Cloud computing-based power grid on-line business client and management method

Abstract

The invention discloses an electric power on-line business client based on cloud computing and a management method. Acquiring electric signals of the power distribution network in real time, and performing harmonic detection; performing harmonic compensation on the detected harmonic waves to obtain a corrected electric signal after the preprocessed electric signal is dynamically compensated by a plurality of harmonic waves; and forming a plurality of user portraits according to the electric signals after harmonic compensation in the step S2, and classifying the user portraits. According to the client and the management method thereof provided by the invention, after original discrete electric signals collected in real time are preprocessed, harmonic detection is carried out; furthermore, harmonic compensation is carried out on malformation of the current signal and the voltage signal caused by harmonic waves in the power grid to obtain a smooth current signal and a smooth voltage signal to carry out power consumption analysis and prediction, so that deviation of an analysis result is avoided. Meanwhile, a user portrait is constructed, different power supply quantity regulation and control targets can be obtained according to different parameters, and then the power supply quantity of the target user is dynamically adjusted according to the power supply quantity regulation and control targets.

Description

Technical Field

[0001] The present invention relates to the technical field of power network business clients, and in particular to a power network business client and a management method based on cloud computing.

[0002] The power grid client is a unified online service portal and the main portal of the ubiquitous power Internet of Things. It integrates five major service scenarios of "residence, electric vehicles, shops, enterprises and institutions, and new energy" to provide users with diversified comprehensive power services such as electricity payment, repair complaints, electric vehicle charging, photovoltaic one-stop service, and energy finance to meet users' all-round and personalized power needs. Among them, the residential scenario mainly serves low-voltage residential customers, focusing on basic functions such as payment and electricity application, and assisting with special services such as energy analysis, points, and shopping malls; the electric vehicle scenario mainly serves electric vehicle users, focusing on professional services such as recharge, charging, and one-stop service; the store scenario mainly serves low-voltage non-residential customers. In addition to providing basic services such as payment and electricity application, it focuses on exclusive services such as electricity bills, energy analysis, and electricity finance; the enterprise scenario mainly serves high-voltage customers, strengthening professional services such as electricity load, electronic invoices, and energy efficiency diagnosis; the new energy scenario mainly serves photovoltaic customers, providing special services such as website construction consultation, photovoltaic installation, and grid-connected electricity charges and subsidy settlement.

[0003] However, in the past, the client had a remote central computer to store user data locally. With the increase in the number of users and the increase in the number of years of use, the local storage space will be limited. In the prior art, there are cloud computing-based power network management or business clients, such as the Chinese invention patent application documents with application numbers CN202410155166.4 and CN202110706829.3. However, the cloud computing power business client in the prior art cannot perform real-time power consumption analysis on user data, characterize and classify user portraits, and form user portrait groups classified by different indicators under different power supply capacity conditions, and then dynamically adjust the power supply. In addition, for the raw data collected in real time, the power consumption is generally calculated after normalization and simple removal of environmental noise. It is not possible to perform harmonic compensation for the distortion of current signals and voltage signals caused by harmonics in the power grid to obtain smooth current signals and voltage signals for power consumption analysis and prediction, resulting in deviations in the analysis results. Background Art Summary of the invention

[0004] The purpose of the present invention is to provide a method for managing client machines on the power grid based on cloud computing to solve the problems in the above-mentioned background technology. The client and its management method provided by the present invention can perform harmonic detection after preprocessing the original discrete electrical signals collected in real time, and further perform harmonic compensation on the distortion of current signals and voltage signals caused by harmonics in the power grid to obtain smooth current signals and voltage signals for power consumption analysis and prediction, thereby avoiding deviations in the analysis results. At the same time, a user portrait is constructed, and different power supply control targets can be obtained according to different parameters, and then the power supply of the target user is dynamically adjusted according to the power supply control target.

[0005] In order to achieve the above-mentioned object, the present invention provides the following technical solution: a method for managing a business client on a power network based on cloud computing, the method comprising the following steps;

[0006] S1: Collect the electrical signals of the distribution network in real time, and perform harmonic detection after preprocessing them under the conditions of endpoint interpolation, continuity and boundary conditions;

[0007] S2: Perform harmonic compensation on the detected harmonics, and control the dynamic compensation signals of multiple harmonics to be injected into the distribution network to obtain the preprocessed electrical signal Y k (f) Corrected electrical signal Y after multiple harmonic dynamic compensation k (t);

[0008] S3: forming a plurality of user portraits according to the harmonic-compensated electrical signal in step S2, and classifying the plurality of user portraits;

[0009] S4: Dynamically adjust the power supply according to the different user portrait groups obtained by classification.

[0010] 2. The method for managing a power network business client based on cloud computing according to claim 1, wherein the step S1 comprises the following steps:

[0011] S11: Real-time collection of electrical signals from the distribution network, including voltage signals and current signals, to obtain the original discrete electrical signal x k (n), n represents the nth discrete sampling point, n = 1, 2, ..., N k ;

[0012] S12: The electrical signal x collected in real time in step S11 k (n) Perform interpolation preprocessing to obtain the preprocessed electrical signal X k (p); p represents the Nth discrete point of the electrical signal that can form a smooth curve after interpolation preprocessing, p = 1, 2, ..., P k ;P kis the total number of discrete points of the electrical signal that can form a smooth curve after interpolation preprocessing; P k >N k ;

[0013] S13: To capture the electrical signal X after preprocessing in step S12 k (p) Harmonic components of different frequencies and time scales, constructing the window function ω k (p), the window function ω k The window length of (p) is P k , that is, the total number of discrete points when the window function is convolved with the electrical signal preprocessed in step S22; k = i or u; when k = i, ω i (p) is the window function for processing current signals; when k = u, ω u (p) is the window function for processing voltage signals;

[0014] S14: The electrical signal X pre-processed in step S12 is k (p) Perform convolution operation with the corresponding window function constructed in step S22 to calculate the convolution result y k (p):

[0015] Where q = 1, 12, .., P k ;

[0016] S15: Perform short-time Fourier transform on the convolution result obtained in step S14 to obtain the frequency domain value Y of the preprocessed electrical signal k (f), the calculation formula for short-time Fourier transform is as follows:

[0017] Where f is the frequency domain value Y of the preprocessed electrical signal obtained by short-time Fourier transform k (f) frequency;

[0018] S16: Perform harmonic detection on the frequency domain value of the preprocessed electrical signal obtained in step S15, and the harmonic frequency of the e-th harmonic in the detected multiple harmonic data is f w，e (t), the harmonic amplitude is A w，e (t) and the harmonic phase is φ w，e (t), wherein t is the t-th moment in the frequency domain after the short-time Fourier transform conversion in step S15; e=1, 2, ..., E; that is, the number of harmonics detected is E in total, and w represents the fundamental harmonic wave.

[0019] Furthermore, the window function of the voltage signal processed in step S12 is:

[0020]

[0021] The window function of the current signal processed in step S12 is:

[0022]

[0023] Among them, ε k is the parameter that controls the shape of the window function. When k = u, ε u =0.5; when k=i,

[0024] Furthermore, in step S12, the electrical signal x collected in real time in step S11 is k (n) Perform interpolation preprocessing to obtain the preprocessed electrical signal X k (p), comprising the following steps:

[0025] S121: Construct discrete point electrical signal set G:

[0026] G={(1,x k (1)), (2, x k (2)),...,(n,x k (n)), ..., (N k -1,x k (N k -1))};

[0027] S122: For any two data (n, x) in the discrete point electrical signal set G constructed in step S221, k (n)) and (n+1, x k (n+1)), construct the interpolation point electrical signal solution equation:

[0028] X k (p) = a n (pn) 3 +b n (pn) 2 +c n (pn)+d n , where a n is the first coefficient, b n is the second coefficient, c n is the third coefficient, d n Solving for error values ​​for electrical signals;

[0029] S123: Under the conditions of endpoint interpolation conditions, continuity conditions and boundary conditions, solve the first coefficient a of the interpolation point electrical signal solution equation constructed in step S122 n , the second coefficient b n , the third coefficient c n And the electrical signal solves the error value d n ;

[0030] During the solution process, construct the solution equation system:

[0031]

[0032] Among them, h n-1 is the first coefficient to be solved, h n-1 = x′ k (n-1)-x′ k (n); h n is the second solution coefficient, h n = x′ k (n)-x′ k (n-1); h n+1 is the third solution coefficient, h n+1 = x′ k (n+1)-x′ k (n); m n-1 is the independent variable of the first system of equations, m n is the independent variable of the second system of equations, m n+1 is the independent variable of the third program group,

[0033] The endpoint interpolation conditions are as follows: k (n) = x k (n), X k (n+1)=x k (n+1)

[0034] The continuity conditions are as follows: Because P k The pre-processed electrical signal data X k (p) is N k The original discrete electrical signal x k (n) After interpolation P k -N k interpolated points and combined with the original x k (n) is formed by combining each time point, so the result of the continuity condition is X k (n) = x k (n);

[0035] The boundary conditions are as follows:

[0036] S124: For (P k -N k) interpolation points, repeat the steps S221-S223, and solve for N k The original discrete electrical signal x k (n) All interpolation point electrical signals X k (p).

[0037] Furthermore, the method for compensating harmonics in step S2 comprises the following steps:

[0038] S21: Determine the e-th harmonic frequency f obtained according to step S16 w，e (t), harmonic amplitude A w,e (t) and harmonic phase φ w，e (t), determine the amplitude A of the dynamic compensation signal of the e-th harmonic com，e (t), phase φ com，e (t) and frequency f com，e (t):

[0039] A com，e (t) = A w，e (t); φ com，e (t) = φ w，e +π;f com，e (t) = f h，e (t);

[0040] S22: Determine the dynamic compensation signal of the harmonics of step S21 for each of the plurality of harmonics;

[0041] S23: According to the dynamic compensation signals of the multiple harmonics obtained in step S22, controlling the dynamic compensation signals of the multiple harmonics to be injected into the distribution network;

[0042] S24: After dynamic compensation of multiple harmonics, a pre-processed electrical signal Y containing a total of E harmonics is obtained k (f) Corrected electrical signal Y after multiple harmonic dynamic compensation k (t):

[0043]

[0044] Among them, A k,0 (t) is the fundamental amplitude of the preprocessed electrical signal; k = i or u, Y i (t) is the corrected current signal after multiple harmonic dynamic compensation, Y u (t) is the corrected voltage signal after multiple harmonic dynamic compensation.

[0045] Furthermore, the method for constructing a user portrait group in step S3 includes the following steps:

[0046] S31: Calculate the corrected current signal Y during the monitored period T i (t) and the corrected voltage signal Y u (t) The average value of the corrected current Y i，avg (t), average value of corrected voltage Y u，avg (t) and the correction current variance σ i , Corrected voltage variance σ u ; t = 1, 2, ..., T;

[0047] S32: Calculate the user's power factor PF according to the calculation result of step S31:

[0048]

[0049] S33: Calculate the maximum current value Y of the user in the monitored time period T according to the calculation result of step S31 i，peak and the maximum voltage value Y u，peak :

[0050] Y i，peak =max |T (Y i (t)); max |T (Y i，avg (t)) is the maximum value function among all the current values ​​at the moment in the monitored period T;

[0051] Y u，peak =max |T (Y u (t)); max |T (Y u (t)) is the maximum value function among all voltage values ​​at the time within the monitored period T;

[0052] S34: Determine the peak power consumption period T of the user within the monitored period T peak ; T peak =t d -t c ;t d is the end time of the peak power consumption period, t c is the starting time of the peak land use period, t c ∈{1, 2, ..., T}, t d ∈{1, 2, ..., T}, and T ≥ t d >t c ≥1;

[0053] S35: Based on the results of steps S31 to S34, construct a user portrait vector R:

[0054] R=[Y i，avg (t), Yu，avg (t), σ i , σ u ,PF,Y i，peak , Y u，peak , T peak ];

[0055] S36: Select a parameter in the user portrait vector R constructed in step S35 as a classification index to classify the user portrait vector.

[0056] Furthermore, the step S36 includes the following steps:

[0057] S361: Select a parameter in the user portrait vector, and select the number of clusters S according to the parameter, and divide the user portrait vectors R of multiple users into different groups G s ; G s is the sth group of user portrait vectors, s = 1, 2, ..., S; the total number of user portrait vectors R is Z, R z is the zth user portrait vector;

[0058] S362: Randomly select the initial cluster center of each group of user portrait vectors: Ce1, Ce2, ...Ce s , ..., Ce S ; Ce s is the user portrait vector G of the sth group s The central user portrait vector of

[0059] S363: Calculate the zth user portrait vector R z and the sth group user portrait vector G s The central user portrait vector Ce s The Euclidean distance is: Among them g is the gth parameter of the user portrait vector R, For the gth parameter o g The zth user portrait vector as an indicator, For the gth parameter o g The sth group user portrait vector G as an indicator s The central user portrait vector of

[0060] o1 is Y i，avg (t); o2 is Y u，avg (t); o3 is σ i ; o4 is σ u ; o5 is PF; o6 is Y i，peak ; o7 is Y u，peak ; o8 is T peak ;

[0061] S364: According to the calculation result of S363, the zth user portrait vector is assigned to the center user portrait vector Ce closest to it. l The user portrait vector G of the first group l , l∈{1, 2, ..., S};

[0062] That is, if D(R z , Ce l )=min{D(R z , Ce1), D(R z , Ce2), ..., D(R z , Ce S )}, then R z ∈G l ;

[0063] S365: For each group of user portrait vectors, relocate and calculate its cluster center to obtain the updated value of the cluster center:

[0064]

[0065] Among them, Ce′ l The user portrait vector G of group I calculated for re-positioning l The cluster center of G is updated to l ) is used to calculate the user portrait vector G of group I l The user portrait vector R contained in z A function of the total number of ;

[0066] S366: Repeat the steps S363-S365 until the cluster center update iteration threshold Q is less than 0.135, then stop the iteration and the user portrait classification is completed;

[0067]

[0068] Furthermore, in step S4, dynamically adjusting the power supply according to different user portrait groups obtained by classification includes the following steps:

[0069] S41: Calculate the load power of each group of users classified according to the user portrait vector:

[0070] Among them, Y i s (t) is the average corrected current of the sth group of users, is the average corrected voltage of the sth group of users, PF s is the power factor of the sth group of users; To calculate the power factor PF of all users in the sth group of users, weighted multiplication is performed, and then the power consumption after normalization of power efficiency is calculated; Power s Represents the total net power load power of all users in the sth group of users;

[0071] S42: If the current signal of the zth user in the sth group of users monitored in real time is preprocessed in step S1 to obtain the corrected current signal Exceeds the real-time current limit threshold of the zth user in the sth group of users Then adjust the real-time supply voltage at the next moment This reduces the amount of power supplied.

[0072] Furthermore, in the step S42, the real-time power supply voltage at the next moment is adjusted. The calculation formula is as follows:

[0073]

[0074] in, is the corrected voltage signal obtained after the voltage signal of the zth user in the sth group of users monitored in real time is preprocessed in step S1; η u is the voltage regulation coefficient, η u =0.63; is the maximum current value of the zth user in the sth group of users within the monitoring period T.

[0075] The present invention also provides a cloud computing-based power network business client using the above method, characterized in that it includes a signal acquisition and harmonic detection module, a harmonic compensation module, a user classification module and a power supply management module;

[0076] The signal acquisition and harmonic detection module is used to collect the electrical signals of the distribution network in real time, and perform harmonic detection after pre-processing them under the conditions of endpoint interpolation conditions, continuity conditions and boundary conditions;

[0077] The harmonic compensation module is used to perform harmonic compensation on the detected harmonics and control the dynamic compensation signals of multiple harmonics to be injected into the distribution network to obtain the pre-processed electrical signal Y k (f) Corrected electrical signal Y after multiple harmonic dynamic compensation k (t);

[0078] The user classification module is used to form multiple user portraits according to the customer's electricity consumption, and the multiple user portraits are classified;

[0079] The power supply management module is used to dynamically adjust the power supply according to different user portrait groups obtained by classification.

[0080] The beneficial effects of the present invention are:

[0081] 1. The method provided by the present invention detects harmonics and then performs harmonic compensation to obtain a pre-processed electrical signal Y k (f) Corrected electrical signal Y after multiple harmonic dynamic compensation k (t) Avoid directly using the collected original discrete voltage and current signals for user load power. The user terminal is connected to the cloud computing platform through the network. The cloud computing platform provides business logic processing and data storage services. Users can access various power service functions through the client. The client can monitor the user's power consumption related signals in real time through cloud computing, and then effectively and dynamically adjust the real-time power supply for different user portrait groups.

[0082] 2. The cloud computing-based power network business client and management method provided by the present application further calculates relevant data on the corrected current signal and voltage signal obtained by harmonic compensation, forms a user portrait vector with the calculated relevant data, and then performs cluster analysis. Different user portrait groups can be obtained according to different selected parameters, and then different parameters can be selected as classification indicators for user portrait groups under different statistics and power supply capacity conditions, and then after the user portrait groups are classified, dynamic power supply adjustment is performed. Therefore, the present application can classify users into different groups, such as peak power users, night power users, etc., and then provide personalized services, dynamically adjust the power supply demand side management on each branch line in the distribution network according to different power consumption requirements, etc.

[0083] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0085] Figure 1 A schematic diagram of the process flow of the power network business client management method based on cloud computing provided by the present invention;

[0086] Figure 2 A schematic diagram of the module structure of a cloud computing-based power network business client provided by the present invention. DETAILED DESCRIPTION

[0087] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0088] like Figure 1 As shown, it is a flow chart of a method for managing a business client on a power network based on cloud computing provided by the present invention, and the method comprises the following steps;

[0089] S1: Collect the electrical signals of the distribution network in real time, and perform harmonic detection after preprocessing them under the conditions of endpoint interpolation, continuity and boundary conditions;

[0090] S2: Perform harmonic compensation on the detected harmonics and control the dynamic compensation signals of multiple harmonics to be injected into the distribution network to obtain the pre-processed electrical signal Y k (f) Corrected electrical signal Y after multiple harmonic dynamic compensation k (t);

[0091] S3: forming a plurality of user portraits according to the harmonic-compensated electrical signal in step S2, and classifying the plurality of user portraits;

[0092] S4: Dynamically adjust the power supply according to the different user portrait groups obtained by classification.

[0093] In the power system, harmonics are caused by nonlinear loads and other reasons, which will have an adverse effect on the stability of power equipment and systems. Although traditional harmonic analysis methods, such as Fourier transform, can effectively identify harmonic components, they are limited in effect when processing non-stationary signals or signals with high harmonic complexity. Therefore, as another preferred embodiment of the present invention, step S1 includes the following steps:

[0094] S11: Real-time collection of electrical signals from the distribution network, including voltage signals and current signals, to obtain the original discrete electrical signal x k (n), n represents the nth discrete sampling point, n = 1, 2, ..., N k ;

[0095] S12: The electrical signal x collected in real time from step S11 k (n) Perform interpolation preprocessing to obtain the preprocessed electrical signal X k(p); p represents the Nth discrete point of the electrical signal that can form a smooth curve after interpolation preprocessing, p = 1, 2, ..., P k ;P k is the total number of discrete points of the electrical signal that can form a smooth curve after interpolation preprocessing; P k >N k ;

[0096] S13: To capture the electrical signal X after preprocessing in step S12 k (p) Harmonic components of different frequencies and time scales, constructing the window function ω k (p), window function ω k The window length of (p) is P k , that is, the total number of discrete points when the window function is convolved with the electrical signal preprocessed in step S22; k = i or u; when k = i, ω i (p) is the window function for processing current signals; when k = u, ω u (p) is the window function for processing voltage signals;

[0097] S14: The electrical signal X pre-processed in step S12 k (p) Perform convolution operation with the corresponding window function constructed in step S22 to calculate the convolution result y k (p):

[0098]

[0099] Where q = 1, 12, .., P k ; That is, q is also a point in the discrete points of the preprocessed electrical signal;

[0100] S15: Perform short-time Fourier transform on the convolution result obtained in step S14 to obtain the frequency domain value Y of the preprocessed electrical signal with frequency f as the independent variable k (f), and based on this, frequency domain analysis is performed and the calculation formula for short-time Fourier transform is as follows:

[0101]

[0102] Where f is the frequency domain value Y of the preprocessed electrical signal obtained by short-time Fourier transform k (f) frequency; through the short-time Fourier transform step S25, the pre-processed electrical signal X can be obtained k Frequency response of harmonics in (p) Y k (f), and then by analyzing Y k (f) The peak and spectrum shape of the harmonics are identified by the frequency distribution and amplitude

[0103] S16: Perform harmonic detection on the frequency domain value of the preprocessed electrical signal obtained in step S15, and the harmonic frequency of the e-th harmonic in the multiple harmonic data obtained by detection is f w，e (t), the harmonic amplitude is A w，e (t) and the harmonic phase is φ w，e (t), where t is the t-th moment in the frequency domain after the short-time Fourier transform in step S15; e=1, 2, ..., E; that is, the number of harmonics detected is E in total, and w represents the fundamental harmonic wave.

[0104] The multi-convolution window function is a powerful tool that can simultaneously and locally process signals in the time and frequency domains. It can effectively extract harmonic features, denoise and standardize voltage signals and / or current signals to eliminate harmonic noise caused by the unstable operating environment of the distribution network and improve the accuracy of the analysis. Adapt to different frequency ranges and time windows. The present invention proposes a harmonic detection method based on a multi-convolution window function, which extracts harmonic components of various orders by designing window functions of different types and widths and convolving them with power signals, and analyzes and suppresses harmonics according to the responses of different window functions.

[0105] Among them, the window function of the voltage signal processed in step S12 is:

[0106]

[0107] The window function of the current signal processed in step S12 is:

[0108]

[0109] Among them, ε k is the parameter that controls the shape of the window function. When k = u, ε u =0.5; when k=i,

[0110] 4. The method for managing a power network business client based on cloud computing according to claim 2 is characterized in that in step S12, the electrical signal x collected in real time in step S11 is k (n) Perform interpolation preprocessing to obtain the preprocessed electrical signal X k (p), comprising the following steps:

[0111] S121: Construct discrete point electrical signal set G:

[0112] G = {(1, x k (1)), (2, x k (2)),...,(n,x k (n)), ..., (N k -1,x k (Nk -1))};

[0113] S122: For any two data (n, x) in the discrete point electrical signal set G constructed in step S221 k (n)) and (n+1, x k (n+1)), construct the interpolation point electrical signal solution equation:

[0114] X k (p) = a n (pn) 3 +b n (pn) 2 +c n (pn)+d n , where a n is the first coefficient, b n is the second coefficient, c n is the third coefficient, d n Solve the error value for the electrical signal; a n , b n 、c n and d n All of them are unknowns to be determined;

[0115] S123: Under the conditions of endpoint interpolation, continuity and boundary conditions, solve the first coefficient a of the interpolation point electrical signal solution equation constructed in step S122 n , the second coefficient b n , the third coefficient c n And the electrical signal solves the error value d n ;

[0116] During the solution process, construct the solution equation system:

