the fault time period, the more the fault time period conforms to the fault signal transmission characteristics.

[0092] In one embodiment of the present invention, a method for obtaining a time period weight includes:

[0093] Calculate the mean of all the fault values ​​to be analyzed corresponding to the same fault monitoring point number to obtain the reference fault value of the fault monitoring point number; count all the reference fault values ​​in sequence according to the order of the fault monitoring point number to obtain the reference fault value sequence of the fault time period; calculate the DTW distance between the fault value sequence to be analyzed and the reference fault value sequence of each fault time period and perform negative correlation mapping to obtain the time period weight of each fault time period. It should be noted that obtaining the DTW distance of two sequences is a prior art well known to those skilled in the art and will not be elaborated here.

[0094] For the above steps, the reference fault value of the fault monitoring point number reflects the expected fault value at the propagation position under normal circumstances, and the reference fault value sequence reflects the attenuation of the fault signal in the overall digital distribution network. The smaller the DTW distance, the higher the similarity between the fault value sequence to be analyzed and the reference fault value sequence, that is, the fault signal in this time period is more consistent with the overall transmission characteristics. The calculated DTW distance is negatively correlated, that is, the smaller the DTW distance, the greater the time period weight assigned. The larger the time period weight of the fault time period, the more the fault time period is consistent with the fault signal transmission characteristics.

[0095] Step S214: forwardly fuse the attention weight of the fault value to be analyzed and the time period weight of the corresponding fault time period to obtain the target weight of the fault value to be analyzed.

[0096] The greater the target weight is, the more the fault value to be analyzed conforms to the performance characteristics of the fault signal.

[0097] It should be noted that forward fusion is an existing technology well known to those skilled in the art, and forward fusion can adopt simple product, arithmetic mean or other suitable fusion methods. In one embodiment of the present invention, a method for obtaining a target weight includes: calculating the attention weight of the fault value to be analyzed and the mean of the time period weight of the fault time period to which it belongs, and obtaining a target weight of the fault value to be analyzed.

[0098] For the above steps, the target weight is the result of comprehensive consideration of the attention weight and the time period weight. The attention weight reflects the degree to which the fault value to be analyzed conforms to the performance characteristics of the fault signal, while the time period weight reflects the degree to which the fault signal conforms to the overall transmission characteristics during the fault time period. By combining these two weights, we get a target weight that can more accurately reflect the true situation of the fault signal.

[0099] Step S215: acquiring a target fault value sequence of the digital distribution network according to the target weights of all the fault values ​​to be analyzed in the fault value sequence to be analyzed in the fault time period.

[0100] The target fault value sequence can more effectively reflect the actual attenuation sequence of digital distribution network faults.

[0101] In one embodiment of the present invention, a method for obtaining a target fault value sequence includes:

[0102] For all the fault values ​​to be analyzed corresponding to the same fault monitoring point number, the target weight is used to perform weighted averaging on all the fault values ​​to be analyzed to obtain the target fault value corresponding to each fault monitoring point number; all the target fault values ​​are counted in sequence according to the order of the fault monitoring point numbers to obtain the target fault value sequence of the digital distribution network.

[0103] For the above steps, the calculated target weights are used to weighted average all the fault values ​​to be analyzed corresponding to each fault monitoring point number to obtain the target fault value corresponding to each fault monitoring point number. Through weighted averaging, those fault values ​​with higher target weights will occupy a larger proportion in the final result, thereby improving the accuracy of the fault value. According to the order of the fault monitoring point numbers from large to small, all target fault values ​​are counted in turn to obtain the target fault value sequence of the digital distribution network. The target fault value sequence can more effectively reflect the actual attenuation sequence of faults in the digital distribution network.

[0104] Step S216: obtaining the noise metric of the digital distribution network in the filtering time period according to the DTW distance between the target fault value sequence of the digital distribution network and the preset standard fault attenuation sequence.

[0105] Considering that the DTW distance between the target fault value sequence and the standard fault attenuation sequence is large, it means that the noise influence is large, so the noise metric value will be high.

[0106] In one embodiment of the present invention, a method for obtaining a noise metric includes:

[0107] The DTW distance between the target fault value sequence of the digital distribution network and the preset standard fault attenuation sequence is calculated and normalized to obtain the noise metric of the digital distribution network in the filtering time period. It should be noted that the preset standard fault attenuation sequence is set based on empirical values. In one embodiment of the present invention, an engineer can select a target fault value sequence stored in a historical database as the preset standard fault attenuation sequence.

[0108] For the above steps, the target fault value sequence can more effectively reflect the actual attenuation sequence of the digital distribution network fault, and the preset standard fault attenuation sequence reflects the attenuation trend of the fault signal in the digital distribution network without noise interference. The larger the noise metric, the more serious the interference of noise on the fault signal; conversely, the smaller the noise metric, the closer the transmission of the fault signal is to the ideal state.

[0109] Step S3: adjusting the smoothing coefficient according to the noise metric to obtain the adjusted smoothing coefficient of the digital distribution network in the filtering time period; filtering the data set in the filtering time period using the adjusted smoothing coefficient to obtain the filtered data set; and performing fault detection based on the filtered data set.

[0110] The noise measurement reflects the possibility of noise in the digital distribution network during the filtering period, and then adjusts the smoothing coefficient. The adjusted smoothing coefficient of the digital distribution network during the filtering period is reasonably set to improve the noise reduction effect and ensure accurate fault monitoring of the digital distribution network.

[0111] When the noise metric is large, it means that the noise interference is high. At this time, the smoothing coefficient needs to be reduced to reduce the smoothing of data details during the filtering process, thereby retaining more fault signal characteristics; when the noise metric is small, it means that the noise interference is low. At this time, the smoothing coefficient can be appropriately increased to further smooth the data and improve the readability of the data and the accuracy of the analysis. Preferably, in one embodiment of the present invention, the method for obtaining the adjusted smoothing coefficient includes:

[0112] Calculate the product of the negative correlation mapping result of the noise metric and the preset adjustment index to obtain the index to be adjusted; calculate the product of the index to be adjusted and the smoothing coefficient to obtain the adjusted smoothing coefficient. In one embodiment of the present invention, the preset adjustment index is set to 1.2, and the smoothing coefficient is determined based on an empirical value. It is set to 0.57 in the present invention, and the implementer can set it according to the implementation scenario. It should be noted that negative correlation mapping is a technical means well known to those skilled in the art, and the negative correlation mapping can be in the form of inverse proportion or negative exponential power, which is not limited here.

[0113] In order to obtain the filtered data, preferably, in one embodiment of the present invention, the method for obtaining the filtered data set includes:

[0114] In the data set of the digital distribution network in the time period to be filtered, based on the exponentially weighted moving average method, each power parameter data is filtered by reasonably setting the adjusted smoothing coefficient of the digital distribution network in the time period to be filtered, and the filtered power parameter data is obtained; all the filtered power parameter data are counted to obtain the filtered data set. It should be noted that the exponentially weighted moving average method is a prior art well known to those skilled in the art and will not be described in detail here.

[0115] Furthermore, the filtered data set is input into the trained neural network model to output the fault detection result. The specific neural network model training process is a technical means well known to those skilled in the art and will not be described in detail here. Professional equipment engineers can also perform fault assessment on the digital distribution network during the filtering period based on the filtered data set.

[0116] In summary, the embodiments of the present invention provide a fault monitoring method and system for a digital distribution network. First, each fault time period is divided from the time period to be filtered; according to the differences in the fault responsiveness of all monitoring points in each fault time period, the noise measurement of the digital distribution network in the time period to be filtered is obtained; according to the noise measurement, the smoothing coefficient is adjusted to obtain the adjusted smoothing coefficient of the digital distribution network in the time period to be filtered; the data set in the time period to be filtered is filtered using the adjusted smoothing coefficient to obtain the filtered data set; and fault detection is performed based on the filtered data set. The present invention improves the noise reduction effect by reasonably setting the adjusted smoothing coefficient of the digital distribution network in the time period to be filtered, so as to ensure accurate fault monitoring of the digital distribution network.

[0117] The present invention also proposes a fault monitoring system for digital distribution network, see Figure 4 , which shows a structural diagram of a fault monitoring system applied to a digital distribution network provided by an embodiment of the present invention. The system includes: a data acquisition module 101, a noise measurement module 102 and a fault detection module 103.

[0118] The data acquisition module 101 is used to acquire a data set of the digital distribution network in the time period to be filtered; the data set includes current value data and voltage value data of each monitoring point.

[0119] The noise measurement module 102 is used to obtain the fault responsiveness of the monitoring point at the sampling moment according to the timing fluctuations of the current value and voltage value corresponding to the monitoring point at the sampling moment, and the correlation between the current value data and the voltage value data corresponding to the preset reference time period of the monitoring point at the sampling moment; divide each fault time period from the time period to be filtered according to the fault responsiveness of all monitoring points at each sampling moment in the time period to be filtered; and obtain the noise measurement of the digital distribution network in the time period to be filtered according to the differences in the fault responsiveness of all monitoring points in each fault time period.

[0120] The fault detection module 103 is used to adjust the smoothing coefficient according to the noise measurement to obtain the adjusted smoothing coefficient of the digital distribution network in the filtering time period; use the adjusted smoothing coefficient to filter the data set in the filtering time period to obtain the filtered data set; and perform fault detection based on the filtered data set.

[0121] It should be noted that: the system provided in the above embodiment is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, the fault monitoring system applied to the digital distribution network provided in the above embodiment and the fault monitoring method embodiment applied to the digital distribution network belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0122] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0123] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A fault monitoring method applied to a digital distribution network, characterized in that: The method comprises: Acquire a data set of the digital distribution network in the time period to be filtered; the data set includes current value data and voltage value data of each monitoring point; According to the fluctuation of the current value and voltage value corresponding to the monitoring point at the sampling moment along with the timing, and the correlation of the current value data and the voltage value data corresponding to the preset reference time period of the monitoring point at the sampling moment, the fault responsiveness of the monitoring point at the sampling moment is obtained; according to the fault responsiveness of all monitoring points at each sampling moment in the time period to be filtered, each fault time period is divided from the time period to be filtered; according to the difference of the fault responsiveness of all monitoring points in each of the fault time periods, the noise measurement of the digital distribution network in the time period to be filtered is obtained; The smoothing coefficient is adjusted according to the noise metric to obtain the adjusted smoothing coefficient of the digital distribution network in the time period to be filtered; the data set in the time period to be filtered is filtered using the adjusted smoothing coefficient to obtain a filtered data set; and fault detection is performed based on the filtered data set.

2. A fault monitoring method applied to a digital distribution network according to claim 1, characterized in that: The method for obtaining the fault responsiveness includes: For any monitoring point, the current value data and the voltage value data of the monitoring point are used as the power parameter data of the monitoring point; according to the average of the upper envelope and the lower envelope of the power parameter data, the baseline data corresponding to the power parameter data is obtained; the baseline data corresponding to the current value data is used as the current baseline data; the baseline data corresponding to the voltage value data is used as the voltage baseline data; Obtaining a first fault indicator at the sampling time according to a difference between the sampling time and the baseline data corresponding to the preset surrounding time period; According to the difference between the sampling moment and the baseline data corresponding to the previous sampling moment, a second fault indicator at the sampling moment is obtained; Obtaining a third fault indicator at the sampling moment according to the absolute value of the Pearson correlation coefficient of the current baseline data and the voltage baseline data corresponding to the preset reference time period of the monitoring point at the sampling moment; The first fault indicator, the second fault indicator and the third fault indicator are forwardly integrated to obtain the fault responsiveness of the monitoring point at the sampling time.

3. A fault monitoring method applied to a digital distribution network according to claim 2, characterized in that: The method for obtaining the first fault indicator includes: For any baseline data corresponding to the power parameter data, within a preset time period around the sampling moment, the mean of all the baseline data is calculated to obtain the overall value around the sampling moment; the absolute value of the difference between the baseline data at the sampling moment and the overall value around the sampling moment is calculated to obtain the local difference value of the baseline data at the sampling moment; The average of the local difference values ​​of the current baseline data and the voltage baseline data at the sampling time is calculated to obtain a first fault indicator at the sampling time.

4. A fault monitoring method applied to a digital distribution network according to claim 2, characterized in that: The method for obtaining the second fault indicator includes: The second fault indicator is obtained according to the second fault indicator acquisition formula, wherein the second fault indicator acquisition formula includes: Among them, Y i is the second fault indicator at sampling time i; ΔU i is the differential value of the voltage baseline data at sampling time i; ΔI i is the difference value of the current baseline data at sampling time i; i-1 is the sampling time before sampling time i; ΔU i-1 is the difference value of the voltage baseline data at sampling time i-1; ΔI i-1 is the differential value of the current baseline data at the sampling time i-1; || is the absolute value symbol; exp() is an exponential function with the natural number e as the base; ε is the preset denominator adjustment value.

5. A fault monitoring method applied to a digital distribution network according to claim 1, characterized in that: Methods for obtaining noise metrics include: The maximum value of the fault response intensity of the monitoring point in the fault time period is used as the fault value to be analyzed of the monitoring point in the fault time period; for any fault time period, all the fault values ​​to be analyzed are sorted in descending order according to the fault values ​​to be analyzed of the monitoring point to obtain the fault monitoring point serial number of the fault value to be analyzed; all the fault values ​​to be analyzed are counted in order according to the order of the fault monitoring point serial numbers to obtain the fault value sequence to be analyzed in the fault time period; In the sequence of fault values ​​to be analyzed in all the fault time periods, obtaining the attention weight of the fault value to be analyzed according to the outlier situation of the fault value to be analyzed; According to the differences of the to-be-analyzed fault value sequences of all fault time periods, obtaining the time period weight of the fault time period; Forward fusion of the attention weight of the fault value to be analyzed and the time period weight of the fault time period to which it belongs, to obtain the target weight of the fault value to be analyzed; Obtaining a target fault value sequence of the digital distribution network according to target weights of all fault values ​​to be analyzed in the fault value sequence to be analyzed in all fault time periods; According to the DTW distance between the target fault value sequence of the digital distribution network and a preset standard fault attenuation sequence, a noise metric of the digital distribution network in the filtering time period is obtained.

6. A fault monitoring method applied to a digital distribution network according to claim 5, characterized in that: The method for obtaining the attention weight includes: Based on a two-dimensional coordinate system, the horizontal axis of the two-dimensional coordinate system is the fault monitoring point number, and the vertical axis of the two-dimensional coordinate system is the fault value to be analyzed. All fault values ​​to be analyzed are mapped into the two-dimensional coordinate system; using the LOF algorithm, the outlier factor of the fault value to be analyzed in the two-dimensional coordinate system is calculated and negative correlation mapping is performed to obtain the attention weight of the fault value to be analyzed.

7. A fault monitoring method applied to a digital distribution network according to claim 5, characterized in that: The method for obtaining the time period weight includes: Calculate the mean of all the fault values ​​to be analyzed corresponding to the same fault monitoring point number to obtain a reference fault value of the fault monitoring point number; count all the reference fault values ​​in sequence according to the order of the fault monitoring point numbers to obtain a reference fault value sequence of the fault time period; The DTW distance between the fault value sequence to be analyzed and the reference fault value sequence in each fault time period is calculated and negative correlation mapping is performed to obtain the time period weight of each fault time period.

8. A fault monitoring method applied to a digital distribution network according to claim 5, characterized in that: The method for obtaining the target fault value sequence includes: For all the fault values ​​to be analyzed corresponding to the same fault monitoring point number, all the fault values ​​to be analyzed are weighted and averaged using the target weight to obtain the target fault value corresponding to each fault monitoring point number; all the target fault values ​​are counted in sequence according to the order of the fault monitoring point numbers to obtain the target fault value sequence of the digital distribution network.

9. A fault monitoring method applied to a digital distribution network according to claim 1, characterized in that: The method for obtaining the adjusted smoothing coefficient includes: The product of the negative correlation mapping result of the noise metric and the preset adjustment index is calculated to obtain the index to be adjusted; and the product of the index to be adjusted and the smoothing coefficient is calculated to obtain the adjusted smoothing coefficient.

10. A fault monitoring system for a digital power distribution network, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of a fault monitoring method applied to a digital distribution network as claimed in any one of claims 1 to 9 are implemented.

Images