CN112348380A - Demand response schedulable capacity probability prediction method and device and electronic equipment - Google Patents
Demand response schedulable capacity probability prediction method and device and electronic equipment Download PDFInfo
- Publication number
- CN112348380A CN112348380A CN202011262389.9A CN202011262389A CN112348380A CN 112348380 A CN112348380 A CN 112348380A CN 202011262389 A CN202011262389 A CN 202011262389A CN 112348380 A CN112348380 A CN 112348380A
- Authority
- CN
- China
- Prior art keywords
- capacity
- response
- sharing
- demand response
- time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000004044 response Effects 0.000 title claims abstract description 251
- 238000000034 method Methods 0.000 title claims abstract description 31
- 230000002776 aggregation Effects 0.000 claims abstract description 73
- 238000004220 aggregation Methods 0.000 claims abstract description 73
- 238000006116 polymerization reaction Methods 0.000 claims abstract description 29
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 28
- 238000000605 extraction Methods 0.000 claims abstract description 23
- 230000006870 function Effects 0.000 claims description 20
- 238000012549 training Methods 0.000 claims description 17
- 238000005457 optimization Methods 0.000 claims description 14
- 230000008901 benefit Effects 0.000 claims description 12
- 238000005315 distribution function Methods 0.000 claims description 11
- 101100455978 Arabidopsis thaliana MAM1 gene Proteins 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 7
- 238000012216 screening Methods 0.000 claims description 7
- 238000011478 gradient descent method Methods 0.000 claims description 6
- 230000001186 cumulative effect Effects 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000005611 electricity Effects 0.000 description 17
- 238000012360 testing method Methods 0.000 description 15
- 238000010586 diagram Methods 0.000 description 9
- 230000000694 effects Effects 0.000 description 4
- 238000003062 neural network model Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000009826 distribution Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 230000001010 compromised effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06315—Needs-based resource requirements planning or analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Data Mining & Analysis (AREA)
- Marketing (AREA)
- Evolutionary Computation (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Artificial Intelligence (AREA)
- General Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Game Theory and Decision Science (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Educational Administration (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
Abstract
The invention provides a demand response schedulable capacity probability prediction method, a demand response schedulable capacity probability prediction device and electronic equipment, wherein the demand response schedulable capacity probability prediction method comprises the following steps: acquiring demand response data of each user subordinate to the load aggregation provider on a historical demand response day; calculating time-sharing aggregation response capacity according to the demand response data; performing feature extraction on the time-sharing polymerization response capacity to obtain a feature parameter of the sample time-sharing polymerization response capacity; taking the sample time-sharing aggregation response capacity characteristic parameter as input, taking the time-sharing aggregation response capacity as output, and performing schedulable capacity estimation according to the characteristic parameter by using a meta-learning algorithm; and performing probability prediction of the demand response schedulable capacity of the load aggregator by using the nonparametric kernel density probability prediction model. And through a meta-learning algorithm and nonparametric kernel density estimation, the small sample probability prediction of the demand response schedulable capacity for the load aggregation provider is realized.
Description
Technical Field
The invention relates to the field of data processing, in particular to a demand response schedulable capacity probability prediction method and device and electronic equipment.
Background
From 2012 onwards, countries have come out relevant notices and opinions about power demand side management, and have emphasized the improvement of the supply and demand balance guarantee level mainly based on demand side management, and gradually form demand side maneuvering peak shaving capacity which occupies about 3% of the maximum electric load. The demand response utilizes price signals and incentive mechanisms to guide users to reduce or transfer loads during peak load periods, relieves the situation of power supply resource shortage, and promotes the flexibility of power system operation. The demand response can prompt the user to actively participate in the load regulation of the power system, so that the power is reasonably used, and the operation pressure of the power system in the peak period of power utilization is reduced, thereby ensuring that the power system can operate more safely, stably and efficiently.
For a single resident user, the load transfer capacity is small, the power utilization rule is uncertain, direct power trading with a system operator is difficult, and the load peak regulation of a power grid is realized. The load aggregator is used as an intermediary between the resident users and the system operator, bears the key task of integrating the resources on the demand side, and can make a bidding strategy with the goal of maximizing the self-market trading benefits. In order to maximize the benefit of marketing trading, the load aggregator needs to accurately predict the aggregate demand response capacity of the user. Meanwhile, due to the fluctuation of the electricity consumption of the user, the fact that the actual demand response of the user does not meet the requirement of the strategy may occur when the strategy is formulated only by the prediction and determination value of the aggregate demand response capacity, and a load aggregator is subjected to economic punishment; or the actual demand response of the user is excessive to meet the policy requirement, the profit of the load aggregator is reduced. Compared with point prediction, probability prediction can predict expected values at a certain time in the future and also can obtain probability distribution information of the expected values, so that more comprehensive reference information is provided for decision making of a load aggregator. The load aggregator can reasonably make a bidding strategy according to the probability prediction information, so that the load aggregator can maximally ensure that the load aggregator does not excessively meet the strategy requirement and is not subjected to economic punishment, and the decision risk is reduced.
