CN113363974B - Method and device for analyzing residential load composition based on accumulated electric quantity low-frequency sampling - Google Patents
Method and device for analyzing residential load composition based on accumulated electric quantity low-frequency sampling Download PDFInfo
- Publication number
- CN113363974B CN113363974B CN202110669092.2A CN202110669092A CN113363974B CN 113363974 B CN113363974 B CN 113363974B CN 202110669092 A CN202110669092 A CN 202110669092A CN 113363974 B CN113363974 B CN 113363974B
- Authority
- CN
- China
- Prior art keywords
- electric quantity
- characteristic vector
- electrical
- actual
- combined
- 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.)
- Active
Links
- 238000005070 sampling Methods 0.000 title claims abstract description 87
- 239000000203 mixture Substances 0.000 title claims abstract description 30
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000004458 analytical method Methods 0.000 claims abstract description 47
- 239000006185 dispersion Substances 0.000 claims description 4
- 238000004422 calculation algorithm Methods 0.000 description 42
- 230000006870 function Effects 0.000 description 18
- 238000012544 monitoring process Methods 0.000 description 14
- 230000003044 adaptive effect Effects 0.000 description 9
- 238000005406 washing Methods 0.000 description 9
- 230000002068 genetic effect Effects 0.000 description 8
- 238000005457 optimization Methods 0.000 description 7
- 230000005611 electricity Effects 0.000 description 6
- 230000035772 mutation Effects 0.000 description 6
- 230000008859 change Effects 0.000 description 5
- 230000001186 cumulative effect Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 230000001052 transient effect Effects 0.000 description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 210000000349 chromosome Anatomy 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 241000209094 Oryza Species 0.000 description 2
- 235000007164 Oryza sativa Nutrition 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 2
- 229910002056 binary alloy Inorganic materials 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000012804 iterative process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 235000009566 rice Nutrition 0.000 description 2
- 241000196324 Embryophyta Species 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000010183 spectrum analysis Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for AC mains or AC distribution networks
-
- 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
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Algebra (AREA)
- Computational Mathematics (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Analysis (AREA)
- Power Engineering (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a resident load composition analysis method and a resident load composition analysis device based on accumulated electric quantity low-frequency sampling, wherein the method comprises the following steps: determining an actual electric quantity differential quantity characteristic vector according to the accumulated electric quantity data of the user; determining a combined electric quantity differential quantity characteristic vector according to the opening time parameter and the closing time parameter of the electrical equipment and each operation state of the electrical equipment in a sampling period, wherein each operation state comprises an electrical equipment opening keeping state, an electrical equipment opening state, an electrical equipment closing state and an electrical equipment opening and closing state; and constructing an objective function according to the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector, and judging the difference between the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector according to the objective function. According to the invention, the accuracy of resident load composition analysis is improved by selecting the accumulated electric quantity when the equipment runs as the equipment load characteristic and constructing the objective function.
Description
Technical Field
The invention relates to the technical field of power systems, in particular to a resident load composition analysis method and device based on accumulated electric quantity low-frequency sampling.
Background
The current Load composition analysis mainly adopts two modes, namely an Intrusive Load Monitoring (ILM) mode and a Non-Intrusive Load Monitoring (NILM) mode, wherein the ILM needs to install an information acquisition device on each piece of electric equipment, the Monitoring result of the scheme has high accuracy, the NILM is originally proposed in 1989 by Hart teaching of Massachusetts Ministry, and the NILM observes the total steady-state power by sampling at the frequency of 1Hz at a total incoming line and realizes the resident Load composition analysis based on the scheme.
On the basis of Hart research, subsequent researchers mainly conduct load analysis research by finding new equipment load characteristics, such as using steady-state harmonic currents, utilizing steady-state harmonic power of equipment, or implementing load composition analysis by extracting V-I trajectories. Compared with the steady-state characteristics, the transient characteristics can better reflect the load change characteristics of the electrical equipment at the switching moment and during the working state conversion, for the extraction of the transient characteristics of the equipment, the wavelet change coefficient can be calculated through indexes to extract the instantaneous power characteristics of the equipment, and the voltage noise when the working state of the equipment is changed is extracted by using spectral analysis, so that the equipment with similar working characteristics can be distinguished.
Intrusive load composition analysis requires a user to install an information acquisition device on each electric device, and has the obvious disadvantages that a large number of monitoring terminals are required and the user is required to cooperate, the mode is complex and has no universality, and in the non-intrusive load composition analysis, compared with a steady-state characteristic, a transient characteristic can better reflect the load change characteristic of the electric device in the switching moment and the working state conversion, but the transient characteristic of the electric device is very short and is usually only a few tenths of seconds or even shorter. Therefore, the higher the sampling frequency is, the more detailed the collected electricity consumption information is, the more accurate the start-stop time and electricity consumption of the user's equipment can be judged, but the high-frequency sampling puts higher requirements on the collection device and the communication, and the large-area modification of the electricity meter and the communication network is required, so that the method generally adopts a load constitution analysis method of accumulated electricity to compare and fit the measured electrical quantity at the bus with the equipment load characteristic data, and the defect that the method requires that the electrical quantity measurement at the bus and the equipment load characteristic data collection keep the sampling frequency consistent, so that the collection device has great frequency constraint.
