CN115935212A - Adjustable load clustering method and system based on longitudinal trend prediction - Google Patents
Adjustable load clustering method and system based on longitudinal trend prediction Download PDFInfo
- Publication number
- CN115935212A CN115935212A CN202211565309.6A CN202211565309A CN115935212A CN 115935212 A CN115935212 A CN 115935212A CN 202211565309 A CN202211565309 A CN 202211565309A CN 115935212 A CN115935212 A CN 115935212A
- Authority
- CN
- China
- Prior art keywords
- load
- adjustable
- state
- matrix
- clustering
- 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
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
An adjustable load clustering method and system based on longitudinal trend prediction divide adjustable load historical data into equal probability state intervals with different output levels, and convert a constructed longitudinal load data matrix into a state variable matrix; calculating state transition matrixes of the loads at different days at each moment to obtain predicted output curves of different loads, and calculating an adjustable rate curve of the corresponding load according to the predicted load and the average value of the historical loads; and clustering the adjustable rate curves of different adjustable loads obtained based on the longitudinal trend prediction by adopting a clustering algorithm, and outputting to obtain a final clustering result. The adjustable load objects with similar adjustable capacity are effectively classified by representing the adjustable capacity of different adjustable loads at different levels through the adjustable rate and clustering according to the adjustable capacity, so that a solid foundation is provided for hierarchical and partitioned aggregation and control scheduling of the adjustable loads, and efficient utilization of regional adjustable load resources is promoted.
Description
Technical Field
The invention belongs to the technical field of energy and power service, and particularly relates to an adjustable load clustering method and system based on longitudinal trend prediction.
Background
With diversified loads such as distributed power supplies and electric vehicles widely connected to an electric power system, the difficulty of load prediction, scheduling control and other work is gradually increased. The users with different adjusting potentials are reasonably classified, and the power utilization behaviors of the users are mastered, so that the method has important significance in the aspects of load prediction, demand side management, power utilization pricing and the like. The user power utilization load has larger uncertainty, and the daily load curve shows the power utilization behavior of the user in one day and shows the transverse characteristic of the load; different daily load curves over a period of time, such as a week or month, may also differ, as may be apparent from the longitudinal nature of the load. The longitudinal difference degrees of different users are obviously different, the adjustment capacities of different types of adjustable loads can be measured by researching the longitudinal trend change of the adjustable loads, and the reasonable classification of the users with different adjustment capacities is facilitated.
Most of the existing research on load clustering aims at the transverse characteristic of the load, for the pretreatment of a daily load curve of multiple days, only an average value is simply taken, or after an abnormal value is eliminated, a certain day is selected as a typical load day, and the like, the consideration on the longitudinal characteristic of the load is lacked, and part of useful information is lost in the pretreatment process. When the daily load curve of a user is greatly changed and the electricity utilization behavior of the user is difficult to characterize by using a typical load curve, the longitudinal fluctuation of the load needs to be considered. By researching the longitudinal trend, corresponding adjustable rate indexes can be established to measure the adjusting capability of different adjustable loads, the adjustable loads with similar adjusting capability in the region can be effectively classified by corresponding clustering methods, and the efficient utilization of region adjustable load resources is promoted.
Chinese patent CN 112884077A "campus short-term load prediction method based on shape dynamic time regression clustering" discloses a method for short-term load prediction, aiming at improving prediction accuracy, but does not relate to load regulation capability analysis.
Disclosure of Invention
In order to solve the defects in the prior art, the invention provides an adjustable load clustering method and system based on longitudinal trend prediction, which are characterized in that acquired historical data of different adjustable loads under a certain time scale are grouped according to the same moment and different days, equal probability state intervals with different output levels are divided through normalization processing, and a constructed longitudinal load data matrix is converted into a state variable matrix; calculating a state transition matrix of the load at each moment and different days, calculating a load predicted value of each moment point of a target day according to the initial state and the transition matrix to obtain predicted output curves of different loads, and calculating an adjustable rate curve of the corresponding load according to the predicted load and the historical load average value; and (3) clustering the adjustable rate curves of different adjustable loads obtained based on longitudinal trend prediction by adopting a Canopy-Kmeans clustering algorithm, and outputting to obtain a final clustering result. According to the invention, the longitudinal trend changes of different adjustable loads are fully considered, the adjustable capacities of different adjustable loads at different levels are represented through the adjustable rate and are clustered according to the adjustable capacity, and the adjustable load objects with similar adjustable capacities are aggregated, so that a solid foundation is provided for hierarchical and regional aggregation and control scheduling of the adjustable loads.
The invention adopts the following technical scheme.
An adjustable load clustering method based on load longitudinal trend prediction specifically comprises the following steps:
step 1, selecting historical adjustable load data of users in different industries, grouping the historical load data according to the same moment and different days, and constructing a longitudinal load data matrix;
step 2, carrying out normalization processing on historical load data, dividing equal probability state intervals with different output levels, and converting the constructed longitudinal load data matrix into a state variable matrix;
step 3, calculating a state transition matrix of the load at different days at each moment, calculating a load predicted value of each moment point of the target day according to the initial state and the transition matrix to obtain predicted output curves of different loads, and calculating an adjustable rate curve of the corresponding load according to the predicted load and the historical load average value;
and 4, clustering the adjustable rate curves of different adjustable loads obtained based on longitudinal trend prediction by adopting a clustering algorithm, and outputting typical user clustering results obtained under different adjustable levels.
Preferably, in step 1, historical adjustable load data with a time scale of m days is selected, the historical load data of each day is divided into n time points at equal time intervals, and a longitudinal load data matrix L is constructed according to the divided load data sequence, wherein the longitudinal load data matrix L is represented by the following formula:
L i ＝[L i ,…L i,j …L i,m ],j＝1,2,3,…,m
wherein L is i A data vector formed by the historical load values of m days at the ith moment; l is i,j The historical load value at the ith time point of the jth day.
Preferably, in step 2, the load data sequence is first tested and processed for outliers by, for example and without limitation, the 3 σ principle, that is, when the load value distribution exceeds the range of (μ -3 σ, μ +3 σ), it is considered as outliers and removed, where μ is the mean value of the load data sequence and σ is the standard deviation of the load data sequence. This process is to prevent the abnormal value from affecting the subsequent state partitioning, thereby affecting the accuracy of prediction.
Normalizing the load data after outlier processing, which is defined as:
wherein L is i,min The minimum value of the historical load values of m days at the ith moment; l is i,max Is the maximum value in the historical load values of m days at the ith moment.
The normalized load value satisfies L' i,j Belongs to (0, 1), and divides the range into K state intervals with the interval length of
The resulting state interval S is as follows:
S＝(S 1 ,S 2 ,…,S k ,…,S K )
s is a total state interval divided by the adjustable load output equal probability; s k The kth substate interval is obtained for the division.
Distributing each historical load value to a corresponding output state interval according to the size of the historical load value, and converting a longitudinal load data matrix into a state matrix E, wherein the state matrix E is as follows:
E i,j ∈S
wherein E is i,j The load state at the ith time on the jth day.
Preferably, in step 3, the state matrix E at the ith time m days i Within h different output states [ E' 1 ,E′ 2 ,E′ 3 ,…,E′ h ]Calculating a state transition probability matrix P i The formula is as follows:
E′ a ,E′ b ∈[E′ 1 ,E′ 2 ,E′ 3 ,…,E′ h ]
wherein, P a,b Is state E' a To state E' b The probability of a transition; n (E' a →E′ b ) Is state E 'in the state matrix at the ith time of m days' a To state E' b Statistic of transitions, N (E' a ) Is state E' a The statistical quantity of (a).
Converting the load state at the ith time of the mth day into a load state probability matrix
As an initial state, calculating a load state probability matrix pi of a corresponding time point of a target m +1 day according to the initial state and a transition matrix 1 The formula is as follows:
π i,1 ,π i,2 ,…,π i,h ∈[0,1]
taking the output state with the highest probability in the load state probability matrix at the ith time point on the m +1 th day as the output state at the point on the prediction day, and taking the median value of the state interval to which the output state is attributed as an output value L' i,m+1 Further obtaining a predicted output curve of the load on the (m + 1) th day;
according to the load predicted value L 'at the ith moment of the m +1 th day' i,m+1 Load mean of calendar history of near m
Calculating corresponding adjustable rate lambda i,m+1 Thereby obtaining the adjustable rate curve of the load target day.
Wherein λ is i,m+1 The adjustable rate of the ith moment of the m +1 th day of the load is used for representing the adjusting capacity of the load point when lambda is i,m+1 > 1 indicates that the point load has upward adjustability when lambda i,m+1 A larger difference value of-1 indicates a stronger adjustability at the time point; when 0 < lambda i,m+1 < 1 indicates that the point load has downward adjustability when the load is 1-lambda i,m+1 A larger difference indicates a greater adjustability.
Preferably, the Canopy-measures clustering algorithm is adopted in the step 4, and the method specifically comprises the following steps:
step 4.1, aiming at different user target day adjustable rate data W N Respectively randomly arranging, and respectively setting initial data sets W 1 ,W 2 ,…W N Selecting a central point P from the initial clustering sample set according to three indexes of change rate, peak-valley difference and average regulation rate N Wherein, in the step (A),
rate of change:
wherein the content of the first and second substances,
average value of the adjustable rate curve, G m Is the maximum value of the adjustable rate curve;
peak-to-valley difference:
Δ＝G m -G n
wherein, G n Is the minimum value of the adjustable rate curve;
step 4.2, selecting the distance closest to the central point as a distance threshold value T 2-N The distance from the center point to the farthest is the distance threshold T 1-N And T is 1-N ＞T 2-N ；
Step 4.3, adding P N As the cluster center point of the first cluster, and the point P N From the initial cluster sample set W N Removing;
step 4.4, from the remaining set of data samples W N In randomly selecting a point Q N Calculating Q N Distances to all known cluster center points are examined, wherein the minimum distance D is examined N : if T is 2-N ≤D N ≤T 1-N Record Q with a weak mark N Represents a point Q N Belongs to the cluster, and Q N Adding into the mixture; if D is N ≤T 2-N Recording the point Q with a strong mark N Represents a point Q N Belongs to the cluster, and Q N From the data sampleThis set S N Deleting; if D is N ＞T 1-N Then Q is obtained N Forming a new cluster, and combining Q N From a set of data samples W N Deleting;
step 4.5, repeat step 4.4 until set W N The number of the elements in the formula is zero;
step 4.6, generating K N Individual cluster center y 1 ，y 2 ，……，y KN Selecting the change rate, the peak-valley difference and the average regulation rate as clustering evaluation indexes;
step 4.7, calculating the similarity between the adjustable rate curve of each user and the clustering evaluation index of the clustering center point, adding the user into the cluster with the highest similarity with the center point, and updating the clustering center point;
and 4.8, iteratively executing the step 4.7 until the iteration step number is 500, and stopping. And obtaining typical user clustering results under different adjustable levels.
An adjustable load clustering system based on load longitudinal trend prediction comprises a user historical load data collection module, an adjustable load prediction module, an adjustable rate calculation module and a clustering analysis module.
The user historical load data collection module selects historical adjustable load data of m days, groups the historical adjustable load data according to different days at a uniform moment, and constructs a longitudinal load data matrix;
the adjustable load prediction module is used for carrying out normalization processing on historical adjustable load data, dividing equal probability intervals with different output levels, and converting a constructed longitudinal load data matrix into a state variable matrix;
the adjustable rate calculation module calculates state transition matrixes of the loads at different days at each moment, calculates load predicted values of target days at each moment point according to the initial states and the transition matrixes to obtain predicted output curves of different loads, and calculates adjustable rate curves of corresponding loads according to the predicted loads and the historical load average value;
and the clustering analysis module is used for clustering the adjustable rate curves of different adjustable loads obtained based on the longitudinal trend prediction by adopting a clustering algorithm and outputting typical user clustering results obtained under different adjustable levels.
Compared with the prior art, the invention has the beneficial effects that:
1. the clustering algorithm is utilized to classify the user data in different industries on the adjustment level and the time scale, classification basis is provided for analyzing the adjustable potential of different types of users, and the obtained typical user curve can be used as the calculation basis for accurately calculating the adjustable capacity of the load in the future.
2. The load-adjustable prediction method can effectively process the possible instability of input data and effectively reduce the influence of the instability of the data on a prediction result.
3. The method utilizes the maximum probability output and the average historical output to establish the adjustable capacity index, does not need to establish a complex equipment load model, and effectively and quickly measures the adjustable capacity of the adjustable load.
Drawings
FIG. 1 is a flow chart of an adjustable load clustering method based on load longitudinal trend prediction according to the present invention;
FIG. 2 is a flow chart of the adjustable load prediction of the present invention;
FIG. 3 is a flow chart of the Canopy-Kmeans clustering algorithm;
FIG. 4 is a diagram illustrating the original longitudinal load data of the No. 1 user in the embodiment during the whole period of 31 days;
FIG. 5 is a schematic diagram of longitudinal load data after normalization processing in the embodiment;
FIG. 6 is a schematic diagram of a longitudinal load prediction curve of user No. 1 at time 0 in the embodiment;
FIG. 7 is a schematic diagram showing an adjustable rate curve for 24 predicted days by the user in the embodiment;
FIG. 8 is a diagram illustrating an initialized user curve clustering center in an embodiment;
FIG. 9 is a diagram illustrating typical user curves after clustering in the embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be described clearly and completely in the following with reference to the accompanying drawings in the embodiments of the present invention. The embodiments described herein are only some embodiments of the invention, and not all embodiments. All other embodiments obtained by a person skilled in the art without any inventive step based on the spirit of the present invention are within the scope of the present invention.
The embodiment 1 of the invention provides an adjustable load clustering method based on load longitudinal trend prediction, which specifically comprises the following steps:
step 1, selecting historical adjustable load data of users in different industries, grouping the historical load data according to the same moment and different days, and constructing a longitudinal load data matrix;
in step 1, selecting historical adjustable load data with a time scale of m days, dividing the historical load data of each day into n time points at equal time intervals, and constructing a longitudinal load data matrix L according to the divided load data sequence, wherein the longitudinal load data matrix L is shown as the following formula:
L i ＝[L i ,…L i,j …L i,m ],j＝1,2,3,…,m
wherein L is i A data vector formed by the historical load values of m days at the ith moment; l is i,j The historical load value at the ith time point of the jth day.
Step 2, carrying out normalization processing on historical load data, dividing equal probability state intervals with different output levels, and converting the constructed longitudinal load data matrix into a state variable matrix;
in step 2, the load data sequence is first tested and processed for outliers, such as but not limited to the 3 σ principle, that is, when the load value distribution exceeds the range of (μ -3 σ, μ +3 σ), it is considered as outliers and removed, where μ is the mean value of the load data sequence and σ is the standard deviation of the load data sequence. This process is to prevent the abnormal value from affecting the subsequent state division, thereby affecting the accuracy of prediction.
Normalizing the load data after outlier processing, and defining the normalization as follows:
wherein L is i,min Is the minimum value in the historical load values of m days at the ith moment; l is i,max Is the maximum value in the historical load values of m days at the ith moment.
The normalized load value meets L' i,j Belongs to (0, 1), the equal probability of the range is divided into K state intervals, the length of the interval is
The resulting state interval S is as follows:
S＝(S 1 ,S 2 ,…,S k ,…,S K )
s is a total state interval divided by the adjustable load output equal probability; s. the k The kth substate interval is obtained for the division.
Distributing each historical load value to a corresponding output state interval according to the size of the historical load value, and converting a longitudinal load data matrix into a state matrix E, wherein the state matrix E is as follows:
E i ＝[E i,1 …E i,j …E i,m ],j＝1,2,3,…,m
E i,j ∈S
wherein E is i,j The load state at the ith time on the jth day.
Step 3, calculating a state transition matrix of the load at different days at each moment, calculating a load predicted value of each moment point of the target day according to the initial state and the transition matrix to obtain predicted output curves of different loads, and calculating an adjustable rate curve of the corresponding load according to the predicted load and the historical load average value;
in step 3, the state matrix E of m days at the ith time i Within h different states of output [ E' 1 ,E′ 2 ,E′ 3 ,…,E′ h ]Computing a state transition probability matrix P i The formula is as follows:
E′ a ,E′ b ∈[E′ 1 ,E′ 2 ,E′ 3 ,…,E′ h ]
wherein, P a,b Is state E' a To state E' b The probability of a transition; n (E' a →E′ b ) Is state E 'in the state matrix at time i of m days' a To state E' b Statistic of transitions, N (E' a ) Is state E' a The statistical quantity of (c).
Load state probability matrix for transforming load state at the ith time of the mth day
As an initial state, calculating a load state probability matrix pi of a corresponding time point of a target m +1 day according to the initial state and a transition matrix 1 The formula is as follows:
π i,1 ,π i,2 ,…,π i,h ∈[0,1]
taking the output state with the highest probability in the load state probability matrix of the ith time point on the m +1 th day as the output state of the point on the predicted day, and taking the median value of the state interval to which the output state is attributed as an output value L' i,m+1 Further obtaining the predicted value of the time point of the day so as to obtain the predicted output curve of the load on the (m + 1) th day;
according to the load predicted value L 'at the ith moment of the m +1 th day' i,m+1 Load mean of calendar history of near m
Calculating the corresponding adjustable rate lambda i,m+1 Thereby obtaining the adjustable rate curve of the load target day. />
Wherein λ is i,m+1 The adjustable rate of the ith moment of the m +1 th day of the load is used for representing the adjusting capacity of the load point when lambda is i,m+1 > 1 indicates that the point load has upward adjustability when lambda i,m+1 A larger difference value of-1 indicates a stronger adjustability at the time point; when 0 < lambda i,m+1 < 1 indicates that the point load has downward adjustability when the load is 1-lambda i,m+1 A larger difference indicates a greater adjustability.
And 4, clustering the adjustable rate curves of different adjustable loads obtained based on the longitudinal trend prediction by adopting a Canopy-Kmeans clustering algorithm, and outputting typical user clustering results obtained under different adjustable levels.
It is worth noting that the clustering algorithm can be implemented in various forms, including K-means, hierarchical clustering, DBSCAN, etc., in order to achieve automatic determination of cluster number and cluster center, and to achieve faster speed and more convenient calculation, the preferred embodiment of the present invention uses Canopy-kmmeans clustering algorithm as an implementation mode. However, this is only a preferred but non-limiting embodiment, and those skilled in the art can obtain the clustering result in any other form under the spirit of the present invention, and all fall within the scope of the present invention.
More particularly, in a preferred but non-limiting embodiment of the invention
Step 4.1, target daily adjustable rate data W for different users N Respectively randomly arranging, and respectively setting initial data sets W 1 ,W 2 ,…W N Selecting a central point P from the initial clustering sample set according to three indexes of change rate, peak-valley difference and average regulation rate N Wherein, in the step (A),
rate of change:
wherein, the first and the second end of the pipe are connected with each other,
average value of the adjustable rate curve, G m Is the maximum value of the adjustable rate curve;
peak-to-valley difference:
Δ＝G m -G n
wherein G is n Is the minimum value of the adjustable rate curve;
step 4.2, selecting the distance closest to the central point as a distance threshold value T 2-N The distance from the center point is the distance threshold T 1-N And T is 1-N ＞T 2-N ；
Step 4.3, adding P N As the cluster center point of the first cluster, and the point P N From the initial cluster sample set W N Removing;
step 4.4, from the remaining set of data samples W N In randomly selecting a point Q N Calculating Q N Distances to all known cluster center points, wherein the minimum distance D is examined N : if T is 2-N ≤D N ≤T 1-N Then record Q with a weak mark N Represents a point Q N Belongs to the cluster, and Q N Adding into the mixture; if D is N ≤T 2-N Recording the point Q with a strong mark N Represents a point Q N Belongs to the cluster, and Q N From a set of data samples S N Deleting; if D is N ＞T 1-N Then Q is N Forming a new cluster, and combining Q N From a set of data samples W N Deleting;
step 4.5, repeat step 4.4 until set W N The number of the elements in the formula is zero;
step 4.6, generating K N Individual cluster center y 1 ，y 2 ，……，y KN Selecting the change rate, the peak-valley difference and the average regulation rate as clustering evaluation indexes;
step 4.7, calculating the similarity between the adjustable rate curve of each user and the clustering evaluation index of the clustering center point, adding the user into the cluster with the highest similarity with the center point, and updating the clustering center point;
and 4.8, iterating the step 4.7 until the number of iteration steps is 500, and stopping. And obtaining typical user clustering results under different adjustable levels.
In a preferred but non-limiting embodiment of the present invention, the adjustable loads of users in different industries in the area are clustered according to the load sample data of 24 users in a certain area; wherein, the area has 24 users, and the sampling point is 24 per day.
Fig. 4 is the raw longitudinal load data of the user No. 1 in the preferred embodiment at 24 hours and 31 days, and it can be seen that the longitudinal load trend of the user No. 1 at different times is approximately similar, and the longitudinal load curve on the working day is relatively smooth, and the non-working day is the valley period of the longitudinal load. Fig. 5 is the longitudinal load data after the normalization processing, and it can be seen from the graph that the similarity of the longitudinal load curve after the normalization processing is more significant. Fig. 6 is a longitudinal load prediction curve of user No. 1 at time 0, and it can be seen that the prediction result is closer to the actual value, and the prediction effect is good. Fig. 7 is a curve of the adjustable rates of the predicted days of the 24 users, and as can be seen from the curve, the adjustable rates of the 24 users are between 0.85 and 1.2, and the maximum adjustable rates of the 24 users are mainly distributed between 6 and 8, that is, the adjustable capacities of the users reach the peak value within the time period.
Fig. 8 is an initialized user curve cluster center, and fig. 9 is a typical user curve after clustering. According to curve analysis, the user curve of type 1 has 1 typical curve, presents obvious "double peak" characteristic, the load adjustability rate is greater than 1, i.e. presents upward adjustability characteristic, the adjustment margin is within 12.5%, the maximum value of the adjustment capability is distributed between 7; the profile for a user of type 2 has 2 typical profiles, one of which exhibits a "unimodal" characteristic and an adjustability rate greater than 1, the load exhibits an upward adjustable characteristic, the adjustment margin is within 15%, the maximum value of the adjustability is distributed between 7 and 00 and 8, the minimum value of the adjustability occurs between 2. According to the method, the adjustable potentials of different users in a certain area can be classified, and technical support is provided for power demand side management and load scheduling.
Embodiment 2 of the present invention provides an adjustable load clustering system based on load longitudinal trend prediction, which operates the adjustable load clustering method based on load longitudinal trend prediction as described in embodiment 1, and includes a user historical load data collection module, an adjustable load prediction module, an adjustable rate calculation module, and a cluster analysis module.
The user historical load data collection module selects historical adjustable load data of m days, groups the historical adjustable load data according to different days at a uniform moment, and constructs a longitudinal load data matrix;
the adjustable load prediction module is used for carrying out normalization processing on historical adjustable load data, dividing equal probability intervals with different output levels, and converting a constructed longitudinal load data matrix into a state variable matrix;
the adjustable rate calculation module calculates state transition matrixes of the loads at different days at each moment, calculates load predicted values of target days at each moment point according to the initial states and the transition matrixes to obtain predicted output curves of different loads, and calculates adjustable rate curves of corresponding loads according to the predicted loads and the historical load average value;
and the clustering analysis module is used for clustering the adjustable rate curves of different adjustable loads obtained based on the longitudinal trend prediction by adopting a clustering algorithm and outputting typical user clustering results obtained under different adjustable levels.
The present disclosure may be systems, methods, and/or computer program products. The computer program product may include a computer-readable storage medium having computer-readable program instructions embodied thereon for causing a processor to implement various aspects of the present disclosure.
The computer-readable storage medium may be a tangible device that can hold and store the instructions for use by the instruction execution device. The computer readable storage medium may be, for example, but not limited to, an electronic memory device, a magnetic memory device, an optical memory device, an electromagnetic memory device, a semiconductor memory device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a Static Random Access Memory (SRAM), a portable compact disc read-only memory (CD-ROM), a Digital Versatile Disc (DVD), a memory stick, a floppy disk, a mechanical coding device, such as punch cards or in-groove projection structures having instructions stored thereon, and any suitable combination of the foregoing. Computer-readable storage media as used herein is not to be construed as transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission medium (e.g., optical pulses through a fiber optic cable), or electrical signals transmitted through electrical wires.
The computer-readable program instructions described herein may be downloaded from a computer-readable storage medium to a respective computing/processing device, or to an external computer or external storage device via a network, such as the internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers and/or edge servers. The network adapter card or network interface in each computing/processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in the respective computing/processing device.
Computer program instructions for carrying out operations of the present disclosure may be assembler instructions, instruction Set Architecture (ISA) instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider). In some embodiments, the electronic circuitry that can execute the computer-readable program instructions implements aspects of the present disclosure by utilizing the state information of the computer-readable program instructions to personalize the electronic circuitry, such as a programmable logic circuit, a Field Programmable Gate Array (FPGA), or a Programmable Logic Array (PLA).
Finally, it should be noted that the above embodiments are only used for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications and equivalents may be made to the embodiments of the invention without departing from the spirit and scope of the invention, which is to be covered by the claims.
Claims (11)
1. An adjustable load clustering method based on longitudinal trend prediction is characterized by comprising the following steps:
step 1, selecting historical adjustable load data of users in different industries, grouping the historical adjustable load data according to the same moment and different days, and constructing a longitudinal load data matrix;
step 2, carrying out normalization processing on historical adjustable load data, dividing equal probability intervals with different output levels, and converting the constructed longitudinal load data matrix into a state variable matrix;
step 3, calculating a state transition matrix of the load at different days at each moment, calculating a load predicted value of each moment point of the target day according to the initial state and the transition matrix to obtain predicted output curves of different loads, and calculating an adjustable rate curve of the corresponding load according to the predicted load and the historical load average value;
and 4, clustering the adjustable rate curves of different adjustable loads obtained based on longitudinal trend prediction by adopting a clustering algorithm to obtain typical user clustering results under different adjustable levels, and thus, carrying out hierarchical partition aggregation and control scheduling on the adjustable loads according to the adjusting capability of the different adjustable loads.
2. The adjustable load clustering method based on longitudinal trend prediction as claimed in claim 1, wherein:
in step 1, selecting historical adjustable load data with a time scale of m days, dividing the historical load data of each day into n time points at equal time intervals, and constructing a longitudinal load data matrix L according to the divided load data sequence, wherein the longitudinal load data matrix L is shown as the following formula:
L i ＝[L i ,…L i,j …L i,m ],j＝1,2,3,…,m
wherein L is i A data vector formed by the historical load values of m days at the ith moment; l is a radical of an alcohol i,j The historical load value of the ith time point of the jth day.
3. The adjustable load clustering method based on longitudinal trend prediction as claimed in claim 1, wherein:
and 2, performing outlier detection and processing on the load data sequence in the step 1, and when the load numerical distribution exceeds the range of (mu-3 sigma, mu +3 sigma), determining that the load numerical distribution is an outlier and removing the outlier, wherein mu is the mean value of the load data sequence, and sigma is the standard deviation of the load data sequence.
4. The adjustable load clustering method based on longitudinal trend prediction as claimed in claim 3, wherein:
in step 2, dividing the load value range after the normalization into K state intervals with equal probability, wherein the interval length is
The resulting state interval S is:
S＝(S 1 ，S 2 ，…，S k ，…，S K )
s is a total state interval divided by the adjustable load output equal probability; s. the k Obtaining a kth sub-state interval for the division;
distributing each historical load value to a corresponding output state interval according to the size of the historical load value, and converting a longitudinal load data matrix into a state matrix E, wherein the state matrix E is as follows:
E i ＝[E i，1 …E i，j …E i，m ]，j＝1，2，3，…，m
E i，j ∈S
wherein E is i，j The load state at the ith time on the jth day.
5. The adjustable load clustering method based on longitudinal trend prediction as claimed in claim 1, wherein:
in step 3, the state matrix E of m days at the ith time i Within h different output states [ E' 1 ，E′ 2 ，E′ 3 ，…，E′ h ]Calculating a state transition probability matrix P i The formula is as follows:
E′ a ,E′ b ∈[E′ 1 ,E′ 2 ,E′ 3 ,…,E′ h ]
wherein, P a,b Is state E' a To state E' b The probability of a transition; n (E' a →E′ b ) Time i of day mState E 'in the state matrix' a To state E' b Statistic of transitions, N (E' a ) Is state E' a The statistical quantity of (a).
6. The adjustable load clustering method based on longitudinal trend prediction as claimed in claim 5, wherein:
converting the load state at the ith time of the mth day into a load state probability matrix
As an initial state, calculating a load state probability matrix pi of a corresponding time point of the target m +1 day according to the initial state and the transition matrix 1 The formula is as follows:
7. the adjustable load clustering method based on longitudinal trend prediction as claimed in claim 6, wherein:
taking the output state with the highest probability in the load state probability matrix at the ith time point on the m +1 th day as the output state at the point on the prediction day, and taking the median value of the state interval to which the output state is attributed as an output value L' i,m+1 Further obtaining a predicted output curve of the load on the m +1 th day;
according to the load predicted value L 'at the ith moment of the m +1 th day' i,m+1 Load mean of calendar history of near m
Calculating corresponding adjustable rate lambda i,m+1 Thereby obtaining an adjustable rate curve of the load target day;
wherein λ is i,m+1 The adjustable rate of the ith moment of the m +1 th day of the load is used for representing the adjusting capacity of the load point when lambda is i,m+1 > 1 indicates that the point load has upward adjustability when lambda i,m+1 A larger difference value of-1 indicates a stronger adjustability at the time point; when 0 < lambda i,m+1 < 1 indicates that the point load has downward adjustability when the load is 1-lambda i,m+1 A larger difference indicates a greater adjustability.
8. The adjustable load clustering method based on longitudinal trend prediction as claimed in claim 1, wherein:
in the step 4, a Canopy-Kmeans clustering method is adopted.
9. An adjustable load clustering system based on longitudinal trend prediction, which is used for realizing the adjustable load clustering method based on longitudinal trend prediction in any one of claims 1-8, and comprises a user historical load data collection module, an adjustable load prediction module, an adjustable rate calculation module and a cluster analysis module; the method is characterized in that:
the user historical load data collection module selects historical adjustable load data of m days, groups the historical adjustable load data according to different days at a uniform moment, and constructs a longitudinal load data matrix;
the adjustable load prediction module is used for carrying out normalization processing on historical adjustable load data, dividing equal probability intervals with different output levels, and converting a constructed longitudinal load data matrix into a state variable matrix;
the adjustable rate calculation module calculates state transition matrixes of the loads at different days at each moment, calculates load predicted values of target days at each moment point according to the initial states and the transition matrixes to obtain predicted output curves of different loads, and calculates adjustable rate curves of corresponding loads according to the predicted loads and the historical load average value;
and the clustering analysis module is used for clustering the adjustable rate curves of different adjustable loads obtained based on the longitudinal trend prediction by adopting a clustering algorithm and outputting typical user clustering results obtained under different adjustable levels.
10. A terminal comprising a processor and a storage medium; the method is characterized in that:
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method according to any one of claims 1 to 8.
11. Computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211565309.6A CN115935212A (en) | 2022-12-07 | 2022-12-07 | Adjustable load clustering method and system based on longitudinal trend prediction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211565309.6A CN115935212A (en) | 2022-12-07 | 2022-12-07 | Adjustable load clustering method and system based on longitudinal trend prediction |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115935212A true CN115935212A (en) | 2023-04-07 |
Family
ID=86653500
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211565309.6A Pending CN115935212A (en) | 2022-12-07 | 2022-12-07 | Adjustable load clustering method and system based on longitudinal trend prediction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115935212A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117239746A (en) * | 2023-11-16 | 2023-12-15 | 国网湖北省电力有限公司武汉供电公司 | Power load prediction method and system based on machine learning |
-
2022
- 2022-12-07 CN CN202211565309.6A patent/CN115935212A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117239746A (en) * | 2023-11-16 | 2023-12-15 | 国网湖北省电力有限公司武汉供电公司 | Power load prediction method and system based on machine learning |
| CN117239746B (en) * | 2023-11-16 | 2024-01-30 | 国网湖北省电力有限公司武汉供电公司 | Power load prediction method and system based on machine learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110232483B (en) | Deep learning load prediction method and device and terminal equipment | |
| CN108805213B (en) | Power load curve double-layer spectral clustering method considering wavelet entropy dimensionality reduction | |
| CN110380444B (en) | Capacity planning method for distributed wind power orderly access to power grid under multiple scenes based on variable structure Copula | |
| Lu et al. | A weekly load data mining approach based on hidden Markov model | |
| CN106067034B (en) | Power distribution network load curve clustering method based on high-dimensional matrix characteristic root | |
| CN114330583B (en) | Abnormal electricity utilization identification method and abnormal electricity utilization identification system | |
| CN111008726B (en) | Class picture conversion method in power load prediction | |
| Lee et al. | Probabilistic wind power forecasting based on the laplace distribution and golden search | |
| CN110796159A (en) | Power data classification method and system based on k-means algorithm | |
| CN115935212A (en) | Adjustable load clustering method and system based on longitudinal trend prediction | |
| CN116307215A (en) | Load prediction method, device, equipment and storage medium of power system | |
| CN114912720A (en) | Memory network-based power load prediction method, device, terminal and storage medium | |
| Shi et al. | Clustering framework based on multi-scale analysis of intraday financial time series | |
| CN115794578A (en) | Data management method, device, equipment and medium for power system | |
| CN116307059A (en) | Power distribution network region fault prediction model construction method and device and electronic equipment | |
| CN111080489A (en) | Load trend prediction method, system and equipment | |
| CN115878960A (en) | Wind power prediction error multi-interval analysis method and system considering output variation trend | |
| Majidpour | Time series prediction for electric vehicle charging load and solar power generation in the context of smart grid | |
| CN116304295A (en) | User energy consumption portrait analysis method based on multivariate data driving | |
| CN111046892A (en) | Abnormity identification method and device | |
| CN115169089A (en) | Wind power probability prediction method and device based on kernel density estimation and copula | |
| CN113627514A (en) | Data processing method and device of knowledge graph, electronic equipment and storage medium | |
| CN112766535A (en) | Building load prediction method and system considering load curve characteristics | |
| Wang et al. | Clustering of Consumption Behavior Dynamics | |
| CN113487080B (en) | Wind speed dynamic scene generation method, system and terminal based on wind speed classification |
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 |