Therefore, the demand response schedulable capacity probability prediction for the load aggregator is more beneficial for the load aggregator to formulate an optimal bidding strategy, reduce risks and maximize market trading benefits.
At present, the probability prediction of the demand response schedulable capacity of the load aggregation business at home and abroad is less researched, and meanwhile, because the demand response implementation years and regions are less, the demand response data samples are less, and the traditional machine learning algorithm is difficult to obtain higher prediction accuracy under the condition of less samples. Therefore, the method for predicting the small sample probability of the demand response schedulable capacity for the load aggregation provider is significant.
Disclosure of Invention
The embodiment of the invention provides a demand response schedulable capacity probability prediction method, which can perform small sample probability prediction of demand response schedulable capacity for a load aggregator.
In a first aspect, an embodiment of the present invention provides a demand response schedulable capacity probability prediction method, including:
acquiring demand response data of each user subordinate to the load aggregation provider on a historical demand response day;
calculating time-sharing aggregation response capacity according to the demand response data;
performing feature extraction on the time-sharing polymerization response capacity to obtain a feature parameter of the sample time-sharing polymerization response capacity;
taking the sample time-sharing aggregation response capacity characteristic parameter as input, taking the time-sharing aggregation response capacity as output, and performing schedulable capacity estimation according to the characteristic parameter by using a meta-learning algorithm;
and performing probability prediction of the demand response schedulable capacity of the load aggregator by using the nonparametric kernel density probability prediction model.
Optionally, the step of calculating the time-sharing aggregate response capacity according to the demand response data specifically includes:
and accumulating all the user demand responses in different demand response time periods to obtain time-sharing aggregate response capacity.
Optionally, the step of performing feature extraction on the time-sharing aggregation response capacity to obtain the feature parameter of the sample time-sharing aggregation response capacity specifically includes:
performing feature extraction on the time-sharing polymerization response capacity;
and screening the extracted features according to the maximum information coefficient to obtain the feature parameters of the sample time-sharing polymerization response capacity.
Optionally, the extracted features include: day-by-day characteristics, time-by-time characteristics, cumulative benefit characteristics.
Optionally, the step of performing schedulable capacity estimation by using a meta-learning algorithm specifically includes:
classifying the sample time-sharing polymerization response capacity according to a preset condition;
grouping the sample time-sharing aggregation response capacity according to the type of a training model;
and using the model-independent meta-learning MAML as a meta-learning algorithm to carry out parameter optimization.
Optionally, the step of performing parameter optimization by using the model-independent meta learning MAML as a meta learning algorithm specifically includes:
and performing parameter optimization on the target parameter by a random gradient descent method according to the initial parameters of the model, the internal learning rate and the task based on the sample time-sharing aggregation response capacity.
Optionally, the specific step of performing probability prediction on the demand response schedulable capacity of the load aggregator by using the non-parametric kernel density probability prediction model includes:
obtaining a probability density function according to the sample quantity of the sample time-sharing aggregation response capacity, the kernel function, the demand response schedulable capacity prediction error and the sample bandwidth;
obtaining a probability distribution function based on the probability density function;
and performing probability prediction of the demand response schedulable capacity of the load aggregator through the probability distribution function.
In a second aspect, an embodiment of the present invention further provides a device for predicting probability of demand response schedulable capacity, where the device includes:
the acquisition module is used for acquiring the demand response data of each user subordinate to the load aggregation provider on the historical demand response day;
the calculation module is used for calculating time-sharing aggregation response capacity according to the demand response data;
the characteristic extraction module is used for carrying out characteristic extraction on the time-sharing polymerization response capacity to obtain a characteristic parameter of the sample time-sharing polymerization response capacity;
the processing module is used for taking the sample time-sharing aggregation response capacity characteristic parameter as input, taking the time-sharing aggregation response capacity as output, and carrying out schedulable capacity estimation according to the characteristic parameter by utilizing a meta-learning algorithm;
and the prediction module is used for performing probability prediction on the demand response schedulable capacity of the load aggregator by using the nonparametric kernel density probability prediction model.
In a third aspect, an embodiment of the present invention provides an electronic device, including: the system comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the computer program to realize the steps of the demand response schedulable capacity probability prediction method provided by the embodiment of the invention.