Disclosure of Invention
The invention aims to provide a method and a device for analyzing the residential load composition based on accumulated electric quantity low-frequency sampling, so as to improve the accuracy of the residential load composition analysis.
In order to achieve the above object, the present invention provides a method for analyzing a load composition of a resident based on low-frequency sampling of accumulated electric quantity, comprising:
determining an actual electric quantity difference characteristic vector according to the user accumulated electric quantity data;
determining a combined electric quantity differential quantity characteristic vector according to the opening time parameter and the closing time parameter of the electrical equipment and each operation state of the electrical equipment in a sampling period, wherein each operation state comprises an electrical equipment opening keeping state, an electrical equipment opening state, an electrical equipment closing state and an electrical equipment opening and closing state;
and constructing an objective function according to the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector, and judging the difference between the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector according to the objective function.
Preferably, the actual electrical quantity difference amount feature vector S (m) is as follows:
S(m)＝[s(1),s(2),…,s(m)]；
where s (m) represents the accumulated charge difference for the mth sampling period.
Preferably, the combined electrical quantity difference characteristic vector E (m) is as follows:
wherein E is i (m) represents the accumulated delta amount of charge vector over the analysis period, and N represents the number of appliance types.
Preferably, the objective function is as follows:
whereinCov (S, E) represents the covariance, σ, of the actual delta-electrical-quantity eigenvector and the combined delta-electrical-quantity eigenvector S 、σ E Singular values, d, representing the actual and combined delta-electrical quantities eigenvectors, respectively S,E A Euclidean distance, r, representing the actual electrical dispersion amount eigenvector and the combined electrical dispersion amount eigenvector S,E Pearson correlation coefficients representing the actual delta charge quantity eigenvector and the combined delta charge quantity eigenvector.
Preferably, the determining a difference between the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector according to the objective function includes:
the smaller the Euclidean distance is, the smaller the distance between the actual electric quantity differential quantity characteristic vector and the combined electric quantity differential quantity characteristic vector is, and otherwise, the larger the distance is;
the Pearson correlation coefficient is biased to 1, which means that the similarity between the actual electrical quantity difference characteristic vector and the combined electrical quantity difference characteristic vector is higher, and vice versa.
The invention also provides a resident load composition analysis device based on accumulated electric quantity low-frequency sampling, which comprises the following components:
the first module is used for determining an actual electric quantity difference characteristic vector according to the accumulated electric quantity data of the user;
the second module is used for determining a combined electric quantity differential quantity characteristic vector according to the opening time parameter and the closing time parameter of the electrical equipment and each operation state of the electrical equipment in a sampling period, wherein each operation state comprises an electrical equipment keeping opening state, an electrical equipment closing state and an electrical equipment opening and closing state;
and the construction module is used for constructing an objective function according to the actual electric quantity differential quantity characteristic vector and the combined electric quantity differential quantity characteristic vector, and judging the difference between the actual electric quantity differential quantity characteristic vector and the combined electric quantity differential quantity characteristic vector according to the objective function.
Preferably, the first module is further configured to construct the actual electrical quantity difference quantity eigenvector S (m) as follows:
S(m)＝[s(1),s(2),…,s(m)]；
where s (m) represents the accumulated charge difference for the mth sampling period.
Preferably, the second module is further configured to construct the combined electrical quantity difference quantity feature vector E (m) as follows:
wherein E is i (m) represents the accumulated electrical delta vector over the analysis period, and N represents the number of appliance types.
Preferably, the constructing module is further configured to construct the objective function as follows:
wherein cov (S, E) represents the covariance of the actual delta-electrical-quantity eigenvector and the combined delta-electrical-quantity eigenvector, σ S 、σ E Singular values, d, representing the actual and combined electrical quantity difference eigenvectors, respectively S,E A Euclidean distance, r, representing the actual delta electrical quantity characteristic vector and the combined delta electrical quantity characteristic vector S,E A Pearson correlation coefficient representing the actual delta electrical quantity eigenvector and the combined delta electrical quantity eigenvector.
Preferably, the building module is further configured to determine a difference between the actual electrical quantity difference characteristic vector and the combined electrical quantity difference characteristic vector, as follows:
the smaller the Euclidean distance is, the smaller the distance between the actual electric quantity differential quantity characteristic vector and the combined electric quantity differential quantity characteristic vector is, and otherwise, the larger the distance is;
the Pearson correlation coefficient tends to 1, which means that the similarity between the actual electrical quantity difference characteristic vector and the combined electrical quantity difference characteristic vector is higher, and vice versa.
The method determines an actual electric quantity difference characteristic vector according to the accumulated electric quantity data of a user, determines a combined electric quantity difference characteristic vector according to an opening time parameter and a closing time parameter of the electric equipment and each operation state of the electric equipment in a sampling period, wherein each operation state comprises an opening state of the electric equipment, a closing state of the electric equipment and an opening and closing state of the electric equipment, constructs a target function according to the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector, judges the difference between the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector according to the target function, and improves the accuracy of resident load composition analysis.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a method for analyzing a load composition of a resident based on low-frequency sampling of accumulated electric quantity according to an embodiment of the present invention;
FIG. 2 is a diagram illustrating the operation of a device during a sampling period according to another embodiment of the present invention;
fig. 3 is a characteristic graph of an accumulated power of a refrigerator according to another embodiment of the present invention;
FIG. 4 is a characteristic curve diagram of the accumulated electric power of the electric rice cooker according to one embodiment of the present invention;
FIG. 5 is a characteristic curve diagram of the accumulated electric quantity of an electric kettle according to another embodiment of the present invention;
fig. 6 is a characteristic graph of an accumulated electric quantity of an air conditioner according to still another embodiment of the present invention;
fig. 7 is a characteristic curve diagram of the accumulated electric quantity of an electric water heater according to an embodiment of the present invention;
FIG. 8 is a graph illustrating a cumulative charge characteristic of a microwave oven according to another embodiment of the present invention;
FIG. 9 is a flow chart of an adaptive genetic algorithm based on optimal individual similarity coefficients according to another embodiment of the present invention;
fig. 10 is a graph of power of a washing machine according to an embodiment of the present invention;
fig. 11 is a graph illustrating an accumulated power of a washing machine according to another embodiment of the present invention;
FIG. 12 is a graph illustrating an iterative evolution of the AGASC algorithm according to another embodiment of the present invention;
fig. 13 is a schematic structural diagram of a residential load composition analysis device based on low-frequency sampling of accumulated electric energy according to an embodiment of the present invention.
Detailed Description
The technical solutions in the present invention will be described clearly and completely with reference to the accompanying drawings, and it is obvious that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be understood that the step numbers used herein are for convenience of description only and are not used as limitations on the order in which the steps are performed.
It is to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the specification of the present invention and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
The terms "comprises" and "comprising" indicate the presence of the described features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
The term "and/or" refers to any and all possible combinations of one or more of the associated listed items and includes such combinations.
Referring to fig. 1, the present invention provides a method for analyzing a load composition of a resident based on low frequency sampling of accumulated electric quantity, comprising:
s101, determining an actual electric quantity differential quantity characteristic vector according to the accumulated electric quantity data of the user.
Specifically, for the accumulated electric quantity at the incoming line of the user, collected n times with a sampling period of Δ t, the collected user accumulated electric quantity data Q (n) is as follows:
Q(n)＝[q(1),q(2),...,q(n)]；
wherein q (n) represents the electrical quantity of the nth sampling point, the actual electrical quantity difference characteristic vector is as follows:
S(m)＝[s(1),s(2),...,s(m)]；
where s (k) = q (k + 1) -q (k) denotes the accumulated charge difference value for the m-th sampling period, and m = n-1.
S102, determining a combined electric quantity difference characteristic vector according to the on-time parameter and the off-time parameter of the electric equipment and each operation state of the electric equipment in a sampling period, wherein each operation state comprises an on-state of the electric equipment, an off-state of the electric equipment and an on-off state of the electric equipment.
Specifically, the accumulated electrical quantity characteristic curve of the electrical equipment is selected and an on/off time parameter is determined, then an accumulated electrical quantity difference vector in an analysis period is determined according to the on/off time parameter of the electrical equipment, and then each element in the accumulated electrical quantity difference vector is obtained, so that the combined electrical quantity difference characteristic vector is as follows:
wherein E is i (m) represents the accumulated delta amount of charge vector over the analysis period, and N represents the number of appliance types.
Referring to fig. 2-8, the device accumulated electricity feature database established by the present invention includes: the refrigerator accumulated power feature, the rice cooker accumulated power feature, the electric kettle accumulated power feature, the air conditioner accumulated power feature, the electric water heater accumulated power feature, and the microwave oven accumulated power feature.
Within a sampling period, the electrical device contains a plurality of operating states: the electric equipment keeps an on state, an off state and an on and off state, and the accumulated electric quantity difference corresponding to each operation state is different in a sampling period, so that a mathematical expression is established for the accumulated electric quantity difference in each operation state as follows:
1) The electrical equipment keeps the on state in the sampling period as follows:
2) The electrical equipment is turned off in the sampling period:
3) The electrical equipment is turned on in a sampling period:
4) The electrical equipment is turned on and off in a sampling period:
wherein, Δ e1 i,k ,Δe2 i,k ,Δe3 i,k ,Δe4 i,k Respectively representing the power consumption of four running states of the electrical equipment i between sampling points k and k +1, a i The working coefficient of the device i is represented as a variable from 0 to 1, represents whether the device works in a sampling period or not,
indicating the cumulative charge of the device iCharacteristic curve, t k ,t k+1 Respectively representing the sampling time points, t k+1 -t k ＝Δt，t i_on Parameter representing the moment of activation of device i, t i_off Representing the off-time parameter of device i.
In a parsing period T h Often includes a plurality of sampling periods, and the accumulated electrical quantity difference vector E of the electrical equipment in the analysis period i Is Δ e1 i,k ,Δe2 i,k ,Δe3 i,k ,Δe4 i,k The number of elements in each vector is the number of sampling cycles, and for the analysis period, there are four cases of accumulated electrical quantity differential quantity vector according to different parameters of the on/off time of the electrical equipment:
1) The device is turned on before the start of monitoring and turned off after the end of monitoring (t) i_on ≤0,t i_off ≥T h )：
E i (m)＝[Δe1 i,1 ,Δe1 i,2 ,...,Δe1 i,m ]；
Where m is the number of sampling cycles in the analysis period, E i (m) represents the accumulated delta-charge vector, T, over the analysis period h Representing a period of resolution.
2) The device is turned on before the start of monitoring and turned off before the end of monitoring (t) i_on ≤0,t i_off <T h )：
E i (m)＝[Δe1 i,1 ,…,Δe1 i,b-1 ,Δe2 i,b ,0,...]；
Wherein b = [ t ] i_off /Δt]+1,[]B represents the sampling period in which the parameter of the closing time of the device i is located, 0, for an integer function<b≤m。
3) The device is turned on after the start of monitoring and turned off after the end of monitoring (t) i_on >0,t i_off ≥T h )：
E i (m)＝[0,...,Δe3 i,a ,Δe1 i,a+1 ,...,Δe1 i,m ]；
Wherein a = [ t ] i_on /Δt]+1,a denotes the sampling period in which the start time parameter of device i is located, 0<b≤m。
4) The equipment is atStarting after starting monitoring and closing before finishing monitoring (t) i_on >0,t i_off <T h )：
E i (m)＝[0,...,Δe3 i,a ,Δe1 i,a+1 ,...,Δe1 i,b-1 ,Δe2 i,b ,0,...]。
S103, constructing an objective function according to the actual electric quantity difference quantity characteristic vector and the combined electric quantity difference quantity characteristic vector, and judging the difference between the actual electric quantity difference quantity characteristic vector and the combined electric quantity difference quantity characteristic vector according to the objective function.
Specifically, an objective function is constructed based on the actual electric quantity difference characteristic vector S (m) and the combined electric quantity difference characteristic vector E (m) obtained in step S102, as follows:
wherein cov (S, E) represents the covariance of the actual delta-electric-quantity eigenvector and the combined delta-electric-quantity eigenvector, σ S 、σ E Singular values, d, representing the actual and combined electrical quantity difference eigenvectors, respectively S,E And the Euclidean distance represents the characteristic vector of the actual electric quantity difference and the characteristic vector of the combined electric quantity difference, wherein the smaller the Euclidean distance is, the smaller the distance between the characteristic vector of the actual electric quantity difference and the characteristic vector of the combined electric quantity difference is, and otherwise, the larger the Euclidean distance is, the larger r is S,E And the Pearson correlation coefficient represents the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector, wherein the Pearson correlation coefficient tends to 1, and the higher the similarity degree of the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector is, and the lower the similarity degree is.
The variables to be solved in the objective function are a working coefficient, an opening time parameter and a closing time parameter, and the working coefficient a is within the analysis period i The constraints need to be satisfied as follows:
a i ＝{0∪1}；
all electrical equipment have their operation cycle, according to the equipment operation time characteristics, this application divides electrical equipment into continuous operation equipment and interval operation equipment two types, and to the interval operation equipment, generally there is great time interval between its operation cycle, can think that this time interval is greater than the analysis period length, so its switch moment parameter constraint is:
in the formula, t i_cycle Indicating the operating cycle of device i.
The continuously-operating equipment is generally in a long-time operating state, the operating interval is not fixed, a plurality of operating cycles may occur in the analysis period, and the characteristic curve of the accumulated electric quantity of the equipment needs to be extrapolated and prolonged, so the parameter constraint of the switching time is as follows:
-T h <t i_on <T h
t i_on +k*t i_cycle ＝t i_off
in the formula, k represents an operation period, k =1,2, \ 8230;. When t is i_off >T h When, defaults to t i_off ＝T h . For continuously operating plants, t i_off Only represents the time when the device enters the run interval and not the actual time when the device is turned off.
Solving the working coefficient, the opening time parameter and the closing time parameter by adopting an Adaptive Genetic Algorithm (AGA), and determining the optimal individual similarity coefficient to measure the individual diversity of the contemporary population by comparing the differences among individuals according to the differences among individuals as follows:
in the formula, the following components are added,
is the individual j in the z-th generation,
for the best individual in the z-th generation, use
To judge whether the individual j is the same as the current optimal individual, M is the population scale, rho z The similarity coefficient of the optimal individual is the proportion of the same number of individuals in the population as the optimal individual.
Considering the optimal individual similarity coefficient, the self-adaptive crossover/variation probability setting principle of the application is as follows:
1) As the number of iterations increases, the crossover probability should gradually decrease, and the mutation probability should gradually increase;
2) Individuals with poor adaptation performance should be endowed with a larger cross probability and a smaller mutation probability, and individuals with good adaptation performance are endowed with cross and mutation probabilities according to individual performance and population distribution conditions;
3) For population with poor population diversity, the mutation probability should be increased and the cross probability should be decreased.
The corresponding implementation method is as follows:
for the setting principle 1), a method of setting an upper limit and a lower limit of the cross/mutation probability is adopted to designate the probability change direction;
for the setting principle 2), the method is realized by adopting a mode of defining a piecewise function according to the size of the fitness value;
for the setting principle 3), by introducing an optimal individual similarity coefficient.
According to the principle and the implementation method, the application provides an adaptive genetic algorithm based on the optimal individual similarity coefficient, and the cross/variation probability is adaptively adjusted as follows:
in the formula, p z,c1 ，p z,c2 Is the upper and lower limits of the crossover probability of z-th generation population individuals, p c_max ，p c_min Is the maximum and minimum cross probability of the algorithm, mgen is the maximum iteration number of the algorithm, p z,m1 ，p z,m2 Is the upper and lower limits of variation probability, p, of z generation population individuals m_max ，p m_min The maximum value and the minimum value of the algorithm variation probability,
is the larger fitness of the individual j and the individual k, f z,avg Is the fitness average of the z-th generation individuals, f z,max Is the maximum fitness of the z-th generation of individuals, p z,(j,k)-c Is the cross probability, p, of the z-th generation of individuals j and k z,j-m The mutation probability of the z-th generation individual j is shown.
Referring to fig. 9, to better fit the search process from wide to fine in the iterative process, the algorithm of the present application is performed in stages, each stage uses a different selection strategy, and before solving, the variable to be solved needs to be chromosome-coded, and the present application uses a binary coding method, in which the working coefficient a is a i Is a variable of 0 to 1, the parameter t of the starting moment i_on Value range of t i_on |<T h If the length of the analysis period is 60 minutes, t is i_on 6 digit value and 1 digit positive and negative digits are needed, and the closing time parameter t i_off In relation to the number k of operating cycles of the device, within the analysis period set out herein, the number of operating cycles of a single device does not exceed 2,the method can be represented by a one-bit binary system, and in summary, the chromosome unit of a single device is a 9-bit binary system, and if there are n electrical devices, one chromosome is an n × 9 matrix.
The invention provides a differential fitting model based on the characteristic that accumulated electric quantity is time cycle type data, the sampling frequency is low, the acquisition difficulty is small, a user-side intelligent electric meter can realize data acquisition without real-time communication feedback, the accumulated electric quantity differential of each device in the sampling period is intercepted, namely the electric quantity of each device in the sampling period is fitted to the total electric quantity between sampling points of a bus, the sampling frequency of the accumulated electric quantity at the bus is lower than that of the accumulated electric quantity characteristic curve of the device in the mode, the consistency constraint of the sampling frequency is broken, aiming at the resident load constitution, an optimization solution based on an optimal individual similarity coefficient is provided on the basis of an adaptive genetic algorithm, the differential fitting model belongs to the problem of high-dimensional multi-constraint linear integer programming, therefore, the patent adopts the adaptive genetic algorithm to solve, starts from the difference between individuals, the difference between the individuals, the optimal individual similarity coefficient (the proportion of the number of the same individual number in the population as the optimal individual is compared with the diversity of the population) is defined to measure the individual diversity of the population, the population diversity evaluation of the population by the individual diversity evaluation of the individual diversity of the population, and the population shortage of the population is improved.
Referring to fig. 10, in another embodiment, an intelligent socket is used to monitor and sample electrical equipment, and it can be seen that, in the power curves collected in fig. 10 when three washing machines with different brands and models but the same capacity are operated, the power characteristics of the washing machines with different brands and models have larger differences in peak values and curve trends under high-frequency sampling.
Referring to fig. 11, based on the three washing machines mentioned in fig. 10, the curves of the accumulated operating power sampled by the three washing machines in minutes are respectively shown, the difference of the accumulated power of the washing machines of different brands and models during operation is much smaller than the power change, the power consumption is approximately the same (about 0.08-0.09kW · h), and the difference of the time section data such as power and current of the electrical devices of different brands and models is large, so high-frequency sampling is required to accurately grasp the load characteristics of the devices, however, if the capacities of the electrical devices of different brands and models are the same, the power consumption is approximately the same, the capacity types of each type of electrical devices are much smaller than the brands and models of the devices, so that an accurate and complete load characteristic database is more easily established by selecting the accumulated power as the load mark.
The device accumulated electric quantity is sampled for 1 minute/time, compared with steady-state power, current and various transient characteristics, the accumulated electric quantity is used as a load characteristic, the sampling frequency can be reduced from the second level or the millisecond level to the minute level, an intelligent socket is installed in tens of users for data acquisition, common capacities of various devices are basically covered, and based on the advantages of high clustering convergence speed and high clustering center interpretation degree of the K-means algorithm, the electric devices are clustered by the operation accumulated electric quantity data of the electric devices, classified according to the capacity, and various clustering centers are taken as electric device accumulated electric quantity characteristic curves under the capacity level.
The invention selects six electrical equipment of a refrigerator, a water heater, an air conditioner, an electric cooker, an electric kettle and a washing machine to build an experimental environment, the accumulated electric quantity characteristic curve of each electrical equipment is shown in an appendix, wherein the air conditioner, the refrigerator and the water heater are continuously operated and the accumulated electric quantity characteristic curve is extrapolated to 2 operation periods. In the experiment, the sampling period of the intelligent electric meter is 15 minutes, the sampling duration is 60 minutes, in the experiment, the washing machine starts to work before monitoring and sampling, the electric kettle is not started, other electric appliances are started to work after monitoring and sampling, table 1 shows the accumulated electric quantity of each sampling point obtained by sampling of the intelligent electric meter, a simulation experiment is carried out on an MATLAB platform, wherein relevant parameters of an AGASC algorithm are set as shown in table 2, the accumulated electric quantity of the sampling points is input, optimization solution is carried out on working coefficients and switching time parameters of electric appliances, the result is shown in table 3 after 1000 times of iterative evolution, and the fitted accumulated electric quantity of the sampling points is obtained as shown in table 4.
TABLE 1 sample points cumulative power
Table 2 AGASC algorithm parameter set
TABLE 3 AGASC Algorithm solution results
TABLE 4 fitting sampling points cumulative electric quantity
| Sampling time point/ |
0 | 15 | 30 | 45 | 60 |
| Cumulative electric quantity/ |
0 | 0.1215 | 0.4457 | 0.9530 | 1.3297 |
According to the parameters, the actual electric quantity of the sampling period and the fitting electric quantity in each sampling period form, the objective function value of the optimal chromosome set is 528.5148, the Euclidean distance is 0.0019, the Pearson correlation coefficient reaches 0.9998, and the characteristic difference between the vector S and the vector E is small.
Please refer to fig. 12, which is an iterative evolution situation of the AGASC algorithm, in the first stage, due to the linear sorting selection, the individual difference is large, but the search range is expanded and the population diversity is maintained, in the second stage, the algorithm increases the probability that the individual with better adaptability is selected by amplifying the fitness of the partial individual sorted in the front and simultaneously reducing the fitness of the partial individual sorted in the back, so as to accelerate the population iteration process, the population begins to win or lose, and in the third stage, the elite retention strategy enables the algorithm to realize probability convergence by retaining the optimal individual of each generation.
In another embodiment, the superiority of the solution performance of the AGASC algorithm of the present invention is further compared, and the above experiments are performed by using the AGA algorithm, the staged fully Adaptive Genetic Algorithm (AGASC) and the dynamic adaptive particle swarm algorithm (DAPSO) according to the parameter settings of table 2, respectively, table 5 is a comparison of calculation results of the algorithms, and it can be found from the analysis of the results of table 5 that the solution result of the AGASC algorithm of the present invention is significantly better than the other three algorithms, compared to the AGASC algorithm, the operation time of the AGASC algorithm of the present invention is slightly longer, but the convergence algebra is less than that of the AGASC algorithm, the operation time of the AGA algorithm is shortest, but the optimization result is worst, and the DAPSO algorithm has the fastest convergence speed, but the optimization effect is not as good as the AGASC algorithm.
TABLE 5 comparison of results calculated by different algorithms
| Algorithm | max F | Convergent algebra | Operation time/s |
| AGASC | 528.5148 | 800 | 76.63 |
| AGAES | 319.1117 | 881 | 68.32 |
| AGA | 226.1062 | 721 | 43.35 |
| DAPSO | 303.7185 | 647 | 77.50 |
The optimal individual similarity coefficients of the AGASC algorithm in the iterative process are compared, so that the similarity of the AGA algorithm among individuals in the later stage is high, the optimal individuals in the population occupy a leader position and are easy to fall into local optimization, while the AGAES algorithm has the advantages that although the optimal individual similarity coefficient in the later stage is not higher than that of the AGA algorithm, a large number of individuals gather to the local optimal individuals due to a roulette selection strategy of fitness amplification/reduction in the middle stage, and the optimal individual similarity coefficient is always kept in the later stage.
In order to measure the accuracy of the load composition analysis of the proposed method, the accuracy of the load composition analysis of the proposed method is measured by the equipment power analysis error, as shown in the following formula:
wherein eq is i Representing the quantity of electricity of the device i, sq, obtained by the algorithm i Representing the actual charge of the device i.
The electrical equipment is randomly started, algorithm verification is carried out for 50 times by using the method, the analysis accuracy of each electrical equipment is recorded, and is compared with the identification accuracy under a basic load characteristic set consisting of P, Q and fundamental current, and the result statistics are shown in a table 6:
TABLE 6 different equipment resolution accuracy
The analysis of the above table shows that the average analysis accuracy of the algorithm adopted by the invention is 84.1%, the accuracy is high, the low-frequency sampling achieves the analysis effect not lower than the high-frequency sampling, the further analysis finds that the analysis accuracy of the algorithm is high for high-power electric appliances such as water heaters and air conditioners and partial nonlinear loads, and the analysis accuracy is low for equipment with the operation period less than the sampling period.
The invention provides a differential fitting model based on the characteristic that accumulated electric quantity is time periodic data, the sampling frequency is low, the acquisition difficulty is small, a user-side intelligent electric meter can realize data acquisition without real-time communication feedback, the total electric quantity between sampling points of a bus is fitted by intercepting accumulated electric quantity differential of an accumulated electric quantity characteristic curve of each device in a sampling period, namely the electric quantity used by each device in the sampling period, in this way, the sampling frequency of the accumulated electric quantity at the bus is lower than that of the accumulated electric quantity characteristic curve of the device, the consistency constraint of the sampling frequency is broken, aiming at the resident load constitution, an optimization method based on an optimal individual similarity coefficient is provided on the basis of an adaptive genetic algorithm to carry out optimization solution, the differential fitting model belongs to the problem of high-dimensional multi-constraint linear integer programming, therefore, the patent adopts the adaptive genetic algorithm to solve, starts from the individual difference, and defines the optimal individual similarity coefficient (the proportion of the number of the same as the optimal individual in the population) to measure the individual diversity of the contemporary population by comparing the individual difference, thereby avoiding the insufficient accuracy of the population analysis diversity evaluation of the population.
Referring to fig. 13, the present invention provides a device for analyzing residential load composition based on low frequency sampling of accumulated electric quantity, comprising:
a first module 11, configured to determine an actual electrical quantity difference characteristic vector according to the user accumulated electrical quantity data.
A second module 12, configured to determine a combined electrical quantity difference characteristic vector according to the on-time parameter and the off-time parameter of the electrical apparatus and each operation state of the electrical apparatus in a sampling cycle, where each operation state includes an on-hold state of the electrical apparatus, an on-state of the electrical apparatus, an off-state of the electrical apparatus, and an on-off state of the electrical apparatus.
A constructing module 13, configured to construct an objective function according to the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector, and determine a difference between the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector according to the objective function.
For the specific definition of the residential load composition analysis device based on the low-frequency sampling of the accumulated electric quantity, reference may be made to the above definition, which is not described herein again. The modules in the device for analyzing the residential load composition based on the low-frequency sampling of the accumulated electric quantity can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
While the foregoing is directed to the preferred embodiment of the present invention, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention.
Claims (6)
1. A resident load composition analysis method based on accumulated electric quantity low-frequency sampling is characterized by comprising the following steps:
determining an actual electrical quantity difference characteristic vector according to the accumulated electrical quantity data of the user, wherein the actual electrical quantity difference characteristic vector comprises: the accumulated electric quantity difference of the sampling period;
determining a combined electric quantity difference characteristic vector according to the on-time parameter and the off-time parameter of the electrical equipment and each operation state of the electrical equipment in a sampling period, wherein each operation state comprises an on-state of the electrical equipment, an off-state of the electrical equipment and an on-off state of the electrical equipment, and the combined electric quantity difference characteristic vector comprises: analyzing the accumulated electric quantity differential quantity vector in the period;
constructing an objective function according to the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector, and judging the difference between the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector according to the objective function, wherein the objective function is as follows:
wherein cov (S, E) represents the covariance of the actual delta-electrical-quantity eigenvector and the combined delta-electrical-quantity eigenvector, σ S 、σ E Singular values, d, representing the actual and combined electrical quantity difference eigenvectors, respectively S,E Represents the realityEuclidean distance, r, of electrical quantity differential quantity eigenvectors and combined electrical quantity differential quantity eigenvectors S,E A Pearson correlation coefficient representing the actual electrical quantity difference characteristic vector and the combined electrical quantity difference characteristic vector, m being a number of sampling cycles within an analysis period, S (i) being the actual electrical quantity difference characteristic vector, E (i) being the combined electrical quantity difference characteristic vector;
the determining a difference between the actual electrical quantity difference characteristic vector and the combined electrical quantity difference characteristic vector according to the objective function includes:
the smaller the Euclidean distance is, the smaller the distance between the actual electric quantity differential quantity characteristic vector and the combined electric quantity differential quantity characteristic vector is, and otherwise, the larger the distance is;
the Pearson correlation coefficient tends to 1, which means that the similarity between the actual electrical quantity difference characteristic vector and the combined electrical quantity difference characteristic vector is higher, and vice versa.
2. The method for analyzing the composition of the load of the residents based on the low-frequency sampling of the accumulated electric quantity according to claim 1, wherein the characteristic vector S (m) of the actual electric quantity difference is as follows:
S(m)＝[s(1),s(2),…,s(m)]；
where s (m) represents the accumulated charge difference for the mth sampling period.
3. The method for analyzing load composition of residents based on low-frequency sampling of accumulated electric quantity according to claim 1, wherein said combined electric quantity difference amount eigenvector E (m) is as follows:
wherein, E i (m) represents the accumulated electrical delta vector over the analysis period, and N represents the number of appliance types.
4. A resident load composition analysis device based on accumulated electric quantity low-frequency sampling is characterized by comprising:
a first module, configured to determine an actual electrical quantity difference characteristic vector according to user accumulated electrical quantity data, where the actual electrical quantity difference characteristic vector includes: the accumulated electric quantity difference of the sampling period;
a second module, configured to determine a combined electrical quantity difference characteristic vector according to an on-time parameter and an off-time parameter of an electrical device and each operation state of the electrical device in a sampling cycle, where each operation state includes an on-state of the electrical device, an off-state of the electrical device, and an on-off state of the electrical device, and the combined electrical quantity difference characteristic vector includes: analyzing the accumulated electric quantity differential quantity vector in the period;
a building module, configured to build an objective function according to the actual electric quantity difference amount eigenvector and the combined electric quantity difference amount eigenvector, and determine a difference between the actual electric quantity difference amount eigenvector and the combined electric quantity difference amount eigenvector according to the objective function, where the objective function is as follows:
wherein cov (S, E) represents the covariance of the actual delta-electrical-quantity eigenvector and the combined delta-electrical-quantity eigenvector, σ S 、σ E Singular values, d, representing the actual and combined electrical quantity difference eigenvectors, respectively S,E A Euclidean distance, r, representing the actual electrical dispersion amount eigenvector and the combined electrical dispersion amount eigenvector S,E A Pearson correlation coefficient representing the actual electrical quantity difference characteristic vector and the combined electrical quantity difference characteristic vector, m being a number of sampling cycles within an analysis period, S (i) being the actual electrical quantity difference characteristic vector, E (i) being the combined electrical quantity difference characteristic vector;
the determining a difference between the actual delta electric quantity feature vector and the combined delta electric quantity feature vector according to the objective function includes:
the smaller the Euclidean distance is, the smaller the distance between the actual electric quantity difference characteristic vector and the combined electric quantity difference characteristic vector is, and otherwise, the larger the distance is;
the Pearson correlation coefficient is biased to 1, which means that the similarity between the actual electrical quantity difference characteristic vector and the combined electrical quantity difference characteristic vector is higher, and vice versa.
5. The device for analyzing the composition of the residential load based on the low-frequency sampling of the accumulated electric quantity according to claim 4, wherein the first module is further configured to construct the actual electric quantity difference quantity eigenvector S (m) as follows:
S(m)＝[s(1),s(2),…,s(m)]；
where s (m) represents the accumulated charge difference for the mth sampling period.
6. The device for analyzing the composition of the residential load based on the low-frequency sampling of the accumulated electric quantity according to claim 4, wherein the second module is further configured to construct the combined electric quantity difference characteristic vector E (m) as follows:
wherein E is i (m) represents the accumulated electrical delta vector over the analysis period, and N represents the number of appliance types.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110669092.2A CN113363974B (en) | 2021-06-16 | 2021-06-16 | Method and device for analyzing residential load composition based on accumulated electric quantity low-frequency sampling |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110669092.2A CN113363974B (en) | 2021-06-16 | 2021-06-16 | Method and device for analyzing residential load composition based on accumulated electric quantity low-frequency sampling |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113363974A CN113363974A (en) | 2021-09-07 |
| CN113363974B true CN113363974B (en) | 2023-01-20 |
Family
ID=77534715
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110669092.2A Active CN113363974B (en) | 2021-06-16 | 2021-06-16 | Method and device for analyzing residential load composition based on accumulated electric quantity low-frequency sampling |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113363974B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113822569B (en) * | 2021-09-17 | 2024-05-24 | 广东电网有限责任公司 | Air conditioning system load calculation method, device, equipment and medium |
| CN116664790B (en) * | 2023-07-26 | 2023-11-17 | 昆明人为峰科技有限公司 | Three-dimensional terrain analysis system and method based on unmanned aerial vehicle mapping |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103675378A (en) * | 2013-09-23 | 2014-03-26 | 东北电力大学 | A non-intruding-type household-used electric load decomposition method and an apparatus |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102135608A (en) * | 2011-01-29 | 2011-07-27 | 四川电力科学研究院 | Electric energy metering and monitoring system of intelligent gateway |
| CN103605757B (en) * | 2013-11-25 | 2017-02-08 | 国家电网公司 | A classification method of high-speed railway power quality data based on support vector machine |
| CN106443233A (en) * | 2016-08-26 | 2017-02-22 | 北京电力经济技术研究院 | Non-invasive steady-state load monitoring method |
| CN106786534B (en) * | 2016-12-28 | 2019-07-26 | 天津求实智源科技有限公司 | Non-invasive power load transient process identification method and system |
| CN107767037B (en) * | 2017-09-30 | 2021-04-02 | 广东电网有限责任公司信息中心 | User electricity utilization composition analysis method |
| CN112083224A (en) * | 2020-06-18 | 2020-12-15 | 曹靖 | Non-invasive real-time intelligent electric meter system based on characteristic quantity matching and identification method |
-
2021
- 2021-06-16 CN CN202110669092.2A patent/CN113363974B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103675378A (en) * | 2013-09-23 | 2014-03-26 | 东北电力大学 | A non-intruding-type household-used electric load decomposition method and an apparatus |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113363974A (en) | 2021-09-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105067857B (en) | Electricity consumption information acquisition system and analysis method | |
| CN105972761B (en) | A non-intrusive air conditioning load monitoring method | |
| CN110504679B (en) | KM matching algorithm-based non-invasive load identification method | |
| CN111564842B (en) | Method for statistical estimation of medium-and-long-term energy consumption in non-invasive electric load monitoring | |
| CN113363974B (en) | Method and device for analyzing residential load composition based on accumulated electric quantity low-frequency sampling | |
| CN110672934A (en) | Non-invasive load identification method, terminal and system | |
| CN110188826A (en) | Non-intrusive detection method of household appliances operating status based on smart meter data | |
| CN111092434B (en) | Residential community power load control method and device based on non-invasive electricity consumption data | |
| CN105911342B (en) | Electric cooker non-intruding discrimination method based on power duration characteristic | |
| CN110288113A (en) | A non-intrusive load intelligent identification system | |
| CN107302217A (en) | Non-intrusion type load decomposition method and system based on V I tracks | |
| CN109633301B (en) | Non-invasive electrical appliance load identification method based on quantum genetic optimization | |
| CN105823948A (en) | Non-invasive resident load identification method | |
| CN104122819A (en) | User habit based intelligent household electricity utilization method and system thereof | |
| CN112909923A (en) | Non-invasive household load behavior recognition device based on DTW algorithm | |
| CN111722028A (en) | A Load Identification Method Based on High Frequency Data | |
| Rathore et al. | Non intrusive load monitoring and load disaggregation using transient data analysis | |
| CN104573858B (en) | A kind of prediction of network load and regulation and control method | |
| CN112862163A (en) | Short-term power load prediction system based on DTW algorithm | |
| CN206370695U (en) | A kind of photovoltaic plant | |
| CN102662325B (en) | Improved adaptive learning tree power supply management method | |
| CN112016587B (en) | Cloud-based collaborative non-intrusive identification method for energy consumption monitoring based on master station feature library technology | |
| Shabbir et al. | Residential DC Load Forecasting Using Long Short-term Memory Network (LSTM) | |
| CN115112989B (en) | Non-invasive load monitoring method based on low-frequency data | |
| CN116951780A (en) | Intelligent central water heater management method and system |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |