CN-122026410-A - Load identification method and system for photovoltaic spontaneous self-use scene of user

CN122026410ACN 122026410 ACN122026410 ACN 122026410ACN-122026410-A

Abstract

The application provides a load identification method and system of a photovoltaic spontaneous utilization scene of a user, and the method comprises the steps of obtaining electric parameter data of the user to be identified and adjacent known photovoltaic users in a set time period, wherein the electric parameter data comprises a current effective value sequence, a power sequence and a harmonic wave sequence, detecting a load switching event of the user to be identified based on the electric parameter data, judging whether photovoltaic access exists or not, responding to the photovoltaic access, calculating the photovoltaic capacity of the user to be identified based on the electric parameter data and the photovoltaic capacity of the adjacent known photovoltaic users, processing the power sequence and the harmonic wave sequence based on the photovoltaic capacity of the user to be identified to separate to obtain a pure load power sequence and a pure load harmonic wave sequence, determining steady state characteristics corresponding to each load switching event based on the pure load power sequence, the pure load harmonic wave sequence and the current effective value sequence, and outputting a load identification result corresponding to each load switching event to a pre-trained deep learning fusion network.

Inventors

LIU XINGQI
ZHANG YUPENG
ZHU ZIXU
HAN YUE
DOU JIAN
QIE SHUANG
SHANG HUAIYING
ZHANG HAINING
WANG HAOTIAN
LI JIADAI
WANG JIADONG
LIU XUAN
MIAO HONGYING
YAN KAI
ZHOU FENGHUA
WANG YANQIN
ZHENG ANGANG
Du Yina
CHEN JI
TANG YUE
WU ZHONGXING

Assignees

中国电力科学研究院有限公司
国网冀北电力有限公司承德供电公司

Dates

Publication Date: 20260512
Application Date: 20251215

Claims (10)

1. The utility model provides a load identification method of a photovoltaic spontaneous self-use scene of a user, which is characterized by comprising the following steps: S1, acquiring electric parameter data of a user to be identified and an adjacent known photovoltaic user in a set time period, wherein the electric parameter data comprises a current effective value sequence, a power sequence and a harmonic sequence; s2, detecting a load switching event of the user to be identified based on the electrical parameter data, and judging whether photovoltaic access exists or not; S3, responding to the judgment that photovoltaic access exists, calculating the photovoltaic capacity of the user to be identified based on the electrical parameter data and the photovoltaic capacity of the adjacent known photovoltaic user, and processing the power sequence and the harmonic sequence based on the photovoltaic capacity of the user to be identified so as to obtain a pure load power sequence and a pure load harmonic sequence in a separation mode; s4, determining steady-state characteristics corresponding to each load switching event based on the pure load power sequence, the pure load harmonic sequence and the current effective value sequence, wherein the steady-state characteristics comprise a waveform chart representing the fluctuation degree of the steady-state characteristics and multidimensional electric characteristics; s5, inputting the waveform diagram and the multidimensional electrical characteristics corresponding to each load switching event into a pre-trained deep learning fusion network, and outputting a load identification result corresponding to each load switching event.
2. The method for identifying the load of the user photovoltaic autonomous and self-service scenario according to claim 1, wherein the step S2 of detecting the load switching event of the user to be identified and judging whether the photovoltaic access exists based on the electrical parameter data comprises: s21, detecting at least one load switching event and corresponding event time based on the current effective value sequence and the harmonic sequence of the user to be identified; S22, cutting power sequences of the user to be identified and the adjacent known photovoltaic users by taking the corresponding event time as a reference for each load switching event, and performing time sequence alignment and splicing on the cut power sequences; S23, normalizing the spliced power sequences of the users to be identified and the spliced power sequences of the adjacent known photovoltaic users, performing correlation analysis on the normalized two power sequences, and judging whether the users to be identified have photovoltaic access or not according to analysis results.
3. The method according to claim 2, wherein calculating the photovoltaic capacity of the user to be identified based on the electrical parameter data and the photovoltaic capacity of the neighboring known photovoltaic users in step S3 comprises: s31, determining a power peak ratio between the two power sequences based on the spliced power sequences of the users to be identified and the spliced power sequences of the adjacent known photovoltaic users; s32, determining the photovoltaic capacity of the user to be identified based on the photovoltaic capacity of the adjacent known photovoltaic users and the power peak ratio.
4. The method for identifying the load of the photovoltaic autonomous and self-service scenario according to claim 2, wherein the step S21 of detecting at least one load switching event and a corresponding event time based on the current effective value sequence and the harmonic sequence of the user to be identified comprises: calculating the difference value of the current effective value sequence of the user to be identified at adjacent sampling moments based on the current effective value sequence of the user to be identified, and judging that a current abrupt change event occurs if the absolute value of the difference value is larger than a first preset threshold value; Calculating the difference value of an nth harmonic sequence at adjacent sampling moments based on the harmonic sequence of the user to be identified, and judging that a harmonic transition event occurs if the absolute value of the difference value is larger than a second preset threshold value, wherein the harmonic sequence of the user to be identified comprises an N-th harmonic sequence, N is a positive integer larger than or equal to 1, and N is a positive integer with a value between 1 and N; and when the moment of the current abrupt event and the harmonic transition event are simultaneously satisfied, marking the moment as the event moment of the load switching event.
5. The method for identifying the load of the user photovoltaic autonomous and self-service scene according to claim 2, wherein in the step S23, correlation analysis is performed on the normalized two power sequences, and determining whether the user to be identified has photovoltaic access according to the analysis result comprises: calculating the correlation coefficients of two power sequences by adopting the pearson correlation coefficient or cosine similarity; And if the correlation coefficient is greater than or equal to a preset correlation threshold, judging that the user to be identified has photovoltaic access.
6. The method for identifying the load of the user photovoltaic autonomous photovoltaic scenario according to claim 1, wherein determining the steady-state feature corresponding to each load switching event in step S4 based on the pure load power sequence, the pure load harmonic sequence and the current effective value sequence comprises: For each load switching event, determining a pre-event steady-state time and a post-event steady-state time based on the current effective value sequence by taking the corresponding event time as a reference; Extracting power values corresponding to the steady-state moment before the event and the steady-state moment after the event from the pure load power sequence, and obtaining steady-state power characteristic data through subtraction; Extracting power values corresponding to the steady-state moment before the event and the steady-state moment after the event from the pure load harmonic sequence, and obtaining steady-state harmonic characteristic data through subtraction; And determining a steady-state characteristic corresponding to a load switching event based on the steady-state power characteristic data and the steady-state harmonic characteristic data, wherein the steady-state characteristic comprises a waveform diagram representing the fluctuation degree of the steady-state characteristic and a multidimensional electrical characteristic.
7. The method for identifying the load of the user photovoltaic spontaneous and self-use scene according to claim 1 or 6, wherein the pre-trained deep learning fusion network comprises an image processing sub-network, a feature processing sub-network and an adaptive fusion module; the image processing sub-network is used for carrying out convolution processing on an input waveform diagram representing the fluctuation degree of the steady-state characteristic and outputting first load class probability distribution; The characteristic processing sub-network is used for carrying out gating attention processing on the input multidimensional electric characteristics and outputting second load category probability distribution; the self-adaptive fusion module is used for extracting a first characteristic vector of the image processing sub-network before outputting probability distribution and a second characteristic vector of the characteristic processing sub-network before outputting probability distribution, fusing the first characteristic vector and the second characteristic vector, generating a dynamic weight vector according to a fusion result, and carrying out weighted summation on the first load type probability distribution and the second load type probability distribution by utilizing the dynamic weight vector to obtain final probability distribution.
8. The method for identifying the load of the user photovoltaic spontaneous use scene according to claim 7, wherein the characteristic processing sub-network comprises a gating unit, a self-attention layer and a multi-layer perceptron classifier; The gating unit is used for processing the input multidimensional electric characteristics to generate corresponding multidimensional gating vectors, and multiplying the multidimensional gating vectors by the multidimensional electric characteristics element by element to obtain feature vectors after gating selection; The self-attention layer is used for calculating the interrelationship among the features based on the feature vector after the gating selection and generating an attention weight vector, and multiplying the attention weight vector and the feature vector after the gating selection element by element to obtain a weighted feature vector; And the multi-layer perceptron classifier is used for classifying the weighted feature vectors and outputting the second probability distribution.
9. A load identification system for a photovoltaic autonomous self-service scenario of a user, the system comprising: the data acquisition module is used for acquiring electric parameter data of a user to be identified and an adjacent known photovoltaic user in a set time period, wherein the electric parameter data comprises a current effective value sequence, a power sequence and a harmonic sequence; The event detection and photovoltaic identification module is used for detecting the load switching event of the user to be identified based on the electrical parameter data and judging whether photovoltaic access exists or not; The capacity estimation and photovoltaic stripping module is used for responding to the judgment that photovoltaic access exists, calculating the photovoltaic capacity of the user to be identified based on the electrical parameter data and the photovoltaic capacity of the adjacent known photovoltaic users, and processing the power sequence and the harmonic sequence based on the photovoltaic capacity of the user to be identified so as to obtain a pure load power sequence and a pure load harmonic sequence in a separation mode; The characteristic extraction module is used for determining steady-state characteristics corresponding to each load switching event based on the pure load power sequence, the pure load harmonic sequence and the current effective value sequence, wherein the steady-state characteristics comprise a waveform chart representing the fluctuation degree of the steady-state characteristics and multidimensional electric characteristics; The load identification module is used for inputting the waveform diagram and the multidimensional electrical characteristics corresponding to each load switching event into the pre-trained deep learning fusion network and outputting the load identification result corresponding to each load switching event.
10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor performs the steps of the method for load identification of a photovoltaic autonomous utility scene of a user according to any of claims 1-8 when the program is executed.

Description

Load identification method and system for photovoltaic spontaneous self-use scene of user Technical Field The application relates to the technical field of power grids, in particular to a load identification method, a system, electronic equipment and a storage medium of a photovoltaic spontaneous self-use scene of a user. Background With the high-speed popularization of distributed photovoltaics, partial users privately install photovoltaics under the condition of unreported or unreported, so that the problems of inaccurate metering of a platform area, distortion of a load model, abnormal line loss, failure in regulation and control and the like are caused. The existing photovoltaic detection method mainly relies on macroscopic data such as voltage abnormality, reverse tide, low-frequency energy analysis and the like, and hidden photovoltaic with small capacity and hidden characteristics in a load is difficult to accurately identify. In addition, photovoltaic access can alter current waveforms and harmonic characteristics, making it difficult for conventional non-invasive load monitoring (NILM) algorithms to accurately resolve load events in a hybrid scenario. Existing NILM methods generally assume that the system contains only user load, without considering the transient-steady state aliasing problem caused by photovoltaic power generation and load superposition. Once photovoltaic access exists, load events are covered by steady-state ripple waves and harmonic waves of the inverter, so that feature extraction is invalid, and identification accuracy cannot be guaranteed. Therefore, in view of the above-described problems, a new method having both photovoltaic identification, photovoltaic lift-off and high-precision load identification capabilities is needed. Disclosure of Invention The embodiment of the application provides a load identification method and a system for a user photovoltaic spontaneous self-use scene, which solve the problems that the traditional NILM cannot separate generation power, harmonic waves and load events under the condition of privately installing photovoltaic, so that false identification and missing identification are caused. In order to achieve the above purpose, the application adopts the following technical scheme: in a first aspect, the present application provides a method for identifying loads of a photovoltaic autonomous self-service scene of a user, the method comprising: S1, acquiring electric parameter data of a user to be identified and an adjacent known photovoltaic user in a set time period, wherein the electric parameter data comprises a current effective value sequence, a power sequence and a harmonic sequence; S2, detecting a load switching event of a user to be identified based on the electrical parameter data, and judging whether photovoltaic access exists or not; S3, responding to the judgment that photovoltaic access exists, calculating the photovoltaic capacity of the user to be identified based on the electrical parameter data and the photovoltaic capacity of the adjacent known photovoltaic users, and processing the power sequence and the harmonic sequence based on the photovoltaic capacity of the user to be identified so as to obtain a pure load power sequence and a pure load harmonic sequence in a separating way; s4, determining steady-state characteristics corresponding to each load switching event based on the pure load power sequence, the pure load harmonic sequence and the current effective value sequence, wherein the steady-state characteristics comprise a waveform chart representing the fluctuation degree of the steady-state characteristics and multidimensional electrical characteristics; s5, inputting the waveform diagram and the multidimensional electrical characteristics corresponding to each load switching event into a pre-trained deep learning fusion network, and outputting a load identification result corresponding to each load switching event. In a second aspect, the present application provides a load identification system for a photovoltaic autonomous utility scenario of a user, the system comprising: the data acquisition module is used for acquiring electric parameter data of a user to be identified and an adjacent known photovoltaic user in a set time period, wherein the electric parameter data comprises a current effective value sequence, a power sequence and a harmonic sequence; the event detection and photovoltaic identification module is used for detecting a load switching event of a user to be identified based on the electrical parameter data and judging whether photovoltaic access exists or not; The capacity estimation and photovoltaic stripping module is used for responding to the judgment that photovoltaic access exists, calculating the photovoltaic capacity of the user to be identified based on the electrical parameter data and the photovoltaic capacity of the adjacent known photovoltaic users, and processing the power sequence an