[0117]

[0118] Among them, h n-1 is the first coefficient to be solved, h n-1 = x′ k (n-1)-x′ k (n); h n is the second solution coefficient, h n = x′ k (n)-x′ k (n-1); h n+1 is the third solution coefficient, h n+1 = x′ k (n+1)-x′ k (n); x′ k (n) is the original discrete electrical signal x acquired in real time k (n) Solve for the first-order derivative with respect to n, x′ k (n+1) and x′ k (n-1) also represents the same calculation method; m n-1 is the independent variable of the first system of equations, m n is the independent variable of the second system of equations, That is, solve for X k (p) second derivative with respect to n; m n+1 is the independent variable of the third program group,

[0119] The endpoint interpolation conditions are as follows: X k (n) = x k (n), X k (n+1)=x k (n+1)

[0120] The continuity conditions are as follows: Because P k The pre-processed electrical signal data X k (p) is N k The original discrete electrical signal x k (n) After interpolation P k -N k interpolated points and combined with the original x k (n) is formed by combining each time point, so the result of the continuity condition is X k (n) = x k (n);

[0121] The boundary conditions are as follows:

[0122] S124: For (P k -N k ) interpolation points, repeat steps S221-S223, and solve for N k The original discrete electrical signal x k (n) All interpolation point electrical signals X k (p).

[0123] Accurate analysis of harmonics is crucial for the accuracy of data obtained by the client regarding power quality monitoring and system stability. However, in the actual sampling process, due to sampling rate limitations or noise interference, the sampling points may be unevenly distributed or not dense enough, resulting in a decrease in the accuracy of harmonic analysis. Interpolation is an effective method for dealing with signal curves with sharp inflection points and not smooth enough due to uneven or insufficient sampling points. It can accurately reconstruct the signal by constructing a smooth curve between known sampling points, thereby improving the accuracy of harmonic analysis. By interpolating the collected current or voltage signals in the power system, the original discrete electrical signal x k (n) Reconstruct the signal and get the interpolation point electrical signal X k (p), and further analyzing the harmonic characteristics based on this smooth curve can accurately determine the amplitude and frequency information of the harmonic components in the original discrete electrical signal, thereby improving the accuracy and resolution of harmonic detection.

[0124] Harmonics are caused by nonlinear loads (such as electronic devices, frequency converters, etc.), which distort current and voltage waveforms and generate harmonic components in the spectrum. In addition, harmonics will affect the power factor of the power system, resulting in a decrease in the power factor. Therefore, it is necessary to perform harmonic compensation on the detected harmonics. Therefore, as another preferred embodiment of the present invention, the method for compensating for harmonics in step S2 includes the following steps:

[0125] S21: Determine the e-th harmonic frequency f obtained according to step S16 w，e (t), harmonic amplitude A w，e (t) and harmonic phase φ w，e (t), determine the amplitude A of the dynamic compensation signal of the e-th harmonic com，e (t), phase φ com，e (t) and frequency f com，e (t):

[0126] A com，e (t) = A w，e (t); φ com，e (t) = φ w，e +π;f com，e (t) = f h，e (t);

[0127] S22: Determine the dynamic compensation signal of the harmonics of step S21 for each of the plurality of harmonics;

[0128] S23: According to the dynamic compensation signals of the multiple harmonics obtained in step S22, controlling the dynamic compensation signals of the multiple harmonics to be injected into the distribution network;

[0129] S24: After dynamic compensation of multiple harmonics, a pre-processed electrical signal Y containing a total of E harmonics is obtainedk (f) Corrected electrical signal Y after multiple harmonic dynamic compensation k (t):

[0130]

[0131] Among them, A k,0 (t) is the fundamental amplitude of the preprocessed electrical signal; k = i or u, Y i (t) is the corrected current signal after multiple harmonic dynamic compensation, Y u (t) is the corrected voltage signal after multiple harmonic dynamic compensation.

[0132] Through harmonic compensation, the power factor can be improved and the consumption of reactive power can be reduced; therefore, harmonic compensation is performed on the harmonics of a specific frequency fh to offset the harmonics of a frequency fh. h (The harmonic amplitude is A h And the harmonic phase is φ h ) on the voltage and current signals collected in real time, and then through harmonic compensation, the distortion of voltage and current signals caused by uncompensated harmonics is avoided, which would lead to inaccurate classification of subsequent user portrait groups.

[0133] As another preferred embodiment of the present invention, the method for constructing a user portrait group in step S3 includes the following steps:

[0134] S31: Calculate the corrected current signal Y within the monitored period T i (t) and the corrected voltage signal Y u (t) The average value of the corrected current Y i，avg (t), average value of corrected voltage Y u，avg (t) and the correction current variance σ i , Corrected voltage variance σ u ; t = 1, 2, ..., T;

[0135] Step S31 calculates the corrected current signal Y within the monitored period T i (t) and the corrected voltage signal Y u (t) The average value of the corrected current Y i，avg (t), average value of corrected voltage Y u，avg (t) and the correction current variance σ i , Corrected voltage variance σ u The formulas are as follows:

[0136]

[0137]

[0138] S32: Calculate the user's power factor PF according to the calculation result of step S31:

[0139]

[0140]

[0141] The user's power factor can effectively reflect the user's power efficiency;

[0142] S33: Calculate the maximum current value Y of the user in the monitored period T according to the calculation result of step S31 i，peak and the maximum voltage value Y u，peak :

[0143] Y i，peak =max |T (Y i (t)); max |T (Y i，avg (t)) is the maximum value function among all the current values ​​at the moment in the monitored period T;

[0144] Y u，peak =max |T (Y u (t)); max |T (Y u (t)) is the maximum value function among all voltage values ​​at the time within the monitored period T;

[0145] S34: Determine the peak power consumption period T of the user within the monitored period T peak ; T peak =t d -t c ;t d is the end time of the peak power consumption period, t c is the starting time of the peak land use period, t c ∈{1, 2, ..., T}, t d ∈{1, 2, ..., T}, and T ≥ t d >t c ≥1;

[0146] S35: Based on the results of steps S31 to S34, construct a user portrait vector R:

[0147] R=[Y i，avg (t), Y u，avg (t), σ i , σ u ,PF,Y i，peak , Y u，peak , T peak ];

[0148] S36: Select a parameter in the user portrait vector R constructed in step S35 as a classification indicator to classify the user portrait vector. By analyzing the user's current and voltage historical data and further classifying them to obtain different user portrait groups, support can be provided for personalized services, demand-side management and other applications. By segmenting user groups, customers can more accurately identify the electricity demand and behavioral characteristics of different user groups, thereby providing more personalized services. For example, office building groups may be more concerned about the stability and reliability of power supply from 9:00 to 18:00, while residential areas may be more concerned about electricity prices and basic electricity demand.

[0149] In addition, understanding the electricity consumption patterns and needs of different user groups can more effectively allocate grid resources and improve the grid's operating efficiency and power supply quality. For example, for high-energy-consuming industrial areas, higher power supply guarantees and more stable power supply may be required. User portraits can help power companies predict and respond to changes in electricity consumption of different user groups, thereby improving the flexibility and adaptability of distribution networks. For example, by analyzing users' electricity consumption habits, peak electricity consumption periods can be predicted and grid load adjustments can be made in advance.

[0150] Step S36 includes the following steps:

[0151] S361: Select a parameter in the user portrait vector, and select the number of clusters S according to the parameter, and divide the user portrait vectors R of multiple users into different groups G s ; G s is the sth group of user portrait vectors, s = 1, 2, ..., S; the total number of user portrait vectors R is Z, R z is the zth user portrait vector;

[0152] S362: Randomly select the initial cluster center of each group of user portrait vectors: Ce1, Ce2, ...Ce s , ..., Ce S ; Ce s is the user portrait vector G of the sth group s The central user portrait vector of

[0153] S363: Calculate the zth user portrait vector R z and the sth group user portrait vector G s The central user portrait vector Ce s The Euclidean distance is: Among them g is the gth parameter of the user portrait vector R, For the gth parameter o g The zth user portrait vector as an indicator, For the gth parameter o g The sth group user portrait vector G as an indicator s The central user portrait vector of

[0154] o1 is Y i，avg (t); o2 is Y u，avg (t); o3 is σ i ; o4 is σ u ; o5 is PF; o6 is Y i，peak ; o7 is Y u，peak ; o8 is T peak ;

[0155] S364: According to the calculation result of S363, the zth user portrait vector is assigned to the center user portrait vector Ce closest to it. l The user portrait vector G of the first group l , l∈{1, 2, ..., S};

[0156] That is, if D(R z , Ce l )=min{D(R z , Ce1), D(R z , Ce2), ..., D(R z , Ce S )}, then R z ∈G l ;

[0157] S365: For each group of user portrait vectors, relocate and calculate its cluster center to obtain the updated value of the cluster center:

[0158]

[0159] Among them, Ce′ l The user portrait vector G of group I calculated for re-positioning l The cluster center of G is updated to l ) is used to calculate the user portrait vector G of group I l The user portrait vector R contained in z A function of the total number of ;

[0160] S366: Repeat steps S363-S365 until the cluster center update iteration threshold Q is less than 0.135, then stop the iteration and the user portrait classification is completed;

[0161]

[0162] In step S4, the power supply is dynamically adjusted according to the different user portrait groups obtained by classification, including the following steps:

[0163] S41: Calculate the load power of each group of users classified according to the user portrait vector:

[0164] Among them, Y i s (t) is the average corrected current of the sth group of users, is the average corrected voltage of the sth group of users, PF s is the power factor of the sth group of users; To calculate the power factor PF of all users in the sth group of users, weighted multiplication is performed, and then the power consumption after normalization of power efficiency is calculated; Power s Represents the total net power load power of all users in the sth group of users;

[0165] S42: If the current signal of the zth user in the sth group of users monitored in real time is preprocessed in step S1 to obtain the corrected current signal Exceeds the real-time current limit threshold of the zth user in the sth group of users (That is, when ), then adjust the real-time supply voltage at the next moment This in turn reduces the amount of power supplied (i.e., reduces the available load power value supplied to the zth user in the sth group).

[0166] In step S42, the real-time power supply voltage at the next moment is adjusted The calculation formula is as follows:

[0167]

[0168]

[0169] in, is the corrected voltage signal obtained after the voltage signal of the zth user in the sth group of users monitored in real time is preprocessed in step S1; η u is the voltage regulation coefficient, η u =0.63; is the maximum current value of the zth user in the sth group of users within the monitoring period T.

[0170] like Figure 2 As shown, the present invention also provides a cloud computing-based power network business client using the above method, the client includes a signal acquisition and harmonic detection module, a harmonic compensation module, a user classification module and a power supply management module;

[0171] The signal acquisition and harmonic detection module is used to collect the electrical signals of the distribution network in real time, and perform harmonic detection after preprocessing them under the conditions of endpoint interpolation, continuity and boundary conditions;

[0172] The harmonic compensation module is used to perform harmonic compensation on the detected harmonics and control the dynamic compensation signals of multiple harmonics to be injected into the distribution network to obtain the pre-processed electrical signal Y k (f) Corrected electrical signal Y after multiple harmonic dynamic compensation k (t);

[0173] A user classification module is used to form multiple user portraits according to the customer's electricity consumption, and multiple user portraits are classified;

[0174] The power supply management module is used to dynamically adjust the power supply according to different user profile groups obtained by classification.

[0175] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0176] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0177] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for managing power network business clients based on cloud computing, characterized in that: The method The steps include: S1: Collect the electrical signals of the distribution network in real time, and perform harmonic detection after preprocessing them under the conditions of endpoint interpolation, continuity and boundary conditions; S2: Perform harmonic compensation on the detected harmonics, and control the dynamic compensation signals of multiple harmonics to be injected into the distribution network to obtain the preprocessed electrical signal Y k (f) Corrected electrical signal Y after multiple harmonic dynamic compensation k (t); S3: forming a plurality of user portraits according to the harmonic-compensated electrical signal in step S2, and classifying the plurality of user portraits; S4: Dynamically adjust the power supply according to the different user portrait groups obtained by classification.

2. The method for managing a power network business client based on cloud computing according to claim 1, characterized in that: The S1 step includes the following steps: S11: Real-time collection of electrical signals from the distribution network, including voltage signals and current signals, to obtain the original discrete electrical signal x k (n), n represents the nth discrete sampling point, n = 1, 2, ..., N k ; S12: The electrical signal x collected in real time in step S11 k (n) Perform interpolation preprocessing to obtain the preprocessed electrical signal X k (p); p represents the Nth discrete point of the electrical signal that can form a smooth curve after interpolation preprocessing, p = 1, 2, ..., P k ;P k is the total number of discrete points of the electrical signal that can form a smooth curve after interpolation preprocessing; P k >N k ; S13: To capture the electrical signal X after preprocessing in step S12 k (p) Harmonic components of different frequencies and time scales, constructing the window function ω k (p), the window function ω k The window length of (p) is P k , that is, the total number of discrete points when the window function is convolved with the electrical signal preprocessed in step S22; k = i or u; when k = i, ω i (p) is the window function for processing current signals; when k = u, ω u (p) is the window function for processing voltage signals; S14: The electrical signal X pre-processed in step S12 is k (p) Perform convolution operation with the corresponding window function constructed in step S22 to calculate the convolution result y k (p): Where q = 1, 12, .., P k ; S15: Perform short-time Fourier transform on the convolution result obtained in step S14 to obtain the frequency domain value Y of the preprocessed electrical signal k (f), the calculation formula for short-time Fourier transform is as follows: Where f is the frequency domain value Y of the preprocessed electrical signal obtained by short-time Fourier transform k (f) frequency; S16: Perform harmonic detection on the frequency domain value of the preprocessed electrical signal obtained in step S15, and the harmonic frequency of the e-th harmonic in the detected multiple harmonic data is f w，e (t), the harmonic amplitude is A w，e (t) and the harmonic phase is φ w，e (t), wherein t is the t-th moment in the frequency domain after the short-time Fourier transform conversion in step S15; e=1, 2, ..., E; that is, the number of harmonics detected is E in total, and w represents the fundamental harmonic wave.

3. The method for managing power network business clients based on cloud computing according to claim 2, characterized in that: The window function of the voltage signal processed in step S12 is: The window function of the current signal processed in step S12 is: Among them, ε k is the parameter that controls the shape of the window function. When k = u, ε u =0.5; when k=i, 4. The method for managing a power network business client based on cloud computing according to claim 2, characterized in that: In the step S12, the electrical signal x collected in real time in the step S11 is k (n) Perform interpolation preprocessing to obtain the preprocessed electrical signal X k (p), comprising the following steps: S121: Construct discrete point electrical signal set G: G={(1,x k (1)),(2,x k (2)),...,(n,x k (n)),...,(N k -1,x k (N k -1))}; S122: For any two data (n, x) in the discrete point electrical signal set G constructed in step S221, k (n)) and (n+1, x k (n+1)), construct the interpolation point electrical signal solution equation: X k (p) = a n (pn) 3 +b n (pn) 2 +c n (pn)+d n , where a n is the first coefficient, b n is the second coefficient, c n is the third coefficient, d n Solving for error values ​​for electrical signals; S123: Under the conditions of endpoint interpolation conditions, continuity conditions and boundary conditions, solve the first coefficient a of the interpolation point electrical signal solution equation constructed in step S122 n , the second coefficient b n , the third coefficient c n And the electrical signal solves the error value d n ; During the solution process, construct the solution equation system: Among them, h n-1 is the first coefficient to be solved, h n-1 = x′ k (n-1)-x′ k (n); h n is the second solution coefficient, h n = x′ k (n)-x′ k (n-1); h n+1 is the third solution coefficient, h n+1 = x′ k (n+1)-x′ k (n); m n-1 is the independent variable of the first system of equations, m n is the independent variable of the second system of equations, m n+1 is the independent variable of the third program group, The endpoint interpolation conditions are as follows: X k (n)=x k (n),X k (n+1)=x k (n+1) The continuity conditions are as follows: Because P k The pre-processed electrical signal data X k (p) is N k The original discrete electrical signal x k (n) After interpolation P k -N k interpolated points and combined with the original x k (n) is formed by combining each time point, so the result of the continuity condition is X k (n) = x k (n); The boundary conditions are as follows: S124: For (P k -N k ) interpolation points, repeat the steps S221-S223, and solve for N k The original discrete electrical signal x k (n) All interpolation point electrical signals X k (p).

5. The method for managing power network business clients based on cloud computing according to claim 2, characterized in that: The method for compensating harmonics in step S2 comprises the following steps: S21: Determine the e-th harmonic frequency f obtained according to step S16 w，e (t), harmonic amplitude A w,e (t) and harmonic phase φ w，e (t), determine the amplitude A of the dynamic compensation signal of the e-th harmonic com，e (t), phase φ com，e (t) and frequency f com，e (t): A com，e (t)=A w，e (t);φ com，e (t)=φ w，e +π;f com，e (t)=f h，e (t); S22: Determine the dynamic compensation signal of the harmonics of step S21 for each of the plurality of harmonics; S23: According to the dynamic compensation signals of the multiple harmonics obtained in step S22, controlling the dynamic compensation signals of the multiple harmonics to be injected into the distribution network; S24: After dynamic compensation of multiple harmonics, a pre-processed electrical signal Y containing a total of E harmonics is obtained k (f) Corrected electrical signal Y after multiple harmonic dynamic compensation k (t): Among them, A k,0 (t) is the fundamental amplitude of the preprocessed electrical signal; k = i or u, Y i (t) is the corrected current signal after multiple harmonic dynamic compensation, Y u (t) is the corrected voltage signal after multiple harmonic dynamic compensation.