In the embodiment of the invention, the demand response data of each user subordinate to the load aggregator on the historical demand response date is obtained; calculating time-sharing aggregation response capacity according to the demand response data; performing feature extraction on the time-sharing polymerization response capacity to obtain a feature parameter of the sample time-sharing polymerization response capacity; taking the sample time-sharing aggregation response capacity characteristic parameter as input, taking the time-sharing aggregation response capacity as output, and performing schedulable capacity estimation according to the characteristic parameter by using a meta-learning algorithm; and performing probability prediction of the demand response schedulable capacity of the load aggregator by using the nonparametric kernel density probability prediction model. And through meta-learning algorithm and non-parameter kernel density estimation, small sample probability prediction of demand response schedulable capacity for the load aggregation provider is realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of a demand response schedulable capacity probability prediction method according to an embodiment of the present invention;
fig. 2 is a flowchart of time-sharing aggregation response capacity feature parameter extraction and screening provided by an embodiment of the present invention;
FIG. 3 is a diagram of a model architecture according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a training model and a testing model provided by an embodiment of the invention;
FIG. 5 is a schematic diagram illustrating a probability prediction result of a demand response schedulable capacity according to an embodiment of the present invention;
FIG. 6 is a schematic structural diagram of a demand response schedulable capacity probability prediction apparatus according to an embodiment of the present invention;
FIG. 7 is a block diagram of a processing module according to an embodiment of the present invention;
FIG. 8 is a block diagram of a prediction module according to an embodiment of the present invention;
fig. 9 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, fig. 1 is a flowchart of a demand response schedulable capacity probability prediction method according to an embodiment of the present invention, as shown in fig. 1, including the following steps:
and S1, acquiring the demand response data of each user subordinate to the load aggregation provider on the historical demand response day.
In the embodiment of the present invention, the demand response means that the power consumer temporarily changes its own power usage pattern according to dynamic changes of power price and power policy, so as to reduce or shift power consumption load for a certain period of time and respond to power supply, thereby ensuring the stability of the power grid system. The above-mentioned demand response day refers to a demand response implementation day, and the above-mentioned historical demand response day refers to a demand response day extracted from the demand response data.
The demand responses described above can be classified into price-based demand responses and incentive-based demand responses. The price-based demand response refers to that a user adjusts the power demand accordingly according to the received power price signal, and specifically may be time-of-use power price response, real-time power price response, spike power price response, and the like. The incentive-based demand response refers to that a user actively reduces the power demand when the system needs to obtain compensation, and specifically can be direct load control, interruptible load, demand-side bidding and emergency power demand response.
The time-of-use electricity price response can be understood as that the fixed electricity price is converted into different price mechanisms in different time periods, for example, the electricity price is reduced in the low ebb of electricity consumption, and the electricity price is increased in the peak of electricity consumption.
The real-time electricity rate response has an electricity rate update period faster than the minute electricity rate, for example, the electricity rate update period is one hour or less, and since the minute electricity rate response cannot cope with the short-term capacity shortage, the real-time electricity rate response may be employed in the case of the short-term capacity shortage.
The peak electricity price response can preset the price at the peak time of electricity utilization, and inform the user in advance for a certain time, thereby having the effect of resisting the sudden peak time of electricity utilization.
The direct load control can be that the load aggregation business remotely controls the on and off of the user equipment to avoid the peak of electricity utilization and inform the user equipment in advance.
The interruptible load can be remotely controlled by the load aggregator to turn on and off the user equipment under the condition of obtaining user consent so as to avoid power utilization peak and inform the user equipment in advance.
The demand-side bidding may be the active participation of the load aggregator in market competition in the form of bidding.
The emergency power demand response described above may be when power system stability is compromised, with the load aggregator providing compensation for the customer to reduce the load, the customer voluntarily electing to participate or give up.
And S2, calculating time-sharing aggregation response capacity according to the demand response data.
In the embodiment of the present invention, the demand response data includes a response capacity of a user, and the time-sharing aggregated response capacity refers to all response capacities in a preset time period. Specifically, the time-sharing polymerization response capacity can be calculated by the following formula:
wherein, the above
For T period aggregate response capacity, f abovei,k,TDemand response capacity for the ith user during the T time period on the kth day.
And S3, performing feature extraction on the time-sharing polymerization response capacity to obtain the feature parameters of the sample time-sharing polymerization response capacity.
In the embodiment of the present invention, the feature extraction is performed on the time-sharing aggregation response capacity, and the extracted feature is a feature that affects the user demand response. And obtaining the characteristic parameters of the sample time-sharing aggregation response capacity according to the characteristics influencing the user demand response.
Specifically, the characteristics affecting the user demand response include a day-by-day characteristic, a time-by-time characteristic and an accumulated benefit characteristic. The above-described features may affect the demand response of the user.
Wherein, the day-by-day characteristics may include: predicting maximum temperature, minimum temperature, maximum humidity, minimum humidity, rainfall, seasonal tags, weekday tags, double holiday tags, incentive money.
The time-by-time features described above may include: the system comprises a sensible temperature, a temperature-humidity index, a cold-humidity index, a human body comfort index, a demand response moment temperature, humidity, wind speed, demand response baseline load power consumption and a moment label.
The cumulative benefit feature may include: and accumulating the highest temperature benefit and the lowest temperature benefit.
Optionally, referring to fig. 2, the step S3 specifically includes the following steps:
and S31, performing feature extraction on the time-sharing aggregation response capacity.
In the embodiment of the present invention, the above-mentioned features may include a day-by-day feature, a time-by-time feature, and an accumulated benefit feature. The above-described features may affect the demand response of the user.
And S32, screening the extracted features according to the maximum information coefficient to obtain the feature parameters of the sample time-sharing aggregation response capacity.
In the embodiment of the present invention, the extracted features may be filtered by a maximum information coefficient calculation formula, which is shown as follows:
wherein, p (x, y) is the joint probability density of x and y, p (x) and p (y) are the edge probability distribution density of x and y, respectively, x and y are two variables with different characteristics, and a and b are the number of the divided grids in the x and y directions, respectively.
Specifically, the maximum information coefficient calculation formula may be understood as that features are gridded according to a preset grid condition, a maximum mutual information value is obtained, the maximum mutual information value is normalized, and a large value of mutual information at different scales is selected as the maximum information coefficient.
In the embodiment of the invention, the extracted features are screened, so that the features with smaller information coefficients can be reduced, and the data volume of the sample time-sharing aggregation response capacity feature parameters is further reduced.
And S4, taking the sample time-sharing aggregation response capacity characteristic parameter as an input, taking the time-sharing aggregation response capacity as an output, and estimating schedulable capacity according to the characteristic parameter by using a meta-learning algorithm.
In the embodiment of the invention, in order to adapt to the meta-learning framework, the sample time-sharing aggregation response capacity is classified according to days, namely
And divided into two groups, p and q. p groups as training models for meta learning, i.e. DtrainQ groups as test models, i.e. Dtest. The training model and the testing model in the meta-learning model both comprise a training set and a testing set. For training models in meta-learning, its training set
Including N classes of samples in the p groups. For each selected category, randomly selecting K examples as training data; test set thereof
N classes are also sampled, each of which randomly selects K instances as test data. Similarly, for a test model in meta-learning, its training set and test set include N classes sampled in q groups, again for each selected classK examples were randomly selected as training/test data.
The meta-learning algorithm is a general algorithm capable of solving various types of tasks, and particularly, the meta-learning algorithm learns a rule for initializing a parameter, and the initialized parameter theta has high sensitivity (namely, the gradient direction is vertical) to an optimal parameter solution theta 1,2,. n of each task in a parameter space, so that the optimal point can be quickly reached along the gradient direction.
The task described above can be understood as a group of samples (i.e., p-group or q-group) that contains N classes, each of which randomly selects K examples as training/testing data. Or taking K training samples and K' testing samples of each type.
Specifically, the meta-learning MAML irrelevant to the model can be used as a meta-learning algorithm to carry out parameter optimization, and the target parameters are subjected to parameter optimization through a random gradient descent method according to the initial parameters, the internal learning rate and the task based on the sample time-sharing aggregation response capacity of the model. For a parametric model f with a parameter thetaθThe parameter updating optimizing formula is as follows:
wherein theta is an initial parameter of the model, theta'iFor the updated parameters, α is the internal learning rate,
is the ith task.
Further, the updated parameter θ 'can be subjected to a random gradient descent method'iPerforming meta-optimization, wherein the formula is as follows:
wherein theta is an initial parameter of the model, theta'iFor the updated parameters, α is the internal learning rate,
for the ith task, β is the meta-learning rate.
And S5, utilizing the nonparametric kernel density probability prediction model to perform probability prediction of the demand response schedulable capacity of the load aggregator.
In the embodiment of the invention, a probability density function can be obtained according to the sample number of the sample time-sharing aggregation response capacity, the kernel function, the demand response schedulable capacity prediction error and the sample bandwidth; obtaining a probability distribution function based on the probability density function; and performing probability prediction of the demand response schedulable capacity of the load aggregator through the probability distribution function.
Specifically, the probability density function obtained by using the non-parametric kernel density estimation can be represented by the following formula:
wherein N is the number of samples; h is the sample bandwidth; k (-) is a kernel function; e.g. of the typeiThe capacity prediction error may be scheduled for the demand response of the ith sample.
The kernel function K (·) described above mainly includes a gaussian kernel, a uniform kernel, a triangular kernel, and an Epanechnikov function.
Further, for probability density function
And (3) integrating to obtain a probability distribution function F (x), wherein gamma is given arbitrarily (gamma is more than 0 and less than 1), and under the confidence coefficient of 1-gamma, the prediction interval of the demand response schedulable capacity is as follows:
wherein,
an aggregate response capacity prediction value;
F-1(y) an inverse function P { x ≦ F of F (x)-1(y) y is γ1、γ2。
In the embodiment of the invention, the demand response data of each user subordinate to the load aggregator on the historical demand response date is obtained; calculating time-sharing aggregation response capacity according to the demand response data; performing feature extraction on the time-sharing polymerization response capacity to obtain a feature parameter of the sample time-sharing polymerization response capacity; taking the characteristic parameter of the sample time-sharing aggregation response capacity as input, taking the time-sharing aggregation response capacity as output, and estimating schedulable capacity according to the characteristic parameter by using a meta-learning algorithm; and performing probability prediction of the demand response schedulable capacity of the load aggregator by using the nonparametric kernel density probability prediction model. And through meta-learning algorithm and non-parameter kernel density estimation, small sample probability prediction of demand response schedulable capacity for the load aggregation provider is realized.
Optionally, the characteristic parameters affecting the user demand response are used as the input of the neural network model, the demand response quantity is used as the output of the neural network model, parameters are updated through the meta-learning MAML irrelevant to the model, and then the non-parametric kernel density model is used for probability prediction, as shown in fig. 3.
It should be noted that both the training data and the test data include training samples and test samples, as shown in fig. 4.
In the embodiment of the invention, in order to specifically explain the effects of the embodiment of the invention, the collected data is used for verification. In the embodiment of the invention, the characteristics influencing the user demand response are used as the input of the neural network model, the demand response quantity is used as the output of the neural network model, parameters are updated through the meta-learning MAML irrelevant to the model, and then the Epanechnikov nuclear density model is used for probability prediction. The prediction results are shown in fig. 5.
As can be seen from fig. 5, through practical tests, the demand response schedulable capacity probability prediction effect provided by the embodiment of the present invention is better, and a 90% confidence interval covers all real aggregated demand response capacities. Therefore, the method has better prediction effect.
It should be noted that the demand response schedulable capacity probability prediction method provided by the embodiment of the present invention may be applied to devices such as a mobile phone, a computer, a server, etc. which can perform demand response schedulable capacity probability prediction.
Referring to fig. 6, fig. 6 is a schematic structural diagram of a demand response schedulable capacity probability prediction apparatus according to an embodiment of the present invention, as shown in fig. 6, the apparatus includes:
an obtaining module 601, configured to obtain demand response data of each user subordinate to the load aggregator on a historical demand response day;
a calculating module 602, configured to calculate a time-sharing aggregation response capacity according to the demand response data;
the feature extraction module 603 is configured to perform feature extraction on the time-sharing aggregation response capacity to obtain a feature parameter of the sample time-sharing aggregation response capacity;
the processing module 604 is configured to use the sample time-sharing aggregation response capacity characteristic parameter as an input, use the time-sharing aggregation response capacity as an output, and perform schedulable capacity estimation according to the characteristic parameter by using a meta-learning algorithm;
and a prediction module 605, configured to perform probability prediction of demand response schedulable capacity of the load aggregator by using the non-parametric kernel density probability prediction model.
Optionally, the calculating module 602 is further configured to accumulate all the user required responses in different demand response time periods to obtain a time-sharing aggregate response capacity.
Optionally, the calculating module 602 specifically includes:
the characteristic extraction unit is used for extracting the characteristics of the time-sharing aggregation response capacity;
and the screening unit is used for screening the extracted features according to the maximum information coefficient to obtain the feature parameters of the sample time-sharing polymerization response capacity.
Optionally, the extracted features include: day-by-day characteristics, time-by-time characteristics, cumulative benefit characteristics.
Optionally, as shown in fig. 7, the processing module 604 specifically includes:
a classifying unit 6041 configured to classify the sample time-sharing aggregation response capacity according to a preset condition; and
a grouping unit 6042, configured to group the sample time-sharing aggregation response capacity according to a training model type;
a parameter unit 6043 for performing parameter optimization using the model-independent meta learning MAML as a meta learning algorithm.
Optionally, the parameter unit 6043 is further configured to perform parameter optimization on the target parameter through a random gradient descent method according to the initial parameter of the model, the internal learning rate, and a task based on the sample time-sharing aggregation response capacity.
Optionally, as shown in fig. 8, the prediction module 605 specifically includes:
a first processing unit 6051, configured to obtain a probability density function according to the sample number of the sample time-sharing aggregation response capacity, a kernel function, a demand response schedulable capacity prediction error, and a sample bandwidth;
a second processing unit 6052, configured to obtain a probability distribution function based on the probability density function;
a prediction unit 6053, configured to perform probability prediction of the demand response schedulable capacity of the load aggregator by using the probability distribution function.
It should be noted that the demand response schedulable capacity probability prediction apparatus provided in the embodiment of the present invention may be applied to devices such as a mobile phone, a computer, and a server that can perform demand response schedulable capacity probability prediction.
The demand response schedulable capacity probability prediction device provided by the embodiment of the invention can realize each process realized by the demand response schedulable capacity probability prediction method in the method embodiment, and can achieve the same beneficial effect. To avoid repetition, further description is omitted here.
Referring to fig. 9, fig. 9 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, as shown in fig. 9, including: a memory 902, a processor 901 and a computer program stored on the memory 902 and executable on the processor 901, wherein:
the processor 901 is used for calling the computer program stored in the memory 902 and executing the following steps:
acquiring demand response data of each user subordinate to the load aggregation provider on a historical demand response day;
calculating time-sharing aggregation response capacity according to the demand response data;
performing feature extraction on the time-sharing polymerization response capacity to obtain a feature parameter of the sample time-sharing polymerization response capacity;
taking the characteristic parameter of the sample time-sharing aggregation response capacity as input, taking the time-sharing aggregation response capacity as output, and estimating schedulable capacity according to the characteristic parameter by using a meta-learning algorithm;
and performing probability prediction of the demand response schedulable capacity of the load aggregator by using the nonparametric kernel density probability prediction model.
Optionally, the step, executed by the processor 901, of calculating the time-sharing aggregation response capacity according to the demand response data specifically includes:
and accumulating all the user required responses in different demand response time periods to obtain time-sharing aggregate response capacity.
Optionally, the step, executed by the processor 901, of performing feature extraction on the time-sharing aggregation response capacity to obtain the sample time-sharing aggregation response capacity specifically includes:
performing feature extraction on the time-sharing polymerization response capacity;
and screening the extracted features according to the maximum information coefficient to obtain the feature parameters of the sample time-sharing polymerization response capacity.
Optionally, the extracted features include: day-by-day characteristics, time-by-time characteristics, cumulative benefit characteristics.
Optionally, the step of performing schedulable capacity estimation by using the meta learning algorithm according to the feature parameters, executed by the processor 901, specifically includes:
classifying the sample time-sharing polymerization response capacity according to a preset condition;
grouping the sample time-sharing aggregation response capacity according to the type of a training model;
and using the model-independent meta-learning MAML as a meta-learning algorithm to carry out parameter optimization.
Optionally, the step of performing parameter optimization by using the model-independent meta learning MAML as a meta learning algorithm executed by the processor 901 specifically includes:
and performing parameter optimization on the target parameter by a random gradient descent method according to the initial parameters of the model, the internal learning rate and the task based on the sample time-sharing aggregation response capacity.
Optionally, the specific steps, executed by the processor 901, of performing the probability prediction of the demand response schedulable capacity of the load aggregator by using the non-parametric kernel density probability prediction model include:
obtaining a probability density function according to the sample quantity of the sample time-sharing aggregation response capacity, the kernel function, the demand response schedulable capacity prediction error and the sample bandwidth;
obtaining a probability distribution function based on the probability density function;
and performing probability prediction of the demand response schedulable capacity of the load aggregator through the probability distribution function.
The electronic device may be a device that can be applied to a mobile phone, a computer, a server, or the like that can perform probability prediction of a demand response schedulable capacity.
The electronic device provided by the embodiment of the invention can realize each process realized by the demand response schedulable capacity probability prediction method in the method embodiment, and can achieve the same beneficial effect, and the details are not repeated here in order to avoid repetition.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
The above disclosure is only for the purpose of illustrating the preferred embodiments of the present invention, and it is therefore to be understood that the invention is not limited by the scope of the appended claims.
Claims (9)
1. A demand response schedulable capacity probability prediction method, comprising the steps of:
acquiring demand response data of each user subordinate to the load aggregation provider on a historical demand response day;
calculating time-sharing aggregation response capacity according to the demand response data;
performing feature extraction on the time-sharing polymerization response capacity to obtain a feature parameter of the sample time-sharing polymerization response capacity;
taking the sample time-sharing aggregation response capacity characteristic parameter as input, taking the time-sharing aggregation response capacity as output, and performing schedulable capacity estimation according to the characteristic parameter by using a meta-learning algorithm;
and performing probability prediction of the demand response schedulable capacity of the load aggregator by using the nonparametric kernel density probability prediction model.
2. The demand response schedulable capacity probability prediction method of claim 1, wherein the step of calculating a time-share aggregated response capacity according to the demand response data specifically comprises:
and accumulating all the user demand responses in different demand response time periods to obtain time-sharing aggregate response capacity.
3. The demand response schedulable capacity probability prediction method of claim 1, wherein the step of performing the feature extraction on the time-sharing aggregate response capacity to obtain the feature parameter of the sample time-sharing aggregate response capacity specifically comprises:
performing feature extraction on the time-sharing polymerization response capacity;
and screening the extracted features according to the maximum information coefficient to obtain the feature parameters of the sample time-sharing polymerization response capacity.
4. The demand-response dispatchable capacity probability prediction method as set forth in claim 1, wherein the extracted characteristic parameters include: day-by-day characteristics, time-by-time characteristics, cumulative benefit characteristics.
5. The demand response schedulable capacity probability prediction method of claim 1, wherein the step of using a meta learning algorithm to estimate schedulable capacity according to the characteristic parameters specifically comprises:
classifying the sample time-sharing polymerization response capacity according to a preset condition;
grouping the sample time-sharing aggregation response capacity according to the type of a training model;
and using the model-independent meta-learning MAML as a meta-learning algorithm to carry out parameter optimization.
6. The demand-response schedulable capacity probability prediction method of claim 5, wherein the step of using model-independent meta-learning MAML as a meta-learning algorithm for parameter optimization specifically comprises:
and performing parameter optimization on the target parameter by a random gradient descent method according to the initial parameters of the model, the internal learning rate and the task based on the sample time-sharing aggregation response capacity.
7. The demand response schedulable capacity probability prediction method of claim 1, wherein the step of performing the probability prediction of the demand response schedulable capacity of the load aggregator using the non-parametric kernel density probability prediction model comprises:
obtaining a probability density function according to the sample quantity of the sample time-sharing aggregation response capacity, the kernel function, the demand response schedulable capacity prediction error and the sample bandwidth;
obtaining a probability distribution function based on the probability density function;
and performing probability prediction of the demand response schedulable capacity of the load aggregator through the probability distribution function.
8. A demand response schedulable capacity probability prediction apparatus, the apparatus comprising:
the acquisition module is used for acquiring the demand response data of each user subordinate to the load aggregation provider on the historical demand response day;
the calculation module is used for calculating time-sharing aggregation response capacity according to the demand response data;
the characteristic extraction module is used for carrying out characteristic extraction on the time-sharing polymerization response capacity to obtain a characteristic parameter of the sample time-sharing polymerization response capacity;
the processing module is used for taking the sample time-sharing aggregation response capacity characteristic parameter as input, taking the time-sharing aggregation response capacity as output, and carrying out schedulable capacity estimation according to the characteristic parameter by utilizing a meta-learning algorithm;
and the prediction module is used for performing probability prediction on the demand response schedulable capacity of the load aggregator by using the nonparametric kernel density probability prediction model.
9. An electronic device, comprising: memory, processor and computer program stored on the memory and executable on the processor, the processor implementing the steps in the demand response schedulable capacity probability prediction method of any of the claims 1 to 7 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011262389.9A CN112348380A (en) | 2020-11-12 | 2020-11-12 | Demand response schedulable capacity probability prediction method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011262389.9A CN112348380A (en) | 2020-11-12 | 2020-11-12 | Demand response schedulable capacity probability prediction method and device and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112348380A true CN112348380A (en) | 2021-02-09 |
Family
ID=74362645
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011262389.9A Pending CN112348380A (en) | 2020-11-12 | 2020-11-12 | Demand response schedulable capacity probability prediction method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112348380A (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103699947A (en) * | 2014-01-16 | 2014-04-02 | 湖南大学 | Meta learning-based combined prediction method for time-varying nonlinear load of electrical power system |
| CN106407627A (en) * | 2016-11-23 | 2017-02-15 | 西南石油大学 | Wind speed probability distribution modeling method and system |
| CN108416695A (en) * | 2018-02-24 | 2018-08-17 | 合肥工业大学 | Electric load probability density prediction technique based on deep learning and system, medium |
| CN109784594A (en) * | 2017-11-10 | 2019-05-21 | 中国电力科学研究院有限公司 | A kind of sale of electricity quotient deferrable load decision-making technique and system |
| CN109858835A (en) * | 2019-02-26 | 2019-06-07 | 合肥工业大学 | The demand response modeling of Load aggregation quotient a kind of and reliability estimation method |
| CN110516882A (en) * | 2019-08-30 | 2019-11-29 | 华北电力大学（保定） | Load agent can use aggregate response capacity prediction methods a few days ago |
| CN110782363A (en) * | 2019-08-15 | 2020-02-11 | 东南大学 | AC/DC power distribution network scheduling method considering wind power uncertainty |
| CN110991693A (en) * | 2019-10-28 | 2020-04-10 | 重庆大学 | Load price-quantity curve aggregation system considering uncertainty of electricity purchase cost |
-
2020
- 2020-11-12 CN CN202011262389.9A patent/CN112348380A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103699947A (en) * | 2014-01-16 | 2014-04-02 | 湖南大学 | Meta learning-based combined prediction method for time-varying nonlinear load of electrical power system |
| CN106407627A (en) * | 2016-11-23 | 2017-02-15 | 西南石油大学 | Wind speed probability distribution modeling method and system |
| CN109784594A (en) * | 2017-11-10 | 2019-05-21 | 中国电力科学研究院有限公司 | A kind of sale of electricity quotient deferrable load decision-making technique and system |
| CN108416695A (en) * | 2018-02-24 | 2018-08-17 | 合肥工业大学 | Electric load probability density prediction technique based on deep learning and system, medium |
| CN109858835A (en) * | 2019-02-26 | 2019-06-07 | 合肥工业大学 | The demand response modeling of Load aggregation quotient a kind of and reliability estimation method |
| CN110782363A (en) * | 2019-08-15 | 2020-02-11 | 东南大学 | AC/DC power distribution network scheduling method considering wind power uncertainty |
| CN110516882A (en) * | 2019-08-30 | 2019-11-29 | 华北电力大学（保定） | Load agent can use aggregate response capacity prediction methods a few days ago |
| CN110991693A (en) * | 2019-10-28 | 2020-04-10 | 重庆大学 | Load price-quantity curve aggregation system considering uncertainty of electricity purchase cost |
Non-Patent Citations (1)
| Title |
|---|
| 孙建波等: "基于非参数核密度估计的风电功率区间预测", 《水电能源科学》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108470233B (en) | Demand response capability assessment method and computing device for smart power grid | |
| CN110782153A (en) | Modeling method and system for comprehensive energy efficiency assessment system of enterprise park | |
| CN114140176B (en) | Adjustable capacity prediction method and device for load aggregation platform | |
| CN111680841A (en) | Short-term load prediction method and system based on principal component analysis and terminal equipment | |
| KR20200128232A (en) | Apparatus for predicting electricity demand and method thereof | |
| CN114372360A (en) | Method, terminal and storage medium for power load prediction | |
| Ribeiro et al. | Customized normalization clustering meth-odology for consumers with heterogeneous characteristics | |
| CN112288172A (en) | Prediction method and device for line loss rate of transformer area | |
| CN113807589A (en) | Rolling optimization method and device for operator energy storage based on model prediction | |
| CN117541314A (en) | Intelligent issuing system, equipment and storage medium for online taxi coupons | |
| CN116628534A (en) | Method for dividing energy dynamic images for park based on electric power big data | |
| CN109034241A (en) | Load cluster control method and system based on support vector machines | |
| CN117691583B (en) | Power dispatching system and method for virtual power plant | |
| CN110717779A (en) | Electric power transaction system, method and application thereof | |
| KR102556093B1 (en) | Reward generation method to reduce peak load of electric power and action control apparatus performing the same method | |
| CN110648248B (en) | Control method, device and equipment for power station | |
| CN117557125A (en) | Method and system for evaluating flexible resource adjustment capability of power system | |
| CN116992216A (en) | Short-term load prediction method based on power user data | |
| CN112348380A (en) | Demand response schedulable capacity probability prediction method and device and electronic equipment | |
| CN116541128A (en) | Load adjusting method, device, computing equipment and storage medium | |
| CN115358441A (en) | New energy cluster consumption intelligent control method and device based on federal learning | |
| CN114759579A (en) | Power grid active power optimization control system, method and medium based on data driving | |
| CN111797924A (en) | Three-dimensional garden portrait method and system based on clustering algorithm | |
| CN118336163B (en) | Intelligent management method applied to energy storage battery and related equipment | |
| CN116957306B (en) | User side response potential evaluation method and system based on resource collaborative interaction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210209 |
|
| RJ01 | Rejection of invention patent application after publication |