CN103020459A - Method and system for sensing multiple-dimension electric utilization activities - Google Patents
Method and system for sensing multiple-dimension electric utilization activities Download PDFInfo
- Publication number
- CN103020459A CN103020459A CN2012105552108A CN201210555210A CN103020459A CN 103020459 A CN103020459 A CN 103020459A CN 2012105552108 A CN2012105552108 A CN 2012105552108A CN 201210555210 A CN201210555210 A CN 201210555210A CN 103020459 A CN103020459 A CN 103020459A
- Authority
- CN
- China
- Prior art keywords
- electricity consumption
- current
- feature
- family
- consumption behavior
- 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.)
- Granted
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a method and a system for sensing multiple-dimension electric utilization activities. The method comprises the following steps of step 1, predefining statistical features of electric utilization equipment under a plurality of dimensions, obtaining a feature value of each statistical feature, and establishing a corresponding statistical model; step 2, installing a collecting unit at an entrance of an electric network, obtaining the current and voltage information of all electric utilization equipment in a real-time way, and comparing and training the feature weight of each statistical feature according to the current and voltage information and input data of an user by a calculating unit; and step 3, carrying out the enlightening type searching and calculation according to the feature values, the statistical models and the feature weights by the calculating unit, so as to obtain the real-time electric utilization activity of each electric utilization equipment, and sending the real-time electric utilization activities to a display unit. The method and the system have the advantages that multiple load features are combined with a series of statistical calculating methods to directly obtain the information of the electric utilization activities of all household equipment, the resolution capability is high, the judgment precision is high, and the electric utilization activities of the similar electric utilization equipment can be effectively distinguished.
Description
Technical field
The present invention relates to the Computer Applied Technology field, particularly a kind of cognitive method and system of various dimensions electricity consumption behavior.
Background technology
IBM in 2008 propose the concept of " the wisdom earth ", the wisdom earth are described as " more thorough perception, more fully interconnect, more deep intellectuality ".U.S. government proposes in the near future also intelligent grid to be brought into schedule as the pith in its " plan of new forms of energy rescue market " at the wisdom earth.Traditional electrical network of knowing with us by comparison, intelligent grid means and obtains as far as possible more information, more pays attention to mutual between the user, and the service that more puts in place is provided by resolving information.The ustomer premises access equipments such as intelligent electric meter directly contact with the user, and guides user electricity consumption behavior is the important embodiment that intelligent grid is different from traditional electrical network.By providing the detailed electricity consumption situation on the relevant devices, can effectively reduce the user to the understanding deviation of electricity consumption behavior, the use habit of optimizing user, thus obtain better power savings.
Therefore, effectively obtaining the detailed power information on the relevant devices in the power utilization environment (family, production environment etc.), is the gordian technique that intelligent grid field user client information gathers.In the situation that does not affect power utilization environment, obtain the monitoring technology of the detailed power information on each equipment from the outside, be called as non-intrusion type load monitoring (NILM) technology.Up to the present, non-intrusion type load monitoring technology mainly comprises two large classes: based on the load monitoring technology of steady-state analysis and load monitoring technology based on transient event.Although these two kinds of technology are all supported the load monitoring of non-intrusion type, satisfy the demand that intelligent grid field user client information gathers, but these two kinds of technology have all been made similar hypothesis to consumer: equipment has metastable running status, determining can be according to known information after the running status, the detailed power information of equipment.
Along with the epoch are progressive, day by day elastification of the behavior of consumer is so that this hypothesis is no longer applicable, for example: in the same time period, the computer of running game program is with respect to the more electric power of simple browsing page consumption, and idle and busy power consumption difference may surpass 30%.Simultaneously, the two dimension of traditional active power, reactive power composition is not enough to distinguish a large amount of similar electrical equipment of current use with electrical feature.Even if introduce higher hamonic wave characteristic set is expanded, a large amount of similar power supply adaptors still can produce the harmonic signal of easily obscuring.
Summary of the invention
The present invention is on the thinking of traditional steady-state analysis, use multiple load characteristic to redefine the steady state (SS) of consumer, in conjunction with series of computation method statistically the whole family front yard power information of Real-time Collection is analyzed, directly obtain the electricity consumption behavioural information on all devices in the family, resolution characteristic is powerful, the judgement precision is high, can effectively distinguish the electricity consumption behavior of similar consumer.
For achieving the above object, the invention provides a kind of cognitive method of various dimensions electricity consumption behavior, comprising:
Step 1, pre-defined under a plurality of dimensions consumer statistical nature and obtain every kind of eigenwert under the statistical nature, and set up corresponding statistical model;
Step 2 is installed collecting unit in the electrical network porch, the current-voltage information of all consumers of Real-time Obtaining, and computing unit is trained the feature weight of each statistical nature according to described current-voltage information and user's input data;
Step 3, described computing unit carries out heuristic search calculating according to described eigenwert, statistical model and feature weight, obtains the real-time electricity consumption behavior of each consumer, and described real-time electricity consumption behavior is sent to display unit.
Further, described collecting unit comprises a current sensor and a voltage sensor, this current sensor is based on the magnetic field induction chip of Hall effect, utilize electromagnetic induction to realize the current data collection of non-intrusion type, the direct place in circuit of this voltage sensor is in parallel with all consumers, measures the voltage on it.
Described computing unit independently uses a processor; Perhaps share a processor with collecting unit; Perhaps share a processor with display unit.
Further, described step 2 comprises:
Step 21, the total current information of voltage of all consumers begins to carry out weight training as training dataset in the family in measurement a period of time,
Step 22 is calculated by described heuristic search, obtains the candidate list that a plurality of candidate result form;
Step 23, the user chooses optimum analysis result according to truth in candidate list;
Step 24 is carried out iterative optimization weight parameter by the information that described candidate list and user are provided, and obtains the feature weight of described all consumers.
Further, the heuristic computing formula in the described step 3 is:
Wherein, t represents character representation total in the family, and c is a kind of electricity condition of using of all electrical equipment in the family, and P (c|t) is illustrated in the situation of known t, and the electricity consumption situation just is the probability of c in the family; The h on right side
iBe the fundamental function of a dimension, λ
iH for correspondence
iWeight parameter, denominator is with the summation of the probability of possible state c ' all in the family in the situation of t.