6. The method for managing power network business clients based on cloud computing according to claim 1, characterized in that: The method for constructing a user portrait group in step S3 includes the following steps: S31: Calculate the corrected current signal Y during the monitored period T i (t) and the corrected voltage signal Y u (t) The average value of the corrected current Y i，avg (t), average value of corrected voltage Y u，avg (t) and the correction current variance σ i , Corrected voltage variance σ u ; t = 1, 2, ..., T; S32: Calculate the user's power factor PF according to the calculation result of step S31: S33: Calculate the maximum current value Y of the user in the monitored time period T according to the calculation result of step S31 i，peak and the maximum voltage value Y u，peak : Y i，peak =max| T (Y i (t)); max| T (Y i，avg (t)) is the maximum value function among all the current values ​​at the moment in the monitored period T; Y u，peak =max| T (Y u (t)); max| T (Y u (t)) is the maximum value function among all voltage values ​​at the time within the monitored period T; S34: Determine the peak power consumption period T of the user within the monitored period T peak ; T peak =t d -t c ;t d is the end time of the peak power consumption period, t c is the starting time of the peak land use period, t c ∈{1, 2, ..., T}, t d ∈{1, 2, ..., T}, and T ≥ t d >t c ≥1; S35: Based on the results of steps S31 to S34, construct a user portrait vector R: R=[Y i，avg (t),Y u，avg (t),σ i ,σ u ,PF,Y i，peak ,Y u，peak ,T peak ]; S36: Select a parameter in the user portrait vector R constructed in step S35 as a classification index to classify the user portrait vector.

7. The method for managing power network business clients based on cloud computing according to claim 6, characterized in that: The step S36 includes the following steps: S361: Select a parameter in the user portrait vector, and select the number of clusters S according to the parameter, and divide the user portrait vectors R of multiple users into different groups G s ; G s is the sth group of user portrait vectors, s = 1, 2, ..., S; the total number of user portrait vectors R is Z, R z is the zth user portrait vector; S362: Randomly select the initial cluster center of each group of user portrait vectors: Ce1, Ce2, ...Ce s , ..., Ce S ; Ce s is the user portrait vector G of the sth group s The central user portrait vector of S363: Calculate the zth user portrait vector R z and the sth group user portrait vector G s The central user portrait vector Ce s The Euclidean distance is: Among them g is the gth parameter of the user portrait vector R, The gth parameter o g The zth user portrait vector as an indicator, For the gth parameter o g The sth group user portrait vector G as an indicator s The central user portrait vector of o1 is Y i，avg (t); o2 is Y u，avg (t); o3 is σ i ; o4 is σ u ; o5 is PF; o6 is Y i，peak ; o7 is Y u，peak ; o8 is T peak ; S364: According to the calculation result of S363, the zth user portrait vector is assigned to the center user portrait vector Ce closest to it. l The user portrait vector G of the first group l , l∈{1, 2, ..., S}; That is, if D(R z , Ce l )=min{D(R z , Ce1), D(R z , Ce2), ..., D(R z , Ce S )}, then R z ∈G l ; S365: For each group of user portrait vectors, relocate and calculate its cluster center to obtain the updated value of the cluster center: Among them, Ce′ l The user portrait vector G of the first group calculated for re-positioning l The cluster center of G is updated to l ) is used to calculate the user portrait vector G of the first group l The user portrait vector R contained in z A function of the total number of ; S366: Repeat the steps S363-S365 until the cluster center update iteration threshold Q is less than 0.135, then stop the iteration and the user portrait classification is completed; 8. The method for managing power network business clients based on cloud computing according to claim 6, characterized in that: In step S4, dynamically adjusting the power supply according to different user portrait groups obtained by classification includes the following steps: S41: Calculate the load power of each group of users classified according to the user portrait vector: Among them, Y i s (t) is the average corrected current of the sth group of users, is the average corrected voltage of the sth group of users, PF s is the power factor of the sth group of users; To calculate the power factor PF of all users in the sth group of users, weighted multiplication is performed, and then the power consumption after normalization of power efficiency is calculated; Power s Represents the total net power load power of all users in the sth group of users; S42: If the current signal of the zth user in the sth group of users monitored in real time is preprocessed in step S1 to obtain the corrected current signal Exceeds the real-time current limit threshold of the zth user in the sth group of users Then adjust the real-time supply voltage at the next moment This reduces the amount of power supplied.

9. The method for managing power network business clients based on cloud computing according to claim 8, characterized in that: In the step S42, the real-time supply voltage at the next moment is adjusted. The calculation formula is as follows: in, is the corrected voltage signal obtained after the voltage signal of the zth user in the sth group of users monitored in real time is preprocessed in step S1; η u is the voltage regulation coefficient, η u =0.63; is the maximum current value of the zth user in the sth group of users within the monitoring period T.

10. A cloud computing-based power network business client using the method according to any one of claims 1 to 9, characterized in that: It includes signal acquisition and harmonic detection module, harmonic compensation module, user classification module and power supply management module; The signal acquisition and harmonic detection module is used to collect the electrical signals of the distribution network in real time, and perform harmonic detection after pre-processing them under the conditions of endpoint interpolation conditions, continuity conditions and boundary conditions; The harmonic compensation module is used to perform harmonic compensation on the detected harmonics and control the dynamic compensation signals of multiple harmonics to be injected into the distribution network to obtain the pre-processed electrical signal Y k (f) Corrected electrical signal Y after multiple harmonic dynamic compensation k (t); The user classification module is used to form multiple user portraits according to the customer's electricity consumption, and the multiple user portraits are classified; The power supply management module is used to dynamically adjust the power supply according to different user portrait groups obtained by classification.

Images