CN113052724A - Non-invasive load electricity consumption information clustering method, device and equipment - Google Patents
Non-invasive load electricity consumption information clustering method, device and equipment Download PDFInfo
- Publication number
- CN113052724A CN113052724A CN202110474324.9A CN202110474324A CN113052724A CN 113052724 A CN113052724 A CN 113052724A CN 202110474324 A CN202110474324 A CN 202110474324A CN 113052724 A CN113052724 A CN 113052724A
- Authority
- CN
- China
- Prior art keywords
- clustering
- clusters
- cluster
- layer
- merging
- 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
- 238000000034 method Methods 0.000 title claims abstract description 68
- 230000005611 electricity Effects 0.000 title claims abstract description 42
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 18
- 230000002776 aggregation Effects 0.000 claims abstract description 13
- 238000005054 agglomeration Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 6
- 230000001960 triggered effect Effects 0.000 claims description 6
- 238000005354 coacervation Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 abstract description 11
- 238000004220 aggregation Methods 0.000 abstract description 5
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 238000004364 calculation method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Economics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Public Health (AREA)
- Probability & Statistics with Applications (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a clustering method of non-invasive load power consumption information, which considers that a Kmeans clustering algorithm can rapidly divide an original data set containing multiple objects into a specified number of clusters with lower calculated amount, so that the application firstly uses the Kmeans clustering algorithm to process an acquisition sequence of the power consumption information of a target electrical appliance to obtain a first preset number of middle-layer clusters, and then uses an aggregation hierarchical clustering method to circularly merge and cluster all the middle-layer clusters on the basis until a clustering result meets a preset iteration cutoff condition, thereby avoiding directly carrying out iterative operation on the original data set, reducing the calculated amount on the basis of less loss of precision, reducing the working pressure of a CPU and improving the classification speed. The invention also discloses a non-invasive load electricity information clustering device and equipment, which have the same beneficial effects as the non-invasive load electricity information clustering method.
Description
Technical Field
The invention relates to the field of information clustering, in particular to a non-invasive load electricity utilization information clustering method, and further relates to a non-invasive load electricity utilization information clustering device and equipment.
Background
In order to decompose and research the non-intrusive load, the power utilization information (including active power, reactive power, voltage, current and the like) of each electrical appliance must be classified, a mature method for classifying the power utilization information of the target electrical appliance is lacked in the prior art, massive calculation is often required during classification, on one hand, the working pressure of a CPU is increased, and on the other hand, the classification speed is slow.
Therefore, how to provide a solution to the above technical problem is a problem that needs to be solved by those skilled in the art.
Disclosure of Invention
The invention aims to provide a clustering method of non-invasive load electricity consumption information, which reduces the operation amount on the basis of less loss precision, reduces the working pressure of a CPU and improves the classification speed; another objective of the present invention is to provide a non-invasive load power consumption information clustering apparatus and device, which reduces the amount of computation on the basis of less loss of precision, reduces the working pressure of the CPU, and increases the classification speed.
In order to solve the technical problem, the invention provides a non-invasive load electricity consumption information clustering method, which comprises the following steps:
acquiring an acquisition sequence of power utilization information of a target electrical appliance;
dividing all objects in the acquisition sequence into a first preset number of middle-layer clusters in advance by adopting a Kmeans clustering algorithm;
performing merging clustering on all the middle-layer clusters by an agglomeration hierarchical clustering method;
judging whether the plurality of clusters after the last merging and clustering meet a preset iteration cutoff condition or not;
if yes, stopping iteration;
if not, taking the plurality of clusters after the last merging and clustering as the middle-layer clusters and executing the step of merging and clustering all the middle-layer clusters by the coacervation hierarchical clustering method.
Preferably, the pre-dividing all the objects in the acquisition sequence into a first preset number of middle-layer clusters by using a Kmeans clustering algorithm specifically includes:
randomly selecting a first preset number of objects in the acquisition sequence as an initial cluster center;
dividing each object except the initial cluster center in the acquisition sequence into the cluster where the initial cluster center closest to the object is located so as to obtain a first preset number of initial clusters;
updating the cluster center of each of the initial clusters;
judging whether the latest cluster center of each initial cluster is the same as the cluster center updated last time;
if not, executing the step of dividing each object except the initial cluster center in the acquisition sequence into the cluster where the initial cluster center closest to the object is located so as to obtain a first preset number of initial clusters;
and if so, taking the first preset number of clusters obtained by the last division as middle-layer clusters.
Preferably, the merging and clustering all the middle-layer clusters by the agglomerative hierarchical clustering method includes:
calculating the proximity between every two middle-layer clusters;
and carrying out cluster merging on the middle-layer clusters according to a second preset number with the minimum proximity.
Preferably, the performing cluster merging on the middle-layer clusters according to a second preset number with the minimum proximity includes:
combining all middle-layer cluster pairs with intersection in the middle-layer clusters by a second preset number pair with the minimum proximity into a middle-layer cluster;
and combining the middle layer cluster pairs with the minimum second preset number of pairs of the middle layer clusters without intersection into a middle layer cluster.
Preferably, the step of judging whether the plurality of clusters after the last merging and clustering meet the preset iteration cutoff condition specifically comprises:
judging whether the minimum cluster distance in the multiple clusters after the last merging and clustering is smaller than a preset threshold value or not;
after the cluster merging is performed on the middle-layer clusters according to the second preset number with the minimum proximity, and before the judgment on whether the minimum cluster distance in the multiple clusters after the last merging and clustering is smaller than the preset threshold value, the clustering method for the non-invasive load electricity utilization information further includes:
and updating the cluster center of each merged middle-layer cluster so as to calculate the cluster distance through the cluster center.
Preferably, the updating of the cluster center of each merged middle-layer cluster specifically includes:
wherein, Cnew(t) is the t-th characteristic value of a certain middle-layer cluster obtained after combination, nuRepresents the number of objects in the u-th cluster, u ∈ [1, q ]]Q is the total number of middle layer clusters used to merge the new cluster, q ∈ [2,2f ∈]F is the second predetermined number, Cu(t) represents the mean of the t-th features of all objects in the u-th cluster.
Preferably, the calculating the proximity between each two middle layer clusters specifically includes:
and calculating the Euclidean distance between every two middle-layer clusters.
Preferably, the first preset number is specifically:
wherein m is the first preset number, and n is the total number of objects in the acquisition sequence.
In order to solve the above technical problem, the present invention further provides a non-invasive load power consumption information clustering apparatus, including:
the acquisition module is used for acquiring an acquisition sequence of the power utilization information of the target electrical appliance;
the pre-dividing module is used for pre-dividing all objects in the acquisition sequence into a first preset number of middle-layer clusters by adopting a Kmeans clustering algorithm;
the merging module is used for merging and clustering all the middle-layer clusters by an agglomeration hierarchical clustering method;
the judging module is used for judging whether the plurality of clusters after the last merging and clustering meet a preset iteration cutoff condition, if so, the terminating module is triggered, and if not, the circulating module is triggered;
the termination module is used for stopping iteration;
and the circulation module is used for taking the plurality of clusters after the last merging and clustering as the middle-layer clusters and executing the step of merging and clustering all the middle-layer clusters by the coacervation hierarchical clustering method.
In order to solve the above technical problem, a non-invasive load power consumption information clustering device includes:
a memory for storing a computer program;
a processor for implementing the steps of the non-intrusive load electricity information clustering method as described above when executing the computer program.
The invention provides a clustering method of non-invasive load power consumption information, which considers that a Kmeans clustering algorithm can rapidly divide an original data set containing multiple objects into a specified number of clusters with lower calculated amount, so that the application firstly uses the Kmeans clustering algorithm to process an acquisition sequence of the power consumption information of a target electrical appliance to obtain a first preset number of middle-layer clusters, and then uses an aggregation hierarchical clustering method to circularly merge and cluster all the middle-layer clusters on the basis until a clustering result meets a preset iteration cutoff condition, thereby avoiding directly carrying out iterative operation on the original data set, reducing the calculated amount on the basis of less loss of precision, reducing the working pressure of a CPU and improving the classification speed.
The invention also provides a non-invasive load electricity information clustering device and equipment, and the clustering device and equipment have the same beneficial effects as the non-invasive load electricity information clustering method.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed in the prior art and the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
Fig. 1 is a schematic flow chart of a non-invasive load electricity consumption information clustering method according to the present invention;
fig. 2 is a schematic structural diagram of a non-invasive load electricity consumption information clustering apparatus provided in the present invention;
fig. 3 is a schematic structural diagram of a non-invasive load power consumption information clustering device provided by the present invention.
Detailed Description
The core of the invention is to provide a clustering method of non-invasive load electricity consumption information, which reduces the operation amount on the basis of less loss precision, reduces the working pressure of a CPU and improves the classification speed; the other core of the invention is to provide a non-invasive load electricity information clustering device and equipment, which reduce the operation amount on the basis of less loss precision, reduce the working pressure of a CPU and improve the classification speed.
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. 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 invention.
Referring to fig. 1, fig. 1 is a schematic flow chart of a method for clustering non-invasive load electricity consumption information provided in the present invention, where the method for clustering non-invasive load electricity consumption information includes:
step S1: acquiring an acquisition sequence of power utilization information of a target electrical appliance;
specifically, since the electricity consumption information is clustered, the electricity consumption information of the target electrical appliance is firstly obtained and used as a data base in the subsequent steps.
Step S2: dividing all objects in the acquisition sequence into a first preset number of middle-layer clusters in advance by adopting a Kmeans clustering algorithm;
specifically, considering that in the prior art, if the distance between each object in the electricity consumption information is directly calculated for clustering, as the number of the objects in the electricity consumption information is huge, the clustering can be completed only with extremely large computation amount, and the applicant considers that the Kmeans clustering algorithm can predetermine a first preset number of initial cluster centers, and then performs cluster generation around the initial cluster centers and optimizes continuously, so that the computation amount can be greatly reduced, therefore, in the application, the Kmeans clustering algorithm is firstly adopted to pre-divide all the objects in the acquisition sequence into a first preset number of middle-layer clusters.
It is worth mentioning that the features of each object in the electricity consumption information mainly include: active power, reactive power, apparent power, voltage, current, and current harmonics, etc., which are not limited herein.
The first preset number may be set autonomously, and is related to the number of objects in the power consumption information.
Step S3: merging and clustering all the middle-layer clusters by an agglomeration hierarchical clustering method;
specifically, the number of the plurality of middle-layer clusters obtained through the Kmeans clustering algorithm is determined from the beginning, and usually the first preset number is larger than the number requirement of the clusters obtained through final clustering, so that the clustering degree of the middle-layer clusters is not enough, the middle-layer clusters need to be further merged and clustered, and the final clustering requirement is met.
Because the middle layer clusters are already clustered to a certain degree, the aggregation level clustering does not consume much calculation amount.
Step S4: judging whether the plurality of clusters after the last merging and clustering meet a preset iteration cutoff condition or not;
specifically, an iteration cutoff condition can be preset for the clustering process of the power consumption information of the target electrical appliance, and judgment can be performed after merging and clustering are performed each time by the aggregation level clustering method, so that iteration is stopped at a proper time and clustering is completed.
Step S5: if yes, stopping iteration;
specifically, iteration can be stopped when the plurality of clusters after the last merging and clustering meet a preset iteration cutoff condition, and the plurality of clusters obtained at this time are the clustering results of the power utilization information of the target electrical appliance.
Step S6: if not, taking the plurality of clusters after the last merging and clustering as middle-layer clusters and executing the step of merging and clustering all the middle-layer clusters by an agglomeration hierarchical clustering method.
Specifically, when the plurality of clusters after the last merging and clustering do not meet the preset iteration cutoff condition, the method may return to step S3 to continue the iterative merging, so as to finally complete the clustering of the non-intrusive load power consumption information.
The invention provides a clustering method of non-invasive load power consumption information, which considers that a Kmeans clustering algorithm can rapidly divide an original data set containing multiple objects into a specified number of clusters with lower calculated amount, so that the application firstly uses the Kmeans clustering algorithm to process an acquisition sequence of the power consumption information of a target electrical appliance to obtain a first preset number of middle-layer clusters, and then uses an aggregation hierarchical clustering method to circularly merge and cluster all the middle-layer clusters on the basis until a clustering result meets a preset iteration cutoff condition, thereby avoiding directly carrying out iterative operation on the original data set, reducing the calculated amount on the basis of less loss of precision, reducing the working pressure of a CPU and improving the classification speed.
On the basis of the above-described embodiment:
as a preferred embodiment, the pre-dividing all objects in the acquisition sequence into a first preset number of middle-layer clusters by using a Kmeans clustering algorithm specifically includes:
randomly selecting a first preset number of objects in an acquisition sequence as an initial cluster center;
dividing each object except the initial cluster center in the acquisition sequence into the cluster where the initial cluster center closest to the object is located so as to obtain a first preset number of initial clusters;
updating the cluster center of each initial cluster;
judging whether the latest cluster center of each initial cluster is the same as the cluster center updated last time;
if not, executing a step of dividing each object except the initial cluster center in the acquisition sequence into the cluster where the initial cluster center closest to the object is located so as to obtain a first preset number of initial clusters;
if yes, the first preset number of clusters obtained by the last division are used as middle-layer clusters.
Specifically, since a first preset number of objects in the acquisition sequence are already used as initial cluster centers, it is necessary to determine cluster affiliations of the objects in the acquisition sequence except the initial cluster centers, in this application, the objects in the acquisition sequence except the initial cluster centers are respectively divided into clusters located closest to the initial cluster centers by a near-priority principle, so as to obtain a first preset number of initial clusters, which is to obtain the first preset number of initial clusters for the first time, and then "determine whether the composition of the objects in the first preset number of clusters can be further optimized" specifically, "determine whether all the cluster centers are the same as the cluster center updated last time", and if not, return to "execute" each object in the acquisition sequence except the initial cluster centers, and respectively dividing the cluster into the clusters where the initial cluster centers closest to the initial cluster centers are located so as to obtain a first preset number of initial clusters, wherein the final condition is that the latest cluster center of each initial cluster is the same as the cluster center updated last time, and at this time, the cluster which has obtained the first preset number of most reasonably clustered electricity utilization information objects is represented.
Specifically, it is assumed that the collection sequence of the power consumption information is X ═ X1,X2,...,Xn]Wherein X isi(i 1, 2.., n) is a single object of the dataset, and any object X isiHas the characteristic ofi1,Xi2...Xir]And r is the total number of features.
Wherein, when Kmeans clustering is carried out on the acquisition sequence:
first, m (a first preset number) objects are randomly selected as initial cluster centers, and the selected initial cluster centers may be recorded as Q ═ Q1,Q2,...,Qj,...,Qm]Then, each object X in the acquisition sequence except the initial cluster center is calculatediDistance d from each initial cluster center in QijThe calculation method is as follows:
Xitrepresenting an object XiIs the t-th feature of (1, n)]，QjtRepresenting an object QjIs the t-th feature of (1, m), j ∈ [, m ∈ [ ]]，t∈[1,r]。
Then, according to the calculated distance, each object is allocated to the cluster where the initial cluster center closest to the object is located, so as to obtain m initial clusters, and then the cluster center of each cluster needs to be updated, in the following updating manner:
wherein, CjCluster center, C, representing the jth clusterj(t) represents the mean of the t characteristic values of all the objects in the j cluster, njIndicates the number of objects within cluster j,
represents the second within cluster j
The t-th feature value of each object,
finally, the cluster center of the first preset number of middle-level clusters finally obtained in the Kmeans clustering process may be represented as C ═ C1,C2,...,Cm]。
As a preferred embodiment, the merged clustering of all the middle-layer clusters by the agglomerative hierarchical clustering method includes:
calculating the proximity between every two middle-layer clusters;
and carrying out cluster merging on the middle-layer clusters according to the second preset number with the minimum proximity.
Specifically, the second preset number in the embodiment of the present invention may be set autonomously, and when the number is set to be greater than 1, more than two clusters may be merged at the same time, so that the merging efficiency is high, and the clustering speed and efficiency are improved.
As a preferred embodiment, the cluster merging of the middle clusters according to the second preset number with the minimum proximity includes:
merging all middle-layer cluster pairs with intersection into a middle-layer cluster from the second preset number of middle-layer clusters with the minimum proximity;
and independently combining the middle-layer cluster pairs without intersection into a middle-layer cluster from the second preset number of middle-layer clusters with the minimum proximity.
Specifically, considering that when more than one pair of middle-layer cluster pairs are merged, there may be an intersection in the middle-layer cluster pairs to be merged, for example, when the second preset number is 2, the middle-layer cluster pairs to be merged include (1, 2) and (2, 4), and since there is an intersection in the two middle-layer cluster pairs, three clusters in the two middle-layer cluster pairs may be merged into one cluster at the same time, thereby further improving the merging efficiency.
As a preferred embodiment, the specific step of judging whether the plurality of clusters after the last merging and clustering satisfy the preset iteration cutoff condition is:
judging whether the minimum cluster distance in the multiple clusters after the last merging and clustering is smaller than a preset threshold value or not;
after the cluster merging is performed on the middle-layer clusters according to the second preset number with the minimum proximity, whether the minimum cluster distance in the multiple clusters after the last merging and clustering is smaller than a preset threshold value is judged, and the non-invasive load electricity utilization information clustering method further includes:
and updating the cluster center of each merged middle-layer cluster so as to calculate the cluster distance through the cluster center.
Specifically, in the merging and clustering process, if excessive merging is performed, the distance between each cluster is inevitably too small, and the situation of excessive clustering is considered, and in order to prevent the situation, in the embodiment of the present invention, the preset iteration cutoff condition is set as: the smallest cluster distance in the clusters after the last merging and clustering is smaller than a preset threshold value, and excessive clustering can be prevented.
Of course, in addition to the condition, the preset iteration cutoff condition may also be of another type, for example, the number of clusters after the last merging and clustering is smaller than a preset value, and the like.
As a preferred embodiment, updating the cluster center of each merged middle-layer cluster specifically includes:
wherein, Cnew(t) is the t-th characteristic value of a certain middle-layer cluster obtained after combination, nuRepresents the number of objects in the u-th cluster, u ∈ [1, q ]]Q is the total number of middle level clusters used to merge the new cluster, q ∈ [2,2f ∈]F is a second predetermined number, Cu(t) represents the mean of the t-th features of all objects in the u-th cluster.
Specifically, the process of updating the cluster center of the merged cluster in the embodiment of the invention is simpler, the calculated amount is less, and the clustering speed and efficiency are improved.
Of course, besides this method, updating the cluster center of the merged cluster may also be implemented in other ways, and the embodiment of the present invention is not limited herein.
As a preferred embodiment, the calculation of the proximity between each two middle-layer clusters is specifically as follows:
the euclidean distance between every two middle layer clusters is calculated.
Specifically, the calculation method of the euclidean distance is simple and efficient.
The process of calculating the euclidean distance between every two middle-layer clusters may be:
wherein, Ci,CjCluster centers that are two different middle tier clusters.
Of course, in addition to the euclidean distance, other distances may be used to represent the proximity between each two middle-layer clusters, and the embodiment of the present invention is not limited herein.
As a preferred embodiment, the first preset number is specifically:
wherein m is a first preset number, and n is the total number of objects in the acquisition sequence.
In particular toIn the embodiment of the present invention, the first predetermined number is specifically pairs
And (4) performing phase rounding to obtain a more reasonable value of m.
Of course, in addition to this mode, the determining mode of the first preset number may also be in other specific forms, and the embodiment of the present invention is not limited herein.
Referring to fig. 2, fig. 2 is a schematic structural diagram of a non-invasive load electricity information clustering apparatus according to another embodiment of the present invention, where the non-invasive load electricity information clustering apparatus includes:
the acquisition module 1 is used for acquiring an acquisition sequence of power utilization information of a target electrical appliance;
the pre-dividing module 2 is used for pre-dividing all objects in the acquisition sequence into a first preset number of middle-layer clusters by adopting a Kmeans clustering algorithm;
the merging module 3 is used for merging and clustering all the middle-layer clusters by an agglomeration hierarchical clustering method;
the judging module 4 is used for judging whether the plurality of clusters after the last merging and clustering meet a preset iteration cutoff condition, if so, the terminating module 5 is triggered, and if not, the circulating module 6 is triggered;
a termination module 5 for stopping the iteration;
and the circulating module 6 is used for taking the plurality of clusters after the last merging and clustering as middle-layer clusters and executing the step of merging and clustering all the middle-layer clusters by an agglomeration hierarchical clustering method.
For the introduction of the non-invasive load electricity consumption information clustering device provided in the embodiment of the present invention, reference is made to the foregoing embodiment of the non-invasive load electricity consumption information clustering method, and the embodiments of the present invention are not described herein again.
Referring to fig. 3, fig. 3 is a schematic structural diagram of a non-invasive load electricity consumption information clustering device according to another embodiment of the present invention, where the non-invasive load electricity consumption information clustering device includes:
a memory 7 for storing a computer program;
and a processor 8, configured to implement the steps of the clustering method for non-intrusive load electricity consumption information as in the foregoing embodiments when executing the computer program.
For the introduction of the non-invasive load electricity consumption information clustering device provided in the embodiment of the present invention, reference is made to the foregoing embodiment of the non-invasive load electricity consumption information clustering method, and the embodiments of the present invention are not described herein again.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description. It is further noted that, in the present specification, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A clustering method for non-intrusive load electricity utilization information is characterized by comprising the following steps:
acquiring an acquisition sequence of power utilization information of a target electrical appliance;
dividing all objects in the acquisition sequence into a first preset number of middle-layer clusters in advance by adopting a Kmeans clustering algorithm;
performing merging clustering on all the middle-layer clusters by an agglomeration hierarchical clustering method;
judging whether the plurality of clusters after the last merging and clustering meet a preset iteration cutoff condition or not;
if yes, stopping iteration;
if not, taking the plurality of clusters after the last merging and clustering as the middle-layer clusters and executing the step of merging and clustering all the middle-layer clusters by the coacervation hierarchical clustering method.
2. The method for clustering non-invasive load electricity consumption information according to claim 1, wherein the pre-dividing all the objects in the acquisition sequence into a first preset number of middle-layer clusters by using a Kmeans clustering algorithm specifically comprises:
randomly selecting a first preset number of objects in the acquisition sequence as an initial cluster center;
dividing each object except the initial cluster center in the acquisition sequence into the cluster where the initial cluster center closest to the object is located so as to obtain a first preset number of initial clusters;
updating the cluster center of each of the initial clusters;
judging whether the latest cluster center of each initial cluster is the same as the cluster center updated last time;
if not, executing the step of dividing each object except the initial cluster center in the acquisition sequence into the cluster where the initial cluster center closest to the object is located so as to obtain a first preset number of initial clusters;
and if so, taking the first preset number of clusters obtained by the last division as middle-layer clusters.
3. The method for clustering non-invasive load electricity information according to claim 2, wherein the performing cluster merging on all the middle-layer clusters by the agglomerative hierarchical clustering method comprises:
calculating the proximity between every two middle-layer clusters;
and carrying out cluster merging on the middle-layer clusters according to a second preset number with the minimum proximity.
4. The method as claimed in claim 3, wherein the clustering the middle-level clusters according to the second predetermined number with the minimum proximity comprises:
combining all middle-layer cluster pairs with intersection in the middle-layer clusters by a second preset number pair with the minimum proximity into a middle-layer cluster;
and combining the middle layer cluster pairs with the minimum second preset number of pairs of the middle layer clusters without intersection into a middle layer cluster.
5. The method for clustering non-invasive load electricity consumption information according to claim 4, wherein the step of judging whether the plurality of clusters after the last merged clustering meet the preset iteration cutoff condition is specifically:
judging whether the minimum cluster distance in the multiple clusters after the last merging and clustering is smaller than a preset threshold value or not;
after the cluster merging is performed on the middle-layer clusters according to the second preset number with the minimum proximity, and before the judgment on whether the minimum cluster distance in the multiple clusters after the last merging and clustering is smaller than the preset threshold value, the clustering method for the non-invasive load electricity utilization information further includes:
and updating the cluster center of each merged middle-layer cluster so as to calculate the cluster distance through the cluster center.
6. The method according to claim 5, wherein the updating the cluster center of each merged middle-layer cluster specifically comprises:
wherein, Cnew(t) is the t-th characteristic value of a certain middle-layer cluster obtained after combination, nuRepresents the number of objects in the u-th cluster, u ∈ [1, q ]]Q is the total number of middle layer clusters used to merge the new cluster, q ∈ [2,2f ∈]F is the second predetermined number, Cu(t) represents the mean of the t-th features of all objects in the u-th cluster.
7. The method as claimed in claim 3, wherein the calculating the proximity between each two middle-layer clusters is specifically:
and calculating the Euclidean distance between every two middle-layer clusters.
8. The method for clustering non-intrusive load electricity information according to any one of claims 2 to 7, wherein the first preset number is specifically:
wherein m is the first preset number, and n is the total number of objects in the acquisition sequence.
9. A clustering device for non-intrusive load electricity information is characterized by comprising:
the acquisition module is used for acquiring an acquisition sequence of the power utilization information of the target electrical appliance;
the pre-dividing module is used for pre-dividing all objects in the acquisition sequence into a first preset number of middle-layer clusters by adopting a Kmeans clustering algorithm;
the merging module is used for merging and clustering all the middle-layer clusters by an agglomeration hierarchical clustering method;
the judging module is used for judging whether the plurality of clusters after the last merging and clustering meet a preset iteration cutoff condition, if so, the terminating module is triggered, and if not, the circulating module is triggered;
the termination module is used for stopping iteration;
and the circulation module is used for taking the plurality of clusters after the last merging and clustering as the middle-layer clusters and executing the step of merging and clustering all the middle-layer clusters by the coacervation hierarchical clustering method.
10. A non-invasive load power consumption information clustering device is characterized by comprising:
a memory for storing a computer program;
a processor for implementing the steps of the method for clustering non-intrusive load electricity information according to any one of claims 1 to 8 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110474324.9A CN113052724A (en) | 2021-04-29 | 2021-04-29 | Non-invasive load electricity consumption information clustering method, device and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110474324.9A CN113052724A (en) | 2021-04-29 | 2021-04-29 | Non-invasive load electricity consumption information clustering method, device and equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113052724A true CN113052724A (en) | 2021-06-29 |
Family
ID=76518067
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110474324.9A Pending CN113052724A (en) | 2021-04-29 | 2021-04-29 | Non-invasive load electricity consumption information clustering method, device and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113052724A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140207398A1 (en) * | 2013-01-23 | 2014-07-24 | Samsung Electronics Co., Ltd | Transient Normalization for Appliance Classification, Disaggregation, and Power Estimation in Non-Intrusive Load Monitoring |
| CN110569876A (en) * | 2019-08-07 | 2019-12-13 | 武汉中原电子信息有限公司 | Non-invasive load identification method and device and computing equipment |
-
2021
- 2021-04-29 CN CN202110474324.9A patent/CN113052724A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140207398A1 (en) * | 2013-01-23 | 2014-07-24 | Samsung Electronics Co., Ltd | Transient Normalization for Appliance Classification, Disaggregation, and Power Estimation in Non-Intrusive Load Monitoring |
| CN110569876A (en) * | 2019-08-07 | 2019-12-13 | 武汉中原电子信息有限公司 | Non-invasive load identification method and device and computing equipment |
Non-Patent Citations (3)
| Title |
|---|
| 张颖超等: "基于模糊层次聚类与模糊推论技术的模糊信息检索方法的研究", 《濮阳职业技术学院学报》 * |
| 徐超等: "基于聚类分析的入侵检测算法", 《指挥控制与仿真》 * |
| 陈晓等: "大数据处理中混合型聚类算法的研究与实现", 《信息网络安全》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Lin et al. | Combining partitional and hierarchical algorithms for robust and efficient data clustering with cohesion self-merging | |
| Smiti et al. | Dbscan-gm: An improved clustering method based on gaussian means and dbscan techniques | |
| Zhou et al. | A clustering-Based KNN improved algorithm CLKNN for text classification | |
| CN109066687A (en) | A kind of electric power system tide calculation method, system and electronic equipment | |
| CN109901932A (en) | A kind of Server Consolidation method based on virtual machine | |
| CN111950620A (en) | User screening method based on DBSCAN and K-means algorithm | |
| CN110543889A (en) | power load hierarchical clustering method and device, computer equipment and storage medium | |
| CN106610977B (en) | Data clustering method and device | |
| Li et al. | Fast density peaks clustering algorithm based on improved mutual K-nearest-neighbor and sub-cluster merging | |
| TWI740895B (en) | Distribution method and device for application attribution service cluster | |
| Tsai et al. | An effective and efficient grid-based data clustering algorithm using intuitive neighbor relationship for data mining | |
| Jain | A hybrid clustering algorithm for data mining | |
| He et al. | Parallel feature selection using positive approximation based on mapreduce | |
| CN114358378A (en) | User side energy storage optimal configuration system and method for considering demand management | |
| CN111797899B (en) | Low-voltage transformer area kmeans clustering method and system | |
| Zhuang et al. | Impact of instance seeking strategies on resource allocation in cloud data centers | |
| CN113052724A (en) | Non-invasive load electricity consumption information clustering method, device and equipment | |
| Xu et al. | Manifold density peaks clustering algorithm | |
| Abd-el Fattah et al. | A TOPSIS based method for gene selection for cancer classification | |
| CN116722589A (en) | Distributed energy power distribution network topological structure selection method and related device | |
| Zhao et al. | K-centers algorithm for clustering mixed type data | |
| Wang et al. | A clustering algorithm based on FDP and DBSCAN | |
| Liu et al. | Accurate grid-based clustering algorithm with diagonal grid searching and merging | |
| Rani et al. | Clustering analysis by Improved Particle Swarm Optimization and K-means algorithm | |
| CN106339413A (en) | Approximate membership query method based on high-dimensional data filter |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210629 |