CN111680590A - Power signal filtering method and system by using contraction gradient - Google Patents

Power signal filtering method and system by using contraction gradient Download PDF

Info

Publication number
CN111680590A
CN111680590A CN202010456924.8A CN202010456924A CN111680590A CN 111680590 A CN111680590 A CN 111680590A CN 202010456924 A CN202010456924 A CN 202010456924A CN 111680590 A CN111680590 A CN 111680590A
Authority
CN
China
Prior art keywords
module
judgment result
contraction
formula
value
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.)
Withdrawn
Application number
CN202010456924.8A
Other languages
Chinese (zh)
Inventor
翟明岳
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong University of Petrochemical Technology
Original Assignee
Guangdong University of Petrochemical Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guangdong University of Petrochemical Technology filed Critical Guangdong University of Petrochemical Technology
Priority to CN202010456924.8A priority Critical patent/CN111680590A/en
Publication of CN111680590A publication Critical patent/CN111680590A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G06F2218/04—Denoising
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R21/00—Arrangements for measuring electric power or power factor
    • G01R21/001—Measuring real or reactive component; Measuring apparent energy
    • G01R21/002—Measuring real component
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Signal Processing (AREA)
  • General Engineering & Computer Science (AREA)
  • Power Engineering (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)

Abstract

The embodiment of the invention discloses a power signal filtering method and system by using a contraction gradient, wherein the method comprises the steps of obtaining a signal sequence S acquired according to a time sequence in step 101, calculating a contraction factor β in step 102, and obtaining a contraction sequence x in step 103SMInitial value of (2)
Figure DDA0002509702530000014
Step 104 creates an iteration control parameter k and assigns it to 0, and creates a contraction step tkAnd is assigned as Δ T; step 105 of finding a gradient vector
Figure DDA0002509702530000011
Step 106 of obtaining a contraction gradient function gk(ii) a Step 107 determines the systolic gradient function gkWhether or not less than
Figure DDA0002509702530000012
Obtaining a first judgment result; step 108, adding 1 to the value of the iterative control parameter k, and calculating the shrinkage vector xSMThe value of the k step; step 109, calculating an adjacent error e; step 110 determines whether the adjacent error e is greater than or equal to a preset threshold0Obtaining a second judgment result; step 111 records the noise-filtered signal sequence SNEWIs concretely provided with
Figure DDA0002509702530000013

Description

Power signal filtering method and system by using contraction gradient
Technical Field
The present invention relates to the field of power, and in particular, to a method and a system for filtering a power signal.
Background
With the development of smart grids, the analysis of household electrical loads becomes more and more important. Through the analysis of the power load, a family user can obtain the power consumption information of each electric appliance and a refined list of the power charge in time; the power department can obtain more detailed user power utilization information, can improve the accuracy of power utilization load prediction, and provides a basis for overall planning for the power department. Meanwhile, the power utilization behavior of the user can be obtained by utilizing the power utilization information of each electric appliance, so that the method has guiding significance for the study of household energy consumption evaluation and energy-saving strategies.
The current electric load decomposition is mainly divided into an invasive load decomposition method and a non-invasive load decomposition method. The non-invasive load decomposition method does not need to install monitoring equipment on internal electric equipment of the load, and can obtain the load information of each electric equipment only according to the total information of the electric load. The non-invasive load decomposition method has the characteristics of less investment, convenience in use and the like, so that the method is suitable for decomposing household load electricity.
In the non-invasive load decomposition algorithm, the detection of the switching event of the electrical equipment is the most important link. The initial switch event detection takes the change value of the active power P as the judgment basis of the switch event detection, and is convenient and intuitive. This is because the power consumed by any one of the electric devices changes, and the change is reflected in the total power consumed by all the electric devices. The method needs to set a reasonable threshold value of the power change value, and also needs to solve the problems existing in the practical application of the event detection method, for example, a large peak appears in the instantaneous power value at the starting time of some electric appliances (the starting current of a motor is far larger than the rated current), which causes the inaccurate steady-state power change value of the electric appliances, thereby influencing the judgment of the detection of the switching event; moreover, the transient process of different household appliances is long or short (the duration and the occurrence frequency of impulse noise are different greatly), so that the determination of the power change value becomes difficult; due to the fact that the active power changes suddenly when the quality of the electric energy changes (such as voltage drop), misjudgment is likely to happen.
Therefore, in the switching event detection process, the actually measured power signal used is often affected by noise, and the switching event detection cannot be performed correctly by using the imperfect power signal. Therefore, how to effectively reconstruct the incomplete power signal and filter the influence of noise is the key to the success of the method. The existing common method has insufficient attention to the problem, and no effective measure is taken to solve the problem.
Disclosure of Invention
In the process of detecting the switching event, the actually measured power signal used is often affected by noise, and the detection of the switching event cannot be correctly performed by using the imperfect power signal. Therefore, how to effectively reconstruct the incomplete power signal and filter the influence of noise is the key to the success of the method. The existing common method has insufficient attention to the problem, and no effective measure is taken to solve the problem.
The invention aims to provide a power signal filtering method and a system by using a contraction gradient. The method has better robustness and simpler calculation.
In order to achieve the purpose, the invention provides the following scheme:
a method of power signal filtering with systolic gradients, comprising:
step 101, acquiring a signal sequence S acquired according to a time sequence;
