Low-voltage-side power consumption load data preprocessing method, device, and storage medium

Abstract

The application discloses a low-voltage-side electric load data preprocessing method and device and a storage medium, and belongs to the technical field of low-voltage-side electric power data processing. The method comprises the following steps: collecting voltage data and current data at an entrance of a monitoring environment in real time; performing metering preprocessing on the voltage data and the current data every preset time period to obtain a characteristic sample point vector; caching the characteristic sample point vector to obtain an original data matrix; calculating a standard data matrix according to the original data matrix; and performing characteristic importance analysis on the standard data matrix to obtain screened characteristic data. Under the premise of guaranteeing the low-voltage electric load state monitoring performance, the application reduces the matrix dimension to effectively reduce the amount of data to be processed, thereby effectively reducing the cost and power consumption of the edge-side product scheme.

Description

Technical Field

[0001] The present invention relates to the field of low-voltage side power data processing, and in particular to a method and device for preprocessing low-voltage side power load data, and a computer-readable storage medium. Background Art

[0002] In the deployment of edge-side applications for monitoring the operating status of power loads on the low-voltage side, there is a dilemma between cost, power consumption and performance. How to effectively reduce the cost and power consumption of edge-side product solutions while ensuring power consumption and performance has become one of the basic issues for the implementation and promotion of the technology.

[0003] In order to obtain more load information, monitoring equipment and methods usually consider a large number of characteristic variables, which directly leads to a surge in the amount of data to be processed. A method is needed to process the raw data to extract important key characteristic information for subsequent condition monitoring. Summary of the Invention

[0004] The main purpose of the present invention is to provide a preprocessing method for low-voltage side power load data, aiming to solve the technical problems in the existing technology of large amount of low-voltage side power data and high cost and power consumption of edge side product solutions.

[0005] To achieve the above object, the present invention provides a method for preprocessing low-voltage side power load data, the method comprising:

[0006] Real-time collection of voltage and current data at the entrance of the monitoring environment;

[0007] Performing metrological preprocessing on the voltage data and the current data at every preset time period to obtain a characteristic sample point vector;

[0008] Buffering the feature sample point vectors to obtain an original data matrix, and calculating a standard data matrix based on the original data matrix;

[0009] Performing feature importance analysis on the standard data matrix to obtain screened feature data.

[0010] Optionally, before the step of performing metering preprocessing on the voltage data and the current data at every preset time period, the method further includes:

[0011] The characteristic sample points are determined according to the requirements of the actual application scenario, and the preset time period is set in units of positive integer cycles.

[0012] Optionally, the characteristic sample points include:

[0013] Current RMS, voltage RMS, current preset subharmonics, voltage preset subharmonics, power factor, fundamental wave active power, fundamental wave reactive power.

[0014] Optionally, the step of caching the feature sample point vector to obtain an original data matrix includes:

[0015] According to the actual application scenario requirements and the sampling interval of the characteristic sample points, the number of sample points to be collected is set;

[0016] The number of the characteristic sample points is used as the internal characteristic number of the characteristic sample point vector, and the original data matrix is ​​determined according to the internal characteristic number and the acquisition quantity.

[0017] Optionally, the step of calculating a standard data matrix based on the original data matrix includes:

[0018] The original data matrix is ​​dimensionally processed, and the standard data matrix is ​​obtained by calculation based on the original data matrix after the dimensionless processing.

[0019] Optionally, the step of performing dimensionless processing on the original data matrix includes:

[0020] The mean and standard deviation of each matrix element of the original data matrix are calculated, and data preprocessing is performed based on the mean and the standard deviation to obtain the standard data matrix.

[0021] Optionally, before the step of performing feature importance analysis on the standard data matrix to obtain screened feature data, the method further includes:

[0022] An autocovariance matrix is ​​obtained according to the standard data matrix, and eigenvalue decomposition is performed on the autocovariance matrix to obtain eigenvalues ​​and eigenvectors.

