CN108959181A - Industrial electrical load decomposition method based on matrix decomposition - Google Patents
Industrial electrical load decomposition method based on matrix decomposition Download PDFInfo
- Publication number
- CN108959181A CN108959181A CN201810961008.2A CN201810961008A CN108959181A CN 108959181 A CN108959181 A CN 108959181A CN 201810961008 A CN201810961008 A CN 201810961008A CN 108959181 A CN108959181 A CN 108959181A
- Authority
- CN
- China
- Prior art keywords
- load
- matrix
- equation
- decomposition
- coefficient matrix
- 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
- 239000011159 matrix material Substances 0.000 title claims abstract description 68
- 238000000034 method Methods 0.000 title claims abstract description 35
- 238000000354 decomposition reaction Methods 0.000 title claims abstract description 29
- 238000012544 monitoring process Methods 0.000 claims abstract description 13
- 238000005457 optimization Methods 0.000 claims description 17
- 230000005611 electricity Effects 0.000 claims description 12
- 241000208340 Araliaceae Species 0.000 claims description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 claims description 2
- 235000008434 ginseng Nutrition 0.000 claims description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses an industrial electric load decomposition method based on matrix decomposition, which comprises the following steps: step S1, load node data are collected through non-invasive monitoring of the load nodes, and a load data matrix is obtained; step S2, carrying out non-negative matrix decomposition on the load data matrix to respectively obtain a base matrix and a coefficient matrix, wherein the base matrix is used for representing a load monitoring base, and the coefficient matrix is used for representing the parameters of the load of each piece of electric equipment in the total load; step S3, establishing a first objective equation for optimizing the basis matrix and the coefficient matrix; and step S4, performing load decomposition according to the optimized coefficient matrix, and acquiring the contribution ratio of each electric device to the total load. Compared with the existing judgment method based on experience, the method provided by the invention can provide more accurate and scientific reference for power load decomposition.
Description
Technical field
The present invention relates to power management techniques field more particularly to a kind of industrial electricity load decomposition based on matrix decomposition
Method.
Background technique
Reasonable demand Side Management is carried out, first have to do is exactly the actual demand of clearly each load bus, and
Actual demand and electric power peak valley state dynamic adjustment power supply strategy according to load bus.In practical application in industry, if right
Each load bus is monitored, and can undoubtedly bring huge workload, while in the case where load bus is excessive, real
The complexity of existing dynamic dispatching algorithm can show exponential growth.Therefore, in practical application in industry, a usually unit
Or a workshop section carries out unified load management, and uses non-intruding mode to the monitoring of load.In this kind of mode
Under, human observer only knows the total quantity and total load curve of electrical equipment, can not but judge the switching shape of each electrical equipment
State.In actual application, specific every the practical of a kind of electrical equipment is only known about and has used electricity condition, could reasonably be each
Class equipment distributes electric power resource, and declares electricity consumption quota.Currently, there is no the standardized burden rate generally used to decompose in industry
Method.
Summary of the invention
Technical problem to be solved by the present invention lies in provide a kind of industrial electricity load decomposition side based on matrix decomposition
Method, effectively to classify to load, the preliminary judgment criterion for realizing industrial user's switching load.
In order to solve the above technical problem, the present invention provides a kind of industrial electricity load decomposition side based on matrix decomposition
Method, comprising:
Step S1 acquires load bus data by non-intruding monitor load bus, obtains load data matrix；
Step S2 carries out Non-negative Matrix Factorization to the load data matrix, obtains basic matrix and coefficient matrix, institute respectively
It states basic matrix and is used to characterize the load of each electrical equipment in total load for characterizing load monitoring base, the coefficient matrix
Parameter；
Step S3 is established for optimizing the first object equation of the basic matrix and the coefficient matrix and solving；
Step S4 carries out load decomposition according to the optimal solution of the coefficient matrix, and obtains each electrical equipment to total load
Contribution proportion.
Wherein, the first object equation are as follows:
S.t., Ui>=0, Vi>=0, i=1,2 ..., m
Wherein, X is load data matrix, and U is basic matrix, and V is coefficient matrix, and subscript i and j indicate the i-th number of acquisition
The data total number of acquisition is indicated according to, m, α > 0 and β > 0 are two balance parameters, for adjusting different items described the
Shared ratio in one target equation, s.t. indicate constraint condition, this black norm of the not Luo Beini of subscript F representing matrix；
The optimal solution of U and V are obtained by optimizing the first object equation.
Wherein, the first object equation is further represented as following second target equation, and uses cross-iteration method
Optimize the second target equation:
S.t., Ui>=0, Vi>=0, i=1,2 ..., m
Wherein, the order of tr () representing matrix.
Wherein, the specific optimization process of the second target equation includes:
Two Lagrange multipliers P and Q are introduced, the Lagrange's equation of the second target equation is expressed as
Three target equations:
By each of third target equation part single optimization, for each specified i, optimization method
It can be expressed as the 4th target equation:
The 4th target equation is further represented as the 5th target equation of following equivalence:
Fixed Vi, by the 5th target equation to UiLocal derviation is sought, following expression is obtained:
Using KKT condition, the U that is optimizediExpression formula are as follows:
Fixed Ui, by the 5th target equation to ViLocal derviation is sought, following expression is obtained:
Using KKT condition, the V that is optimizediExpression formula are as follows:
Wherein, in coefficient matrix after optimization, if certain is classified as 0, or it is lower than a certain threshold value, then the corresponding electricity consumption of the column is set
It is standby to be in off state, it is otherwise normal operating conditions.
Wherein, in coefficient matrix after optimization, for not for 0 or not less than a certain threshold value column, it is returned
After one change processing, contribution proportion of the value as corresponding electrical equipment in total load.
Wherein, load decomposition is carried out according to the coefficient matrix after optimization in the step S4, specifically included: by total load point
Solution is the sum of the load of the electricity consumption startup equipment in normal operating conditions.
The beneficial effect of the embodiment of the present invention is: providing a kind of effective load for non-intrusion type electric power monitoring
Decomposition method, compared to the existing judgment method based on experience, method proposed by the invention can be decomposed for electric load and be provided
More accurate and more scientific reference.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is a kind of flow diagram of the industrial electricity load decomposition method based on matrix decomposition of the embodiment of the present invention.
Specific embodiment
The explanation of following embodiment be with reference to attached drawing, can be to the specific embodiment implemented to the example present invention.
It please referring to shown in Fig. 1, the embodiment of the present invention provides a kind of industrial electricity load decomposition method based on matrix decomposition,
The load decomposition that can be used in non-intrusive electrical load monitoring, comprising the following steps:
Step S1 acquires load bus data by non-intruding monitor load bus, obtains load data matrix；
Step S2 carries out Non-negative Matrix Factorization to the load data matrix, obtains basic matrix and coefficient matrix, institute respectively
Basic matrix is stated for load monitoring base, the coefficient matrix is used to characterize ginseng of the load of each electrical equipment in total load
Number；
Step S3 establishes first object equation, optimizes the basic matrix and the coefficient matrix；
Step S4 carries out load decomposition according to the optimization solution of the coefficient matrix, and obtains each electrical equipment to total load
Contribution proportion.
Specifically, in step S1, a data matrix is formed in the monitor value of a period of time T each variable of internal loading entrance
X, it is assumed that load monitoring base is U, and parameter of the load of each electrical equipment in total load is that V then in the ideal case has
Following relationship:
X≈UVT (1)
In actual industrial production, some variations may occur in the load of different periods electrical equipment, but giving
These variations are again metastable in the fixed time.Therefore, more accurate U and V in order to obtain can carry out m monitoring analysis,
All indicate there is X in training data using different subscripts each time1, X2..., Xm, accordingly, different bases have U1,
U2..., UmAnd V1, V2..., Vm.In order to distinguish different electrical equipments, the corresponding parameter V of different electrical equipments is answered
As different as possible, the difference between different V is bigger, then each V could correspond to a different electrical equipment, otherwise
Similar V, which will cause equipment, to be obscured, and can not normally be distinguished.Based on this, the present embodiment establishes following first object equation:
S.t., Ui>=0, Vi>=0, i=1,2 ..., m (2)
Wherein, subscript i and j indicates i-th part of data of acquisition, and m indicates the data total number of acquisition, and α > 0 and β > 0 are
Two balance parameters, for adjusting different items ratio shared in target equation；The s.t. that abridges indicates subject to, i.e.,
Constraint condition；This black norm (Frobenius Norm) of the not Luo Beini of subscript F representing matrix.By optimizing above-mentioned first object
Optimal U and V can be obtained in equation.
First object equation (2) can be further represented as following second target equation (3):
S.t., Ui>=0, Vi>=0, i=1,2 ..., m (3)
Wherein, the order of tr () representing matrix.
It is individually right since the second target equation (3) is not convex to two variables of U and V, but when one of fixed
U is individually convex to V.Therefore, the present embodiment optimizes the second target equation (3) using a kind of cross-iteration method.Specifically
U is considered as a constant first and goes to update V, V is then considered as a constant again and removes iteration U, successively repeatedly, until second by ground
Target equation converges to only.The specific optimization process of second target equation (3) is as follows:
Firstly, introducing two Lagrange multipliers P and Q, the Lagrange's equation of the second target equation (3) is written as:
Since third target equation (4) is related to summing, if will with each of formula part single optimization, for every
One specified i, optimization method can be written as:
4th target equation (5) can further be written as follow equivalent form:
In order to optimize Ui, V fixed firsti, by the 5th target equation (6) to UiLocal derviation is sought, following expression is obtained:
Using KKT (Karush-Kuhn-Tucher) condition, the U optimizediExpression formula are as follows:
Similarly, in order to optimize Vi, U fixed firsti, by the 5th target equation (6) to ViLocal derviation is sought, following expression is obtained:
Using KKT condition, the V that is optimizediExpression formula are as follows:
Finally, for the load monitoring data of certain time, U can be decomposed intoiWithProduct.It is each due to X
Column represent the observation at a moment, and every a line represents an observational characteristic, UiEach column represent an observation moment
The observation index of a certain electrical equipment, ViRespective column represent UiColumn representated by electrical equipment in practical total load
Coefficient, i.e. contribution of the electrical equipment to total load.Therefore, ViIt is classified as 0 in if, or is lower than a certain threshold value, then the column are corresponding
Electrical equipment be in off state, be otherwise normal operating conditions.For the equipment of normal work, to not being 0 or not low
After the column of a certain threshold value are normalized, contribution proportion of the value as corresponding electrical equipment in total load.
Due to when carrying out load monitoring, it is known which a total of possible electrical equipment, ViColumns just
It is coefficient corresponding to possible electrical equipment, i.e., contribution proportion of each possible electrical equipment in total load, if low
In the threshold value, just illustrate that corresponding electrical equipment is not powered on, naturally, total load is just decomposed into normal operating conditions
The sum of the load of electrical equipment.
By above description it is found that the beneficial effect of the embodiment of the present invention is, provided for non-intrusion type electric power monitoring
A kind of effective load decomposition method, compared to the existing judgment method based on experience, method energy proposed by the invention
It is decomposed for electric load and more accurate and more scientific reference is provided.
The above disclosure is only the preferred embodiments of the present invention, cannot limit the right model of the present invention with this certainly
It encloses, therefore equivalent changes made in accordance with the claims of the present invention, is still within the scope of the present invention.
Claims (7)
1. a kind of industrial electricity load decomposition method based on matrix decomposition characterized by comprising
Step S1 acquires load bus data by non-intruding monitor load bus, obtains load data matrix；
Step S2 carries out Non-negative Matrix Factorization to the load data matrix, obtains basic matrix and coefficient matrix, the base respectively
Matrix is used to characterize ginseng of the load of each electrical equipment in total load for characterizing load monitoring base, the coefficient matrix
Number；
Step S3 establishes the first object equation for optimizing the basic matrix and the coefficient matrix；
Step S4 carries out load decomposition according to the coefficient matrix after optimization, and obtains each electrical equipment to the contribution ratio of total load
Example.
2. the method according to claim 1, wherein the first object equation are as follows:
S.t., Ui>=0, Vi>=0, i=1,2 ..., m
Wherein, X is load data matrix, and U is basic matrix, and V is coefficient matrix, and subscript i and j indicate i-th part of data of acquisition, m
Indicate the data total number of acquisition, α > 0 and β > 0 are two balance parameters, for adjusting different items in first mesh
Ratio shared in equation is marked, s.t. indicates constraint condition, this black norm of the not Luo Beini of subscript F representing matrix；
The optimal solution of U and V are obtained by optimizing the first object equation.
3. according to the method described in claim 2, it is characterized in that, the first object equation is further represented as following second
Target equation, and the second target equation is optimized using cross-iteration method:
S.t., Ui>=0, Vi>=0, i=1,2 ..., m
Wherein, the order of tr () representing matrix.
4. according to the method described in claim 3, it is characterized in that, the specific optimization process of the second target equation includes:
Two Lagrange multipliers P and Q are introduced, the Lagrange's equation of the second target equation is expressed as third mesh
Mark equation:
By each of third target equation part single optimization, for each specified i, optimization method can be with
It is expressed as the 4th target equation:
The 4th target equation is further represented as the 5th target equation of following equivalence:
Fixed Vi, by the 5th target equation to UiLocal derviation is sought, following expression is obtained:
Using KKT condition, the U that is optimizediExpression formula are as follows:
Fixed Ui, by the 5th target equation to ViLocal derviation is sought, following expression is obtained:
Using KKT condition, the V that is optimizediExpression formula are as follows:
5. according to the method described in claim 4, it is characterized in that, in coefficient matrix after optimization, if certain is classified as 0 or low
In a certain threshold value, then the corresponding electrical equipment of the column is in off state, and is otherwise normal operating conditions.
6. according to the method described in claim 4, it is characterized in that, in coefficient matrix after optimization, for not being 0 or not
Lower than the column of a certain threshold value, after it is normalized, contribution of the value as corresponding electrical equipment in total load
Ratio.
7. according to the method described in claim 5, it is characterized in that, being carried out in the step S4 according to the coefficient matrix after optimization
Load decomposition specifically includes: total load is decomposed into the sum of the load of the electricity consumption startup equipment in normal operating conditions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810961008.2A CN108959181A (en) | 2018-08-22 | 2018-08-22 | Industrial electrical load decomposition method based on matrix decomposition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810961008.2A CN108959181A (en) | 2018-08-22 | 2018-08-22 | Industrial electrical load decomposition method based on matrix decomposition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108959181A true CN108959181A (en) | 2018-12-07 |
Family
ID=64473009
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810961008.2A Pending CN108959181A (en) | 2018-08-22 | 2018-08-22 | Industrial electrical load decomposition method based on matrix decomposition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108959181A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106096726A (en) * | 2016-05-31 | 2016-11-09 | 华北电力大学 | A kind of non-intrusion type load monitoring method and device |
| CN108257044A (en) * | 2017-09-19 | 2018-07-06 | 济南大学 | A kind of non-intrusion type load decomposition method based on steady-state current model |
-
2018
- 2018-08-22 CN CN201810961008.2A patent/CN108959181A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106096726A (en) * | 2016-05-31 | 2016-11-09 | 华北电力大学 | A kind of non-intrusion type load monitoring method and device |
| CN108257044A (en) * | 2017-09-19 | 2018-07-06 | 济南大学 | A kind of non-intrusion type load decomposition method based on steady-state current model |
Non-Patent Citations (1)
| Title |
|---|
| 饶竹一等: "基于矩阵分解的非侵入式负荷分解算法", 《城市建设理论研究（电子版）》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105186525B (en) | Power Network Partitioning method under wind power integration | |
| CN106934497A (en) | Wisdom cell power consumption real-time predicting method and device based on deep learning | |
| CN111401664A (en) | Robust optimization scheduling method and device for comprehensive energy system | |
| CN111340299A (en) | Multi-objective optimization scheduling method for micro-grid | |
| CN105630648B (en) | Data center's intelligent control method and system based on multidimensional data deep learning | |
| CN105186570A (en) | Micro power grid power supply control method and device | |
| Hu et al. | Edge intelligence for real-time data analytics in an IoT-based smart metering system | |
| CN108876091A (en) | A kind of virtual plant realized based on software definition power grid | |
| CN109946968B (en) | Distributed energy matching system and method in building | |
| CN107611976A (en) | A kind of intelligent electric power utilization system and its implementation | |
| Pan et al. | Near‐optimal control of a stochastic vegetation‐water system with reaction diffusion | |
| CN109150603A (en) | The automatic expansion method of cloud server terminal and device | |
| CN107134790A (en) | A kind of GA for reactive power optimization control sequence based on big data determines method | |
| CN108959181A (en) | Industrial electrical load decomposition method based on matrix decomposition | |
| CN112001578A (en) | Generalized energy storage resource optimization scheduling method and system | |
| CN111342463A (en) | Electric load auxiliary identification method based on multivariate external data | |
| CN110599032A (en) | Deep Steinberg self-adaptive dynamic game method for flexible power supply | |
| CN116137445A (en) | Island network distribution method, device, equipment and medium based on distributed power supply | |
| CN109313673A (en) | The operation method and Related product of network model | |
| CN113673141A (en) | Energy router modeling and optimization control method based on data driving | |
| CN102438325B (en) | Resource scheduling method based on cognitive radio terminal reconfiguration system | |
| CN117236650B (en) | Water-fertilizer integrated intelligent pump house control method | |
| CN211207384U (en) | Energy management system for field operation mobile microgrid | |
| CN109063889A (en) | A kind of load power prediction technique of active distribution network | |
| CN117575308B (en) | Risk assessment method, device and equipment for distributed power distribution network and storage medium |
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: 20181207 |