A low-voltage environment anti-interference intelligent electric energy meter state evaluation method and system

Abstract

The present application relates to the technical field of data processing, in particular to a kind of low-voltage environment anti-interference intelligent electric energy meter state evaluation method and system, comprising: obtaining current data, and then obtaining current jump set;According to the data in current and historical jump set, obtain the data variation degree in current and historical jump set, and then obtain the load jump rate of current jump set;Based on the load jump rate of current jump set, determine the invariant of current jump set, and then determine whether there is low-frequency noise in current jump set, determine the low-frequency cutoff frequency and high-frequency cutoff frequency of current jump set;Based on the low-frequency cutoff frequency and high-frequency cutoff frequency of current jump set, get the data after denoising;The data after denoising is input into electric energy meter state evaluation network, and the electric energy meter state evaluation result is obtained.The present application can reduce the power consumption of data denoising process in low-voltage environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a low-voltage environment anti-interference intelligent electric energy meter state evaluation method and system. BACKGROUND

[0002] Low-voltage conditions are more common in remote areas or during peak electricity consumption. Traditional electric energy meters may not accurately measure in low-voltage environments, resulting in discrepancies between actual electricity consumption and metering data. Accurate measurement of intelligent electric energy meters can accurately measure the electricity consumption of users in low-voltage environments, avoid electricity disputes caused by measurement errors, and protect the legitimate rights and interests of users. For example, for small businesses, if the electric energy meter underestimates in low-voltage conditions, they may face the situation of having to pay a large amount of electricity bills later, affecting operating costs and financial planning; accurate measurement can help users clearly understand their electricity consumption and arrange their electricity usage rationally. Accurate electric energy measurement data is the basis for electric power enterprises to analyze electricity consumption, predict load, plan power grids, etc. Accurate measurement of intelligent electric energy meters in low-voltage environments can provide more accurate and reliable electricity data for electric power enterprises, helping them more accurately understand user electricity consumption patterns, optimize power grid operation, and improve power supply reliability and power quality.

[0003] Existing problems: In low-voltage environments, electric energy meters often detect a large amount of noise data, which can interfere with the measurement accuracy of electric energy meters and lead to inaccurate operation state judgments. Specifically, in low-voltage environments, current and voltage fluctuations are more frequent due to low robustness, especially when the load changes significantly. This type of interference is unavoidable low-frequency interference in low-voltage conditions. At the same time, high-frequency electromagnetic interference usually comes from surrounding equipment such as power tools, industrial equipment, etc. The switching action of such equipment, drive motors, power switches, etc. can generate high-frequency noise, which becomes electromagnetic interference. When such interference is not filtered out, the data with noise can affect the judgment of the electric energy meter state evaluation. To solve this problem, a band-pass filter is usually used to filter out noise and only retain the frequency available for analysis. However, under the requirement of band-pass filter denoising, the high and low frequency noise filtering process consumes high power, which cannot meet the low power consumption requirement in low-voltage special environments. SUMMARY

[0004] The present application provides a low-voltage environment anti-interference intelligent electric energy meter state evaluation method and system to solve the existing problems.

[0005] The low-voltage environment anti-interference intelligent electric energy meter state evaluation method and system of the present application adopts the following technical solutions:

[0006] One embodiment of the present application provides a state evaluation method of an anti-interference smart electric energy meter in a low-voltage environment, which comprises the following steps:

[0007] obtaining current data, and obtaining all current jump sets based on the current data;

[0008] obtaining a data variation degree in each current jump set according to data in the jump set;

[0009] obtaining a data variation degree in each jump set in the past, and obtaining a load jump rate of each current jump set by using the data variation degree in each current jump set and the data variation degree in the historical jump set corresponding to the current jump set;

[0010] determining an invariable quantity of each current jump set based on the load jump rate of the jump set, and determining whether there is low-frequency noise in each current jump set according to the invariable quantity of the jump set;

[0011] in the case that there is low-frequency noise in the current jump set, taking the lowest frequency value corresponding to the jump set as a low-frequency cutoff frequency of the jump set, obtaining a random value of the jump set, and obtaining a high-frequency cutoff frequency of the jump set according to the random value of the jump set;

[0012] in the case that there is no low-frequency noise in the current jump set, taking 0 as the low-frequency cutoff frequency of the jump set, and taking the highest frequency value corresponding to the jump set as the high-frequency cutoff frequency of the jump set;

[0013] performing denoising on data in the jump set based on the low-frequency cutoff frequency and the high-frequency cutoff frequency of each current jump set, to obtain denoised data;

[0014] inputting the denoised data into an electric energy meter state evaluation network to obtain an electric energy meter state evaluation result.

[0015] Further, the step of obtaining current data and obtaining all current jump sets based on the current data comprises the following specific steps:

[0016] collecting voltage data or current data of the current day according to a preset collection frequency, and taking the voltage data or current data of the current day as the current data;

[0017] dividing the current data into different jump sets according to a preset division time length, and obtaining all current jump sets.

