CN108054749B - Non-invasive power load decomposition method and device - Google Patents

Non-invasive power load decomposition method and device Download PDF

Info

Publication number
CN108054749B
CN108054749B CN201711207723.9A CN201711207723A CN108054749B CN 108054749 B CN108054749 B CN 108054749B CN 201711207723 A CN201711207723 A CN 201711207723A CN 108054749 B CN108054749 B CN 108054749B
Authority
CN
China
Prior art keywords
current value
sample data
intra
electric equipment
time period
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
Application number
CN201711207723.9A
Other languages
Chinese (zh)
Other versions
CN108054749A (en
Inventor
刘松
田洁
刘鹏
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.)
Yangzhong Intelligent Electrical Institute North China Electric Power University
North China Electric Power University
Original Assignee
Yangzhong Intelligent Electrical Institute North China Electric Power University
North China Electric Power University
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 Yangzhong Intelligent Electrical Institute North China Electric Power University, North China Electric Power University filed Critical Yangzhong Intelligent Electrical Institute North China Electric Power University
Priority to CN201711207723.9A priority Critical patent/CN108054749B/en
Publication of CN108054749A publication Critical patent/CN108054749A/en
Application granted granted Critical
Publication of CN108054749B publication Critical patent/CN108054749B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • 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]
    • 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
    • H02J2310/00—The network for supplying or distributing electric power characterised by its spatial reach or by the load
    • H02J2310/10—The network having a local or delimited stationary reach
    • H02J2310/12—The local stationary network supplying a household or a building
    • H02J2310/14—The load or loads being home appliances
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02B—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO BUILDINGS, e.g. HOUSING, HOUSE APPLIANCES OR RELATED END-USER APPLICATIONS
    • Y02B70/00—Technologies for an efficient end-user side electric power management and consumption
    • Y02B70/30—Systems integrating technologies related to power network operation and communication or information technologies for improving the carbon footprint of the management of residential or tertiary loads, i.e. smart grids as climate change mitigation technology in the buildings sector, including also the last stages of power distribution and the control, monitoring or operating management systems at local level
    • Y02B70/3225—Demand response systems, e.g. load shedding, peak shaving
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S20/00—Management or operation of end-user stationary applications or the last stages of power distribution; Controlling, monitoring or operating thereof
    • Y04S20/20—End-user application control systems
    • Y04S20/222—Demand response systems, e.g. load shedding, peak shaving

Abstract

The invention discloses a non-invasive power load decomposition method and a non-invasive power load decomposition device, which are used for solving the problems that the existing non-invasive power load decomposition method has high requirement on equipment sampling frequency and low load decomposition efficiency and accuracy. The non-intrusive power load decomposition method comprises the following steps: acquiring the current sum of each sampling point of each piece of electrical equipment to be monitored in a monitoring time period; taking the current sum and the time variable as input layer variables of a neural network model, and taking the working state of each electric device as an output layer variable; determining the working state of each electric device in the monitoring time period by utilizing the neural network model according to the current sum and the time variable; the neural network model is a forward feedback BP neural network model which is trained according to the obtained sampling data of each electric device in a first preset time period.