step 102 calculates a shrinkage factor β, as
Figure BDA0002509702510000021
Wherein λ isminIs matrix B ═ S-m0]T[S-m0]A non-zero minimum eigenvalue of; lambda [ alpha ]maxIs the maximum eigenvalue of the matrix B; m is0Is the mean of the signal sequence S;
step 103 finds the puncture sequence xSMInitial value of (2)
Figure BDA0002509702510000022
The formula is obtained as
Figure BDA0002509702510000023
Wherein sigma0Is the mean square error of the signal sequence S;
step 104 creates an iteration control parameter k and assigns it to 0, and creates a contraction step tkAnd assigned as deltat. Wherein Δ T is a sampling interval of the signal sequence S;
step 105 of finding a gradient vector
Figure BDA0002509702510000024
The formula is obtained as
Figure BDA0002509702510000025
Wherein the content of the first and second substances,
Figure BDA0002509702510000026
for the shrinkage vector xSMThe k step number of
Figure BDA0002509702510000027
The 1 st element of (a);
Figure BDA0002509702510000028
for the shrinkage vector xSMThe k step number of
Figure BDA0002509702510000029
The 2 nd element of (1);
Figure BDA00025097025100000210
for the shrinkage vector xSMThe k step number of
Figure BDA00025097025100000211
The (N-1) th element of (a);
Figure BDA00025097025100000212
for the shrinkage vector xSMThe k step number of
Figure BDA00025097025100000213
The nth element of (a);
step 106 of obtaining a contraction gradient functionNumber gkThe formula is obtained as
Figure BDA00025097025100000214
Step 107 determines the systolic gradient function gkWhether or not less than
Figure BDA00025097025100000215
Obtaining a first judgment result; if the first judgment result shows the contraction gradient function gkIs less than
Figure BDA00025097025100000216
Said shrinking step tkUpdated to β tk(ii) a If the first judgment result shows the contraction gradient function gkGreater than or equal to
Figure BDA00025097025100000217
Said shrinking step tkUpdating to 0;
step 108, adding 1 to the value of the iterative control parameter k, and calculating the shrinkage vector xSMThe k step value is calculated by the formula
Figure BDA00025097025100000218
Step 109 calculates the adjacent error e by the formula
Figure BDA00025097025100000219
Step 110 determines whether the adjacent error e is greater than or equal to a preset threshold0And obtaining a second judgment result. If the second judgment result shows that the approximation error e is greater than or equal to the preset threshold value0Returning to the step 105, the step 106, the step 107, the step 108, the step 109 and the step 110, and adding 1 to the value of the iterative control parameter k; until the second judgment result shows that the approximation error e is smaller than the preset threshold value0. The preset threshold value is0＝0.001；
Step 111Recording a noise-filtered signal sequence SNEWIs concretely provided with
Figure BDA00025097025100000220
A power signal filtering system utilizing a systolic gradient, comprising:
the module 201 acquires a signal sequence S acquired in time sequence;
the module 202 calculates the shrinkage factor β as
Figure BDA00025097025100000221
Wherein λ isminIs matrix B ═ S-m0]T[S-m0]A non-zero minimum eigenvalue of; lambda [ alpha ]maxIs the maximum eigenvalue of the matrix B; m is0Is the mean of the signal sequence S;
module 203 finds the puncture sequence xSMInitial value of (2)
Figure BDA00025097025100000222
The formula is obtained as
Figure BDA00025097025100000223
Wherein sigma0Is the mean square error of the signal sequence S;
the module 204 creates an iteration control parameter k and assigns a value of 0, and creates a contraction step tkAnd assigned as deltat. Wherein Δ T is a sampling interval of the signal sequence S;
module 205 finds a gradient vector
Figure BDA0002509702510000031
The formula is obtained as
Figure BDA0002509702510000032
Wherein the content of the first and second substances,
Figure BDA0002509702510000033
for the shrinkage vector xSMThe k step number of
Figure BDA0002509702510000034
The 1 st element of (a);
Figure BDA0002509702510000035
for the shrinkage vector xSMThe k step number of
Figure BDA0002509702510000036
The 2 nd element of (1);
Figure BDA0002509702510000037
for the shrinkage vector xSMThe k step number of
Figure BDA0002509702510000038
The (N-1) th element of (a);
Figure BDA0002509702510000039
for the shrinkage vector xSMThe k step number of
Figure BDA00025097025100000310
The nth element of (a);
module 206 finds the systolic gradient function gkThe formula is obtained as
Figure BDA00025097025100000311
Module 207 determines the systolic gradient function gkWhether or not less than
Figure BDA00025097025100000312
Obtaining a first judgment result; if the first judgment result shows the contraction gradient function gkIs less than
Figure BDA00025097025100000313
Said shrinking step tkUpdated to β tk(ii) a If the first judgment result shows the contraction gradient function gkGreater than or equal to
Figure BDA00025097025100000314
Said shrinking step tkUpdating to 0;
the module 208 calculates the shrinkage vector x by adding 1 to the value of the iterative control parameter kSMThe k step value is calculated by the formula
Figure BDA00025097025100000315
The module 209 evaluates the adjacent error e to the formula
Figure BDA00025097025100000316
The module 210 determines whether the adjacent error e is greater than or equal to a predetermined threshold0And obtaining a second judgment result. If the second judgment result shows that the approximation error e is greater than or equal to the preset threshold value0Then returning to said module 205, said module 206, said module 207, said module 208, said module 209 and said module 210, the value of said iterative control parameter k is increased by 1; until the second judgment result shows that the approximation error e is smaller than the preset threshold value0. The preset threshold value is0＝0.001；
Module 211 records the noise-filtered signal sequence SNEWIs concretely provided with
Figure BDA00025097025100000317
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
in the process of detecting the switching event, the actually measured power signal used is often affected by noise, and the detection of the switching event cannot be correctly performed by using the imperfect power signal. Therefore, how to effectively reconstruct the incomplete power signal and filter the influence of noise is the key to the success of the method. The existing common method has insufficient attention to the problem, and no effective measure is taken to solve the problem.
The invention aims to provide a power signal filtering method and a system by using a contraction gradient. The method has better robustness and simpler calculation.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments will be briefly described below. It is obvious that the drawings in the following description are only some embodiments of the invention, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.
FIG. 1 is a schematic flow diagram of the process of the present invention;
FIG. 2 is a schematic flow chart of the system of the present invention;
FIG. 3 is a flow chart illustrating an embodiment of the present invention.
Detailed Description
The technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the 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.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
FIG. 1 is a schematic flow chart of a power signal filtering method using a systolic gradient
FIG. 1 is a flow chart illustrating a power signal filtering method using a systolic gradient according to the present invention. As shown in fig. 1, the method for filtering a power signal by using a systolic gradient specifically includes the following steps:
step 101, acquiring a signal sequence S acquired according to a time sequence;
step 102 calculates a shrinkage factor β, as
Figure BDA0002509702510000041
Wherein λ isminIs matrix B ═ S-m0]T[S-m0]A non-zero minimum eigenvalue of; lambda [ alpha ]maxIs the maximum eigenvalue of the matrix B; m is0Is the mean of the signal sequence S;
step 103 finds the puncture sequence xSMInitial value of (2)
Figure BDA0002509702510000042
The formula is obtained as
Figure BDA0002509702510000043
Wherein sigma0Is the mean square error of the signal sequence S;
step 104 creates an iteration control parameter k and assigns it to 0, and creates a contraction step tkAnd assigned as deltat. Wherein Δ T is a sampling interval of the signal sequence S;
step 105 of finding a gradient vector
Figure BDA0002509702510000044
The formula is obtained as
Figure BDA0002509702510000045
Wherein the content of the first and second substances,
Figure BDA0002509702510000046
for the shrinkage vector xSMThe k step number of
Figure BDA0002509702510000047
The 1 st element of (a);
Figure BDA0002509702510000048
for the shrinkage vector xSMThe k step number of
Figure BDA0002509702510000049
The 2 nd element of (1);
Figure BDA00025097025100000410
is said shrinkageVector xSMThe k step number of
Figure BDA00025097025100000411
The (N-1) th element of (a);
Figure BDA00025097025100000412
for the shrinkage vector xSMThe k step number of
Figure BDA00025097025100000413
The nth element of (a);
step 106 of obtaining a contraction gradient function gkThe formula is obtained as
Figure BDA00025097025100000414
Step 107 determines the systolic gradient function gkWhether or not less than
Figure BDA00025097025100000415
Obtaining a first judgment result; if the first judgment result shows the contraction gradient function gkIs less than
Figure BDA00025097025100000416
Said shrinking step tkUpdated to β tk(ii) a If the first judgment result shows the contraction gradient function gkGreater than or equal to
Figure BDA00025097025100000417
Said shrinking step tkUpdating to 0;
step 108, adding 1 to the value of the iterative control parameter k, and calculating the shrinkage vector xSMThe k step value is calculated by the formula
Figure BDA00025097025100000418
Step 109 calculates the adjacent error e by the formula
Figure BDA00025097025100000419
Step 110 determines whether the adjacent error e is greater than or equal to a preset threshold0And obtaining a second judgment result. If the second judgment result shows that the approximation error e is greater than or equal to the preset threshold value0Returning to the step 105, the step 106, the step 107, the step 108, the step 109 and the step 110, and adding 1 to the value of the iterative control parameter k; until the second judgment result shows that the approximation error e is smaller than the preset threshold value0. The preset threshold value is0＝0.001；
Step 111 records the noise-filtered signal sequence SNEWIs concretely provided with
Figure BDA0002509702510000051
FIG. 2 is a schematic diagram of a power signal filtering system using a systolic gradient
FIG. 2 is a schematic diagram of a power signal filtering system using a systolic gradient according to the present invention. As shown in fig. 2, the power signal filtering system using the systolic gradient includes the following structures:
the module 201 acquires a signal sequence S acquired in time sequence;
the module 202 calculates the shrinkage factor β as
Figure BDA0002509702510000052
Wherein λ isminIs matrix B ═ S-m0]T[S-m0]A non-zero minimum eigenvalue of; lambda [ alpha ]maxIs the maximum eigenvalue of the matrix B; m is0Is the mean of the signal sequence S;
module 203 finds the puncture sequence xSMInitial value of (2)
Figure BDA0002509702510000053
The formula is obtained as
Figure BDA0002509702510000054
Wherein sigma0Is the mean square error of the signal sequence S;
module 204 creating an iteration control parameter k, assigning the iteration control parameter k to be 0, and creating a contraction step length tkAnd assigned as deltat. Wherein Δ T is a sampling interval of the signal sequence S;
module 205 finds a gradient vector
Figure BDA0002509702510000055
The formula is obtained as
Figure BDA0002509702510000056
Wherein the content of the first and second substances,
Figure BDA0002509702510000057
for the shrinkage vector xSMThe k step number of
Figure BDA0002509702510000058
The 1 st element of (a);
Figure BDA0002509702510000059
for the shrinkage vector xSMThe k step number of
Figure BDA00025097025100000510
The 2 nd element of (1);
Figure BDA00025097025100000511
for the shrinkage vector xSMThe k step number of
Figure BDA00025097025100000512
The (N-1) th element of (a);
Figure BDA00025097025100000513
for the shrinkage vector xSMThe k step number of
Figure BDA00025097025100000514
The nth element of (a);
module 206 finds the systolic gradient function gkThe formula is obtained as
Figure BDA00025097025100000515
Module 207 determines the systolic gradient function gkWhether or not less than
Figure BDA00025097025100000516
Obtaining a first judgment result; if the first judgment result shows the contraction gradient function gkIs less than
Figure BDA00025097025100000517
Said shrinking step tkUpdated to β tk(ii) a If the first judgment result shows the contraction gradient function gkGreater than or equal to
Figure BDA00025097025100000518
Said shrinking step tkUpdating to 0;
the module 208 calculates the shrinkage vector x by adding 1 to the value of the iterative control parameter kSMThe k step value is calculated by the formula
Figure BDA00025097025100000519
The module 209 evaluates the adjacent error e to the formula
Figure BDA00025097025100000520
The module 210 determines whether the adjacent error e is greater than or equal to a predetermined threshold0And obtaining a second judgment result. If the second judgment result shows that the approximation error e is greater than or equal to the preset threshold value0Then returning to said module 205, said module 206, said module 207, said module 208, said module 209 and said module 210, the value of said iterative control parameter k is increased by 1; until the second judgment result shows that the approximation error e is smaller than the preset threshold value0. The preset threshold value is0＝0.001；
Module 211 records the noise-filtered signal sequence SNEWIs concretely provided with
Figure BDA00025097025100000521
The following provides an embodiment for further illustrating the invention
FIG. 3 is a flow chart illustrating an embodiment of the present invention. As shown in fig. 3, the method specifically includes the following steps:
step 301, acquiring a signal sequence S acquired according to a time sequence;
step 302 calculates a shrinkage factor β, as
Figure BDA0002509702510000061
Wherein λ isminIs matrix B ═ S-m0]T[S-m0]A non-zero minimum eigenvalue of; lambda [ alpha ]maxIs the maximum eigenvalue of the matrix B; m is0Is the mean of the signal sequence S;
step 303 finds the puncture sequence xSMInitial value of (2)
Figure BDA0002509702510000062
The formula is obtained as
Figure BDA0002509702510000063
Wherein sigma0Is the mean square error of the signal sequence S;
step 304 creates an iteration control parameter k and assigns it to 0, and creates a contraction step tkAnd assigned as deltat. Wherein Δ T is a sampling interval of the signal sequence S;
step 305 finds a gradient vector
Figure BDA0002509702510000064
The formula is obtained as
Figure BDA0002509702510000065
Wherein the content of the first and second substances,
Figure BDA0002509702510000066
for the shrinkage vector xSMThe k step number of
Figure BDA0002509702510000067
The 1 st element of (a);
Figure BDA0002509702510000068
for the shrinkage vector xSMThe k step number of
Figure BDA0002509702510000069
The 2 nd element of (1);
Figure BDA00025097025100000610
for the shrinkage vector xSMThe k step number of
Figure BDA00025097025100000611
The (N-1) th element of (a);
Figure BDA00025097025100000612
for the shrinkage vector xSMThe k step number of
Figure BDA00025097025100000613
The nth element of (a);
step 306 finds the systolic gradient function gkThe formula is obtained as
Figure BDA00025097025100000614
Step 307 determines the systolic gradient function gkWhether or not less than
Figure BDA00025097025100000615
Obtaining a first judgment result; if the first judgment result shows the contraction gradient function gkIs less than
Figure BDA00025097025100000616
Said shrinking step tkUpdated to β tk(ii) a If the first judgment result shows the contraction gradient function gkGreater than or equal to
Figure BDA00025097025100000617
Said shrinking step tkUpdating to 0;
step 308, adding 1 to the value of the iterative control parameter k, and calculating the shrinkage vector xSMThe k step value is calculated by the formula
Figure BDA00025097025100000618
Step 309 solves for the adjacent error e, the formula being
Figure BDA00025097025100000619
Step 310 determines whether the adjacent error e is greater than or equal to a predetermined threshold0And obtaining a second judgment result. If the second judgment result shows that the approximation error e is greater than or equal to the preset threshold value0Returning to said step 305, said step 306, said step 307, said step 308, said step 309 and said step 310, and adding 1 to the value of said iterative control parameter k; until the second judgment result shows that the approximation error e is smaller than the preset threshold value0. The preset threshold value is0＝0.001；
Step 311 records the noise-filtered signal sequence SNEWIs concretely provided with
Figure BDA00025097025100000620
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. For the system disclosed by the embodiment, the description is simple because the system corresponds to the method disclosed by the embodiment, and the relevant part can be referred to the method part for description.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.

Claims (2)

1. The method for filtering a power signal by using a systolic gradient is characterized by comprising the following steps:
step 101, acquiring a signal sequence S acquired according to a time sequence;
step 102 calculates a shrinkage factor β, as
Figure FDA0002509702500000011
Wherein λ isminIs matrix B ═ S-m0]T[S-m0]A non-zero minimum eigenvalue of; lambda [ alpha ]maxIs the maximum eigenvalue of the matrix B; m is0Is the mean of the signal sequence S;
step 103 finds the puncture sequence xSMInitial value of (2)
Figure FDA0002509702500000012
The formula is obtained as
Figure FDA0002509702500000013
Figure FDA0002509702500000014
Wherein sigma0Is the mean square error of the signal sequence S;
step 104 creates an iteration control parameter k and assigns it to 0, and creates a contraction step tkAnd assigned as deltat. Wherein Δ T is a sampling interval of the signal sequence S;
step 105 of finding a gradient vector
Figure FDA00025097025000000123
The formula is obtained as
Figure FDA0002509702500000015
Figure FDA0002509702500000016
Wherein the content of the first and second substances,
Figure FDA0002509702500000017
for the shrinkage vector xSMThe k step number of
Figure FDA0002509702500000018
The 1 st element of (a);
Figure FDA0002509702500000019
for the shrinkage vector xSMThe k step number of
Figure FDA00025097025000000110
The 2 nd element of (1);
Figure FDA00025097025000000111
for the shrinkage vector xSMThe k step number of
Figure FDA00025097025000000112
The (N-1) th element of (a);
Figure FDA00025097025000000113
for the shrinkage vector xSMThe k step number of
Figure FDA00025097025000000114
The nth element of (a);
step 106 of obtaining a contraction gradient function gkThe formula is obtained as
Figure FDA00025097025000000115
Step 107 determines the systolic gradient function gkWhether or not less than
Figure FDA00025097025000000116
Obtaining a first judgment result; if the first judgment result shows the contraction gradient function gkIs less than
Figure FDA00025097025000000117
Said shrinking step tkUpdated to β tk(ii) a If the first judgment result shows the contraction gradient function gkGreater than or equal to
Figure FDA00025097025000000118
Said shrinking step tkUpdating to 0;
step 108, adding 1 to the value of the iterative control parameter k, and calculating the shrinkage vector xSMThe k step value is calculated by the formula
Figure FDA00025097025000000119
Step 109 calculates the adjacent error e by the formula
Figure FDA00025097025000000120
Step 110 determines whether the adjacent error e is greater than or equal to a preset threshold0And obtaining a second judgment result. If the second judgment result shows that the approximation error e is greater than or equal to the preset threshold value0Returning to the step 105, the step 106, the step 107, the step 108, the step 109 and the step 110, and adding 1 to the value of the iterative control parameter k; until the second judgment result shows that the approximation error e is smaller than the preset threshold value0. The preset threshold value is0＝0.001；
Step 111 records the noise-filtered signal sequence SNEWIs concretely provided with
Figure FDA00025097025000000121
2. The power signal filtering system using a systolic gradient is characterized by comprising:
the module 201 acquires a signal sequence S acquired in time sequence;
module 202 calculates the contraction causeβ, the calculation formula is
Figure FDA00025097025000000122
Wherein λ isminIs matrix B ═ S-m0]T[S-m0]A non-zero minimum eigenvalue of; lambda [ alpha ]maxIs the maximum eigenvalue of the matrix B; m is0Is the mean of the signal sequence S;
module 203 finds the puncture sequence xSMInitial value of (2)
Figure FDA0002509702500000021
The formula is obtained as
Figure FDA0002509702500000022
Figure FDA0002509702500000023
Wherein sigma0Is the mean square error of the signal sequence S;
the module 204 creates an iteration control parameter k and assigns a value of 0, and creates a contraction step tkAnd assigned as deltat. Wherein Δ T is a sampling interval of the signal sequence S;
module 205 finds a gradient vector
Figure FDA00025097025000000221
The formula is obtained as
Figure FDA0002509702500000024
Figure FDA0002509702500000025
Wherein the content of the first and second substances,
Figure FDA0002509702500000026
for the shrinkage vector xSMThe k step number of
Figure FDA0002509702500000027
The 1 st element of (a);
Figure FDA0002509702500000028
for the shrinkage vector xSMThe k step number of
Figure FDA0002509702500000029
The 2 nd element of (1);
Figure FDA00025097025000000210
for the shrinkage vector xSMThe k step number of
Figure FDA00025097025000000211
The (N-1) th element of (a);
Figure FDA00025097025000000212
for the shrinkage vector xSMThe k step number of
Figure FDA00025097025000000213
The nth element of (a);
module 206 finds the systolic gradient function gkThe formula is obtained as
Figure FDA00025097025000000214
Module 207 determines the systolic gradient function gkWhether or not less than
Figure FDA00025097025000000215
Obtaining a first judgment result; if the first judgment result shows the contraction gradient function gkIs less than
Figure FDA00025097025000000216
Said shrinking step tkUpdated to β tk(ii) a If the first judgment result shows the contraction gradient function gkGreater than or equal to
Figure FDA00025097025000000217
Said shrinking step tkUpdating to 0;
the module 208 calculates the shrinkage vector x by adding 1 to the value of the iterative control parameter kSMThe k step value is calculated by the formula
Figure FDA00025097025000000218
The module 209 evaluates the adjacent error e to the formula
Figure FDA00025097025000000219
The module 210 determines whether the adjacent error e is greater than or equal to a predetermined threshold0And obtaining a second judgment result. If the second judgment result shows that the approximation error e is greater than or equal to the preset threshold value0Then returning to said module 205, said module 206, said module 207, said module 208, said module 209 and said module 210, the value of said iterative control parameter k is increased by 1; until the second judgment result shows that the approximation error e is smaller than the preset threshold value0. The preset threshold value is0＝0.001；
Module 211 records the noise-filtered signal sequence SNEWIs concretely provided with
Figure FDA00025097025000000220
CN202010456924.8A 2020-05-26 2020-05-26 Power signal filtering method and system by using contraction gradient Withdrawn CN111680590A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010456924.8A CN111680590A (en) 2020-05-26 2020-05-26 Power signal filtering method and system by using contraction gradient