[0023] Optionally, the step of performing feature importance analysis on the standard data matrix to obtain screened feature data includes:

[0024] Selecting a preset number of target eigenvalues ​​and inner vectors of the eigenvectors corresponding to the target eigenvalues ​​from the eigenvalues, and obtaining a transformation matrix based on the target eigenvalues ​​and the inner vectors;

[0025] A filtered data information matrix is ​​obtained based on the standard data matrix and the transformation matrix, and the characteristic data is stored in the data information matrix.

[0026] In addition, to achieve the above-mentioned purpose, the present invention also provides a preprocessing device for low-voltage side power load data, and the preprocessing device for low-voltage side power load data includes: a memory, a processor, and a preprocessing program for low-voltage side power load data stored on the memory and runnable on the processor. When the preprocessing program for low-voltage side power load data is executed by the processor, the steps of the preprocessing method for low-voltage side power load data as described above are implemented.

[0027] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, on which a preprocessing program for low-voltage side power load data is stored. When the preprocessing program for low-voltage side power load data is executed by a processor, the steps of the preprocessing method for low-voltage side power load data as described above are implemented.

[0028] The embodiments of the present invention propose a method, device, and computer-readable storage medium for preprocessing low-voltage power load data. These methods collect voltage and current data at the entrance of the monitoring environment in real time; perform metering preprocessing on the voltage and current data at preset time intervals to obtain a feature sample point vector; cache the feature sample point vector to obtain an original data matrix, and calculate a standard data matrix based on the original data matrix; and perform feature importance analysis on the standard data matrix to obtain filtered feature data. While ensuring the performance of low-voltage power load status monitoring, the matrix dimension is reduced to effectively reduce the amount of data to be processed, thereby effectively reducing the cost and power consumption of edge-side product solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a schematic diagram of the terminal structure of the hardware operating environment involved in the embodiment of the present invention;

[0030] Figure 2 The figure is a flow chart of an embodiment of a method for preprocessing low-voltage side power load data according to the present invention.

[0031] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0032] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0033] Reference Figure 1 , Figure 1 This is a schematic diagram of the terminal structure of the hardware operating environment involved in the embodiment of the present invention.

[0034] like Figure 1As shown, the terminal device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the user interface 1003 may optionally include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (WIreless-FIdelity, WI-FI) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM) memory, or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0035] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the terminal device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0036] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a data storage module, a network communication module, a user interface module, and a pre-processing program for low-voltage side power load data.

[0037] exist Figure 1 In the terminal device shown, the network interface 1004 is mainly used for data communication with other devices; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the terminal device of the present invention can be set in the terminal device, and the terminal device calls the preprocessing program of the low-voltage side power load data stored in the memory 1005 through the processor 1001 and performs the following operations:

[0038] Real-time collection of voltage and current data at the entrance of the monitoring environment;

[0039] Performing metrological preprocessing on the voltage data and the current data at every preset time period to obtain a characteristic sample point vector;

[0040] Buffering the feature sample point vectors to obtain an original data matrix, and calculating a standard data matrix based on the original data matrix;

[0041] Performing feature importance analysis on the standard data matrix to obtain screened feature data.

[0042] Furthermore, the processor 1001 may call a pre-processing program for the low-voltage side power load data stored in the memory 1005, and further perform the following operations:

[0043] Before the step of performing metering preprocessing on the voltage data and the current data at every preset time period, the method further includes:

[0044] The characteristic sample points are determined according to the requirements of the actual application scenario, and the preset time period is set in units of positive integer cycles.

[0045] Furthermore, the processor 1001 may call a pre-processing program for the low-voltage side power load data stored in the memory 1005, and further perform the following operations:

[0046] The characteristic sample points include:

[0047] Current RMS, voltage RMS, current preset subharmonics, voltage preset subharmonics, power factor, fundamental wave active power, fundamental wave reactive power.

[0048] Furthermore, the processor 1001 may call a pre-processing program for the low-voltage side power load data stored in the memory 1005, and further perform the following operations:

[0049] The step of caching the feature sample point vector to obtain the original data matrix includes:

[0050] According to the actual application scenario requirements and the sampling interval of the characteristic sample points, the number of sample points to be collected is set;

[0051] The number of the characteristic sample points is used as the internal characteristic number of the characteristic sample point vector, and the original data matrix is ​​determined according to the internal characteristic number and the acquisition quantity.

[0052] Furthermore, the processor 1001 may call a pre-processing program for the low-voltage side power load data stored in the memory 1005, and further perform the following operations:

[0053] The step of calculating the standard data matrix based on the original data matrix includes:

[0054] The original data matrix is ​​dimensionally processed, and the standard data matrix is ​​obtained by calculation based on the original data matrix after the dimensionless processing.

[0055] Furthermore, the processor 1001 may call a pre-processing program for the low-voltage side power load data stored in the memory 1005, and further perform the following operations:

[0056] The step of performing dimensionless processing on the original data matrix comprises:

[0057] The mean and standard deviation of each matrix element of the original data matrix are calculated, and data preprocessing is performed based on the mean and the standard deviation to obtain the standard data matrix.

[0058] Furthermore, the processor 1001 may call a pre-processing program for the low-voltage side power load data stored in the memory 1005, and further perform the following operations:

[0059] Before the step of performing feature importance analysis on the standard data matrix to obtain screened feature data, the method further includes:

[0060] An autocovariance matrix is ​​obtained according to the standard data matrix, and eigenvalue decomposition is performed on the autocovariance matrix to obtain eigenvalues ​​and eigenvectors.

[0061] Furthermore, the processor 1001 may call a pre-processing program for the low-voltage side power load data stored in the memory 1005, and further perform the following operations:

[0062] The step of performing feature importance analysis on the standard data matrix to obtain filtered feature data includes:

[0063] Selecting a preset number of target eigenvalues ​​and inner vectors of the eigenvectors corresponding to the target eigenvalues ​​from the eigenvalues, and obtaining a transformation matrix based on the target eigenvalues ​​and the inner vectors;

[0064] A filtered data information matrix is ​​obtained based on the standard data matrix and the transformation matrix, and the characteristic data is stored in the data information matrix.

[0065] The embodiment of the present invention provides a method for preprocessing low-voltage side power load data, referring to Figure 2 , Figure 2 This is a flow chart of a first embodiment of a method for preprocessing low-voltage side power load data according to the present invention.

[0066] In this embodiment, the preprocessing method of the low-voltage side power load data includes:

[0067] Step S10: collecting voltage data and current data at the entrance of the monitoring environment in real time.

[0068] Generally speaking, the function of a transformer is to reduce the voltage of electricity transmitted at high voltage (to save materials and reduce losses) (sometimes it is boosted, usually at power plant substations) to meet the standard for conventional electrical equipment. Therefore, on a conventional step-down transformer, the input end (also called the primary) is the high-voltage side, and the output end (also called the secondary) is the low-voltage side. In this embodiment, the entrance to the monitoring environment is the single-phase power connection point for home users. The edge computing used in this embodiment is a distributed computing architecture. In this architecture, the computation of applications, data, and services is moved from network central nodes (such as cloud servers) to logically edge nodes (such as data collection devices) within the network. In other words, edge computing breaks down large services that were originally handled entirely by central nodes into smaller, more manageable components, which are then distributed to edge nodes for processing. Edge nodes are closer to user terminals, which can speed up data processing and transmission and reduce latency. The application in this embodiment is deployed on end-node devices, which belongs to the edge computing model. However, since the voltage and current data collected at the entrance of the monitoring environment are of many types and huge in quantity, they need to be preprocessed to reduce the data volume and increase the speed of edge computing.

[0069] Step S20: performing metrological preprocessing on the voltage data and the current data at every preset time period to obtain a characteristic sample point vector.

[0070] Step S30: caching the feature sample point vectors to obtain an original data matrix, and calculating a standard data matrix based on the original data matrix.

[0071] Step S40: performing feature importance analysis on the standard data matrix to obtain filtered feature data.

[0072] Preprocessing refers to the calculation of the basic feature quantities required by the algorithm, namely the feature sample point vector, from the basic voltage and current data obtained through sampling.

[0073] A feature sample point vector is a one-dimensional matrix, and multiple continuous feature sample point vectors form a two-dimensional matrix. The two-dimensional matrix is ​​used as the original data matrix, and then dimensionless processing is performed on the original data matrix to obtain the standard data matrix.

[0074] Among them, dimension refers to the basic property of physical quantity. Physical quantities can be divided into basic quantities and derived quantities. Basic quantities are physical quantities with independent dimensions, and derived quantities refer to physical quantities whose dimensions can be expressed as a combination of the dimensions of basic quantities; all derived quantities can be derived from basic quantities, thereby establishing a functional relationship between the entire physical quantities, and this functional relationship is usually called a measurement system. The International System of Units (SI for short) has selected a measurement system consisting of 7 basic quantities, and derived quantities can be derived using these 7 basic quantities. The dimensions of the 7 basic quantities are respectively expressed by length L, mass M, time T, current intensity I, temperature Θ, amount of substance n, and light intensity J. Then, for the dimension of any derived quantity, dimA=L α M β T γ I δ Θ ε n ζ J η , which is the general formula for dimensions. The exponents α, β, γ, etc. in the formula are called dimension indices. A physical quantity with all exponents equal to zero is called a dimensionless quantity.

[0075] The feature importance analysis is performed on the standard data matrix obtained by dimensionless processing to obtain the filtered feature data.

[0076] In this embodiment, voltage and current data at the entrance of the monitoring environment are collected in real time; the voltage and current data are pre-processed at preset intervals to obtain a feature sample point vector; the feature sample point vector is cached to obtain a raw data matrix, and a standard data matrix is ​​calculated based on the raw data matrix; and feature importance analysis is performed on the standard data matrix to obtain filtered feature data. While ensuring the performance of low-voltage power load status monitoring, the matrix dimension is reduced to effectively reduce the amount of data to be processed, thereby effectively reducing the cost and power consumption of edge-side product solutions.

[0077] Optionally, before the step of performing metering preprocessing on the voltage data and the current data at every preset time period, the method further includes:

[0078] The characteristic sample points are determined according to the requirements of the actual application scenario, and the preset time period is set in units of positive integer cycles.

[0079] Before the voltage data and current data are metered and preprocessed at preset time intervals, characteristic sample points are determined according to the requirements of the actual application scenario, that is, according to the requirements of the actual application scenario for real-time algorithm processing. The characteristic sample points in this embodiment include but are not limited to: current effective value, voltage effective value, current preset order (1-31) harmonics, voltage preset order (1-31) harmonics, power factor, fundamental wave active power, fundamental wave reactive power, etc.

[0080] And according to the needs of the actual application scenario, that is, according to the requirements of the actual application scenario for real-time algorithm processing, the preset time period is set in units of positive integer cycles, that is, the preset time period is set in units of N cycles, where the parameter N is a positive integer and is an adjustable parameter. Among them, the time it takes for the alternating current to complete a complete change (i.e., a sine waveform) is called a cycle. The time required to complete each cycle of change is called a period. It is represented by the symbol T, and the unit is seconds. The number of cycles of the alternating current change per second is called frequency. It is represented by the symbol f, and the unit is Hertz or cycle / second. The number of cycles is the frequency, that is, the number of fluctuations in the fluctuation process per unit time, that is, the regular frequency of a cycle of fluctuation process.

[0081] Optionally, the characteristic sample points include:

[0082] Current RMS, voltage RMS, current preset subharmonics, voltage preset subharmonics, power factor, fundamental wave active power, fundamental wave reactive power.

[0083] Sample feature points include but are not limited to: current effective value, voltage effective value, current preset subharmonics, voltage preset subharmonics, power factor, fundamental active power, fundamental reactive power, etc. Sample feature points can be increased or decreased according to the actual application scenario requirements. Because different devices require different features for detection and identification, the corresponding feature information can be increased or decreased according to the requirements for the detection device category in different application scenarios.

[0084] The effective value is also called the "root mean square value", a value used to measure the size of alternating current. When alternating current passes through a resistor, the heat generated in one cycle is equal to the heat generated by direct current passing through the resistor in the same period of time. The value of this direct current is the effective value of the alternating current.

[0085] Harmonic current is the collective name for all sinusoidal components whose frequencies are integer multiples of the original periodic current frequency when a non-sinusoidal periodic current function is expanded according to the Fourier series. A harmonic current with a frequency equal to k times the original periodic current frequency is called a kth harmonic current, and harmonic currents with k greater than 1 are also collectively referred to as higher-order harmonic currents. A periodic signal can be decomposed into a linear superposition of a DC component c0 and sinusoidal signals of different frequencies using Fourier transform:

[0086]

[0087] ,in, is the expression of the mth harmonic, c m Represents the amplitude of the mth harmonic, its angular frequency is mω, and its initial phase is Its effective value is c m / √2. When m=1, is the expression of the fundamental component, its angular frequency is ω, and the initial phase is Its root mean square value, c1 / √2, is called the fundamental effective value. ω / 2π is the frequency of the fundamental component, also known as the fundamental frequency. The frequency of the fundamental component is equal to the frequency of the AC signal. The frequency of the mth harmonic is an integer multiple (m times) of the fundamental frequency. Harmonic current is a collective term for sinusoidal components whose frequencies are integer multiples of the original periodic current frequency. Power supply system harmonics are defined as the Fourier series decomposition of periodic non-sinusoidal electrical quantities. In addition to components with the same frequency as the grid's fundamental frequency, a series of components greater than the grid's fundamental frequency are obtained. These components are called harmonics. The ratio of the harmonic frequency to the fundamental frequency (n = fn / f1) is called the harmonic order. Non-integer multiples of harmonics sometimes exist in power grids, known as non-harmonics or fractional harmonics. Harmonics are actually a type of interference that "pollutes" the power grid. The field of electrical engineering primarily studies the generation, transmission, measurement, harm, and suppression of harmonics, whose frequency range is generally 2 ≤ n ≤ 40. The 1st to 31st harmonics of voltage and current are selected because the harmonics in this range can basically represent the main characteristic information and meet application requirements.

[0088] Active power, also known as average power, is the instantaneous power of AC. The instantaneous power of AC is not a constant value; the average value over a cycle is called active power. It refers to the power consumed by the resistor in a circuit, represented by the letter P, and is measured in watts (W). P = U * I * cosφ, which is the rated voltage multiplied by the rated current multiplied by the power factor. Apparent power, on the other hand, is the product of voltage and current in a circuit with resistance and reactance. It is represented by the letter S and is measured in volt-amperes (VA). S = UI.

[0089] Reactive power refers to the energy storage components in a circuit with inductors and capacitors converting power supply energy into magnetic (or electric) field energy during half a cycle and returning the stored magnetic (or electric) field energy to the power supply during the other half. They merely exchange energy with the power supply, not actually consume it. The amplitude of the rate of energy exchange with the power supply is called reactive power. It is represented by the letter Q and is measured in vars (Var). Q = U*I*sinφ.

[0090] The power factor refers to the ratio of active power to apparent power in an AC circuit. For a given voltage and power, the higher this value, the better the efficiency of user electrical equipment and the more fully utilized the power generation equipment. In a DC circuit, active power is calculated by multiplying voltage by current. However, in an AC circuit, voltage multiplied by current is apparent power, and the portion of power that actually performs work (i.e., active power) is less than the apparent power. The ratio of active power to apparent power is called power factor, expressed as cosφ. COSφ = P / S. Fundamental power factor (DPF)—the ratio of fundamental active power to apparent power—is the cosine of the angle between the fundamental current and the fundamental voltage (Cosφ). Full power factor (PF)—the calculation uses the entire RMS value, meaning it includes the fundamental and all harmonics. If harmonics are present, the PF will be less than the DPF.

[0091] Optionally, the step of caching the feature sample point vector to obtain an original data matrix includes:

[0092] According to the actual application scenario requirements and the sampling interval of the characteristic sample points, the number of sample points to be collected is set;

[0093] The number of the characteristic sample points is used as the internal characteristic number of the characteristic sample point vector, and the original data matrix is ​​determined according to the internal characteristic number and the acquisition quantity.

[0094] In this embodiment, the original data matrix O is defined as follows:

[0095]

[0096] , where m is the number of characteristic sample points used as the intrinsic characteristic number of the characteristic sample point vector, and n is the number of characteristic sample points collected as set by the device input. In this embodiment, there are seven intrinsic characteristics: current RMS value, voltage RMS value, current preset subharmonic, voltage preset subharmonic, power factor, fundamental active power, and fundamental reactive power.

[0097] Optionally, the step of calculating a standard data matrix based on the original data matrix includes:

[0098] The original data matrix is ​​dimensionally processed, and the standard data matrix is ​​obtained by calculation based on the original data matrix after the dimensionless processing.

[0099] When calculating the standard data matrix S based on the original data matrix O, the original data matrix O is first dimensionally processed, and the standard data matrix S is calculated based on the original data matrix O after the dimensionless processing.

[0100] Optionally, the step of performing dimensionless processing on the original data matrix includes:

[0101] The mean and standard deviation of each matrix element of the original data matrix are calculated, and data preprocessing is performed based on the mean and the standard deviation to obtain the standard data matrix.

[0102] The specific calculation process of dimensionless processing of the original data matrix O is:

[0103] First, calculate the average of each matrix element of the original data matrix O,

[0104]

[0105] , and then calculate the standard deviation of each matrix element of the original data matrix O,

[0106]

[0107] , then perform data preprocessing based on the above mean and standard deviation,

[0108]

[0109] , get the standard data matrix S,

[0110]

[0111] Optionally, before the step of performing feature importance analysis on the standard data matrix to obtain screened feature data, the method further includes:

[0112] An autocovariance matrix is ​​obtained according to the standard data matrix, and eigenvalue decomposition is performed on the autocovariance matrix to obtain eigenvalues ​​and eigenvectors.

[0113] After obtaining the standard data matrix S, and before performing feature importance analysis on the standard data matrix S to obtain the filtered feature data, the specific calculation process for obtaining the eigenvalues ​​and eigenvectors is as follows:

[0114] According to the standard data matrix S, the autocovariance matrix C is obtained.

[0115]

[0116] Then, the autocovariance matrix C is subjected to eigenvalue decomposition to obtain the eigenvalue λ and eigenvector P, thereby obtaining the eigenvalue λ and eigenvector P that support the main feature analysis.

[0117] C m×m =P T λP

[0118]

[0119] Optionally, the step of performing feature importance analysis on the standard data matrix to obtain screened feature data includes:

[0120] Selecting a preset number of target eigenvalues ​​and inner vectors of the eigenvectors corresponding to the target eigenvalues ​​from the eigenvalues, and obtaining a transformation matrix based on the target eigenvalues ​​and the inner vectors;

[0121] A filtered data information matrix is ​​obtained based on the standard data matrix and the transformation matrix, and the characteristic data is stored in the data information matrix.

[0122] After obtaining the eigenvalue λ and eigenvector P, the feature importance analysis is performed on the standard data matrix S to obtain the filtered feature data. The specific calculation process is as follows:

[0123] Select the preset number k target eigenvalues ​​with the largest weight among the eigenvalues ​​λ, and the inner vector of the eigenvector P corresponding to the target eigenvalue, so as to obtain the transformation matrix P for reducing the amount of data based on the target eigenvalue and the inner vector k×n .

[0124] According to the standard data matrix S and transformation matrix P k×n , get the filtered data information matrix, store the feature data in the data information matrix, and get the final filtered data information S * :

[0125] S * =P k×n ×(S m×n ) T

[0126] In addition, an embodiment of the present invention also provides a preprocessing device for low-voltage side power load data, which includes: a memory, a processor, and a preprocessing program for low-voltage side power load data stored in the memory and runnable on the processor. When the preprocessing program for low-voltage side power load data is executed by the processor, the steps of the preprocessing method for low-voltage side power load data as described above are implemented.

[0127] In addition, an embodiment of the present invention also provides a computer-readable storage medium, on which a preprocessing program for low-voltage side power load data is stored. When the preprocessing program for low-voltage side power load data is executed by a processor, the steps of the preprocessing method for low-voltage side power load data as described above are implemented.

[0128] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or system comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or system. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or system comprising the element.

[0129] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0130] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better embodiment. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0131] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for preprocessing low-voltage side power load data, characterized in that: The method for preprocessing low-voltage side power load data comprises the following steps: Real-time collection of voltage and current data at the entrance of the monitoring environment; Performing metrological preprocessing on the voltage data and the current data at every preset time period to obtain a characteristic sample point vector; caching the feature sample point vectors to obtain an original data matrix, performing dimensionless processing on the original data matrix, and performing calculation based on the dimensionless processed original data matrix to obtain a standard data matrix; Obtain an autocovariance matrix according to the standard data matrix, and perform eigenvalue decomposition on the autocovariance matrix to obtain eigenvalues ​​and eigenvectors; Selecting a preset number of target eigenvalues ​​and inner vectors of the eigenvectors corresponding to the target eigenvalues ​​from the eigenvalues, and obtaining a transformation matrix based on the target eigenvalues ​​and the inner vectors; A filtered data information matrix is ​​obtained based on the standard data matrix and the transformation matrix, and characteristic data is stored in the data information matrix.

2. The method for preprocessing low-voltage side power load data according to claim 1, characterized in that: Before the step of performing metering preprocessing on the voltage data and the current data at every preset time period, the method further includes: The characteristic sample points are determined according to the requirements of the actual application scenario, and the preset time period is set in units of positive integer cycles.

3. The method for preprocessing low-voltage side power load data according to claim 2, characterized in that: The characteristic sample points include: Current RMS, voltage RMS, current preset subharmonics, voltage preset subharmonics, power factor, fundamental wave active power, fundamental wave reactive power.

4. The method for preprocessing low-voltage side power load data according to claim 3, characterized in that: The step of caching the feature sample point vector to obtain the original data matrix includes: According to the actual application scenario requirements and the sampling interval of the characteristic sample points, the number of sample points to be collected is set; The number of the characteristic sample points is used as the internal characteristic number of the characteristic sample point vector, and the original data matrix is ​​determined according to the internal characteristic number and the acquisition quantity.

5. The method for preprocessing low-voltage side power load data according to claim 1, characterized in that: The step of performing dimensionless processing on the original data matrix comprises: The mean and standard deviation of each matrix element of the original data matrix are calculated, and data preprocessing is performed based on the mean and the standard deviation to obtain the standard data matrix.

6. A preprocessing device for low-voltage side power load data, characterized in that: The preprocessing device for the low-voltage side power load data includes: a memory, a processor, and a preprocessing program for the low-voltage side power load data stored in the memory and runnable on the processor. The preprocessing program for the low-voltage side power load data is configured to implement the steps of the preprocessing method for the low-voltage side power load data as described in any one of claims 1 to 5.

7. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a preprocessing program for low-voltage side power load data, which, when executed by a processor, implements the steps of the preprocessing method for low-voltage side power load data according to any one of claims 1 to 5.

Images