[0018] Further, the step of obtaining a data variation degree in each current jump set according to data in the jump set comprises the following specific steps:

[0019] For each current jump set, obtain all data in the jump set, and the total number of data in the jump set;

[0020] Calculate the difference between each data in the jump set and each other data in the jump set, and take the maximum difference as the maximum data difference in the jump set;

[0021] Using the Isolation Forest algorithm, obtain the isolation value of each data in the jump set;

[0022] Obtain the normalized value of the isolation value of each data;

[0023] Obtain the number of data in the jump set whose normalized value of the isolation value is greater than the preset isolation value threshold, and take the number of data whose normalized value of the isolation value is greater than the preset isolation value threshold as the inverse proportion coefficient;

[0024] Based on the total number of data in the jump set, the maximum data difference, the isolation value of each data, and the inverse proportion coefficient, obtain the data variation degree in each current jump set.

[0025] Further, the use of the data variation degree in each current jump set, and the data variation degree in the historical jump set corresponding to the current jump set, obtains the load jump rate of each current jump set, including the specific steps as follows:

[0026] For each current jump set, obtain the historical jump set corresponding to the jump set, and the number of historical jump sets;

[0027] Obtain the fault tolerance coefficient; based on the data variation degree of each current jump set, the data variation degree of each historical jump set corresponding to the jump set, the number of historical jump sets and the fault tolerance coefficient, obtain the load jump rate of each current jump set.

[0028] Further, the determination of the invariable quantity of each current jump set based on the load jump rate of each current jump set includes the specific steps as follows:

[0029] For each current jump set, obtain the normalized value of the load jump rate of the jump set, and determine the trend score of the jump set as the normalized value of the load jump rate of the jump set;

[0030] Obtain other jump sets adjacent to each current jump set and having high trend scores, and take the other jump sets as adjacent jump sets of each current jump set;

[0031] Based on the trend score of each current jump set, the trend score of the adjacent jump set and the number of adjacent jump sets, determine the invariable quantity of each current jump set.

[0032] Further, the determining whether low-frequency noise exists in each current hop set according to the invariant of each current hop set comprises the following specific steps:

[0033] For each current hop set, the hop set with the invariant less than or equal to the preset invariant threshold is determined as existing low-frequency noise, and the hop set with the invariant greater than the preset invariant threshold is determined as not existing low-frequency noise.

[0034] Further, in the case that low-frequency noise exists in the current hop set, the lowest frequency value corresponding to the hop set is taken as the low-frequency cutoff frequency of the hop set, the random value of the hop set is obtained, and the high-frequency cutoff frequency of the hop set is obtained according to the random value of the hop set, which comprises the following specific steps:

[0035] In the case that low-frequency noise exists in the current hop set, the lowest frequency value after the fast Fourier transform of the hop set is obtained, and the lowest frequency value is determined as the low-frequency cutoff frequency of the hop set.

[0036] Based on the data in the hop set, the data entropy value in the hop set is obtained, and the data entropy value is determined as the random value of the hop set.

[0037] The highest frequency value after the fast Fourier transform of the hop set is obtained.

[0038] The high-frequency cutoff frequency of the hop set is determined according to the normalized value of the random value of the hop set and the highest frequency value after the fast Fourier transform of the hop set.

[0039] Further, in the case that low-frequency noise does not exist in the current hop set, 0 is taken as the low-frequency cutoff frequency of the hop set, and the highest frequency value corresponding to the hop set is taken as the high-frequency cutoff frequency of the hop set, which comprises the following specific steps:

[0040] In the case that low-frequency noise does not exist in the current hop set, the low-frequency cutoff frequency of the hop set is determined as 0, the highest frequency value after the fast Fourier transform of the hop set is obtained, and the highest frequency value is determined as the high-frequency cutoff frequency of the hop set.

[0041] Further, the obtaining of the data variation degree in each historical hop set comprises the following specific steps:

[0042] For each historical hop set, the total number of data, the maximum value of data difference, the isolated value of each data, and the inverse ratio coefficient in the hop set are obtained.

[0043] Based on the total number of data in the jump set, the maximum value of data difference, the isolated value of each data, and the inverse ratio coefficient, the data variation degree in each jump set in the history is obtained.

[0044] An embodiment of the present application provides a low-voltage environment anti-interference intelligent electric energy meter state evaluation system, the system comprises the following modules:

[0045] A data acquisition module is configured to obtain current data, and based on the current data, obtain all current jump sets;

[0046] A data denoising module is configured to obtain the data variation degree in each current jump set according to the data in each current jump set, obtain the data variation degree in each jump set in the history, obtain the load jump rate of each current jump set by using the data variation degree in each current jump set and the data variation degree in the historical jump set corresponding to the current jump set, determine the invariable quantity of each current jump set based on the load jump rate of each current jump set, determine whether there is low-frequency noise in each current jump set according to the invariable quantity of each current jump set, and further determine the low-frequency cutoff frequency and the high-frequency cutoff frequency of each current jump set, and perform denoising on the data in the jump set based on the low-frequency cutoff frequency and the high-frequency cutoff frequency of each current jump set, to obtain denoised data.

[0047] A data evaluation module is configured to input the denoised data into an electric energy meter state evaluation network to obtain an electric energy meter state evaluation result.

[0048] The technical scheme of the present application has the following beneficial effects: the embodiment of the present application provides a low-voltage environment anti-interference intelligent electric energy meter state evaluation method and system, which divides the collected data into multiple sets, analyzes the change of data in each set, obtains the noise condition of each set, and further adaptively adjusts the bandpass cutoff frequency range of each set according to the noise condition. The present application dynamically adjusts the noise filtering process, realizes the function of bandpass low-frequency cutoff frequency being 0 when there is no obvious fluctuation in the low-frequency region, thereby reducing the power consumption of the data denoising process, achieving the purpose of low power consumption of the intelligent electric energy meter, and improving the accuracy of the evaluation result, and preventing the error of the state evaluation result caused by the existence of noise data. BRIEF DESCRIPTION OF DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without any creative effort.

[0050] Figure 1 A low-voltage environment anti-interference intelligent electric energy meter state evaluation method according to an embodiment of the present application;

[0051] Figure 2 A low-voltage environment anti-interference intelligent electric energy meter state evaluation system according to an embodiment of the present application. DETAILED DESCRIPTION

[0052] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined object, the following describes in detail the specific implementation, structure, features and effects of a low-voltage environment anti-interference intelligent electric energy meter state evaluation method and system according to the present application, with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0053] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0054] The following specifically describes the specific scheme of a low-voltage environment anti-interference intelligent electric energy meter state evaluation method and system provided by the present application with reference to the accompanying drawings.

[0055] Please refer to Figure 1 which shows a low-voltage environment anti-interference intelligent electric energy meter state evaluation method step flow chart provided by an embodiment of the present application. The method includes the following steps:

[0056] Step S001: Obtain current data; based on the current data, obtain all the current jump sets.

[0057] The present embodiment proposes a low-voltage environment anti-interference intelligent electric energy meter state evaluation method, which first divides the collected data into multiple sets, analyzes the change of data in each set, obtains the noise condition of each set, and then adjusts the band-pass cutoff frequency range of each set according to the noise condition. The present application dynamically adjusts the noise filtering process, realizes the function of band-pass low-frequency cutoff frequency being 0 when there is no obvious fluctuation in the low-frequency region, thereby reducing the power consumption of the data denoising process, achieving the purpose of low-power consumption of the intelligent electric energy meter, and improving the accuracy of the evaluation result, preventing the error of the state evaluation result caused by the existence of noise data.

[0058] In this embodiment, the voltage data or current data of the current day is collected according to a preset collection frequency, and the voltage data or current data of the current day is taken as current data; the current data is divided into different jump sets according to a preset division time length, and all the current jump sets are obtained.

[0059] It should be noted that the smart electric energy meter can measure power parameters such as voltage, current and power, and the low-voltage electric energy meter is used in a power system with an alternating voltage of 220V to 380V.

[0060] The preset collection frequency is determined according to specific conditions, which is not limited here. The preset division time length is determined according to specific conditions, which is not limited here.

[0061] For example, the real-time voltage data of the current day is collected at a collection frequency of 5 times per second: ; wherein, The voltage data at time t is represented by V (volt).

[0062] Then, the voltage data every 15 minutes in the current day is taken as a jump set, so that multiple jump sets can be obtained.

[0063] Step S002: According to the data in each current jump set, the data variation degree in each current jump set is obtained.

[0064] It should be noted that in a low-voltage environment, the robustness of voltage and current to noise is poor, resulting in a large error in the state evaluation result of the electric energy meter obtained by analyzing the data. In this environment, noise can be divided into low-frequency noise and high-frequency noise according to the frequency. Among them, the low-frequency noise mainly refers to the instantaneous change caused by the small amplitude fluctuation of voltage and current; the high-frequency noise refers to part of the electromagnetic noise in this environment. In the denoising process, in order to remove the low-frequency noise and the high-frequency noise at the same time, a band-pass filter is selected, which has the effect of removing the low-frequency noise and the high-frequency noise at the same time.

[0065] At the same time, through the analysis of the actual denoising effect, it is found that the high-frequency noise is caused by: due to the operation of electrical equipment (such as switching power supply, wireless communication equipment, frequency converter, etc.), they will produce high-frequency electromagnetic radiation or signal interference. These signals usually have strong randomness and high frequency.

[0066] Low frequency noise: in low voltage state, due to the change of load (such as motor starting, lighting equipment switching, etc.), the fluctuation of current and voltage presents periodicity. Such load fluctuation is usually closely related to the power grid frequency, often in the form of low frequency fluctuation in the electric energy meter data. In particular, periodic load changes, such as air conditioning, power tools, etc., will cause periodic current changes. Therefore, low frequency noise has high periodicity in this environment, and low frequency noise does not exist every moment.

[0067] Therefore, the embodiment of the present application adjusts the low frequency cutoff frequency of the band pass filter part adaptively in this case, with the purpose of reducing the power consumption of the device in the filtering process on the basis of ensuring that the low frequency noise does not affect the real data. Based on this purpose, the embodiment of the present application first needs to analyze whether the band pass low frequency cutoff frequency needs to be dynamically changed, i.e. the invariance of the band pass cutoff frequency; when the invariance is low, it proves that the voltage and current in the low frequency part have high variability at this time, and the periodic change of the corresponding low frequency region is higher, i.e. the trend score of the data is higher. The trend score of the data is obtained by analyzing the real-time voltage and current data load jump and other specific changes.

[0068] Therefore, on the basis of obtaining the voltage and current data, the embodiment of the present application needs to obtain the load jump rate by analyzing the data variation.

[0069] Further explanation is as follows: under normal circumstances, voltage does not exist frequent fluctuation, even if the current environment is low voltage detection environment. However, when the load changes, the corresponding power consumption increases or decreases instantaneously, which will cause voltage change and often has periodicity. Under low voltage, especially in the environment of the power consumption equipment with fixed start-stop mode, the voltage and current fluctuation presents regular periodicity. For example, the current and voltage fluctuation of air conditioner has obvious periodicity when starting and stopping.

[0070] Therefore, when the voltage data in a jump set appears instantaneous increase or decrease, and does not change for a long time, it can be proved that there is possible load change in the jump set; at the same time, in the same n th jump set in the history day, or n-1, n+1 jump set, instantaneous increase or decrease also appears. Then it can be proved that: in the time corresponding to the jump set, there is a load switch with high jump rate.

[0071] In this embodiment, the data variation degree in each jump set of the current day is calculated first, and then the load jump rate is obtained based on the data variation degree.

[0072] In this embodiment, for each current jump set, all data within the jump set and the total number of data within the jump set are obtained; the difference between each data in the jump set and each other data in the jump set is calculated, and the largest difference is taken as the maximum value of data difference within the jump set; the isolated value of each data in the jump set is obtained using the isolated forest algorithm; the normalized value of the isolated value of each data is obtained; the number of data in the jump set whose normalized isolated value is greater than a preset isolated value threshold is obtained, and the number of data whose normalized isolated value is greater than the preset isolated value threshold is taken as an inverse coefficient; based on the total number of data in the jump set, the maximum value of data difference, the isolated value of each data, and the inverse coefficient, the degree of data change within each current jump set is obtained.

[0073] The preset isolated value threshold should be set according to the specific situation, and no specific limit is set here.

[0074] Specifically, the expression for the degree of data change within the nth jump set is as follows:

[0075]

[0076] in, Indicates the first The degree of data change within a jump set; This indicates that all voltage data within this transition set are used as samples, and the time... The voltage value is the isolated value calculated by the isolated forest algorithm (the larger the value, the more isolated the value). This represents the total number of voltage value elements within the transition set; This represents the maximum voltage difference within the jump set; This indicates the number of voltage value elements within the jump set whose isolated (normalized) value is greater than 0.8.

[0077] It should be noted that by analyzing the isolation of voltage data within a set of switching events, the difference between the voltage value at that time and other voltage values ​​within the same set of switching events can be quantified. The higher the isolation, the greater this difference, and therefore the higher the corresponding data variability, and the more likely the voltage fluctuation is caused by load switching.

[0078] Therefore, the sum The higher the value, the greater the degree of data variation. Furthermore, in the former case, the maximum voltage difference is used as a coefficient to quantify the instantaneous voltage variability after load changes; the larger the voltage difference, the higher this instantaneous variability.

[0079] The quantity P serves as an inverse proportionality coefficient. In order to quantify the short-term nature of load switching, if most voltage values ​​in a set of switching events are unstable and not caused by physically complex switching, then the data variability will be lower, because it will not match the voltage fluctuations caused by the actual load changes that we need to find.

[0080] Step S003: Obtain the degree of data change in each historical jump set; using the degree of data change in each current jump set and the degree of data change in the historical jump sets corresponding to the current jump set, obtain the load jump rate of each current jump set.

[0081] It should be noted that the calculation process for the degree of data change within each historical jump set is consistent with that for each current jump set. Furthermore, after obtaining the degree of data change, analyzing and comparing historical data yields the load jump rate, which quantifies the logic. Simultaneously, within the same historical jump set (n-1, n+1), a momentary increase or decrease also occurred. This proves that within the time frame corresponding to that jump set, there exists a load switching event with a relatively high jump rate.

[0082] In this embodiment, for each current jump set, the corresponding historical jump set and the number of historical jump sets are obtained; the fault tolerance coefficient is obtained; and the load jump rate of each current jump set is obtained based on the data change degree of each current jump set, the data change degree of each historical jump set corresponding to the current jump set, the number of historical jump sets, and the fault tolerance coefficient.

[0083] Specifically, the expression for the load hopping rate of the nth hopping set is as follows:

[0084]

[0085] in, Indicates the first The load hopping rate of a hopping set; Indicates the first The degree of data change in the nth jump set of the current day; That is to say, the first The degree of data change in the historical jump set of days (historical days); Indicates the number of historical days included in the calculation; represents the fault tolerance coefficient, and j represents the difference between the number of days in the historical data and the current data.

[0086] It should be noted that: First, through the analysis of the first... Tianhe Di The difference analysis of the data variation degree of the same jump set voltage data within the day, i.e. the difference value The model considers the time corresponding to the first jump set, and further increases the fault tolerance section of the previous and next jump sets, and sets the fault tolerance coefficient i. This is because the switching time of some loads (such as computer room air conditioners) may have certain deviation, so two jump sets of fault tolerance (half an hour) are reserved. The fault tolerance coefficient takes values -1, 0, and 1.

[0087] Secondly, the load jump rate considering the historical data variation degree is obtained by comparing and calculating the difference of the data variation degree within J days. The higher the jump rate is, the higher the trend score corresponding to the voltage data is, i.e. the voltage data has certain periodic fluctuation, and the periodic fluctuation indicates the occurrence of low-frequency noise.

[0088] Step S004: Based on the load jump rate of each current jump set, the invariance of each current jump set is determined; and based on the invariance of each current jump set, it is determined whether there is low-frequency noise in each current jump set.

[0089] In the embodiment, for each current jump set, the normalized value of the load jump rate of the jump set is obtained, the normalized value of the load jump rate of the jump set is determined as the trend score of the jump set, other jump sets adjacent to each current jump set and having high trend scores are obtained, and the other jump sets are taken as the adjacent jump sets of each current jump set; based on the trend score of each current jump set, the trend score of the adjacent jump set, and the number of the adjacent jump sets, the invariance of each current jump set is determined. For each current jump set, the jump set with the invariance less than or equal to the preset invariance threshold is determined as the jump set with low-frequency noise, and the jump set with the invariance greater than the preset invariance threshold is determined as the jump set without low-frequency noise.

[0090] It should be noted that when the trend score of the jump set is greater than the preset trend score threshold, it is determined that the jump set has a high trend score. The preset trend score threshold is set according to specific circumstances, which is not limited here.

[0091] The preset invariance threshold is set according to specific circumstances, which is not limited here.

[0092] Specifically, the data trend score is determined based on the load jump rate, and the expression of the data trend score is as follows:

[0093]

[0094] wherein, The trend score of the nth jump set, which is the normalized load jump rate.

[0095] Further, the invariable quantity corresponding to the current jump set (the nth) and the low frequency cutoff frequency thereof are determined.

[0096] It should be noted that: the invariable quantity refers to the fact that, on the basis of a higher trend score, there is a higher trend score in the subsequent jump set. The expression of the invariable quantity is as follows:

[0097]

[0098] Wherein, W represents the invariable quantity corresponding to the current jump set (the nth); N represents the number of adjacent jump sets participating.

[0099] It should be noted that: for the nth jump set, if the trend score of the (n+1)th jump set is higher, and the trend score of the (n+2)th jump set is not high, then the (n+1)th jump set participates in the calculation of the invariable quantity, and the number of N is 2; if the trend score of the (n+2)th jump set is also high, and the trend score of the (n+3)th jump set is not high, then the number of N is 3.

[0100] The smaller the mean value (the greater the variable), the more concentrated the load change in a time period. Here, it is considered that the invariable quantity is low when it is less than or equal to 0.3. For example, during off-peak hours, in a low-voltage environment, air conditioning load, indoor lighting, multimedia, and external motors are turned off one after another. Therefore, in a plurality of adjacent jump sets, there is a high trend score. Because the voltage fluctuation caused by the load switching is high at this time, there is low-frequency noise at this time, and the low-frequency noise is high. At this time, the band-pass low-frequency cutoff frequency after FFT (Fast Fourier Transform) in the jump set is the lowest frequency value . Thus, the low-frequency cutoff frequency is determined when the invariable quantity is low.

[0101] Conversely, when the invariable quantity is greater than 0.3, that is, when the invariable quantity is high (the variable is low), it is proved that the voltage data change at this time does not have concentration and noise removability, because the corresponding voltage fluctuation noise is almost non-existent. If the band-pass low-frequency cutoff frequency is still set at this time, it will increase the power consumption of the device. Therefore, in this case, the low-frequency cutoff frequency is turned off to reduce the power consumption of the device.

[0102] Step S005: In the case where there is low-frequency noise in the current jump set, the lowest frequency value of the jump set corresponding to the jump set is taken as the low-frequency cutoff frequency of the jump set, the random value of the jump set is obtained, and the high-frequency cutoff frequency of the jump set is obtained according to the random value of the jump set.

[0103] In the case of low frequency noise in the current jump set in this embodiment, the lowest frequency value of the jump set after fast Fourier transform is obtained, and the lowest frequency value is determined as the low frequency cutoff frequency of the jump set; based on the data in the jump set, the data entropy value in the jump set is obtained, and the data entropy value is determined as the random value of the jump set; the highest frequency value of the jump set after fast Fourier transform is obtained; according to the normalized value of the random value of the jump set and the highest frequency value of the jump set after fast Fourier transform, the high frequency cutoff frequency of the jump set is determined.

[0104] It should be noted that: when the invariant is less than or equal to 0.3, it indicates that there is low frequency noise in the current jump set, and the lowest frequency value of the jump set after FFT (Fast Fourier Transform) is as the low frequency cutoff frequency.

[0105] Continue to analyze when the above invariant is satisfied, determine the high frequency noise and the corresponding band-pass high frequency cutoff frequency. High frequency noise will disturb the digital signal of the electric energy meter, will introduce unnecessary error, and will affect the state evaluation of the electric energy meter. Because of its strong randomness and violent fluctuation, it must be filtered in time to ensure data accuracy. Therefore, for the high frequency part, its main feature is the randomness it shows, and randomness generally cannot be accurately self-adaptive controlled. In order to prevent the band-pass filter from not filtering the high frequency noise completely when there is higher randomness, here the highest frequency value of the current jump set is obtained by FFT as the high frequency cutoff frequency of the band-pass filter, and the high frequency cutoff frequency is improved according to the data randomness.

[0106] Specifically, the random value represents the data randomness, and the expression of the random value of the nth jump set is as follows:

[0107]

[0108] Among them, the random value of the nth jump set, refers to the voltage element entropy value in the jump set, the larger the value, the higher the randomness, that is, the higher the high frequency noise energy at this time, and the higher the high frequency electromagnetic radiation or signal interference in the electrical operation process.

[0109] Then the high frequency cutoff frequency at this time is:

[0110]

[0111] Among them, the high frequency cutoff frequency; the higher the randomness, the higher the corresponding high frequency cutoff frequency. Norm represents normalization.

[0112] At this point, in the case that there is low-frequency noise in the current jump set, the cut-off frequency range of the jump set is obtained as .

[0113] Step S006: In the case that there is no low-frequency noise in the current jump set, 0 is taken as the low-frequency cut-off frequency of the jump set, and the highest frequency value corresponding to the jump set is taken as the high-frequency cut-off frequency of the jump set.

[0114] In the case that there is no low-frequency noise in the current jump set, the embodiment determines that the low-frequency cut-off frequency of the jump set is 0, obtains the highest frequency value of the jump set after fast Fourier transform, and determines the highest frequency value as the high-frequency cut-off frequency of the jump set.

[0115] It should be noted that when the invariant is greater than 0.3, it indicates that there is no low-frequency noise in the current jump set, and the highest frequency value of the jump set after FFT (Fast Fourier Transform) is The low-frequency cut-off frequency is 0 as the high-frequency cut-off frequency.

[0116] At this point, in the case that there is no low-frequency noise in the current jump set, the cut-off frequency range of the jump set is obtained as .

[0117] It should be noted that when there is much low-frequency noise generated by voltage fluctuation, which is generally the switching state of the load, then whether there is subsequent electromagnetic interference is faced, and when the low-frequency noise is 0, it is proved that the load is closed, and then the high-frequency cut-off frequency can be set to .

[0118] Step S007: Based on the low-frequency cut-off frequency and the high-frequency cut-off frequency of each current jump set, the data in the jump set is denoised to obtain denoised data.

[0119] At this point, the cut-off frequency range of the real-time dynamic low-power denoising band-pass filter is obtained. The cut-off frequency range is determined once in each jump set to realize efficient denoising and meet the demand of low power consumption.

[0120] Step S008: The denoised data is input into the electric energy meter state evaluation network to obtain an electric energy meter state evaluation result.

[0121] The low-power de-noising data result is obtained, the state of the electric energy meter is determined according to the noise-free data, and an evaluation result is obtained: the de-noising data is input into the electric energy meter state evaluation network, key information such as voltage value and fluctuation condition is obtained from the de-noised voltage data through the feature extraction module, and the evaluation model determines whether the electric energy meter is normal according to the stability of the voltage, whether the voltage is too high or too low and other characteristics. Finally, according to the data, it is determined whether the electric energy meter is in a normal state, an abnormal state or a fault state. Specifically, the normal state indicates that the electric energy meter is in a normal working state; the abnormal state indicates that the electric energy meter is disturbed by the power grid or abnormal load fluctuation occurs; and the fault state indicates that the electric energy meter has a serious fault or the power grid has a problem, which needs to be further checked.

[0122] In a low-voltage environment, the robustness of voltage and current to noise is poor, resulting in a large error in the state evaluation result of the electric energy meter obtained by analyzing the data. The embodiment of the present application performs adaptive band-pass filtering processing on the original data to dynamically filter environmental noise, thereby improving the reliability of the electric energy meter state evaluation result and the quality of the power grid.

[0123] In the de-noising process, in a low-voltage state, the source of high-frequency noise is relatively stable, and the high-frequency threshold of the band-pass is also relatively stable; on the contrary, the low-frequency part has a low frequency and periodicity. In order to meet this periodicity, the embodiment of the present application sets a dynamic band-pass cutoff frequency and dynamically adjusts the filtering parameters according to the data form and data characteristics, thereby improving the filtering effect.

[0124] Please refer to Figure 2 which shows a block diagram of an anti-interference intelligent electric energy meter state evaluation system in a low-voltage environment according to an embodiment of the present application. The system includes the following modules:

[0125] The data acquisition module 100 is configured to obtain current data, and obtain all current jump sets based on the current data.

[0126] The data de-noising module 200 is configured to obtain the data variation degree in each current jump set according to the data in each current jump set, obtain the data variation degree in each historical jump set, obtain the load jump rate of each current jump set by using the data variation degree in each current jump set and the data variation degree in the historical jump set corresponding to the current jump set, determine the invariable quantity of each current jump set based on the load jump rate of each current jump set, determine whether there is low-frequency noise in each current jump set according to the invariable quantity of each current jump set, and further determine the low-frequency cutoff frequency and the high-frequency cutoff frequency of each current jump set, and de-noise the data in the jump set based on the low-frequency cutoff frequency and the high-frequency cutoff frequency of each current jump set to obtain de-noised data.

[0127] The data evaluation module 300 is configured to input the de-noised data into an electric energy meter state evaluation network to obtain an electric energy meter state evaluation result.

[0128] It should be noted that the system further comprises a data transmission module (not shown in the figure), which is configured to transmit data from the data acquisition module to the data de-noising module, and transmit data from the data de-noising module to the data evaluation module.

[0129] Thus, the present application is completed.

[0130] To sum up, in the embodiment of the present application, the low-voltage environment anti-interference intelligent electric energy meter state evaluation method and system are proposed, which first divides the collected data into multiple sets, analyzes the change of data in each set, obtains the noise condition of each set, and then adaptively adjusts the band-pass upper and lower cutoff frequency of each set according to the noise condition. The present application dynamically adjusts the noise filtering process, realizes the function of band-pass low-frequency cutoff frequency being 0 when there is no obvious fluctuation in the low-frequency region, thereby reducing the power consumption of the data de-noising process, achieving the purpose of low-power consumption of the intelligent electric energy meter, and improving the accuracy of the evaluation result, preventing the error of the state evaluation result caused by the existence of noise data.

[0131] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. within the principles of the present application shall be included in the protection scope of the present application.

Claims

1. A low-voltage environment anti-interference intelligent electric energy meter state evaluation method, characterized in that, The method comprises the following steps: obtaining current data; based on the current data, obtaining all current jump sets; obtaining the data variation degree in each current jump set according to the data in each current jump set; obtaining the data variation degree in each historical jump set; obtaining the load jump rate of each current jump set by using the data variation degree in each current jump set and the data variation degree in the historical jump set corresponding to the current jump set; based on the load jump rate of each current jump set, determining the invariance of each current jump set; determining whether there is low-frequency noise in each current jump set according to the invariance of each current jump set; in the case that there is low-frequency noise in the current jump set, taking the lowest frequency value corresponding to the jump set as the low-frequency cutoff frequency of the jump set, obtaining the random value of the jump set, and obtaining the high-frequency cutoff frequency of the jump set according to the random value of the jump set; in the case that there is no low-frequency noise in the current jump set, taking 0 as the low-frequency cutoff frequency of the jump set, and taking the highest frequency value corresponding to the jump set as the high-frequency cutoff frequency of the jump set; based on the low-frequency cutoff frequency and the high-frequency cutoff frequency of each current jump set, denoising the data in the jump set to obtain denoised data; inputting the denoised data into the electric energy meter state evaluation network to obtain an electric energy meter state evaluation result.

2. The state evaluation method of the anti-interference smart electric energy meter in a low-voltage environment according to claim 1, characterized in that, The specific steps of obtaining the current data and obtaining all current jump sets based on the current data are as follows: collecting voltage data or current data of the current day according to a preset collection frequency, and taking the voltage data or current data of the current day as the current data; dividing the current data into different jump sets according to a preset division time length to obtain all current jump sets.

3. The state evaluation method of the anti-interference smart electric energy meter in a low-voltage environment according to claim 2, characterized in that, The specific steps of obtaining the data variation degree in each current jump set according to the data in each current jump set are as follows: for each current jump set, obtaining all data in the jump set and the total number of data in the jump set; calculating the difference between each data in the jump set and each other data in the jump set, and taking the maximum difference as the maximum data difference in the jump set; obtaining the isolation value of each data in the jump set by using the isolation forest algorithm; obtaining the normalized value of the isolation value of each data; obtaining the number of data whose normalized value of the isolation value is greater than a preset isolation value threshold in the jump set, and taking the number of data whose normalized value of the isolation value is greater than the preset isolation value threshold as the inverse proportion coefficient; based on the total number of data in the jump set, the maximum data difference, the isolation value of each data, and the inverse proportion coefficient, obtaining the data variation degree in each current jump set.

4. The state evaluation method of the anti-interference smart electric energy meter in a low-voltage environment according to claim 3, characterized in that, The specific steps of obtaining the load jump rate of each current jump set by using the data variation degree in each current jump set and the data variation degree in the historical jump set corresponding to the current jump set are as follows: for each current jump set, obtaining the historical jump set corresponding to the jump set and the number of historical jump sets; Obtaining a fault tolerance coefficient; obtaining a load jump rate of each current jump set based on a data variation degree of each current jump set, a data variation degree of each historical jump set corresponding to the jump set, a number of the historical jump sets, and the fault tolerance coefficient.

5. The state evaluation method of the anti-interference smart electric energy meter in a low-voltage environment according to claim 4, characterized in that, The specific steps of determining the invariance of each current jump set based on the load jump rate of each current jump set include the following. For each current jump set, obtaining a normalized value of the load jump rate of the jump set, and determining the normalized value of the load jump rate of the jump set as a trend score of the jump set; Obtaining other jump sets adjacent to each current jump set and having high trend scores, and taking the other jump sets as adjacent jump sets of each current jump set; Determining the invariance of each current jump set based on the trend score of each current jump set, the trend scores of the adjacent jump sets, and the number of the adjacent jump sets.

6. The state evaluation method of the anti-interference smart electric energy meter in a low-voltage environment according to claim 5, characterized in that, The specific steps of determining whether low-frequency noise exists in each current jump set based on the invariance of each current jump set include the following. For each current jump set, determining a jump set with an invariance less than or equal to a preset invariance threshold as existing low-frequency noise, and determining a jump set with an invariance greater than the preset invariance threshold as not existing low-frequency noise.

7. The state evaluation method of the anti-interference smart electric energy meter in a low-voltage environment according to claim 6, characterized in that, The specific steps of obtaining a random value of the jump set and obtaining a high-frequency cutoff frequency of the jump set based on the random value of the jump set in the case that low-frequency noise exists in the current jump set include the following. In the case that low-frequency noise exists in the current jump set, obtaining a lowest frequency value of the jump set after fast Fourier transform, and determining the lowest frequency value as a low-frequency cutoff frequency of the jump set; Obtaining a data entropy value in the jump set based on data in the jump set, and determining the data entropy value as a random value of the jump set; Obtaining a highest frequency value of the jump set after fast Fourier transform; Determining a high-frequency cutoff frequency of the jump set based on a normalized value of the random value of the jump set and the highest frequency value of the jump set after fast Fourier transform.

8. The state evaluation method of the anti-interference smart electric energy meter in a low-voltage environment according to claim 6, characterized in that, The specific steps of determining a low-frequency cutoff frequency of the jump set as 0 and obtaining a highest frequency value of the jump set after fast Fourier transform, and determining the highest frequency value as a high-frequency cutoff frequency of the jump set in the case that low-frequency noise does not exist in the current jump set include the following. The specific steps of obtaining a data variation degree in each historical jump set include the following.

9. The low-voltage environment anti-interference intelligent electric energy meter state evaluation method according to claim 1, characterized in that, For each historical jump set, obtaining a total number of data in the jump set, a maximum value of data differences, an isolated value of each data, and an inverse ratio coefficient; Obtaining a data variation degree in each historical jump set based on the total number of data in the jump set, the maximum value of data differences, the isolated value of each data, and the inverse ratio coefficient. ​ 10. An anti-interference intelligent electric energy meter state evaluation system in a low-voltage environment, characterized in that, The system comprises the following modules: A data acquisition module is configured to acquire current data, and based on the current data, acquire all current jump sets; A data denoising module is configured to acquire a data variation degree in each current jump set according to data in each current jump set, acquire a data variation degree in each historical jump set, acquire a load jump rate of each current jump set by using the data variation degree in each current jump set and the data variation degree in a historical jump set corresponding to the current jump set, and determine an invariable quantity of each current jump set based on the load jump rate of each current jump set; Based on the invariable quantity of each current jump set, it is determined whether low-frequency noise exists in each current jump set, and then a low-frequency cutoff frequency and a high-frequency cutoff frequency of each current jump set are determined; and based on the low-frequency cutoff frequency and the high-frequency cutoff frequency of each current jump set, data in the jump set is denoised to obtain denoised data; A data evaluation module is configured to input the denoised data into an electric energy meter state evaluation network to acquire an electric energy meter state evaluation result.

Images