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010456924.8A CN111680590A (en) 2020-05-26 2020-05-26 Power signal filtering method and system by using contraction gradient

Publications (1)

Publication Number Publication Date
CN111680590A true CN111680590A (en) 2020-09-18

Family

ID=72434819

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010456924.8A Withdrawn CN111680590A (en) 2020-05-26 2020-05-26 Power signal filtering method and system by using contraction gradient

Country Status (1)

Country Link
CN (1) CN111680590A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112307924A (en) * 2020-10-25 2021-02-02 广东石油化工学院 Power signal filtering method and system by using conversion learning algorithm
CN112347922A (en) * 2020-11-06 2021-02-09 华北电力大学 Power signal filtering method and system by using Hankerl matrix
CN112362967A (en) * 2020-10-25 2021-02-12 广东石油化工学院 Power signal filtering method and system by utilizing KL divergence

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112307924A (en) * 2020-10-25 2021-02-02 广东石油化工学院 Power signal filtering method and system by using conversion learning algorithm
CN112362967A (en) * 2020-10-25 2021-02-12 广东石油化工学院 Power signal filtering method and system by utilizing KL divergence
CN112347922A (en) * 2020-11-06 2021-02-09 华北电力大学 Power signal filtering method and system by using Hankerl matrix
CN112347922B (en) * 2020-11-06 2022-02-08 华北电力大学 Power signal filtering method and system by using Hankerl matrix

Similar Documents

Publication Publication Date Title
CN111680590A (en) Power signal filtering method and system by using contraction gradient
CN108918932B (en) Adaptive filtering method for power signal in load decomposition
CN109145825B (en) Coherent noise filtering method and system
CN111666870A (en) Power signal reconstruction method and system by utilizing quadratic constraint
CN108918929B (en) Power signal self-adaptive filtering method in load decomposition
CN112434567B (en) Power signal filtering method and system by using noise jitter property
CN110221119B (en) Load switch event detection method and system based on power and akie fusion information
CN111639606A (en) Power signal filtering method and system utilizing Dantzig total gradient minimization
CN110196354B (en) Method and device for detecting switching event of load
CN110542855B (en) Load switch event detection method and system based on discrete cosine transform
CN111830405A (en) Load switch event detection method and system by using frequency difference
CN110244115B (en) Load switch event detection method and system based on signal connectivity
CN109241874A (en) Power signal filtering method in Energy Decomposition
CN110187167B (en) Method and device for detecting load switch event based on manifold classification
CN110702981A (en) Load switch event detection method and system using classification tree
CN112270282B (en) Power signal filtering method and system by utilizing matrix spectral mode
CN110749841A (en) Load switch event detection method and system by utilizing conversion space factor
CN111665389A (en) Power signal filtering method and system by utilizing random projection
CN112307986B (en) Load switch event detection method and system by utilizing Gaussian gradient
CN112347922B (en) Power signal filtering method and system by using Hankerl matrix
CN110514884B (en) Power signal filtering method and system based on delay vector
CN110705426B (en) Power signal filtering method and system by using deblurring operator
CN110784019A (en) Power signal filtering method and system by using Mercer projection
CN112183410A (en) Power signal filtering method and system using DS theory
CN112307924A (en) Power signal filtering method and system by using conversion learning algorithm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20200918

WW01 Invention patent application withdrawn after publication