For achieving the above object, the present invention also provides a kind of sensory perceptual system of various dimensions electricity consumption behavior, comprising:
Pretreatment module, pre-defined statistical nature and the eigenwert of obtaining under every kind of statistical nature at consumer under a plurality of dimensions, and set up corresponding statistical model;
Training module is installed collecting unit in the electrical network porch, and the current-voltage information of all consumers of Real-time Obtaining, computing unit are trained the feature weight of each statistical nature according to described current-voltage information and user's input Data Comparison
Computing module, described computing unit carry out heuristic search according to described eigenwert, statistical model and feature weight and calculate, and obtain the real-time electricity consumption behavior of each consumer, and described real-time electricity consumption behavior is sent to display unit.
Described collecting unit comprises a current sensor and a voltage sensor, this current sensor is based on the magnetic field induction chip of Hall effect, utilize electromagnetic induction to realize the current data collection of non-intrusion type, the direct place in circuit of this voltage sensor is in parallel with all consumers, measures the voltage on it.
Described computing unit independently uses a processor; Perhaps share a processor with collecting unit; Perhaps share a processor with display unit.
Described training module comprises:
The weight training preparation module, the current-voltage information of each consumer in measurement a period of time begins to carry out weight training as training dataset,
The candidate collection acquisition module calculates by described heuristic search, obtains the candidate list that a plurality of candidate result form;
Choose module, the user chooses optimum analysis result according to truth in candidate list;
Weight obtains module, by described candidate list and the information that the user provides being carried out iterative optimization weight parameter, the feature weight of described all consumers of acquisition.
Heuristic computing formula in the described computing module is:
Wherein, t represents character representation total in the family, and c is a kind of electricity condition of using of all electrical equipment in the family, and P (c|t) is illustrated in the situation of known t, and the electricity consumption situation just is the probability of c in the family; The h on right side
iBe the fundamental function of a dimension, λ
iH for correspondence
iWeight parameter, denominator is with the summation of the probability of possible state c ' all in the family in the situation of t.
Technique effect:
1) the present invention uses a large amount of statistical nature models that possess physical significance from different dimensions that equipment is described with electricity condition, and it organically can be combined, powerful resolution characteristic is provided: existing non-intrusion type load monitoring technology uses the same class physical quantity to carry out the description of equipment state mostly, afterwards again according to actual electricity consumption situation identification electrical equipment with the conversion between electricity condition or the state, finish the analysis of power information.These class methods have a direct defective, and when a plurality of equipment had similar physical features in the family, analysis precision can significantly decrease.And the present invention sets up the statistical nature of a plurality of different dimensions and is combined into unified computation model equipment electricity consumption behavior, and analytic process does not rely on the electrical feature of using of single dimension, can effectively distinguish the electricity consumption behavior that like device produces.
2) the present invention is in electricity consumption behavior perception, the computing method of heuristic search have been used, effective statistical nature model in conjunction with various dimensions, greatly improve and judge precision with electricity condition: existing non-intrusion type load monitoring technology only can be used the electrical feature of using of single dimension when carrying out equipment electricity consumption behavioural analysis.And the present invention uses heuristic search Direct Analysis equipment electricity condition under the unified quantization that the various dimensions statistical nature consists of represents, the electricity consumption behavior situation of change of effectively following the trail of each equipment can bring up to 95% with the mean accuracy of power information analysis.
Describe the present invention below in conjunction with the drawings and specific embodiments, but not as a limitation of the invention.
Description of drawings
Fig. 1 is the cognitive method process flow diagram of various dimensions electricity consumption behavior of the present invention;
Fig. 2 is electricity consumption behavior sensory perceptual system algorithm flow synoptic diagram;
Fig. 3 is feature set establishment stage process flow diagram;
Fig. 4 is parameter learning stage process flow diagram;
Fig. 5 is behavioural analysis stage process flow diagram;
Fig. 6 is statistical nature environment synoptic diagram to be measured;
Fig. 7 is the electricity consumption behavior sensory perceptual system synoptic diagram of household internal power utilization environment;
Fig. 8 is the sensory perceptual system synoptic diagram of various dimensions electricity consumption behavior of the present invention.
Embodiment
Fig. 1 is the cognitive method process flow diagram of various dimensions electricity consumption behavior of the present invention.As shown in Figure 1, the method comprises:
Step 1, pre-defined under a plurality of dimensions consumer statistical nature and obtain every kind of eigenwert under the statistical nature, and set up corresponding statistical model;
Step 2 is installed collecting unit in the electrical network porch, the current-voltage information of all consumers of Real-time Obtaining, and computing unit is trained the feature weight of each statistical nature according to described current-voltage information and user's input data;
Step 3, described computing unit carries out heuristic search calculating according to described eigenwert, statistical model and feature weight, obtains the real-time electricity consumption behavior of each consumer, and described real-time electricity consumption behavior is sent to display unit.
Further, described step 2 comprises:
Step 21, the total current information of voltage of all consumers begins to carry out weight training as training dataset in the family in measurement a period of time,
Step 22 by the heuristic search identical with step 3, obtains the candidate list that a plurality of candidate result form;
Step 23, the user chooses optimum analysis result according to truth in candidate list;
Step 24 is carried out iterative optimization weight parameter by the information that described candidate list and user are provided, and obtains the feature weight of described all consumers.
Further, the heuristic computing formula in the described step 3 is:
Wherein, t represents character representation total in the family, and c is a kind of electricity condition of using of all electrical equipment in the family, and P (c|t) is illustrated in the situation of known t, and the electricity consumption situation just is the probability of c in the family; The h on right side
iBe the fundamental function of a dimension, λ
iH for correspondence
iWeight parameter, denominator is with the summation of the probability of possible state c ' all in the family in the situation of t.
Fig. 8 is the sensory perceptual system synoptic diagram of various dimensions electricity consumption behavior of the present invention.As shown in Figure 8, this system comprises:
Described training module 200 comprises:
The weight training preparation module, the current-voltage information of each consumer in measurement a period of time begins to carry out weight training as training dataset,
The candidate collection acquisition module calculates by described heuristic search, obtains the candidate list that a plurality of candidate result form;
Choose module, the user chooses optimum analysis result according to truth in candidate list;
Weight obtains module, by described candidate list and the information that the user provides being carried out iterative optimization weight parameter, the feature weight of described all consumers of acquisition.
Heuristic computing formula in the described computing module 300 is:
Wherein, t represents character representation total in the family, and c is a kind of electricity condition of using of all electrical equipment in the family, and P (c|t) is illustrated in the situation of known t, and the electricity consumption situation just is the probability of c in the family; The h on right side
iBe the fundamental function of a dimension, λ
iH for correspondence
iWeight parameter.
Wherein collecting unit is main data input device, and electricity consumption behavior sensory perceptual system obtains the current-voltage information in the whole family by near the domestic electric network entrance collecting unit being installed.Current sensor in the collecting unit can use the magnetic field induction chip based on Hall effect, utilizes the current data collection of the principle realization non-intrusion type of electromagnetic induction.And voltage sensor directly access the mains circuit can be in parallel with all consumers, measure the voltage on it.Using in the electrical feature of setting up, frequency domain character has requirement for the frequency of image data, in one embodiment of the invention: when algorithm uses Short Time Fourier Transform to calculate eigenwert in the frequency domain space, require to comprise in each cycle 256 sampled points, the sample frequency of 12.8kHz namely need to be arranged the alternating current of 50Hz.
Computing unit is the core of whole electrical energy consumption analysis system, and the current-voltage information of whole family will at first be converted into various dimensions eigenwert (observing the set of the eigenwert in the dimension in difference) herein, and then analyze the real-time power information that obtains each equipment.Computing unit obtains the circuit data of Real-time Collection from collecting unit, after calculating analysis result is delivered to display unit.In an embodiment of the present invention, possessing arbitrarily the equipment of enough computing powers can be as computing unit in the system.Therefore computing unit can independently be realized, also can share a processor with collecting unit, consists of the novel collecting device of similar intelligent electric meter; Can also share a processor with display unit, consist of the novel display device of similar intelligent terminal.
Display unit is the formant of system and user interactions, except basic demonstration and interactive function, also needs and can add up and analyze real-time power information, and the storage and management function of supported data.Display unit gets access to the real-time power information of each consumer in the family from computing unit, historical power information and the real-time power information of each equipment and whole family are upgraded, and these contents are saved in database or other media; On the other hand, display unit need to be realized the effective User Interface of a cover, comprises patterned display interface, and the form with power information can be understood to be converted into the user feeds back to the user; Simultaneously the user can be to the partial feedback of analyzing distortion to electricity consumption behavior sensory perceptual system, thereby improves the precision with post analysis.
Technological core of the present invention concentrates in the employed behavior perception algorithm of computing unit.Fig. 2-the 5th, the process flow diagram of behavior perception algorithm among the present invention.Algorithm flow can be divided into three phases (such as Fig. 2): feature set establishment stage, parameter learning stage, behavioural analysis stage.
The feature set establishment stage is the necessary stage of setting up various dimensions statistical nature model for each consumer.This stage begins elder generation according to the measurement capability of collecting unit, is set in needed characteristic set in the whole computation process.In this one-phase, electric current and voltage signal when at first gathering the individual equipment normal operation; Then the data that collect are converted to predefined various dimensions eigenwert; Set up corresponding statistical models according to these eigenwerts.For the equipment with a plurality of running statuses (such as: the larger electrical equipment of electricity consumption situation difference under the different conditions such as air-conditioning, micro-wave oven) can set up respectively statistical model for each running status.This stage is finished by equipment supplier or third party professional institution usually, in order to can access more accurate result.The eigenwert of each equipment that obtains in this stage will be stored into database, use in the stage of back.
The parameter learning stage is the weight that is used for each feature of balance for the consumer training.This stage should finish in subscriber household.The user normally uses the electrical equipment in the family in this one-phase, the electric current and the voltage signal that produce at the domestic electric network entrance during a plurality of equipment normal operation of system acquisition; Use initial weight (predefined default value) to carry out electrical energy consumption analysis, obtain the candidate list that a plurality of candidate result form; By with user's input ratio pair, iterative optimization weight parameter realizes that error rate minimizes training.Each feature weight that obtains in this stage will be stored into database, use in next stage.
The behavioural analysis stage is the Main Stage of carrying out equipment electricity consumption behavioural analysis.In this one-phase, at first need from database, to take out feature and the parameter thereof of all known devices in the family, and choose suitable statistical analysis technique according to the structure of characteristic set; Then gather electric current and the voltage signal of whole family in the electrical network porch, the data that need simultaneously to collect are converted to predefined eigenwert; In the situation of known total power information, by the computing method of heuristic search, infer in the current whole family the most possible electricity condition of using on each equipment; Obtain thus power information actual on each equipment, these communications carry out last processing to display unit the most at last.
The below introduces various dimensions electricity consumption behavior sensory perceptual system in detail, introduces respectively three Main Stage in the present embodiment in conjunction with process flow diagram:
The feature set establishment stage: at first need define the characteristic set of use when this stage begins, the characteristic set that uses in the present embodiment is the inherent feature of electrical equipment itself, so that feature can be reused in different families, comprising:
The harmonic current feature: the independent probability of each order harmonic current value that equipment produces distributes, and the value of the harmonic current of equipment is described;
Correlative character between each order harmonics: the joint probability distribution during the equipment operation between the different order harmonic currents is described the relation between the harmonic current of equipment;
Electricity is led feature: the electricity of use equipment when operation equipment is led probability distribution, and the characteristic of the physical circuit of device interior is described.
The equipment supplier finishes the feature of each equipment is measured in this stage.After the equipment supplier begins the feature set establishment stage, as shown in Figure 3 at first according to the measurement capability of collecting unit, the characteristic set that initialization need to be set up.Then according to the characteristic set of above-mentioned definition, dispose the collecting device of enough accuracy and finish building of test environment, Devices to test, deployed environment namely are installed.In this environment, operation Devices to test, the current and voltage data in the recording unit operational process.For the complex apparatus with different running statuses, need to the running status that it is all measure respectively.Therefore need to judge whether to exist not measure running status, do not measure running status if exist, then equipment is adjusted to steady operation under this running status, begin simultaneously image data, according to image data, the characteristic set of computing equipment repeats deterministic process, do not measure running status until do not exist, then preserve the characteristic set of all acquisitions.So far the feature set establishment stage finishes.
After measurement is finished, according to characterizing definition before, set out by original current and voltage data, calculate the statistical model of character pair.In the present embodiment, the process of calculating harmonic current feature is: the primary current data are carried out discrete Fourier transformation, obtain each harmonic.Harmonic wave to different orders uses respectively the fitting of distribution method to calculate its probability distribution.This distribution is exactly the foundation of calculating the harmonic current feature.Any time, the harmonic current feature was exactly the long-pending logarithm value of the parameter probability valuing of each operation electrical equipment.
Log-linear model is used for the statistical nature (h) with each dimension, and the weight parameter (λ) by feature combines, and finally obtains the most possible statistical probability with electricity condition, and its formula is as follows:
Wherein, t represents character representation total in the family, and c is a kind of electricity condition of using of all electrical equipment in the family.P (c|t) is illustrated in the situation of known t, and the electricity consumption situation just is the probability of c in the family; The h on right side
iBe the fundamental function of a dimension, λ
iH for correspondence
iWeight parameter, denominator is with the summation of the probability of possible state c ' all in the family in the situation of t.The physical significance of this formula is that the electricity consumption situation just is proportional to the linear sum of the logarithm value of each multidimensional characteristic in the family for the probability of particular state when overall electricity consumption situation, therefore is called log-linear model, namely sets up the process of statistical nature model.
The parameter learning stage (as shown in Figure 4): this one-phase in this stage need to be finished in subscriber household, the user at first needs to obtain from the equipment supplier statistical nature model of each consumer, it is imported sensory perceptual system, namely be written into the basic configuration situation of consumer in the family, be written into the statistical nature model under each state of each equipment.System will gather all devices in a period of time in the family and normally use the total current of (not necessarily working simultaneously) and voltage signal as training dataset, and the data that then need to collect are converted to the character representation of predefined various dimensions.Set out by initial weight, carry out weight training.System sets beginning from predefined general initial parameter, and by calculate the candidate state set of a plurality of optimums with similar searching method of behavioural analysis stage (being described in detail) in next stage, the set sizes of present embodiment setting is 100.To each constantly namely, by current weight, calculate the candidate state set that obtains 100 optimums.After calculating candidate collection first, submit to the user, allow it select the situation of an optimum.Afterwards, according to user's selection, choose the weight of certain feature and adjust, thereby so that the candidate result tabulation trends towards the truth of user's input.Namely begin the process of iterative optimization characteristic parameter.This process is chosen a feature at every turn, and it is independent variable that the parameter lambda of this feature is regarded as, by changing this parameter, can change the order of whole candidate collection, adjusts the value of parameter, so that the ordering of candidate collection is according to identical as much as possible with user's input.Choose successively afterwards each feature, carry out iteration optimization.Finish one take turns tuning after, analyze quality and whether improve, if improved, then return the weight training step, until quality no longer improves.All weight parameter of each equipment that will obtain at last will be stored into database, record current feature weight, and as the optimal weights in the family, to use in next stage, the parameter learning stage finishes.
The behavioural analysis stage (as shown in Figure 5): statistical nature model and the weight parameter thereof of from database, taking out all devices in the family; Open collecting device, obtain the real-time electricity consumption data in the family; Electricity consumption characteristic set according to setting up in advance turns to characteristic of correspondence with the data that collect and represents; In conjunction with feature and the weight information of each known equipment, calculate the most probable electricity consumption situation of current each equipment by didactic post search; The result is imported display unit into, finishes the post-processed such as visual, statistics; Judge whether to receive halt command, if do not receive, then repeat above-mentioned steps until receive halt command, the behavioural analysis stage finishes.In the stage, the character representation of known population calculates in the whole family electricity consumption situation most possible on each equipment by the method for heuristic post search.The concrete grammar of the method for heuristic post search is as follows: at first take the starting point of a running status as searching for of whole appliance system, this state can be chosen the last result phase of analyzing, and also can choose the state that all electrical equipment all cut out; At first calculate the value of the log-linear model of original state; Then from initial system state, calculate each and only have an electrical equipment to be in value corresponding to the running status of system state of different operating state with original state, only keep top n state (N is the constant that the post search is set).So calculate every one deck of search tree, until the characteristic model value in the new levels does not improve.In the state that all search, what the characteristic model value was the highest is exactly the most possible duty of each equipment.Electricity consumption situation by each equipment that obtains is derived more detailed power information, and final step is that these communications are carried out visualization processing and carry out statistical study to display unit.
The below is a specific embodiment of the present invention:
One. set up characteristic set, as follows:
The current harmonics feature: the independent probability of each order harmonic current value that equipment produces distributes;
The eigenwert of input is the harmonic current of all orders of all electrical equipment;
Be output as the logarithm value of a statistical probability, be the lucky therewith identical probability of eigenwert of input of all electrical equipment.
Formula is:
Function h wherein
1The computation model that is used for expression current harmonics feature, c
IjIt is the current value of the j subharmonic of i electrical equipment.Suppose total n operating electrical equipment, the m order harmonics.P
IjThe independent probability that is the j subharmonic of i electrical equipment distributes.Total feature value such as following formula.
Consider the current harmonics feature of correlativity between each order harmonics: the joint probability distribution during the equipment operation between the different order harmonic currents.
The eigenwert of input is the harmonic current of all orders of all electrical equipment;
Be output as the logarithm value of a statistical probability, be the lucky therewith identical probability of eigenwert of input of all electrical equipment.
Formula is:
Function h wherein
2The computation model that is used for expression current harmonics correlative character, c
IjIt is the current value of the j subharmonic of i electrical equipment.Suppose total n operating electrical equipment, the m order harmonics.P
iThat correlation probabilities on the current harmonics of i electrical equipment distributes.Total feature value such as following formula.
Electricity is led feature: the electricity of equipment is led probability distribution during the operation of use equipment, and basic definition is the same, and electric current is changed to electricity and leads
Electricity is led and is defined as follows:
The real part of admittance when the value that electricity is led is equipment work, the calculating of getting real part represents with function Re.Similar with top two kinds of characterizing definitions, electricity is led the probability distribution that its electricity is led when being characterized as the equipment operation, and wherein v (t) represents respectively t electric current and voltage constantly with i (t).
Input feature vector is that the electricity of all devices is led;
Be output as the logarithm value of a statistical probability, for the electricity that all electrical equipment are inputted is just therewith led identical probability.
Wherein, function h
3Be used for the computation model that the expression electricity is led feature, G
iIt is the electric conductivity value (being calculated by current/voltage) of i electrical equipment.P '
iThe electricity that is i electrical equipment is led distribution.
To each electrical equipment, set up the feature corresponding to each statistical model.
After determining the statistical nature that needs to measure, can begin each equipment is carried out the calculating of statistical distribution.
The installation testing environment obtains raw data.
Fig. 6 is statistical nature environment synoptic diagram to be measured, according to Fig. 6, suitable current/voltage sensor is installed between data acquisition system (DAS), Devices to test and civil power, builds experimental situation: the data of the collection of processing unit processes data acquisition system (DAS).Voltage and current data when afterwards, the electricity measurer normal operation treated in record.The data that gather this moment are the ifq circuit data, can represent respectively t electric current and voltage constantly with v (t) and i (t).
1) according to definition, counting statistics distributes.
According to the definition of three features, we at first are converted to harmonic wave with original circuit data and represent (frequency domain)
To moment T, with and subsequent the data of one-period carry out Fourier transform, can obtain electric current in this cycle that T begins constantly and the harmonic wave of voltage and represent have
Wherein, I
k(T) being the value of the harmonic current on k rank, is a complex value.N is the sampling number in the one-period, and n represents the number of electrical equipment.
The computing method of voltage are identical with electric current.
Electricity is led
Each cycle to the raw data that collects so calculates, and just can obtain a large amount of harmonic current data.
According to characterizing definition, calculate respectively the independent distribution probability P of harmonic current
Ij, the joint distribution probability P
i, and the electric distribution probability P ' that leads
i
Arrive this, finish even if the statistical model of equipment is set up
Two. in user's family, adjust weight parameter
The weight parameter adjusting stage at first needs to finish the installment work of analytic system in subscriber household when beginning.Fig. 7 is the electricity consumption behavior sensory perceptual system synoptic diagram of household internal power utilization environment.System installs collecting unit according to Fig. 7 in the electrical network porch of subscriber household and gets final product, and collecting unit is from electrical network Acquisition Circuit data (power information), and computing unit sends to display unit according to the result that described circuit data compute dependent data obtains.
The below simply introduces the impact of weight parameter on analyzing:
Use log-linear model for a plurality of features are combined us effectively, formula is as follows:
Wherein: h
iThe statistical nature of definition for it, and λ
iBe character pair h
iWeight parameter.The circuit information that t and c represent respectively to have recorded, the circuit state that analyzes by analytic system.In different features, bring different physical significance (meaning of t only just can show) into when the feature of existence condition probability.For example: to the current harmonics feature, the harmonic current of each order of all devices under this circuit state of c, t is nonsensical.In following formula, the effect of denominator is normalization, and c ' wherein represents all possible situation, when we will obtain most possible c, so that during P (c|t) maximization, can remove denominator, has:
Therefore only obtain so that each characteristic line and maximum state c are exactly the state that system most possibly is in.
And weight λ has just indicated the significance level between the feature.
For example: two different system state c1 and c2 are arranged; Its characteristic of correspondence value such as following table:
? | c1 | c2 |
h1 | 1 | 3 |
h2 | 3 | 1 |
h3 | 1 | 2 |
As: λ gets 1 entirely, and then c2 is more excellent; λ is that (1,0.5,0.5) then c1 is more excellent.Different parameters can obtain different optimal situation
Adjusting the task that the weight parameter stage will finish is exactly, with parameter adjustment to a suitable state, so that the optimal situation of calculating by log-linear model is identical as much as possible with the electricity consumption situation of reality.
After environmental structure is finished, formally begin the parameter learning stage.
1. initial weight is set, and initial weight can all get 1.
2. the circuit data of system acquisition long period comprises the data in a plurality of moment.
3. by the method (seeing the 4th part) of search, constantly calculate a tabulation for each, deposit so that the large as far as possible candidate state of P (c|t).
4. give the user with these tabulations, from each tabulation, select the candidate state the most similar to actual electricity consumption situation by the user.
5. begin iteration weight optimization algorithm, algorithm flow is as follows:
Input: a series of candidate state tabulations, the optimum state that each tabulation is corresponding, the value of each feature in the tabulation under each state, initial weight
Output: the weight after the optimization:
Algorithm:
1) choose successively each feature, carry out 2) step.
2) adjust this feature respective weights so that in all candidate lists, the optimum state that calculates and user input identical as much as possible.
Finish one take turns optimization after, according to new weight calculation candidate list.As in each tabulation optimum input identically with the user, or without any optimization, then withdraw from optimizing process.Otherwise, get back to 1) and step proceeds.
The below describes this process with an example:
Suppose, have three moment, three features have three states in each candidate list constantly.Three feature h1, h2, the value of h3 (shown in the table is the value of one group of hypothesis) as shown in the table:
The user inputs optimum state and is respectively: the candidate 1, and the candidate 5, and the candidate 7
Algorithm is by, λ
1=λ
2=λ
3=1 beginning
First round iteration:
Selected characteristic 1 is brought in the summation formula proper λ into
11 o'clock, constantly 1 and 3 can get optimum, we can suppose λ
1=2.
Selected characteristic 2 is brought in the summation formula, can obtain 2〉λ
21.5 o'clock, all constantly can obtain optimum.
We can suppose λ
2=1.75;
All all satisfy constantly to this.Optimizing process can stop.Final weight is respectively
λ
1=2;λ
2=1.75;λ
3=1
6. output optimized parameter, the parameter learning stage finishes.
Three. the electrical energy consumption analysis stage
After having finished model foundation and parameter training, system begins electrical energy consumption analysis.The hardware installation method in this stage is identical with the upper parameter training stage, and collecting unit is installed in the electrical network porch.
In this stage, analytic system is analyzed each circuit data constantly.Calculate respectively in the family optimum electricity consumption situation, in the square frame, the method can also be used for calculating optimum candidate list to computing method below.
Input: the characteristic model of each electrical equipment, the weight parameter of each model, total current and voltage data in a certain moment family.
Output: this is (or a plurality of) optimum state of all electrical equipment of whole family front yard constantly.
Algorithm flow:
With total current and voltage data according to characterizing definition, be converted to frequency domain, represent with the harmonic wave form.
2. set up three empty queue Closed, Tmp and Open, be respectively applied to record possible state, the state that buffer memory is to be launched, the state that record is launching.Set up a Hash table H, the state that record had launched
3. choose an original state S
0, the strategy of choosing has two kinds usually, keeps the optimum state in a upper moment, perhaps chooses full electrical equipment closed condition, calculates S
0Probability and with S
0Put into Tmp and H;
4. all elements among the formation Tmp is imported Open, empty Tmp;
5. from Open, choose a state S
i, calculate all states that are adjacent (adjacent definition is to only have the duty of equipment difference), the part with not appearing in these states among the Hash table H joins among Tmp and the H.With S
iMove to the Closed from Open
6. repeating step 5, until Open is empty.
7. be empty such as Tmp, then stop search, enter step 9.Otherwise to all states among the Tmp, calculate the value of each feature under each state, obtain final probability by log-linear model.According to the probability size, these states are sorted, only keep top n state (N is predefined search width).
8. if institute's stateful probability was all less than the optimal value among the Closed during Tmp showed.Then stop search, enter step 9.Otherwise get back to step 4, continue search.
All states among the Tmp are moved among the Closed, from Closed, select (a plurality of) state of maximum probability as the output of algorithm.
The below (only uses a kind of statistical nature: the current characteristic with correlativity with a simplified example in this example, current harmonics represents with a tri-vector, three dimensions is expressed as the coordinate system of xyz, electrical appliance state only has two kinds on switch in this example, F represents to close, T represents out), the compare analyzing process is demonstrated;
Supposing has three electrical equipment now, is respectively A, B, C.Feature on it is respectively:
The distribution of current of electrical equipment A is the even distribution on the x axle, is a class normal distribution on the yz plane, and its probability function can be expressed as, and a is constant:
The distribution of current of electrical equipment B is the even distribution on the y axle, is a class normal distribution on the xz plane, and its probability function can be expressed as, and b is constant:
The distribution of current of electrical equipment C is the even distribution on the z axle, is a class normal distribution on the xy plane, and its probability function can be expressed as, and c is constant:
Therefore the account form of feature is:
c
iIt is the current vector of i electrical equipment.Only has a feature, so log-linear model just equals the exponential quantity of this feature.
After simplifying, maximized function to be expressed as
The below begins search procedure, and a state in the process represents with three-dimensional boolean vector, such as (FTF).The width of supposing search is 2
1. calculate the total current harmonic wave and represent, suppose that the total current vector is (1,2,3)
2. initialize queue Open, Closed and Tmp, initialization Hash H
3. original state is chosen the state S that whole electrical equipment all cut out
0(FFF), obviously under total current is not 0 situation, S
0Probability is 0, and objective function is equivalent to; With S
0Add among Tmp and the H
4. with the whole states among the Tmp, namely S
0Join among the Open;
5. choose the state (FFF) among the Open, adjacent states have (TFF), (FTF), (FFT).Put into Tmp and H, (FFF) put into the Closed table
6. to all state computation objective functions among the Tmp, and keep the first two state
(TFF) obvious all electric currents all are that electrical equipment A produces under the state.So target function value is-13;
(FTF) obvious all electric currents all are that electrical equipment B produces under the state.So target function value is-10;
(FFT) obvious all electric currents all are that electrical equipment A produces under the state.So target function value is-5;
Keep at last (FTF), (FFT),
7. continue search, all states among the Tmp are added among the Open
8. choose the state (FTF) among the Open, adjacent having among the H (TTF), (FTT) of not appearing at simultaneously.Put into Tmp and H, (FTF) put into the Closed table
9. choose the state (FFT) among the Open, adjacently do not appear at simultaneously having among the H (TFT).Put into Tmp and H, (FFT) put into the Closed table
10. to all state computation objective functions among the Tmp, and keep the first two state
(TTF) electric current all is to be produced by electrical equipment A and electrical equipment B under the state, is respectively (1,0,1.5) according to characteristic formula A and the B stream that powers on that is easy to get, when (0,2,1.5) in maximum probability.Target function value is-4.5;
In like manner can get, (FTT) B and the C stream that powers on is respectively (0.5,2,0) under the state, when (0.5,0,3) in maximum probability.Target function value is-0.5;
In like manner can get, (TFT) A and the C stream that powers on is respectively (1,1,0) under the state, when (0,1,3) in maximum probability.Target function value is-1;
Keep at last (FTT), (TFT),
11. continue search, all states among the Tmp added among the Open
12. choose the state (FTT) among the Open, adjacently do not appear at simultaneously having among the H (TTT).Put into Tmp and H, (FTT) put into the Closed table
13. choose the state (TFT) among the Open, there is not the adjacent state that does not appear at simultaneously among the H, (TFT) put into the Closed table
14. to all state computation objective functions among the Tmp, and keep the first two state
(TTT) electric current is to be produced by electrical equipment A, electrical equipment B and electrical equipment C under the state, is respectively (1,0,0) according to characteristic formula A, B and the C stream that powers on that is easy to get, (0,2,0), when (0,0,3) in maximum probability.Target function value is 0;
15. continue search, all states among the Tmp added among the Open
16. choose the state (TTT) among the Open, there is not the adjacent state that does not appear at simultaneously among the H, (TTT) put into the Closed table
17.Tmp table is for empty, search stops, and to all state orderings in the Closed table, obtaining optimum state is (TTT), and suboptimum is (FTT)
Obviously this result and our understanding also are identical, and on three dimensions, when total current was (1,2,3), three electrical equipment all should be in open mode to the current vector of A, B, three electrical equipment of C naturally respectively.
Certainly; the present invention also can have other various embodiments; in the situation that does not deviate from spirit of the present invention and essence thereof; those of ordinary skill in the art work as can make according to the present invention various corresponding changes and distortion, but these corresponding changes and distortion all should belong to the protection domain of the appended claim of the present invention.
Claims (10)
1. the cognitive method of a various dimensions electricity consumption behavior is characterized in that, comprising:
Step 1, pre-defined under a plurality of dimensions consumer statistical nature and obtain every kind of eigenwert under the statistical nature, and set up corresponding statistical model;
Step 2 is installed collecting unit in the electrical network porch, the current-voltage information of all consumers of Real-time Obtaining, and computing unit is trained the feature weight of each statistical nature according to described current-voltage information and user's input data;
Step 3, described computing unit carries out heuristic search calculating according to described eigenwert, statistical model and feature weight, obtains the real-time electricity consumption behavior of each consumer, and described real-time electricity consumption behavior is sent to display unit.
2. the cognitive method of various dimensions electricity consumption behavior as claimed in claim 1, it is characterized in that, described collecting unit comprises a current sensor and a voltage sensor, this current sensor is based on the magnetic field induction chip of Hall effect, utilize electromagnetic induction to realize the current data collection of non-intrusion type, the direct place in circuit of this voltage sensor is in parallel with all consumers, measures the voltage on it.
3. the cognitive method of various dimensions electricity consumption behavior as claimed in claim 1 is characterized in that, described computing unit independently uses a processor; Perhaps share a processor with collecting unit; Perhaps share a processor with display unit.
4. the cognitive method of various dimensions electricity consumption behavior as claimed in claim 1 is characterized in that, described step 2 comprises:
Step 21, the total current information of voltage of all consumers begins to carry out weight training as training dataset in the family in measurement a period of time;
Step 22 is calculated by described heuristic search, obtains the candidate list that a plurality of candidate result form;
Step 23, the user chooses optimum analysis result according to truth in candidate list;
Step 24 is carried out iterative optimization weight parameter by the information that described candidate list and user are provided, and obtains the feature weight of described all consumers.
5. the cognitive method of various dimensions electricity consumption behavior as claimed in claim 1 is characterized in that, the heuristic computing formula in the described step 3 is:
Wherein, t represents character representation total in the family, and c is a kind of electricity condition of using of all electrical equipment in the family, and P (c|t) is illustrated in the situation of known t, and the electricity consumption situation just is the probability of c in the family; The h on right side
iBe the fundamental function of a dimension, λ
iH for correspondence
iWeight parameter, denominator is with the summation of the probability of possible state c ' all in the family in the situation of t.
6. the sensory perceptual system of a various dimensions electricity consumption behavior is characterized in that, comprising:
Pretreatment module, pre-defined statistical nature and the eigenwert of obtaining under every kind of statistical nature at consumer under a plurality of dimensions, and set up corresponding statistical model;
Training module is installed collecting unit in the electrical network porch, and the current-voltage information of all consumers of Real-time Obtaining, computing unit are trained the feature weight of each statistical nature according to described current-voltage information and user's input Data Comparison;
Computing module, described computing unit carry out heuristic search according to described eigenwert, statistical model and feature weight and calculate, and obtain the real-time electricity consumption behavior of each consumer, and described real-time electricity consumption behavior is sent to display unit.
7. the sensory perceptual system of various dimensions electricity consumption behavior as claimed in claim 6, it is characterized in that, described collecting unit comprises a current sensor and a voltage sensor, this current sensor is based on the magnetic field induction chip of Hall effect, utilize electromagnetic induction to realize the current data collection of non-intrusion type, the direct place in circuit of this voltage sensor is in parallel with all consumers, measures the voltage on it.
8. the sensory perceptual system of various dimensions electricity consumption behavior as claimed in claim 6 is characterized in that, described computing unit independently uses a processor; Perhaps share a processor with collecting unit; Perhaps share a processor with display unit.
9. the sensory perceptual system of various dimensions electricity consumption behavior as claimed in claim 6 is characterized in that, described training module comprises:
The weight training preparation module, the current-voltage information of each consumer in measurement a period of time begins to carry out weight training as training dataset;
The candidate collection acquisition module calculates by described heuristic search, obtains the candidate list that a plurality of candidate result form;
Choose module, the user chooses optimum analysis result according to truth in candidate list;
Weight obtains module, by described candidate list and the information that the user provides being carried out iterative optimization weight parameter, the feature weight of described all consumers of acquisition.
10. the sensory perceptual system of various dimensions electricity consumption behavior as claimed in claim 6 is characterized in that, the heuristic computing formula in the described computing module is:
Wherein, t represents character representation total in the family, and c is a kind of electricity condition of using of all electrical equipment in the family, and P (c|t) is illustrated in the situation of known t, and the electricity consumption situation just is the probability of c in the family; The h on right side
iBe the fundamental function of a dimension, λ
iH for correspondence
iWeight parameter, denominator is with the summation of the probability of possible state c ' all in the family in the situation of t.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210555210.8A CN103020459B (en) | 2012-12-19 | 2012-12-19 | A kind of cognitive method of various dimensions electricity consumption behavior and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210555210.8A CN103020459B (en) | 2012-12-19 | 2012-12-19 | A kind of cognitive method of various dimensions electricity consumption behavior and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103020459A true CN103020459A (en) | 2013-04-03 |
CN103020459B CN103020459B (en) | 2015-12-09 |
Family
ID=47969057
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210555210.8A Active CN103020459B (en) | 2012-12-19 | 2012-12-19 | A kind of cognitive method of various dimensions electricity consumption behavior and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103020459B (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103985063A (en) * | 2014-05-16 | 2014-08-13 | 苏州鸣伦电子科技有限公司 | Imaging feedback method of power demand side consumption state |
CN105320834A (en) * | 2014-07-29 | 2016-02-10 | 财团法人交大思源基金会 | Method for calculating number of people based on using state of electric appliance and monitoring system thereof |
CN105652118A (en) * | 2015-12-29 | 2016-06-08 | 国家电网公司 | Load instantaneous energy feature-based power grid electric energy load monitoring method |
CN106569026A (en) * | 2015-10-09 | 2017-04-19 | 腾讯科技(深圳)有限公司 | Power consumption device power consumption statistics method and system |
CN106789163A (en) * | 2016-11-25 | 2017-05-31 | 华电智连信达科技(北京)有限公司 | A kind of network equipment power information monitoring method, device and system |
CN108090647A (en) * | 2016-11-23 | 2018-05-29 | 财团法人资讯工业策进会 | Method for obtaining load operation probability of power consumers and method for obtaining load operation probability of power consumer group |
CN108492134A (en) * | 2018-03-07 | 2018-09-04 | 国网四川省电力公司 | The big data user power utilization behavior analysis system integrated based on multicycle regression tree |
CN109299082A (en) * | 2018-09-13 | 2019-02-01 | 北京中科闻歌科技股份有限公司 | A kind of big data analysis method and system |
CN109782086A (en) * | 2018-12-25 | 2019-05-21 | 武汉中原电子信息有限公司 | A kind of non-intruding load recognition methods based on the analysis of various dimensions signal |
CN109813978A (en) * | 2018-12-25 | 2019-05-28 | 武汉中原电子信息有限公司 | A kind of non-intruding load-type recognition methods of variation characteristic between comprehensive transient characteristic and stable state |
CN110516743A (en) * | 2019-08-28 | 2019-11-29 | 珠海格力智能装备有限公司 | Identification method and device of electric equipment, storage medium and processor |
CN110995543A (en) * | 2019-12-18 | 2020-04-10 | 云南大学 | Non-invasive method for monitoring abnormal internet surfing behavior of minors |
CN113311740A (en) * | 2021-03-25 | 2021-08-27 | 珠海荣邦智能科技有限公司 | Control method and device for electrical equipment |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102323469A (en) * | 2011-07-27 | 2012-01-18 | 四川大学 | System for monitoring state of harmonic load |
-
2012
- 2012-12-19 CN CN201210555210.8A patent/CN103020459B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102323469A (en) * | 2011-07-27 | 2012-01-18 | 四川大学 | System for monitoring state of harmonic load |
Non-Patent Citations (2)
Title |
---|
赵春柳: "基于稳态谐波分析的非侵入式电力负荷分解法的应用研究", 《中国优秀硕士学位论文全文数据库-工程科技II辑》, vol. 2011, no. 2, 15 December 2011 (2011-12-15), pages 042 - 535 * |
黎鹏: "非侵入式电力负荷分解与监测", 《中国博士学位论文全文数据库-工程科技II辑》, vol. 2010, no. 12, 15 December 2010 (2010-12-15) * |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103985063A (en) * | 2014-05-16 | 2014-08-13 | 苏州鸣伦电子科技有限公司 | Imaging feedback method of power demand side consumption state |
CN105320834B (en) * | 2014-07-29 | 2018-11-02 | 财团法人交大思源基金会 | Method for calculating number of people based on using state of electric appliance and monitoring system thereof |
CN105320834A (en) * | 2014-07-29 | 2016-02-10 | 财团法人交大思源基金会 | Method for calculating number of people based on using state of electric appliance and monitoring system thereof |
CN106569026A (en) * | 2015-10-09 | 2017-04-19 | 腾讯科技(深圳)有限公司 | Power consumption device power consumption statistics method and system |
CN105652118B (en) * | 2015-12-29 | 2019-02-26 | 国家电网公司 | A kind of power grid power budget monitoring method based on load instantaneous energy feature |
CN105652118A (en) * | 2015-12-29 | 2016-06-08 | 国家电网公司 | Load instantaneous energy feature-based power grid electric energy load monitoring method |
CN108090647A (en) * | 2016-11-23 | 2018-05-29 | 财团法人资讯工业策进会 | Method for obtaining load operation probability of power consumers and method for obtaining load operation probability of power consumer group |
CN108090647B (en) * | 2016-11-23 | 2021-10-01 | 财团法人资讯工业策进会 | Method for obtaining load operation probability of power consumers and method for obtaining load operation probability of power consumer group |
CN106789163A (en) * | 2016-11-25 | 2017-05-31 | 华电智连信达科技(北京)有限公司 | A kind of network equipment power information monitoring method, device and system |
CN106789163B (en) * | 2016-11-25 | 2020-08-18 | 华电智连信达科技(北京)有限公司 | Method, device and system for monitoring power utilization information of network equipment |
CN108492134A (en) * | 2018-03-07 | 2018-09-04 | 国网四川省电力公司 | The big data user power utilization behavior analysis system integrated based on multicycle regression tree |
CN109299082A (en) * | 2018-09-13 | 2019-02-01 | 北京中科闻歌科技股份有限公司 | A kind of big data analysis method and system |
CN109299082B (en) * | 2018-09-13 | 2021-09-14 | 北京中科闻歌科技股份有限公司 | Big data analysis method and system |
CN109782086A (en) * | 2018-12-25 | 2019-05-21 | 武汉中原电子信息有限公司 | A kind of non-intruding load recognition methods based on the analysis of various dimensions signal |
CN109782086B (en) * | 2018-12-25 | 2020-12-18 | 武汉中原电子信息有限公司 | Non-intrusive load identification method based on multi-dimensional signal analysis |
CN109813978B (en) * | 2018-12-25 | 2021-04-20 | 武汉中原电子信息有限公司 | Non-intrusive load type identification method integrating transient characteristic and change characteristic between stable states |
CN109813978A (en) * | 2018-12-25 | 2019-05-28 | 武汉中原电子信息有限公司 | A kind of non-intruding load-type recognition methods of variation characteristic between comprehensive transient characteristic and stable state |
CN110516743A (en) * | 2019-08-28 | 2019-11-29 | 珠海格力智能装备有限公司 | Identification method and device of electric equipment, storage medium and processor |
CN110995543A (en) * | 2019-12-18 | 2020-04-10 | 云南大学 | Non-invasive method for monitoring abnormal internet surfing behavior of minors |
CN110995543B (en) * | 2019-12-18 | 2022-11-01 | 云南大学 | Non-invasive method for monitoring abnormal internet surfing behavior of minors |
CN113311740A (en) * | 2021-03-25 | 2021-08-27 | 珠海荣邦智能科技有限公司 | Control method and device for electrical equipment |
Also Published As
Publication number | Publication date |
---|---|
CN103020459B (en) | 2015-12-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103020459B (en) | A kind of cognitive method of various dimensions electricity consumption behavior and system | |
Angelis et al. | NILM applications: Literature review of learning approaches, recent developments and challenges | |
CN110097297B (en) | Multi-dimensional electricity stealing situation intelligent sensing method, system, equipment and medium | |
CN109145949A (en) | Non-intrusive electrical load monitoring and decomposition method and system based on integrated study | |
Choksi et al. | Feature based clustering technique for investigation of domestic load profiles and probabilistic variation assessment: Smart meter dataset | |
Li et al. | Energy data generation with wasserstein deep convolutional generative adversarial networks | |
CN104966161B (en) | A kind of power quality recorder data calculation and analysis methods based on gauss hybrid models | |
Verdolini et al. | The effects of expert selection, elicitation design, and R&D assumptions on experts' estimates of the future costs of photovoltaics | |
CN103268519A (en) | Electric power system short-term load forecast method and device based on improved Lyapunov exponent | |
Luo et al. | Service recommendation in smart grid: vision, technologies, and applications | |
Cortes-Robles et al. | Fast-training feedforward neural network for multi-scale power quality monitoring in power systems with distributed generation sources | |
Kong et al. | Refined peak shaving potential assessment and differentiated decision-making method for user load in virtual power plants | |
CN109840691A (en) | Non-intrusion type subitem electricity estimation method based on deep neural network | |
Zhao et al. | Short-term microgrid load probability density forecasting method based on k-means-deep learning quantile regression | |
CN103018611A (en) | Non-invasive load monitoring method and system based on current decomposition | |
Blanco et al. | Stochastic harmonic load model of residential users based on measurements | |
CN112614004A (en) | Method and device for processing power utilization information | |
Brusokas et al. | HeatFlex: Machine learning based data-driven flexibility prediction for individual heat pumps | |
Awasthi et al. | Smart Grid Sensor Monitoring Based on Deep Learning Technique with Control System Management in Fault Detection | |
CN106295877B (en) | Method for predicting electric energy consumption of smart power grid | |
Patidar et al. | A hybrid system of data-driven approaches for simulating residential energy demand profiles | |
Bianchi et al. | Load forecasting in district heating networks: Model comparison on a real-world case study | |
Elombo et al. | Residential load variability and diversity at different sampling time and aggregation scales | |
Chou et al. | Simulating long-term energy consumption prediction in campus buildings through enhanced data augmentation and metaheuristic-optimized artificial intelligence | |
CN116384622A (en) | Carbon emission monitoring method and device based on electric power big data |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |