CN115841338A - Method and device for determining abnormal electricity utilization behavior and non-volatile storage medium - Google Patents
Method and device for determining abnormal electricity utilization behavior and non-volatile storage medium Download PDFInfo
- Publication number
- CN115841338A CN115841338A CN202211537869.0A CN202211537869A CN115841338A CN 115841338 A CN115841338 A CN 115841338A CN 202211537869 A CN202211537869 A CN 202211537869A CN 115841338 A CN115841338 A CN 115841338A
- Authority
- CN
- China
- Prior art keywords
- target
- information data
- electricity utilization
- target object
- electricity
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The application discloses a method and a device for determining abnormal electricity utilization behaviors and a nonvolatile storage medium. Wherein, the method comprises the following steps: acquiring power utilization information data of a target object; extracting target electricity utilization information data in the electricity utilization information data, wherein the target electricity utilization information data is used for indicating the electricity utilization behavior of a target object; processing the quantized value of the target electricity utilization information data by adopting a pre-trained clustering model to obtain electricity utilization behavior data of a target object; and determining whether the target object has abnormal electricity utilization behavior according to the electricity utilization behavior data of the target object. The method and the device solve the technical problems that a large amount of manpower and material resources are consumed for determining the electricity stealing user by adopting a manual inspection mode, and the electricity stealing is difficult to effectively identify in a high-tech manner.
Description
Technical Field
The application relates to the field of electricity stealing behavior detection, in particular to a method and a device for determining abnormal electricity using behavior and a nonvolatile storage medium.
Background
At present, the electric larceny mode is various, the means is concealed strongly, the electricity is stolen through private circuits, modification metering devices and other high-tech equipment which adopts high-frequency electromagnetic waves at present by the tradition, and the electric larceny is difficult to effectively discover. The traditional method for determining the electricity stealing users by adopting a manual inspection mode consumes a large amount of manpower and material resources, and is difficult to effectively identify the high-tech electricity stealing.
In view of the above problems, no effective solution has been proposed.
Disclosure of Invention
The embodiment of the application provides a method and a device for determining abnormal electricity utilization behaviors and a nonvolatile storage medium, and aims to at least solve the technical problems that a large amount of manpower and material resources are consumed and high-tech electricity stealing is difficult to effectively identify by means of determining electricity stealing users in a manual inspection mode.
According to an aspect of an embodiment of the present application, there is provided a method for determining abnormal power consumption behavior, including: acquiring power utilization information data of a target object; extracting target electricity utilization information data in the electricity utilization information data, wherein the target electricity utilization information data is used for indicating electricity utilization behaviors of the target object; processing the quantized value of the target electricity utilization information data by adopting a pre-trained clustering model to obtain electricity utilization behavior data of a target object; and determining whether the target object has abnormal electricity utilization behavior according to the electricity utilization behavior data of the target object.
Optionally, the target electricity consumption information data includes at least one of: voltage, reactive power, active power, power factor, and electrical quantity; storing target electricity consumption information data by a two-dimensional matrix comprising:
wherein n represents the number of target objects, m represents the number of power consumption information data of each target object respectively collected in the same time period, and x ij And target electricity utilization information data representing the ith target object in the j time period.
Optionally, extracting the target electricity consumption information data from the electricity consumption information data includes: the two-dimensional matrix storing the target electricity utilization information data is represented in a column vector mode, wherein the column vector is (x) 1 ,x 2 ...x m ) (ii) a And expressing each element in the column vector by adopting a coordinate transformation method to obtain a linear combination of each element in the column vector:
wherein, mu ij The coefficient variables satisfy the following conditions:
COV(F i ,F j )＝0,i≠ji,j＝1,2...m
wherein [ mu ] k1 ,μ k2 …μ km ]As a main component F k Characteristic vector of (1), COV (F) i ,F j ) =0 means that the two principal components are not related to each other; and processing the linear combination of each element in the column vector according to a principal component analysis method to obtain target electricity utilization information data.
Optionally, the processing the quantized value of the target power consumption information data by using a pre-trained clustering model to obtain the power consumption behavior data of the target object includes: and clustering the target electricity utilization information data by adopting a pre-trained clustering model, and outputting an electricity utilization behavior curve of the target object.
Optionally, determining whether the target object has an abnormal electricity consumption behavior according to the electricity consumption behavior data of the target object includes: determining a first similarity between the target electricity consumption information data and the electricity consumption information data included in the electricity consumption behavior curve; determining a second similarity between the power consumption behavior curve and the historical power consumption behavior curve of the corresponding target object; and determining whether the target object has abnormal electricity utilization behavior according to the first similarity and the second similarity.
Optionally, determining whether the target object has an abnormal electricity consumption behavior according to the first similarity and the second similarity includes: determining a weighted sum of the first similarity and the second similarity; if the weighted sum is larger than a preset threshold value, determining that the target object does not have abnormal electricity utilization behavior; and if the weighted sum is smaller than a preset threshold value, determining that the target object possibly has abnormal electricity utilization behavior.
Optionally, before processing the quantized value of the target electricity consumption information data by using a pre-trained clustering model, the method further includes: carrying out iterative optimization on the quantized value of the target electricity utilization information data by utilizing an improved quantum revolving door artificial fish swarm algorithm to obtain a target quantized value; and taking the target quantized value as an initial centroid of a K-means clustering algorithm, and clustering the quantized value of the target power utilization information data to obtain a pre-trained clustering model.
According to another aspect of the embodiments of the present application, there is also provided an apparatus for determining abnormal electricity consumption behavior, including: the acquisition module is used for acquiring the electricity utilization information data of the target object; the extraction module is used for extracting target electricity utilization information data in the electricity utilization information data; the processing module is used for processing the quantized value of the target electricity utilization information data by adopting a pre-trained clustering model to obtain electricity utilization behavior data of a target object; and the determining module is used for determining whether the target object has abnormal electricity utilization behavior according to the electricity utilization behavior data of the target object.