Description

Non-invasive power load decomposition method and device
Technical Field
The invention relates to the technical field of electric power, in particular to a non-invasive electric power load decomposition method and device.
Background
With the development of smart power grids, the demand of society on power supplies is increasing day by day, and power load decomposition plays an increasingly important role in energy conservation and operation management of modern smart power grids. Through power load decomposition, household power users can obtain power consumption information of all kinds of electric equipment in time, the users can know the power consumption of all kinds of electric equipment in more detail, and the users can be helped to make reasonable energy-saving plans, so that the consumption of electric energy is reduced and the electricity charge expenditure is reduced on the premise that the normal production and life of the users are not influenced. The power load decomposition can also provide detailed user power utilization data for the power department, so that the accuracy of power load prediction can be improved, and data support can be provided for the power department to control power utilization.
The power load decomposition is divided into two modes of intrusive power load decomposition and non-intrusive power load decomposition. Invasive power load decomposes and need to install the sensor for every consumer, gathers each consumer's load power consumption information, and this kind of mode equipment and maintenance cost are higher, need get into the load inside, and the establishment and the transformation of circuit can bring inconvenience for user's life, and the sensor probably influences the operation of consumer simultaneously, and the reliability is lower. Compared with the intrusive power load decomposition, the non-intrusive power load decomposition mode has the advantages of less investment, low cost, easiness in installation and convenience in use, and is very suitable for load decomposition of household power.
At present, the load characteristics of electric equipment used for power load decomposition mainly include two types: the method has the advantages of stable state characteristics and transient state characteristics, high sampling frequency is needed for obtaining the transient state characteristics of the electric equipment, the requirement on monitoring equipment is high, the capital investment is high, the method is not suitable for being applied to families, and the practicability is not high. In addition, in the case that the existing non-intrusive power load decomposition method based on the steady-state characteristics is used for multi-use electrical equipment, if the accuracy of decomposition is required to be ensured, the calculation amount and the calculation time are increased.
Therefore, how to solve the problem that the existing non-intrusive power load decomposition method has high requirement on the sampling frequency and improves the efficiency and accuracy of load decomposition is one of the problems to be solved in the prior art.
Disclosure of Invention
The invention provides a non-invasive power load decomposition method and a non-invasive power load decomposition device, which are used for solving the problems that the existing non-invasive power load decomposition method has high requirement on equipment sampling frequency and low load decomposition efficiency and accuracy.
The embodiment of the invention provides a non-intrusive power load decomposition method, which comprises the following steps:
acquiring the current sum of each sampling point of each piece of electrical equipment to be monitored in a monitoring time period;
taking the current sum and the time variable as input layer variables of a neural network model, and taking the working state of each electric device as an output layer variable;
determining the working state of each electric device in the monitoring time period by utilizing the neural network model according to the current sum and the time variable;
the neural network model is a forward feedback BP neural network model which is trained according to the obtained sampling data of each electric device in a first preset time period.
Optionally, before the operating state of each of the electric devices is used as an output layer variable, the method further includes:
and determining the number of the types of the working states of the electric equipment by using a preset algorithm according to the current value sample data acquired in the second preset time period for each electric equipment.
Preferably, the current value sample data is current value data actually monitored by the electrical equipment within the second preset time period; the preset algorithm is a K-means clustering algorithm;
determining the number of the types of the working states of the electric equipment by using a preset algorithm according to current value sample data acquired in a second preset time period, wherein the method specifically comprises the following steps:
determining a current probability density curve of the electric equipment by using a current probability density function according to the current value sample data;
determining the value range of the number k of the clustering center points according to the peak points in the probability density curve;
determining the number of the optimal clustering central points from the value range of the number k of the clustering central points according to the intra-class distance and the inter-class distance of the current value sample data;
and determining the number of the optimal cluster center points as the number of the working state types of the electric equipment.
Preferably, the probability density function of the current is:
Figure BDA0001484028410000031
wherein p isX(x) Representing a probability density of a current value x of the consumer;
x represents a current value of the consumer;
i represents the current value actually monitored by the electric equipment;
pr [ X ═ X ] represents a probability value that the current value of the electric device is X.
Preferably, the optimal cluster center point number is determined by the following formula:
Figure BDA0001484028410000032
wherein k isoptRepresenting the number of the optimal cluster center points;
k represents the number of the cluster center points, and the value range of k is k ∈ [ k [)min,kmax]；
inside (k) represents an intra-class distance of the current value sample data;
and (k) representing the inter-class distance of the current value sample data.
Preferably, the intra-class distance of the current value sample data is the maximum value of the intra-class distances of the clusters of the sample data; the intra-class distance of the cluster is the minimum value of the average distance between each object in the cluster and each other object in the cluster; and
calculating an intra-class distance of the current value sample data by the following formula:
Figure BDA0001484028410000033
wherein inside (k) represents an intra-class distance of the current value sample data;
Figure BDA0001484028410000034
represents the intra-class distance of the ith cluster;
i＝1,2,...,k，j＝1,2,...|Ci|，p＝1,2,...|Ci|；
Cirepresents the ith cluster;
|Cii indicates that the category belongs to CiThe number of objects of (1);
xj、xpto belong to class CiThe object of (a); and
calculating the inter-class distance of the current value sample data by the following formula:
Figure BDA0001484028410000041
wherein m is 1,2,., k, n is 1,2,., k, and q is not equal to r;
Cmdenotes the m-th cluster, CnRepresents the nth cluster;
xqto belong to class CmObject of (1), xrTo belong to class CnThe object of (1).
Optionally, the method further comprises:
for each electric device, determining a corresponding energy consumption value of the electric device in the monitoring time period by the following formula:
Figure BDA0001484028410000042
wherein the content of the first and second substances,
Figure BDA0001484028410000043
representing a power consumption amount of the powered device;
u represents the operating voltage of the consumer;
k represents the number of the working state types of the electric equipment;
Isindicate the electricity utilization equipmentPreparing a working current value in a working state s;
fsand indicating the frequency of the working state s of the electric equipment in the monitoring time period.
The embodiment of the invention provides a non-intrusive power load decomposition device, which comprises:
the acquisition unit is used for acquiring the current sum of each sampling point of each piece of electric equipment to be monitored in a monitoring time period;
the processing unit is used for taking the current sum and the time variable as input layer variables of a neural network model and taking the working state of each piece of electric equipment as output layer variables;
the first determining unit is used for determining the working state of each piece of electric equipment in the monitoring time period by utilizing the neural network model according to the current sum and the time variable;
the neural network model is a forward feedback BP neural network model which is trained according to the obtained sampling data of each electric device in a first preset time period.
Optionally, the apparatus further comprises:
and the second determining unit is used for determining the number of the working state types of the electric equipment by using a preset algorithm according to current value sample data acquired in a second preset time period for each electric equipment before the working state of each electric equipment is taken as an output layer variable.
Preferably, the current value sample data is current value data actually monitored by the electrical equipment within the second preset time period; the preset algorithm is a K-means clustering algorithm;
the second determining unit is specifically configured to determine a current probability density curve of the electrical device according to the current value sample data by using a probability density function of a current; determining the value range of the number k of the clustering center points according to the peak points in the probability density curve; determining the number of the optimal clustering central points from the value range of the number k of the clustering central points according to the intra-class distance and the inter-class distance of the current value sample data; and determining the number of the optimal cluster center points as the number of the working state types of the electric equipment.
Preferably, the probability density function of the current is:
Figure BDA0001484028410000051
wherein p isX(x) Representing a probability density of a current value x of the consumer;
x represents a current value of the consumer;
i represents the current value actually monitored by the electric equipment;
pr [ X ═ X ] represents a probability value that the current value of the electric device is X.
Preferably, the second determining unit is specifically configured to determine the optimal number of cluster center points by the following formula:
Figure BDA0001484028410000052
wherein k isoptRepresenting the number of the optimal cluster center points;
k represents the number of the cluster center points, and the value range of k is k ∈ [ k [)min,kmax]；
inside (k) represents an intra-class distance of the current value sample data;
and (k) representing the inter-class distance of the current value sample data.
Preferably, the intra-class distance of the current value sample data is the maximum value of the intra-class distances of the clusters of the sample data; the intra-class distance of the cluster is the minimum value of the average distance between each object in the cluster and each other object in the cluster; and
the second determining unit is specifically configured to calculate an intra-class distance of the current value sample data by using the following formula:
Figure BDA0001484028410000061
wherein inside (k) represents an intra-class distance of the current value sample data;
Figure BDA0001484028410000062
represents the intra-class distance of the ith cluster;
i＝1,2,...,k，j＝1,2,...|Ci|，p＝1,2,...|Ci|；
Cirepresents the ith cluster;
Cii indicates that the category belongs to CiThe number of objects of (1);
xj、xpto belong to class CiThe object of (a); and
calculating the inter-class distance of the current value sample data by the following formula:
Figure BDA0001484028410000063
wherein m is 1,2,., k, n is 1,2,., k, and q is not equal to r;
Cmdenotes the m-th cluster, CnRepresents the nth cluster;
xqto belong to class CmObject of (1), xrTo belong to class CnThe object of (1).
Optionally, the apparatus further comprises:
a third determining unit, configured to determine, for each electric device, an energy consumption value corresponding to the electric device in the monitoring time period according to the following formula:
Figure BDA0001484028410000071
wherein the content of the first and second substances,
Figure BDA0001484028410000072
representing a power consumption amount of the powered device;
u represents the operating voltage of the consumer;
k represents the number of the working state types of the electric equipment;
Isthe working current value of the electric equipment in the working state s is represented;
fsand indicating the frequency of the working state s of the electric equipment in the monitoring time period.
The beneficial effects of the invention include:
the non-invasive power load decomposition method provided by the embodiment of the invention obtains the current sum of each sampling point of each electric device to be monitored in a monitoring time period, takes the current sum and a time variable as input layer variables of a neural network model, takes the working state of each electric device as an output layer variable, and determines the working state of each electric device in the monitoring time period by using the neural network model according to the current sum and the time variable, wherein the neural network model is a forward feedback BP neural network model trained according to the obtained sampling data of each electric device in a first preset time period, according to the process of the embodiment of the invention, the current sum and the time variable of the electric device are taken as input nodes of the trained BP neural network model, and the working state of the electric device is taken as an output node to perform power load decomposition, because the current is a load steady-state characteristic, and the data sampling frequency adopted by the experiment is low, the time is used as another input node of the neural network model, the electricity utilization behaviors and habits of users using each type of electric equipment at different time points can be analyzed according to electricity utilization data statistics, the accuracy of working state identification is improved, and the efficiency and the accuracy of load decomposition are improved.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention and not to limit the invention. In the drawings:
fig. 1 is a schematic view of an application scenario of a non-intrusive power load decomposition method according to an embodiment of the present invention;
fig. 2 is a schematic flow chart illustrating an implementation of a non-intrusive power load splitting method according to an embodiment of the present invention;
FIG. 3 is a schematic structural diagram of a BP neural network model according to an embodiment of the present invention;
FIG. 4 is a graph comparing frequency of three kinds of electric devices according to an embodiment of the present invention;
fig. 5 is a schematic flow chart illustrating an implementation of determining the number of the operating state types of the electric devices according to the embodiment of the present invention;
FIG. 6 is a graph of energy consumption prediction accuracy using a conventional neural network model that does not take into account time factors and a neural network model proposed by an embodiment of the present invention;
fig. 7 is a schematic structural diagram of a non-intrusive power load splitting apparatus according to an embodiment of the present invention.
Detailed Description
The invention provides a non-intrusive power load decomposition method and device, and aims to solve the problems that the existing non-intrusive power load decomposition method has high requirement on equipment sampling frequency and low load decomposition efficiency and accuracy.
The implementation principle of the non-intrusive power load decomposition method and device provided by the embodiment of the invention is as follows: obtaining the current sum of each sampling point of each electric device to be monitored in a monitoring time period, using the current sum and a time variable as input layer variables of a neural network model, using the working state of each electric device as output layer variables, and determining the working state of each electric device in the monitoring time period by using the neural network model according to the current sum and the time variable, wherein the neural network model is a forward feedback BP neural network model trained according to the obtained sampling data of each electric device in a first preset time period, according to the process of the embodiment of the invention, using the current sum and the time variable of the electric device as input nodes of the trained BP neural network model, using the working state of the electric device as output nodes to carry out power load decomposition, and the current is a load steady-state characteristic, and the data sampling frequency adopted by the experiment is low, the time is used as another input node of the neural network model, the power utilization behaviors and habits of users using each type of power utilization equipment at different time points can be analyzed according to the power utilization data statistics, the accuracy of the identification of the working state is improved, and the efficiency and the accuracy of the load decomposition are improved.
The preferred embodiments of the present invention will be described below with reference to the accompanying drawings of the specification, it being understood that the preferred embodiments described herein are merely for illustrating and explaining the present invention, and are not intended to limit the present invention, and that the embodiments and features of the embodiments in the present invention may be combined with each other without conflict.
As shown in fig. 1, which is a schematic view of an application scenario of the non-intrusive power load decomposition method according to the embodiment of the present invention, the application scenario may include an electric device 10, an intelligent electric meter 11, an intelligent gateway 12, and a computer 13, where an electric information demonstration system 131 and a MySQL database 132 are installed on the computer 13, each electric device 10 is connected to one intelligent electric meter 11 through a conversion plug, and the above devices together form an electric information acquisition system. The smart meter 11 may collect the power consumption information of the power consumption device in real time, and the power consumption information may include, but is not limited to, the following information: voltage, current, active power. The smart meter 11 transmits the collected power utilization information to the smart gateway, wherein the data collection frequency may be 1/60 Hz. The electricity consumption information demonstration system 131 can collect electricity consumption information of the electricity consumption devices in the intelligent gateway 12 at regular time, store the data into the basic data table of the MySQL database 132, generate an electricity consumption data report of each electricity consumption device, and meanwhile, a user can conveniently observe real-time electricity consumption data of each electricity consumption device, so that each intelligent electric meter 11 is ensured to be in a normal working state, and if the intelligent electric meter 11 breaks down, the reason of the fault can be timely found and checked or replaced.
It should be noted that, in the embodiment of the present invention, the smart gateway 12 may be a smart gateway based on a ZigBee protocol (ZigBee protocol), the smart meter 11 may be a smart meter based on a ZigBee protocol, and the electric device 10 may include: electric network equipment such as a refrigerator, a water dispenser, an air conditioner, a television, a network printer and the like.
As shown in fig. 2, which is a schematic flow chart of an implementation of the non-intrusive power load splitting method according to the embodiment of the present invention, the method may include the following steps:
and S21, acquiring the current sum of each sampling point of each electric device to be monitored in the monitoring time period.
During specific implementation, aiming at each electric device to be detected, the computer terminal sends an electric information query request to the electric device, receives electric information of each sampling point in the monitoring time period returned by the electric device through the intelligent gateway based on the ZigBee protocol, and calculates the current sum of each sampling point in the monitoring time period according to the current value of each sampling point in the electric information.
In another embodiment of the present invention, each electrical device may report the electrical information to the computer actively according to a preset period, or may report the electrical information when a certain trigger condition is met, where the trigger condition may be, but is not limited to, a preset time interval, for example, the preset time interval is 10s, and each network device reports the electrical information to the computer every 10 s.
It should be noted that the monitoring time period and the sampling point may be set by a user according to needs, and the embodiment of the present invention is not limited thereto. For example, the monitoring period is set to 24 hours a day, with each minute as a sampling point.
And S22, taking the current sum and the time variable as input layer variables of a neural network model, and taking the working state of each electric device as an output layer variable.
In this step, the Neural Network model is a Back Propagation Neural Network (Back) model trained according to the obtained sampling data of each electric device in a first preset time period. The neural network can approximate to complex nonlinear mapping with any precision by automatically learning the input and output sample pairs of the system.
In the embodiment of the invention, a three-layer BP neural network model can be adopted to simulate a power load decomposition model, and the BP neural network is trained according to the obtained sampling data of each electric device in a first preset time period.
Fig. 3 is a schematic structural diagram of a BP neural network model adopted in the embodiment of the present invention. The BP neural network consists of an input layer, a hidden layer and an output layer, and the contents of each layer are sequentially designed as follows:
firstly, the number of input neurons, i.e., the number m' of input nodes, of an input layer of the BP neural network model is 2, and the sum of currents of each sampling point of each electric device in a first preset time period and a time variable are respectively used as input layer variables. In specific implementation, time characteristics influencing energy consumption behaviors of a user are analyzed, differences of use conditions of different electric equipment are found after working characteristics of the household electric equipment are counted and observed, the time characteristics of the electric equipment are used as new characteristic quantities, and the new characteristic quantities and total load current are used as input layer variables of a BP neural network model. Because of the lack of types of household appliances and the short monitoring time in the simulated experimental environment, the embodiment of the present invention may use a public data set ampds (the almanac of Minutely Power dataset) to perform an experiment, and the first preset time period may be set by a user, for example, 1 year, as shown in fig. 4, which is a graph comparing usage frequency of each electric device with three electric devices, such as a washing machine, a television, and a dishwasher. In fig. 4, the abscissa is time, and the ordinate represents the number of times of use of the electric device, taking 24 hours a day as an example, it can be seen from fig. 4 that the on and off working conditions of the three electric devices, i.e., the washing machine, the television and the dishwasher, are significantly different, for example, the on time of the television is mostly concentrated around 12 pm, the dishwasher is frequently turned on between 8 pm and 10 pm, the washing machine is often turned on at night and in the morning, which is closely related to the behavior habit of the family, and therefore, the embodiment of the present invention utilizes the difference of the on condition of the electric device, takes time as an input variable, utilizes the powerful autonomous learning function of the neural network model, and deeply excavates hidden information in data.
In the specific implementation, the time variable in the input layer variable may be in minutes, and the period is 24 hours, that is, each minute is used as a sampling point. The sum of the currents at each sampling point of each consumer during the first preset time period is obtained from the AMPDs data set collected during one year.
Secondly, the number of neurons of the output layer of the BP neural network model, that is, the number n' of output layer nodes is the number of electric devices participating in electric load decomposition, and the operating state of each electric device is used as an output layer variable.
In specific implementation, assuming that the electrical devices involved in load splitting have 9 common household electrical appliances, namely, a basic light (BME), a Clothes Dryer (CDE), a washing machine (CWE, Clothes Washer), a Dishwasher (DWE, Dishwasher), a refrigerator (FGE, kitchen fridge), an air conditioner (FRE), a Heat Pump (HPE, Heat Pump), a Television (TVE), and a fireplace (WOE, Wall Oven), with respect to the used AMPDs dataset, the number of neurons n' in the output layer is designed to be 9.
'
Third, the number of hidden layer neurons, i.e., the number of hidden layer nodes/can be determined using the following empirical formula:
Figure BDA0001484028410000121
wherein m 'is the number of nodes of the input layer, n' is the number of nodes of the output layer, and a is a constant between 1 and 10.
And finally determining the number of nodes of the hidden layer by utilizing a step-by-step pruning method on the basis of referring to an empirical formula. In the embodiment of the present invention, the number of hidden layer nodes may be: l' ═ 7. In a specific implementation process, the number of hidden layer nodes may be set according to actual experience, and is not limited herein.
Preferably, before training the BP neural network model, the number of the operating state classes of each electric device needs to be determined.
Specifically, before the working state of each electric device is used as an output layer variable, the number of the working state types of each electric device is determined by using a preset algorithm according to current value sample data acquired in a second preset time period.
In specific implementation, the current value sample data is current value data actually monitored by the electrical equipment within the second preset time period, the preset algorithm may be a K-means (K-means) clustering algorithm, and the K-means algorithm is a simple and effective distance-based clustering algorithm and can classify data objects under the condition of a given number of clustering central points. In practice, the types k of the operating states of different electric devices are different, and therefore, how to determine the number of the cluster center points is an urgent problem to be solved. The second preset time period may be selected by the user according to the need, and is not limited herein.
Specifically, for each electric device, the number of the operating state categories of the electric device may be determined according to the process shown in fig. 5, which may include the following steps:
and S31, determining a current probability density curve of the electric equipment by using a current probability density function according to the current value sample data.
Specifically, a current value with a relatively high occurrence probability, that is, the number of peak points appearing in a probability density curve of the current, may be selected using a probability density function of the current, where the probability density function of the current is expressed as:
Figure BDA0001484028410000122
wherein p isX(x) Representing a probability density of a current value x of the consumer; x represents a current value of the consumer; i represents the current value actually monitored by the electric equipment; pr [ X ═ X]Representing the probability value of the current value x of the consumer.
And calculating the probability density of each current value of the electric equipment in the current value sample data according to the probability density function formula, and generating a probability density curve.
And S32, determining the value range of the number k of the clustering center points according to the peak points in the probability density curve.
Specifically, according to the number of peak points appearing in the probability density curve, the value range of the number k of the cluster center points for classifying the current value sample data is determined to be k ∈ [ k [min,kmax]. For example, the current value appearing after counting the current distribution of a certain electric device has {0,1.5,3,3.6}, the probability of corresponding appearance is {0.6,0.15,0.15,0.1}, and the range of k can be determined as [3,4 ]]。
And S33, determining the number of the optimal cluster central points from the value range of the number k of the cluster central points according to the intra-class distance and the inter-class distance of the current value sample data.
Specifically, the optimal number of cluster center points is a k value when the ratio of the intra-class distance to the inter-class distance is a minimum value, and the optimal number of cluster center points can be determined by the following formula:
Figure BDA0001484028410000131
wherein k isoptRepresenting the number of the optimal cluster center points;
k represents the number of the cluster center points, and the value range of k is k ∈ [ k [)min,kmax]；
inside (k) represents an intra-class distance of the current value sample data;
and (k) representing the inter-class distance of the current value sample data.
The intra-class distance of the current value sample data is the maximum value of the intra-class distance of each cluster of the sample data; the intra-class distance of the cluster is the minimum of the average distances between each object in the cluster and the other objects in the cluster.
Specifically, the intra-class distance of the current value sample data may be calculated by the following formula:
Figure BDA0001484028410000141
wherein inside (k) represents an intra-class distance of the current value sample data;
Figure BDA0001484028410000142
represents the intra-class distance of the ith cluster;
i＝1,2,...,k，j＝1,2,...|Ci|，p＝1,2,...|Ci|；
Cirepresents the ith cluster;
|Cii indicates that the category belongs to CiThe number of objects of (1);
xj、xpto belong to class CiThe object of (1).
The inter-class distance of the current value sample data may be calculated by the following formula:
Figure BDA0001484028410000143
wherein m is 1,2,., k, n is 1,2,., k, and q is not equal to r;
Cmdenotes the m-th cluster, CnRepresents the nth cluster;
xqto belong to class CmObject of (1), xrTo belong to class CnThe object of (1).
And S34, determining the number of the optimal cluster center points as the number of the working state types of the electric equipment.
Specifically, the determined optimal number of the cluster center points is the number of the working state types of the electric equipment.
Further, the BP neural network model is trained.
Specifically, in the implementation of the present invention, the transfer function of the hidden layer and the transfer function of the output layer may adopt a Sigmoid function, the training function may adopt a trailing dx function, and a Mean square error function (MSE, Mean Sum of Squares) is used as a performance target for model training to train the BP neural network model. It should be noted that the transfer function and the training function are not limited to the above two.
And then, taking the BP neural network model after training as a non-invasive power load decomposition model.
And S23, determining the working state of each electric device in the monitoring time period by using the neural network model according to the current sum and the time variable.
In specific implementation, the obtained current sum and time variable of each sampling point of each electric device to be monitored in the monitoring time period are used as input layer variables of the BP neural network model after training in the above steps, the working state of each electric device is used as output layer variables, power load decomposition is carried out, and the working state of each electric device corresponding to each sampling point is obtained.
Further, for each electric device, the corresponding energy consumption value of the electric device in the monitoring time period may be determined by the following formula:
Figure BDA0001484028410000151
wherein the content of the first and second substances,
Figure BDA0001484028410000152
representing a power consumption amount of the powered device;
u represents the operating voltage of the consumer;
k represents the number of the working state types of the electric equipment;
Isthe working current value of the electric equipment in the working state s is represented;
fsand indicating the frequency of the working state s of the electric equipment in the monitoring time period.
In specific implementation, the working voltage U is 220V and IsIs determined according to the cluster center, i.e. the current value of the cluster center corresponding to the working state s of the electrical device.
The embodiment of the invention can evaluate the power load decomposition method from two aspects of load decomposition accuracy and energy consumption prediction accuracy.
The accuracy indexes for evaluating the load decomposition method are many, the most widely used method is that the basic accuracy (accuracisure) and the F value (F-score) are provided, the F value can more accurately balance the performance of a NILM (Non-intrusive load monitoring) method, but only aims at the problem of two classifications, and the method is not suitable for multi-state electrical appliances. The electrical appliances selected by the experiment comprise multi-state electrical appliances, so that the most basic accuracy can be selected to evaluate the effectiveness of the method, and the calculation formula of the accuracy is as follows:
sampling point number/total sampling point number with correct accuracy rate of state identification
The public data sets AMPDs are used for carrying out experiments, 9 common electric appliances are selected and compared with a traditional power load decomposition method based on a neural network model and other methods using the same data, and the experimental results show that the state recognition accuracy can be remarkably improved by the method disclosed by the invention, and as shown in the table 1, the state recognition accuracy of the traditional neural network model method and the method disclosed by the embodiment of the invention is high.
Figure BDA0001484028410000161
On the basis of completing state identification, the accuracy of energy consumption prediction of each electrical appliance is calculated by using the following formula:
Figure BDA0001484028410000162
Figure BDA0001484028410000163
is the predicted power consumption value and y represents the actual power consumption value. The traditional neural network model without considering the time factor and the neural network model provided by the invention are respectively used for carrying out experiments, and the corresponding energy consumption prediction accuracy is calculated, as shown in fig. 6, the energy consumption prediction accuracy of the method provided by the invention is superior to that of the traditional method.
The non-invasive power load decomposition method provided by the embodiment of the invention comprises the steps of obtaining the current sum of each sampling point of each electric device to be monitored in a monitoring time period, using the current sum and a time variable as input layer variables of a neural network model, using the working state of each electric device as output layer variables, and determining the working state of each electric device in the monitoring time period by using the neural network model according to the current sum and the time variable, wherein the neural network model is a forward feedback BP neural network model trained according to the obtained sampling data of each electric device in a first preset time period. The working state of the electric equipment is used as an output node to carry out power load decomposition, the working state of the current value is marked, a probability density function and a clustering algorithm are combined, the ratio of the distance in a class to the distance between the classes is used as a criterion for judging the clustering effectiveness, the optimal state marking effect is achieved, the current is used as the load steady-state characteristic and can be obtained only by low-frequency sampling, time is used as another input node of a neural network model, the electricity utilization behaviors and habits of each kind of electric equipment used by a user at different time points can be statistically analyzed according to electricity utilization data, the accuracy of working state identification is improved, and the efficiency and the accuracy of load decomposition are improved.
Based on the same inventive concept, the embodiment of the invention also provides a non-intrusive power load decomposition device, and as the principle of the device for solving the problems is similar to the non-intrusive power load decomposition method, the implementation of the device can refer to the implementation of the method, and repeated details are not repeated.
As shown in fig. 7, which is a schematic structural diagram of a non-intrusive power load splitting apparatus according to an embodiment of the present invention, the non-intrusive power load splitting apparatus may include:
the acquiring unit 41 is configured to acquire a current sum of each sampling point of each piece of electrical equipment to be monitored in a monitoring time period;
the processing unit 42 is configured to use the current sum and the time variable as input layer variables of a neural network model, and use the working states of the electrical devices as output layer variables;
a first determining unit 43, configured to determine, according to the sum of currents and a time variable, a working state of each electrical device in the monitoring time period by using the neural network model;
the neural network model is a forward feedback BP neural network model which is trained according to the obtained sampling data of each electric device in a first preset time period.
Optionally, the apparatus further comprises:
and the second determining unit is used for determining the number of the working state types of the electric equipment by using a preset algorithm according to current value sample data acquired in a second preset time period for each electric equipment before the working state of each electric equipment is taken as an output layer variable.
Preferably, the current value sample data is current value data actually monitored by the electrical equipment within the second preset time period; the preset algorithm is a K-means clustering algorithm;
the second determining unit is specifically configured to determine a current probability density curve of the electrical device according to the current value sample data by using a probability density function of a current; determining the value range of the number k of the clustering center points according to the peak points in the probability density curve; determining the number of the optimal clustering central points from the value range of the number k of the clustering central points according to the intra-class distance and the inter-class distance of the current value sample data; and determining the number of the optimal cluster center points as the number of the working state types of the electric equipment.
Preferably, the probability density function of the current is:
Figure BDA0001484028410000181
wherein p isX(x) Electric device for indicationThe prepared current value is the probability density of x;
x represents a current value of the consumer;
i represents the current value actually monitored by the electric equipment;
pr [ X ═ X ] represents a probability value that the current value of the electric device is X.
Preferably, the second determining unit is specifically configured to determine the optimal number of cluster center points by the following formula:
Figure BDA0001484028410000182
wherein k isoptRepresenting the number of the optimal cluster center points;
k represents the number of the cluster center points, and the value range of k is k ∈ [ k [)min,kmax]；
inside (k) represents an intra-class distance of the current value sample data;
and (k) representing the inter-class distance of the current value sample data.
Preferably, the intra-class distance of the current value sample data is the maximum value of the intra-class distances of the clusters of the sample data; the intra-class distance of the cluster is the minimum value of the average distance between each object in the cluster and each other object in the cluster; and
the second determining unit is specifically configured to calculate an intra-class distance of the current value sample data by using the following formula:
Figure BDA0001484028410000191
wherein inside (k) represents an intra-class distance of the current value sample data;
Figure BDA0001484028410000192
represents the intra-class distance of the ith cluster;
i＝1,2,...,k，j＝1,2,...|Ci|，p＝1,2,...|Ci|；
Cito representThe ith cluster;
|Cii indicates that the category belongs to CiThe number of objects of (1);
xj、xpto belong to class CiThe object of (a); and
calculating the inter-class distance of the current value sample data by the following formula:
Figure BDA0001484028410000193
wherein m is 1,2,., k, n is 1,2,., k, and q is not equal to r;
Cmdenotes the m-th cluster, CnRepresents the nth cluster;
xqto belong to class CmObject of (1), xrTo belong to class CnThe object of (1).
Optionally, the apparatus further comprises:
a third determining unit, configured to determine, for each electric device, an energy consumption value corresponding to the electric device in the monitoring time period according to the following formula:
Figure BDA0001484028410000194
wherein the content of the first and second substances,
Figure BDA0001484028410000195
representing a power consumption amount of the powered device;
u represents the operating voltage of the consumer;
k represents the number of the working state types of the electric equipment;
Isthe working current value of the electric equipment in the working state s is represented;
fsand indicating the frequency of the working state s of the electric equipment in the monitoring time period.
For convenience of description, the above parts are separately described as modules (or units) according to functional division. Of course, the functionality of the various modules (or units) may be implemented in the same or in multiple pieces of software or hardware in practicing the invention.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.

Claims (10)

1. A method of non-intrusive power load splitting, comprising:
acquiring the current sum of each sampling point of each piece of electrical equipment to be monitored in a monitoring time period;
taking the current sum and the time variable as input layer variables of a neural network model, and taking the working state of each electric device as an output layer variable;
determining the working state of each electric device in the monitoring time period by utilizing the neural network model according to the current sum and the time variable;
the neural network model is a forward feedback BP neural network model which is trained according to the obtained sampling data of each electric device in a first preset time period;
before the operating states of the electric devices are used as output layer variables, the method further comprises the following steps:
for each electric equipment, determining the number of the types of the working states of the electric equipment by using a preset algorithm according to current value sample data acquired in a second preset time period; the current value sample data is current value data actually monitored by the electric equipment in the second preset time period; the preset algorithm is a K-means clustering algorithm;
determining the number of the types of the working states of the electric equipment by using a preset algorithm according to current value sample data acquired in a second preset time period, wherein the method specifically comprises the following steps:
determining a current probability density curve of the electric equipment by using a current probability density function according to the current value sample data;
determining the value range of the number k of the clustering center points according to the peak points in the probability density curve;
determining the number of optimal clustering center points from the value range of the number k of the clustering center points according to the intra-class distance and the inter-class distance of the current value sample data, wherein the number of the optimal clustering center points is the k value when the ratio of the intra-class distance to the inter-class distance is the minimum value;
and determining the number of the optimal cluster center points as the number of the working state types of the electric equipment.
2. The method of claim 1, wherein the probability density function of the current is:
Figure FDA0002440458790000011
wherein p isX(x) Representing a probability density of a current value x of the consumer;
x represents a current value of the consumer;
i represents the current value actually monitored by the electric equipment;
pr [ X ═ X ] represents a probability value that the current value of the electric device is X.
3. The method of claim 1, wherein the optimal cluster center point number is determined by the following formula:
Figure FDA0002440458790000021
wherein k isoptRepresenting the number of the optimal cluster center points;
k represents a clusterThe number of central points, k, is in the range of k ∈ [ k ]min,kmax]；
inside (k) represents an intra-class distance of the current value sample data;
and (k) representing the inter-class distance of the current value sample data.
4. The method of claim 3, wherein the intra-class distance of the current value sample data is a maximum value of the intra-class distances of the clusters of the sample data; the intra-class distance of the cluster is the minimum value of the average distance between each object in the cluster and each other object in the cluster; and
calculating an intra-class distance of the current value sample data by the following formula:
Figure FDA0002440458790000022
wherein inside (k) represents an intra-class distance of the current value sample data;
Figure FDA0002440458790000023
represents the intra-class distance of the ith cluster;
i＝1,2,...,k，j＝1,2,...|Ci|，p＝1,2,...|Ci|；
Cirepresents the ith cluster;
|Cii indicates that the category belongs to CiThe number of objects of (1);
xj、xpto belong to class CiThe object of (a); and
calculating the inter-class distance of the current value sample data by the following formula:
Figure FDA0002440458790000031
wherein m is 1,2,., k, n is 1,2,., k, and q is not equal to r;
Cmdenotes the m-th cluster, CnIs shown asn clusters;
xqto belong to class CmObject of (1), xrTo belong to class CnThe object of (1).
5. The method of any one of claims 1 to 4, further comprising:
for each electric device, determining a corresponding energy consumption value of the electric device in the monitoring time period by the following formula:
Figure FDA0002440458790000032
wherein the content of the first and second substances,
Figure FDA0002440458790000033
representing a power consumption amount of the powered device;
u represents the operating voltage of the consumer;
k represents the number of the working state types of the electric equipment;
Isthe working current value of the electric equipment in the working state s is represented;
fsand indicating the frequency of the working state s of the electric equipment in the monitoring time period.
6. A non-intrusive electrical load splitting apparatus, comprising:
the acquisition unit is used for acquiring the current sum of each sampling point of each piece of electric equipment to be monitored in a monitoring time period;
the processing unit is used for taking the current sum and the time variable as input layer variables of a neural network model and taking the working state of each piece of electric equipment as output layer variables;
the first determining unit is used for determining the working state of each piece of electric equipment in the monitoring time period by utilizing the neural network model according to the current sum and the time variable;
the neural network model is a forward feedback BP neural network model which is trained according to the obtained sampling data of each electric device in a first preset time period;
the device, still include:
the second determining unit is used for determining the number of the working state types of the electric equipment by using a preset algorithm according to current value sample data acquired in a second preset time period for each electric equipment before the working state of each electric equipment is used as an output layer variable; the current value sample data is current value data actually monitored by the electric equipment in the second preset time period; the preset algorithm is a K-means clustering algorithm;
the second determining unit is specifically configured to determine a current probability density curve of the electrical device according to the current value sample data by using a probability density function of a current; determining the value range of the number k of the clustering center points according to the peak points in the probability density curve; determining the number of optimal clustering center points from the value range of the number k of the clustering center points according to the intra-class distance and the inter-class distance of the current value sample data, wherein the number of the optimal clustering center points is the k value when the ratio of the intra-class distance to the inter-class distance is the minimum value; and determining the number of the optimal cluster center points as the number of the working state types of the electric equipment.
7. The apparatus of claim 6, wherein the probability density function of the current is:
Figure FDA0002440458790000041
wherein p isX(x) Representing a probability density of a current value x of the consumer;
x represents a current value of the consumer;
i represents the current value actually monitored by the electric equipment;
pr [ X ═ X ] represents a probability value that the current value of the electric device is X.
8. The apparatus of claim 6,
the second determining unit is specifically configured to determine the optimal number of cluster center points by using the following formula:
Figure FDA0002440458790000042
wherein k isoptRepresenting the number of the optimal cluster center points;
k represents the number of the cluster center points, and the value range of k is k ∈ [ k [)min,kmax]；
inside (k) represents an intra-class distance of the current value sample data;
and (k) representing the inter-class distance of the current value sample data.
9. The apparatus according to claim 8, wherein the intra-class distance of the current value sample data is a maximum value of the intra-class distances of the clusters of the sample data; the intra-class distance of the cluster is the minimum value of the average distance between each object in the cluster and each other object in the cluster; and
the second determining unit is specifically configured to calculate an intra-class distance of the current value sample data by using the following formula:
Figure FDA0002440458790000051
wherein inside (k) represents an intra-class distance of the current value sample data;
Figure FDA0002440458790000052
represents the intra-class distance of the ith cluster;
i＝1,2,...,k，j＝1,2,...|Ci|，p＝1,2,...|Ci|；
Cirepresents the ith cluster;
|Cii indicates that the category belongs to CiThe number of objects of (1);
xj、xpto belong to class CiThe object of (a); and
calculating the inter-class distance of the current value sample data by the following formula:
Figure FDA0002440458790000053
wherein m is 1,2,., k, n is 1,2,., k, and q is not equal to r;
Cmdenotes the m-th cluster, CnRepresents the nth cluster;
xqto belong to class CmObject of (1), xrTo belong to class CnThe object of (1).
10. The apparatus of any of claims 6 to 9, further comprising:
a third determining unit, configured to determine, for each electric device, an energy consumption value corresponding to the electric device in the monitoring time period according to the following formula:
Figure FDA0002440458790000054
wherein the content of the first and second substances,
Figure FDA0002440458790000055
representing a power consumption amount of the powered device;
u represents the operating voltage of the consumer;
k represents the number of the working state types of the electric equipment;
Isthe working current value of the electric equipment in the working state s is represented;
fsand indicating the frequency of the working state s of the electric equipment in the monitoring time period.
CN201711207723.9A 2017-11-27 2017-11-27 Non-invasive power load decomposition method and device Active CN108054749B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711207723.9A CN108054749B (en) 2017-11-27 2017-11-27 Non-invasive power load decomposition method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711207723.9A CN108054749B (en) 2017-11-27 2017-11-27 Non-invasive power load decomposition method and device

Publications (2)

Publication Number Publication Date
CN108054749A CN108054749A (en) 2018-05-18
CN108054749B true CN108054749B (en) 2020-08-18

Family

ID=62120612

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711207723.9A Active CN108054749B (en) 2017-11-27 2017-11-27 Non-invasive power load decomposition method and device

Country Status (1)

Country Link
CN (1) CN108054749B (en)

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108847665B (en) * 2018-06-26 2020-04-28 广东石油化工学院 Non-invasive load decomposition method and system
CN109116100B (en) * 2018-07-09 2019-09-27 清华大学 It is a kind of based on coding-decoding structure electric load electricity consumption decomposition method
CN109165604A (en) * 2018-08-28 2019-01-08 四川大学 The recognition methods of non-intrusion type load and its test macro based on coorinated training
CN111242161B (en) * 2019-12-27 2023-08-25 国网山西省电力公司大同供电公司 Non-invasive non-resident user load identification method based on intelligent learning
CN113033775B (en) * 2021-03-10 2023-08-18 南方电网数字电网研究院有限公司 Non-invasive load identification network architecture based on supervised learning
CN113138894B (en) * 2021-04-14 2022-07-15 哈尔滨工业大学 Experimental equipment monitoring method based on power parameter monitoring and screen information identification
CN113408210B (en) * 2021-07-12 2023-08-01 内蒙古电力(集团)有限责任公司乌兰察布供电分公司 Deep learning-based non-invasive load decomposition method, system, medium and equipment

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101282040A (en) * 2008-05-09 2008-10-08 天津大学 Method for real time sorting non-intrusion type electric load
CN105429135A (en) * 2015-12-08 2016-03-23 河南许继仪表有限公司 Distinguishing and decision-making method and system for noninvasive power load decomposition
CN106096726A (en) * 2016-05-31 2016-11-09 华北电力大学 A kind of non-intrusion type load monitoring method and device

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101282040A (en) * 2008-05-09 2008-10-08 天津大学 Method for real time sorting non-intrusion type electric load
CN105429135A (en) * 2015-12-08 2016-03-23 河南许继仪表有限公司 Distinguishing and decision-making method and system for noninvasive power load decomposition
CN106096726A (en) * 2016-05-31 2016-11-09 华北电力大学 A kind of non-intrusion type load monitoring method and device

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Design of non-intrusive power load decomposition and monitoring system based on multi-agent and visual instrument;Dongtao Wang等;《2010 International Conference on Power System Technology》;20101213;第1-4页 *
计及用电模式的居民负荷梯度提升树分类识别方法;王守相等;《电力系统及其自动化学报》;20170930;第29卷(第9期);第27-33页 *

Also Published As

Publication number Publication date
CN108054749A (en) 2018-05-18

Similar Documents

Publication Publication Date Title
CN108054749B (en) Non-invasive power load decomposition method and device
CN106096726B (en) A kind of non-intrusion type load monitoring method and device
Zhai et al. Appliance flexibility analysis considering user behavior in home energy management system using smart plugs
Dinesh et al. Residential power forecasting using load identification and graph spectral clustering
Liu et al. Energy disaggregation of appliances consumptions using ham approach
CN105514993B (en) A kind of electric load decomposition method and device
CN109813978A (en) A kind of non-intruding load-type recognition methods of variation characteristic between comprehensive transient characteristic and stable state
Chen et al. Smartsim: A device-accurate smart home simulator for energy analytics
Macedo et al. Typification of load curves for DSM in Brazil for a smart grid environment
CN103017290A (en) Air conditioner electric energy control device and air conditioner electric energy management method
CN202254110U (en) Air conditioner power control device
CN104680339B (en) A kind of household electrical appliance dispatching method based on Spot Price
CN103489045A (en) Demand response load optimization potential evaluation method based on multi-scene design
CN104537438A (en) Forecast and monitoring method for peak-hour power usage
WO2014126131A1 (en) Lifestyle behavior estimation system, lifestyle behavior estimation device, lifestyle behavior estimation program, and recording medium
JP2016077090A (en) Energy saving support system and energy saving support method
CN105868073A (en) Data center energy saving strategy implementation method based on time series analysis
Eskander et al. A complementary unsupervised load disaggregation method for residential loads at very low sampling rate data
CN112396098A (en) Non-embedded apartment electrical appliance load identification and analysis method, system and application
CN108767847B (en) Non-invasive electric load decomposition method and device
Amayri et al. Flexibility characterization of residential electricity consumption: A machine learning approach
Waczowicz et al. Demand response clustering-how do dynamic prices affect household electricity consumption?
Zhou et al. Analysis of the influence of the stay-at-home order on the electricity consumption in Chinese university dormitory buildings during the COVID-19 pandemic
Ming et al. Typical commercial user adjustable potential assessment method
Lin et al. The pattern recognition of residential power consumption based on HMM

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