According to another aspect of the embodiments of the present application, there is also provided a non-volatile storage medium, in which a program is stored, and when the program runs, a device in which the non-volatile storage medium is located is controlled to perform the above method for determining abnormal power consumption behavior.
According to still another aspect of the embodiments of the present application, there is also provided an electronic device, including: a memory and a processor for executing a program stored in the memory, wherein the program when executed performs the above method of determining abnormal power usage behavior.
In the embodiment of the application, the method comprises the steps of acquiring power utilization information data of a target object; extracting target electricity utilization information data in the electricity utilization information data, wherein the target electricity utilization information data is used for indicating the electricity utilization behavior of a target object; processing the quantized value of the target electricity utilization information data by adopting a pre-trained clustering model to obtain electricity utilization behavior data of a target object; the method comprises the steps of determining whether a target object has abnormal power utilization behaviors according to power utilization behavior data of the target object, performing cluster analysis on power utilization characteristics of users to obtain typical load curves of power consumption of different users, combining the typical load curves with historical power utilization data of the users to establish a power utilization abnormality detection model, and rapidly determining suspected power-stealing users to perform key investigation according to results output by the model, so that a large amount of manpower and material resources are saved for workers who investigate the power-stealing users, economic losses caused by the power-stealing behaviors are reduced, the technical effect of improving efficiency and accuracy of marketing inspection work is achieved, and the technical problems that a large amount of manpower and material resources are consumed for determining the power-stealing users by adopting a manual inspection mode and high-tech power stealing is difficult to effectively identify are solved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
fig. 1 shows a hardware configuration block diagram of a computer terminal (or mobile device) for implementing a method of determining abnormal electricity usage behavior;
FIG. 2 is a flow chart of a method of determining abnormal electricity usage behavior according to an embodiment of the present application;
fig. 3 is a block diagram of a device for determining abnormal electricity consumption behavior according to an embodiment of the present application.
Detailed Description
In order to make the technical solutions better understood by those skilled in the art, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only partial embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
It should be noted that the terms "first," "second," and the like in the description and claims of this application and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the application described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
For a better understanding of the embodiments of the present application, technical terms referred to in the embodiments of the present application are explained as follows:
in the related art, most power enterprises still use the traditional statistical analysis method or use simple threshold determination to detect abnormal power utilization so as to determine whether users have abnormal power utilization behaviors. A routing inspection method, a line loss analysis method, an electric quantity analysis method and a power factor analysis method are conventional methods commonly used for detecting power utilization abnormity of power enterprises at present.
Inspection and inspection method: an inspector of the power enterprise timely discovers whether the user has abnormal power utilization behaviors by regularly or randomly inspecting the service condition of the electric energy metering equipment.
Line loss analysis method: the power enterprise inspects different areas at fixed points by analyzing the line loss rate of the different areas. If the line loss rate of a certain area is too high, the area is subjected to key investigation, the investigation range is further narrowed, and then the power users with abnormal power utilization behaviors are checked.
Electric quantity analysis method: the power consumption abnormality judgment is carried out by the power enterprises through analyzing related power consumption parameters of metering equipment such as an electric energy meter. And if the related power utilization parameters are far deviated from the normal range, judging that the user is an abnormal power utilization user.
Power factor analysis method: and the power supply enterprise judges whether the abnormal power utilization behavior exists in the user by analyzing the power factor fed back by the power metering equipment. And if the power factor fed back on the metering equipment shows a stable state, the user is a normal power user. On the contrary, if the fluctuation of the power factor fed back by the electric energy metering device is relatively large, the user has abnormal electricity utilization behavior, and the user is highly likely to be an abnormal electricity user.
The above four methods are conventional methods for the power company to determine whether the user has abnormal power consumption behaviors (such as power stealing, power leakage, etc.). The methods have the problems of low efficiency, large time consumption, inaccurate judgment, missed judgment, erroneous judgment, information redundancy and the like. The abnormal electricity consumption behavior judgment efficiency of the power enterprises is low, and the waste of a large amount of electricity consumption data is caused, so that the power enterprises cannot dig out valuable information by analyzing the data.
In order to solve the problem, in the embodiment of the application, a principal component analysis method is adopted to perform dimensionality reduction processing on massive power consumption information, clustering processing is performed on user load data after dimensionality reduction, a K-means algorithm optimized by an improved quantum revolving door artificial fish swarm algorithm is adopted to perform clustering processing, a typical user load characteristic curve is obtained, and different types of load characteristics and influences on power grid operation are analyzed. Compared with the prior art, the method and the device have the advantages that the abnormal electricity utilization behaviors of the power consumers are diagnosed by adopting a principal component analysis method and a K-means algorithm optimized by improving the artificial fish swarm algorithm of the quantum revolving door, so that the data dimension is reduced, the processing efficiency is improved, the accuracy of the abnormal electricity utilization detection and the accuracy of the electricity stealing behavior detection can be improved, and the problems can be solved, and are explained in detail below.
In accordance with an embodiment of the present application, there is provided an embodiment of a method of determining abnormal power usage behavior, it being noted that the steps illustrated in the flowchart of the drawings may be performed in a computer system such as a set of computer-executable instructions and that, although a logical order is illustrated in the flowchart, in some cases, the steps illustrated or described may be performed in an order different than presented herein.
The method provided by the embodiment of the application can be executed in a mobile terminal, a computer terminal or a similar operation device. Fig. 1 shows a hardware configuration block diagram of a computer terminal (or mobile device) for implementing the method of determining abnormal power usage behavior. As shown in fig. 1, the computer terminal 10 (or mobile device 10) may include one or more (shown as 102a, 102b, … …,102 n) processors 102 (the processors 102 may include, but are not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 104 for storing data, and a transmission module 106 for communication functions. Besides, the method can also comprise the following steps: a display, an input/output interface (I/O interface), a Universal Serial BUS (USB) port (which may be included as one of the ports of the BUS), a network interface, a power source, and/or a camera. It will be understood by those skilled in the art that the structure shown in fig. 1 is only an illustration and is not intended to limit the structure of the electronic device. For example, the computer terminal 10 may also include more or fewer components than shown in FIG. 1, or have a different configuration than shown in FIG. 1.
It should be noted that the one or more processors 102 and/or other data processing circuitry described above may be referred to generally herein as "data processing circuitry". The data processing circuitry may be embodied in whole or in part in software, hardware, firmware, or any combination thereof. Further, the data processing circuit may be a single stand-alone processing module, or incorporated in whole or in part into any of the other elements in the computer terminal 10 (or mobile device). As referred to in the embodiments of the application, the data processing circuit acts as a processor control (e.g. selection of a variable resistance termination path connected to the interface).
The memory 104 may be used to store software programs and modules of application software, such as program instructions/data storage devices corresponding to the method for determining abnormal electricity consumption behavior in the embodiment of the present application, and the processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, implements the above-mentioned method for determining abnormal electricity consumption behavior of the application program. The memory 104 may include high speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include memory located remotely from the processor 102, which may be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The transmission device 106 is used for receiving or transmitting data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a Network adapter (NIC) that can be connected to other Network devices through a base station to communicate with the internet. In one example, the transmission device 106 can be a Radio Frequency (RF) module, which is used to communicate with the internet in a wireless manner.
The display may be, for example, a touch screen type Liquid Crystal Display (LCD) that may enable a user to interact with a user interface of the computer terminal 10 (or mobile device).
In the foregoing operating environment, an embodiment of the present application provides a method for determining an abnormal electricity consumption behavior, as shown in fig. 2, the method includes the following steps:
step S202, acquiring power utilization information data of the target object.
And step S204, extracting target electricity utilization information data in the electricity utilization information data, wherein the target electricity utilization information data is used for indicating the electricity utilization behavior of the target object.
According to an optional embodiment of the present application, the target electricity consumption data is a main influence index for determining whether the user has abnormal electricity consumption behavior in the electricity information data.
Step S206, processing the quantized value of the target electricity utilization information data by adopting a pre-trained clustering model to obtain electricity utilization behavior data of a target object;
and step S208, determining whether the target object has abnormal electricity utilization behavior according to the electricity utilization behavior data of the target object.
Through the steps, the typical load curve of power consumption of different users is obtained by carrying out cluster analysis on the power consumption characteristics of the users, the typical load curve is combined with the historical power consumption data of the users to establish a power consumption abnormity detection model, and suspected power stealing users can be rapidly determined to carry out key investigation according to the result output by the model, so that the technical effects of saving a large amount of manpower and material resources for workers who investigate the power stealing users, reducing economic loss caused by power stealing behaviors and improving the efficiency and accuracy of marketing inspection work are achieved.
According to an optional embodiment of the present application, the target electricity consumption information data in step S204 includes at least one of the following: voltage, reactive power, active power, power factor, and electrical quantity.
The target electricity consumption information Data in step S204 includes voltage, reactive power, active power, power factor, electric quantity, and the like generated by the smart meter every 15min, and the adopted Data is a SCADA (Supervisory Control And Data Acquisition) real-time Data source table of a power grid Data Acquisition And monitoring Control system, and 96 Data points are acquired each day.
It is understood that the above 96 data points are collected for illustration purposes only, and in the implementation process, the number of the collected data points is less than 96 or more than 96, so as to implement the function of diagnosing whether the abnormal power consumption exists in the user.
As an alternative embodiment of the present application, the target electricity information data is stored by the following two-dimensional matrix:
wherein n represents the number of target objects, m represents the number of power consumption information data of each target object respectively collected in the same time period, and x ij And target electricity utilization information data representing the ith target object in the j time period.
In the present application, a matrix is used to represent the dimension of load data, and assuming that there are n users in a region and load values of m data points are collected every day, we use an n × m matrix to represent all load data, as shown in equation (1):
in the formula (1), x ij Representing the load value of the ith user in the j period.
According to another alternative embodiment of the present application, the step S204 is executed to extract the target electricity consumption information data from the electricity consumption information data, and the method is implemented by: the two-dimensional matrix storing the target electricity utilization information data is represented in a column vector mode, wherein the column vector is (x) 1 ,x 2 ...x m ) (ii) a And expressing each element in the column vector by adopting a coordinate transformation method to obtain a linear combination of each element in the column vector:
wherein, mu ij As a coefficient variable, the coefficient variable satisfies the following condition:
COV(F i ,F j )＝0,i≠ji,j＝1,2...m
wherein [ mu ] k1 ,μ k2 …μ km ]As a main component F k Characteristic vector of (1), COV (F) i ,F j ) =0 indicates that the two principal components do not correlate with each other; and processing the linear combination of each element in the column vector according to a principal component analysis method to obtain target electricity utilization information data.
In this step, a principal component analysis method is used to extract the main influence indexes in the electricity consumption information data.
Principal Component Analysis (PCA), a statistical method, converts a set of variables that may have correlation into a set of linearly uncorrelated variables by orthogonal transformation, and the set of converted variables is called Principal components.
The sample set contains n samples, each sample has m dimensions, and the data set matrix is shown in formula (1). Processing the data set in a coordinate transformation mode to obtain the mathematical expressions of the main components as follows:
in the formula (2), the data matrix is expressed as (x) by using a column vector 1 ,x 2 ...x m )。
And expressing each principal component by adopting a coordinate transformation method to obtain a formula as follows:
in formula (3): mu.s ij For coefficient variables, the following formula is satisfied:
COV(F i ,F j )＝0,i≠ji,j＝1,2...m (5)
vector [ mu ] in formula (3-4) k1 ,μ k2 …μ km ]As a main component F k The feature vector of (2). COV (F) in formula (5) i ,F j ) And =0 indicates that the two principal components do not correlate with each other. Therefore, the principle of equation (3) is to replace the variables in the original data sample with mutually independent abstract synthesis components to form a linear combination of the original variables. Then, the variance of each principal component is calculated to carry out descending arrangement, and the principal component with the largest variance is changed into F 1 By analogy, if F 1 If not all information in the original data can be included, F is set 2 And (4) the first k principal components are included until the first k principal components can be used for including most information of the original data, and then the first k principal components are output. The algorithm is specifically realized by the following steps:
calculating a principal component analysis method:
step 1: preprocessing the data to obtain a normalization result;
step 2: computing the correlation coefficient (covariance) r of any two variables between the original data samples ij ＝COV(x i ,x j ) The formula is as follows:
the correlation coefficient is expressed as a covariance matrix form:
and step 3: establishing a characteristic equation of lambda I-R =0, and calculating a characteristic value lambda of a correlation coefficient matrix i (i =1,2 … m) and feature vector a i ＝[a i1 ,a i2 …a im ]And arranges the eigenvalues (representing the variance magnitudes) and the corresponding eigenvectors from large to small. The key point of the principal component analysis is to find the eigenvalues and eigenvectors of the correlation coefficient matrix.
And 4, step 4: the contribution rate and the cumulative contribution rate of each principal component are calculated from equations (9) and (10). When the principal component analysis method analyzes the original data, the variance of each principal component gradually decreases, and the accumulated variance of the first principal components can reach more than 90% of the original data, that is, more than 90% of the information of the original data can be restored, so that the first g principal components with larger accumulated contribution rate are selected as the final result of the principal component analysis method.
The respective principal component variance contribution ratio formula:
the calculation formula of the cumulative variance contribution rate is as follows:
and 5: and outputting the latest data result obtained by the principal component analysis method, wherein n rows and g columns are provided.
The principal component analysis method is realized in two ways, one is to set the number of the principal components in advance and then judge whether the principal components are effective or not according to whether the contribution rate of the principal components meets the requirement or not; the other is that the number is not established in advance, and the expected cumulative contribution rate is adopted for determination. We set the desired cumulative variance contribution rate to 90% herein, then extract the top n principal component outputs that can reach 90% cumulative variance. The dimension reduction is performed in a second way.
In some optional embodiments of the present application, before the step S206 is executed to process the quantized value of the target electricity consumption information data by using the pre-trained clustering model, it is further required to perform iterative optimization on the quantized value of the target electricity consumption information data by using an improved quantum revolving door artificial fish swarm algorithm to obtain a target quantized value; and taking the target quantized value as an initial centroid of a K-means clustering algorithm, and clustering the quantized value of the target power utilization information data to obtain a pre-trained clustering model.
The method is characterized in that a principal component analysis method and a K-means clustering method optimized by an improved quantum revolving door artificial fish swarm algorithm are fused to analyze massive load data, the principal component analysis method is used for reducing dimensions, and the improved quantum revolving door artificial fish swarm algorithm is used for optimizing the K-means clustering method and classifying users. In the invention, load data is collected every 15 minutes to obtain a data point dimension of 96 in one day, and an n multiplied by 96 original data matrix is generated, and considering that the daily load of a user has strong regularity, the trend and the regular condition of daily load change can be generally described by using a plurality of points, and the purpose of applying a principal component analysis method is to improve the data value by using a dimension reduction idea and reduce the algorithm complexity and the time cost on one hand; on the other hand, typical load users and atypical load users are divided, the atypical load users are classified into one class, subsequent operation is not performed, and the efficiency of the clustering algorithm can be improved.
The K-means clustering algorithm is a distance-based clustering algorithm, the distance is used as an evaluation index of similarity, and the closer the distance between two data points is, the greater the similarity is. The distance formula for calculating the distance between the samples comprises Euclidean distance, manhattan distance, cosine similarity and the like, wherein the Euclidean distance is most commonly used. The basic idea of the algorithm is as follows: selecting K data elements as an initial clustering centroid; assigning each element to its nearest cluster centroid according to the distance between the element and the cluster center to which it belongs; the above operations are repeated until all elements have been allocated. In each iteration process, the clustering centroid is recalculated according to the existing elements in each class, and the operation is repeated until the termination condition is met.
a) Sample set
Wherein d is the sample dimension, K is the cluster number, and T is the maximum iteration number.
b) Let I =1, randomly select K initial clustering centroids c from X 1 (I),c 2 (I),…,c k (I)。
c) Calculating Euclidean distance d (x) from each data element to the K cluster centroids i ,c j (I) ); (i =1,2, …, N; j =1,2, …, K) if d (x) is satisfied i ,c j (I))＝arg,om{(x i ,c j (I) J =1,2, …, K }, then x i C belonging to the centre of mass j Is denoted as x i ∈c j 。
d) Each cluster centroid ci (I + 1), I =1,2, …, K, is again calculated.
Wherein n is i Indicating the number of elements in the ith class.
e) If for any i e {1,2, …, K }, c i (I+1)＝c i (I) All are true, or when the maximum number of iterations is reached, the algorithm terminates and the current c i (I) (i =1,2, …, K) is the final clustering centroid, and the clustering result is output; otherwise, returning to execute the step b).
The artificial fish school algorithm principle is as follows:
a) And (5) initializing a fish school. Let the artificial fish school individual be x i (i =1,2, …, n), the current food concentration of the artificial fish school is represented as Y = f (x), where Y is the objective function value and the distance between individual artificial fish schools is represented as d i,j ＝x i -x j I and j are numbers of two artificial fish schools, visual represents the sensing distance of the artificial fish schools, namely the visual field range, step represents the maximum step length of the movement of the artificial fish schools, delta is the crowding degree, and try _ number is the number of attempts.
b) Foraging behavior. Let the current artificial fish school be x i Randomly selecting another artificial fish group x in the perception range j If Y is in solving the maximum problem i ＜Y j (in the very Small problem Y i ＞Y j ) Then, go one step forward in the direction; otherwise, other artificial fish shoal x is randomly selected again j And judging whether the forward condition is met. After repeating try _ number times in this way, if the condition is not satisfied, the step is randomly moved.
x next ＝x i +rand×step×(x j -x i ) (13)
c) Clustering behavior. Let the current artificial fish school be x j Exploring the current domain (i.e., d) i,j < visual) of the partner number n f And center position x c If Y is c /n f ＞δY i If the food is more in the center of the partner and the center of the partner is not too crowded, the food is further moved to the center position of the partner; otherwise, foraging is performed.
d) And (5) rear-end collision behavior. Let the current artificial fish school be x j Exploring the current domain (i.e., d) i,j < visual) of the partner number n f And Y in partner j Is the largest partner x j If Y is c /n f ＞δY i Denotes a partner x j With higher food concentration and less crowding around, then towards buddy x j One step forward; otherwise, foraging behavior is executed.
e) Random behavior. The random behavior is simpler to implement by selecting a state within the field of view and then moving in that direction. It is actually a default behavior of foraging, i.e. x i Next position x of next ＝x i + rand × visual, rand being [ -1,1]The random number of (2).
Quantum artificial fish swarm algorithm
In quantum space, a qubit, i.e. a qubit, is the smallest unit of information, and the state of a qubit can be represented as
Where α and β are called the probability magnitudes of the qubits and satisfy the normalization condition | α $ 2 +|β| 2 And =1. Let α = cos θ, β = sin θ, i.e. the qubits be denoted [ cos θ sin θ] T Where θ represents the phase of the qubit.
Encoding of quantum artificial fish shoal
Applying the characteristics of the quantum to the artificial fish school, setting the population scale of the artificial fish school of the quantum as N, and the position of the artificial fish school as the probability amplitude P i To express, each probability amplitude is mapped to the solution space of the optimization problem by unit space, and each probability amplitude corresponds to oneOptimization variable x of solution space i . The quantum artificial fish school coding mode is as follows:
wherein, P ic Is the cosine position; p is Is a sinusoidal position; theta ij Is the rotation angle; m is a spatial dimension; i belongs to {1,2, …, N }; j is the same as {1,2, …, m }.
Method for transforming solution space
Let the feasible solution x of the optimization problem i Is expressed as [ cos θ ] ij sinθ ij ] T Corresponding solution variable is
And defines a domain [ a ] j ,b j ]Then using linear transformation can be obtained.
Position update of quantum artificial fish school
The quantum artificial fish school updating mode generally adopts a quantum gate for updating, the commonly used quantum gate is a quantum revolving gate, and the formula (19) is as follows:
wherein, the first and the second end of the pipe are connected with each other,
the probability amplitude of the jth qubit for the t +1 th iteration; />
The rotation angle is typically obtained by looking up a table of the magnitude and direction.
Quantum artificial fish school variation
In some multi-peak function optimization problems, the algorithm tends to fall into a local optimum. In order to increase the diversity of artificial fish schools in quantum artificial fish schools, mutation operations are introduced, and a common mutation operation is realized by a quantum NOT gate.
Quantum artificial fish school algorithm process
a) Initializing the artificial fish school population scale of quantum as N, setting step length, visual field range visual, try times try _ number, crowding factor delta, variation probability Pm and the like.
b) And calculating the fitness value of each artificial fish school and recording the fitness value in the bulletin board.
c) And performing four actions of foraging, clustering, rear-end collision and random movement, evaluating, determining a target position, and updating by using a quantum revolving door.
d) Randomly generating a random number of 0-1, comparing the random number with the mutation probability Pm, and performing mutation operation by using a quantum not gate according with the condition.
e) And (5) calculating the adaptability value of each artificial fish school again, and updating the bulletin board.
f) And c, stopping iteration and outputting a result until the condition is met or the maximum iteration number is reached, otherwise, turning to the step c.
Improved quantum artificial fish swarm algorithm
Improved quantum revolving door
In the quantum artificial fish swarm algorithm, the rotation angle Δ θ of the quantum revolving gate is obtained by looking up a table, and is a discrete and fixed value. The angle and direction of Δ θ rotation may directly affect the accuracy and convergence speed of the algorithm, and the search range of the artificial fish school after the rotation operation is limited. A method of dynamically updating a rotation angle is presented.
Wherein ITER is the total iteration number; t is the current iteration number; delta theta p (t) is the position rotation angle increment amplitude obtained after the tth foraging action is carried out; delta theta c (t) obtaining the increment amplitude of the rotation angle of the central position of the whole artificial fish school after the t-th clustering action; delta theta g And (t) is the increment amplitude of the rotation angle of the current optimal position in the whole artificial fish school after the t-th rear-end collision behavior. And dynamically combining the results of the foraging behavior, the clustering behavior and the rear-end collision behavior with the iteration number. The position of the next generation artificial fish school is adjusted according to the foraging result of the current t generation, the central position of the colony after the colony clustering and the optimal position of the colony, and the method has the advantages that: a) The rotation angle is dynamically adjusted according to the obtained result, and the rotation angle is not required to be obtained in a table look-up mode; b) The number of parameters in the traditional quantum artificial fish school is reduced, and calculation is facilitated; c) After the rotation angle calculation is performed every time, the global optimal state and the local optimal state of the artificial fish school are combined with each other, and random behaviors are contained in the foraging behaviors, so that the defect of falling into the local optimal state is better avoided. Therefore, the improved delta theta can more reasonably adjust the direction and the size of the artificial fish school to be stepped in the next step.
Mutation strategy
The method adopts a Hadamard gate-H gate to perform mutation operation on the quantum artificial fish school. The H-gate transform matrix is such that if the quantum state α |0 > + β |1 is written as a vector form [ α β [ ]] T Then the output of the H gate is:
h gate is Hibert space transform, which transforms the space based on |0 > and |1 > into the space based on | plus > and | -. Due to H 2 =1, if a qubit is converted twice in succession, the qubit will switch to the original state, i.e. no quanta are appliedAnd (6) performing logical operation.
Under the condition of ensuring that the optimal value is not ignored, the diversity of the fish school is increased more obviously, and the obtained result is applied to the next step of the quantum fish school algorithm, and the specific variation operation of the gate H is as follows:
wherein, theta ij Is the angle of rotation.
K-means clustering based on improved quantum artificial fish swarm algorithm
The improved quantum artificial fish swarm algorithm is fused with the K-means algorithm, each artificial fish swarm represents the mass center of one group of K-means clusters, the fitness function value of each artificial fish swarm is calculated by using the following fitness function, and the improved quantum artificial fish swarm algorithm is adopted for iterative optimization. And finally, outputting an optimal group of quantum artificial fish schools as the initial mass center of the K-means clustering, and replacing the mode of randomly selecting the mass center in the traditional K-means clustering algorithm.
Fitness function
In clustering algorithms, the ultimate goal is to maximize the data similarity within a class and minimize the similarity between classes. Therefore, a fitness function is typically used that is:
wherein K represents the number of clusters;
calculating the average distance from each point in the class to the centroid of the class; omega i 、ω j Representing the centroids of class i and class j; i omega 1 -ω 2 || 2 Generations are distances between various types of centroids. If the clustering result generated by the algorithm changes towards the minimum distance within a class (maximum intra-class similarity) and the maximum distance between classes (minimum inter-class similarity), the smaller the fitness value, i.e. the smaller the fitness valueThe better the clustering effect. />
The algorithm flow is as follows:
a) Inputting M data to be clustered, and setting iteration times T and cluster number K.
b) Initializing a quantum fish swarm with the population size of N, setting all quantum bits of the artificial fish swarm to pi/4, representing that the quantum bits are overlapped at equal probability in the initial period, and setting step length, visual field range visual, try times try _ number, crowding factor delta and variation probability P m And so on.
c) The fitness value of each artificial fish school is calculated according to the formula (24) and recorded in the bulletin board.
d) And performing four actions of foraging, clustering, rear-end collision and random movement, comparing, and updating by using an improved quantum revolving door after determining the target position.
e) Randomly generating a random number of 0-1 and correlating with the mutation probability P m And (4) comparing, and carrying out mutation operation by using a quantum H gate according with the condition.
f) And (4) performing solution space transformation on the varied quantum artificial fish school by using an equation (17) to obtain a solution of the actual problem.
g) And (4) calculating the adaptability value of the artificial fish shoal according to the formula (23), comparing the adaptability value with the value in the bulletin board, and using the adaptability value as the next-generation artificial fish shoal.
h) And (4) performing reverse conversion by using a conversion formula (17) of the quantum space and the solution space again to obtain a new generation of quantum artificial fish school, returning to the step d), and circulating until the maximum trial times are reached or the optimal value in the bulletin board is not changed any more.
i) And selecting an optimal value, namely taking K clustering centers as initial centroids of K-means to perform K-means clustering operation, circularly reaching the maximum iteration number T, finishing the algorithm and outputting a clustering result.
According to an optional embodiment of the present application, step S206 is executed to process the quantized value of the target electricity consumption information data by using a pre-trained clustering model to obtain the electricity consumption behavior data of the target object, and the method is implemented by the following steps: and clustering the target electricity utilization information data by adopting a pre-trained clustering model, and outputting an electricity utilization behavior curve of the target object.
In the step, the main influence indexes of the electricity utilization information obtained through principal component analysis are used as input data of the optimized K-means algorithm of the improved quantum revolving door artificial fish swarm algorithm, and the input data are output as an electricity utilization behavior curve of the power consumer.
As an alternative embodiment of the present application, the step S208 of determining whether the target object has the abnormal electricity consumption behavior according to the electricity consumption behavior data of the target object includes the following steps: determining a first similarity between the target electricity utilization information data and the electricity utilization information data included in the electricity utilization behavior curve; determining a second similarity between the power consumption behavior curve and the historical power consumption behavior curve of the corresponding target object; and determining whether the target object has abnormal electricity utilization behavior according to the first similarity and the second similarity.
For the improved quantum revolving door artificial fish swarm algorithm optimized K-means clustering method, the criterion of whether the data is classified into one category is the Euclidean distance of two load curves, and the change trends of the load data classified into one category are not always consistent. Therefore, according to the principle of abnormal electricity detection, the similarity of the load change trends of the two needs to be judged through an algorithm. The similarity of the values and the change trends between the two is measured through a similarity matching algorithm. The load data is D = [ D = [) 1 ,d 2 …d 96 ]And a typical load curve for the class to which it belongs is R = [ R ] 1 ,r 2 …r 96 ]The formula is as follows:
in the formula (25), the reaction mixture,
similarity matching algorithm is applied to judge similarity sim of the two 1 The closer to 1 indicates the greater the similarity of the two trends. />
For the similarity matching problem of the user load and the historical load curve, the difference of the numerical change of the current curve and the historical curve is more concerned here. The historical curve is represented by the mean value of the workload of the working days under the conditions of similar weather and temperature in the same month in the last year. Such a history can be characterized and trend-changing by contemporaneous load. Therefore, the distance index is selected to evaluate the similarity between the user load curve and the historical load curve. Setting user load data as D = [ D = 1 ,d 2 …d 96 ]The historical data is H = [ H ] 1 ,h 2 …h 96 ]The formula is as follows:
for similarity sim between load curve and historical curve 2 In other words, a value closer to 1 indicates a higher degree of similarity.
In an optional embodiment, determining whether the target object has the abnormal electricity utilization behavior according to the first similarity and the second similarity includes the following steps: determining a weighted sum of the first similarity and the second similarity; if the weighted sum is larger than a preset threshold value, determining that the target object does not have abnormal electricity utilization behavior; and if the weighted sum is smaller than a preset threshold value, determining that the target object possibly has abnormal electricity utilization behavior.
After similarity analysis, a final evaluation index of power utilization abnormality detection is obtained through weighting calculation, and the formula is as follows:
judge＝θ 1 sim 1 +θ 2 sim 2 (27)
in the formula (27), θ 1 ，θ 2 -are the weights of the similarity indicators, respectively, and satisfy theta 1 +θ 2 ＝1。
And finally, judging abnormal power utilization by setting a proper threshold value. And when the evaluation index judge is larger than the threshold value, the electricity utilization of the user is normal, and when the evaluation index judge is smaller than the threshold value, the suspicion of abnormal electricity utilization exists. The method comprises the steps of conducting preliminary judgment on a user in an online monitoring mode, then sending professional workers to the user site to conduct detailed screening detection to finally judge whether electricity stealing behaviors exist in the user. Although the evaluation index can not completely determine whether the electricity utilization of one user is abnormal or not, the detection range and the detection proportion can be greatly reduced, and material resources and manpower are saved.
According to the technical scheme, a principal component analysis method is adopted to perform dimensionality reduction processing on the massive power utilization information to obtain main influence characteristics; and clustering the user load data subjected to dimensionality reduction, performing clustering by adopting a K-means algorithm optimized by an improved quantum revolving door artificial fish swarm algorithm to obtain a typical user load characteristic curve, and analyzing different types of load characteristics and influences on power grid operation. Compared with the prior art, the method and the device have the advantages that the abnormal electricity utilization behaviors of the power consumers are diagnosed by adopting a principal component analysis method and a K-means algorithm optimized by improving the artificial fish swarm algorithm of the quantum revolving door, so that the data dimension is reduced, the processing efficiency is improved, and the accuracy of the abnormal electricity utilization detection and the electricity stealing behavior detection can be improved.
Fig. 3 is a block diagram of an apparatus for determining abnormal electricity consumption behavior according to an embodiment of the present application, and as shown in fig. 3, the apparatus includes:
and the obtaining module 30 is used for obtaining the electricity utilization information data of the target object.
And the extraction module 32 is used for extracting the target electricity utilization information data in the electricity utilization information data.
And the processing module 34 is configured to process the quantized value of the target power consumption information data by using a pre-trained clustering model to obtain power consumption behavior data of the target object.
And the determining module 36 is configured to determine whether the target object has an abnormal electricity consumption behavior according to the electricity consumption behavior data of the target object.
It should be noted that each module in the above-mentioned device for determining abnormal electricity consumption behavior may be a program module (for example, a set of program instructions for implementing a certain specific function), or may be a hardware module, and in the latter case, it may be represented in the following form, but is not limited thereto: the above modules are all represented by one processor, or the functions of the above modules are realized by one processor.
It should be noted that, reference may be made to the description related to the embodiment shown in fig. 2 for a preferred implementation of the embodiment shown in fig. 3, and details are not described here again.
The embodiment of the application also provides a nonvolatile storage medium, wherein the nonvolatile storage medium stores a program, and the method for determining the abnormal electricity utilization behavior is executed by controlling the equipment where the nonvolatile storage medium is located when the program runs.
The nonvolatile storage medium stores a program for executing the following functions: acquiring power utilization information data of a target object; extracting target electricity utilization information data in the electricity utilization information data, wherein the target electricity utilization information data is used for indicating the electricity utilization behavior of a target object; processing the quantized value of the target electricity utilization information data by adopting a pre-trained clustering model to obtain electricity utilization behavior data of a target object; and determining whether the target object has abnormal electricity utilization behavior according to the electricity utilization behavior data of the target object.
An embodiment of the present application further provides an electronic device, including: a memory and a processor for executing a program stored in the memory, wherein the program when executed performs the above method of determining abnormal power usage behavior.
The processor is used for running a program for executing the following functions: acquiring power utilization information data of a target object; extracting target electricity utilization information data in the electricity utilization information data, wherein the target electricity utilization information data is used for indicating the electricity utilization behavior of a target object; processing the quantized value of the target electricity utilization information data by adopting a pre-trained clustering model to obtain electricity utilization behavior data of a target object; and determining whether the target object has abnormal electricity utilization behavior according to the electricity utilization behavior data of the target object.
The above-mentioned serial numbers of the embodiments of the present application are merely for description and do not represent the merits of the embodiments.
In the above embodiments of the present application, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.
In the embodiments provided in the present application, it should be understood that the disclosed technology can be implemented in other ways. The above-described apparatus embodiments are merely illustrative, and for example, the division of the units may be a logical division, and in actual implementation, there may be another division, for example, multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or may not be executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, units or modules, and may be in an electrical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and other various media capable of storing program codes.
The foregoing is only a preferred embodiment of the present application and it should be noted that those skilled in the art can make several improvements and modifications without departing from the principle of the present application, and these improvements and modifications should also be considered as the protection scope of the present application.
Claims (10)
1. A method of determining abnormal electricity usage behavior, comprising:
acquiring power utilization information data of a target object;
extracting target electricity utilization information data in the electricity utilization information data, wherein the target electricity utilization information data is used for indicating electricity utilization behaviors of the target object;
processing the quantitative value of the target electricity utilization information data by adopting a pre-trained clustering model to obtain electricity utilization behavior data of the target object;
and determining whether the target object has abnormal electricity utilization behavior according to the electricity utilization behavior data of the target object.
2. The method of claim 1,
the target electricity information data comprises at least one of the following: voltage, reactive power, active power, power factor, and electrical quantity;
storing the target electricity consumption information data by a two-dimensional matrix of:
wherein n represents the number of the target objects, m represents the number of the power consumption information data of each of the target objects respectively collected in the same time period, and x ij Target electricity consumption information data representing the ith target object in the j time period。
3. The method according to claim 2, wherein extracting the target electricity consumption information data from the electricity consumption information data comprises:
representing a two-dimensional matrix storing the target electricity utilization information data in a column vector mode, wherein the column vector is (x) 1 ,x 2 ...x m )；
Expressing each element in the column vector by adopting a coordinate transformation method to obtain a linear combination of each element in the column vector:
wherein, mu ij Is a coefficient variable that satisfies the following condition:
COV(F i ,F j )＝0,i≠ji,j＝1,2...m
wherein [ mu ] k1 ,μ k2 …μ km ]As a main component F k Characteristic vector of (1), COV (F) i ,F j ) =0 indicates that the two principal components do not correlate with each other;
and processing the linear combination of each element in the column vector according to a principal component analysis method to obtain the target electricity utilization information data.
4. The method according to claim 3, wherein the processing the quantized value of the target power consumption information data by using a pre-trained clustering model to obtain the power consumption behavior data of the target object comprises:
and clustering the target power utilization information data by adopting a pre-trained clustering model, and outputting a power utilization behavior curve of the target object.
5. The method of claim 4, wherein determining whether the target object has abnormal electricity consumption behavior according to the electricity consumption behavior data of the target object comprises:
determining a first similarity between the target electricity consumption information data and the electricity consumption information data included in the electricity consumption behavior curve;
determining a second similarity between the power consumption behavior curve and a corresponding historical power consumption behavior curve of the target object;
and determining whether the target object has abnormal electricity utilization behavior according to the first similarity and the second similarity.
6. The method of claim 5, wherein determining whether the target object has abnormal electricity usage behavior according to the first similarity and the second similarity comprises:
determining a weighted sum of the first similarity and the second similarity;
if the weighted sum is larger than a preset threshold value, determining that the target object does not have abnormal electricity utilization behavior;
and if the weighted sum is smaller than the preset threshold value, determining that the target object possibly has abnormal electricity utilization behavior.
7. The method of claim 1, wherein before processing the quantized values of the target electricity consumption information data using a pre-trained clustering model, the method further comprises:
carrying out iterative optimization on the quantized value of the target electricity utilization information data by utilizing an improved quantum revolving door artificial fish swarm algorithm to obtain a target quantized value;
and clustering the quantized values of the target electricity utilization information data by taking the target quantized values as the initial mass center of a K-means clustering algorithm to obtain the pre-trained clustering model.
8. An apparatus for determining abnormal electricity usage behavior, comprising:
the acquisition module is used for acquiring the electricity utilization information data of the target object;
the extraction module is used for extracting target electricity utilization information data in the electricity utilization information data;
the processing module is used for processing the quantized value of the target power consumption information data by adopting a pre-trained clustering model to obtain power consumption behavior data of the target object;
and the determining module is used for determining whether the target object has abnormal electricity utilization behavior according to the electricity utilization behavior data of the target object.
9. A non-volatile storage medium, wherein a program is stored in the non-volatile storage medium, and when the program runs, a device where the non-volatile storage medium is located is controlled to execute the method for determining abnormal electricity consumption behavior according to any one of claims 1 to 7.
10. An electronic device, comprising: a memory and a processor for executing a program stored in the memory, wherein the program when executed performs the method of determining abnormal power usage of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211537869.0A CN115841338A (en) | 2022-12-02 | 2022-12-02 | Method and device for determining abnormal electricity utilization behavior and non-volatile storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211537869.0A CN115841338A (en) | 2022-12-02 | 2022-12-02 | Method and device for determining abnormal electricity utilization behavior and non-volatile storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115841338A true CN115841338A (en) | 2023-03-24 |
Family
ID=85577888
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211537869.0A Pending CN115841338A (en) | 2022-12-02 | 2022-12-02 | Method and device for determining abnormal electricity utilization behavior and non-volatile storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115841338A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116777124A (en) * | 2023-08-24 | 2023-09-19 | 国网山东省电力公司临沂供电公司 | Power stealing monitoring method based on user power consumption behavior |
-
2022
- 2022-12-02 CN CN202211537869.0A patent/CN115841338A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116777124A (en) * | 2023-08-24 | 2023-09-19 | 国网山东省电力公司临沂供电公司 | Power stealing monitoring method based on user power consumption behavior |
| CN116777124B (en) * | 2023-08-24 | 2023-11-07 | 国网山东省电力公司临沂供电公司 | Power stealing monitoring method based on user power consumption behavior |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhao et al. | Improving event-based non-intrusive load monitoring using graph signal processing | |
| He et al. | Non-intrusive load disaggregation using graph signal processing | |
| Massaferro et al. | Fraud detection in electric power distribution: An approach that maximizes the economic return | |
| Figueiredo et al. | Electrical signal source separation via nonnegative tensor factorization using on site measurements in a smart home | |
| US8756181B2 (en) | System and method employing a self-organizing map load feature database to identify electric load types of different electric loads | |
| Van Der Laan et al. | Gene expression analysis with the parametric bootstrap | |
| CN109740694A (en) | A kind of smart grid inartful loss detection method based on unsupervised learning | |
| CN110990461A (en) | Big data analysis model algorithm model selection method and device, electronic equipment and medium | |
| Feng et al. | Data mining for abnormal power consumption pattern detection based on local matrix reconstruction | |
| Golay et al. | A new estimator of intrinsic dimension based on the multipoint Morisita index | |
| Zhao et al. | Analysis of users’ electricity consumption behavior based on ensemble clustering | |
| Saha et al. | Comprehensive NILM framework: device type classification and device activity status monitoring using capsule network | |
| CN110796159A (en) | Power data classification method and system based on k-means algorithm | |
| Narasimhan et al. | Contribution-based clustering algorithm for content-based image retrieval | |
| Dent et al. | Variability of behaviour in electricity load profile clustering; Who does things at the same time each day? | |
| CN115841338A (en) | Method and device for determining abnormal electricity utilization behavior and non-volatile storage medium | |
| Li et al. | Distance measures in building informatics: An in-depth assessment through typical tasks in building energy management | |
| CN112463848A (en) | Method, system, device and storage medium for detecting abnormal user behavior | |
| Hosseini et al. | A study on Markovian and deep learning based architectures for household appliance-level load modeling and recognition | |
| Xu et al. | Cooperative co-evolutionary algorithm for multi-objective optimization problems with changing decision variables | |
| Bao et al. | Feature selection method for nonintrusive load monitoring with balanced redundancy and relevancy | |
| Tirelli et al. | Unsupervised machine learning approaches to the q-state Potts model | |
| Yin et al. | Non-intrusive load monitoring by load trajectory and multi feature based on DCNN | |
| CN107274025B (en) | System and method for realizing intelligent identification and management of power consumption mode | |
| Wang et al. | Application of clustering technique to electricity customer classification for load forecasting |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |