CN110648248B - Control method, device and equipment for power station - Google Patents
Control method, device and equipment for power station Download PDFInfo
- Publication number
- CN110648248B CN110648248B CN201910839181.XA CN201910839181A CN110648248B CN 110648248 B CN110648248 B CN 110648248B CN 201910839181 A CN201910839181 A CN 201910839181A CN 110648248 B CN110648248 B CN 110648248B
- Authority
- CN
- China
- Prior art keywords
- data
- load
- historical
- environmental
- estimation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
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
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- 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"
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The invention discloses a control method of a power station, wherein a part of historical load data and historical environment data with the highest similarity to two types of estimated data of a target day can be screened out, and the part of data has the highest similarity to the estimated data of the target day, so that the accuracy of a prediction result can be improved by training a limit learning machine through the part of data, and a worker arranges the scheduling output of the power station of the target day according to the accurate prediction result, so that the supply and demand can be more balanced, the waste of electric energy is reduced, and the operation cost of a power grid is reduced. The invention also discloses a control device and equipment of the power station, and the control device and equipment have the same beneficial effects as the control method of the power station.
Description
Technical Field
The invention relates to the field of load scheduling of a power system, in particular to a control method of a power station, and further relates to a control device and equipment of the power station.
Background
In the field of load scheduling of power systems, scheduling personnel can generally arrange scheduling output of one day when a target day arrives according to load prediction data of the target day so as to enable supply and demand of power grid electric energy to be more balanced and reduce power grid operation cost.
Therefore, how to provide a solution to the above technical problem is a problem that needs to be solved by those skilled in the art.
Disclosure of Invention
The invention aims to provide a control method of a power station, which can balance the supply and demand of electric energy in a power grid, reduce the waste of the electric energy and reduce the operation cost of the power grid; another object of the present invention is to provide a control apparatus and device for a power station, which can balance the supply and demand of electric energy in a power grid, reduce the waste of electric energy, and reduce the operation cost of the power grid.
In order to solve the above technical problem, the present invention provides a method for controlling a power plant, including:
acquiring historical load data and historical environment data of each day in a past preset time period;
acquiring environment estimation data and load estimation data of a target day to be predicted;
determining data with the highest similarity with the environmental estimation data and the load estimation data in a preset proportion from the daily historical load data and the historical environmental data as training samples;
predicting the load data of the target day according to the extreme learning machine trained by the training sample;
and controlling the output of the power station on the target day according to the predicted load data.
Preferably, after determining a preset proportion of data with the highest similarity to the environmental estimation data and the load estimation data as a training sample in the daily historical load data and the historical environmental data, and before predicting the load data of the target day according to the extreme learning machine trained by the training sample, the method for controlling the power plant further includes:
adding a preset number of load data and environmental data of extreme days with largely inaccurate predictions to the training samples.
Preferably, after determining a preset proportion of data with the highest similarity to the environmental estimation data and the load estimation data from the daily historical load data and the historical environmental data as training samples, and before predicting the load data of the target day according to the extreme learning machine trained by the training samples, the method for controlling a power plant further includes:
adding a preset number of load data and environment data of typical scenes corresponding to the target day into the training sample.
Preferably, after obtaining the environmental estimated data and the load estimated data of the target day to be predicted, before determining data with the highest similarity to the environmental estimated data and the load estimated data in a preset proportion from the daily historical load data and the historical environmental data as a training sample, the method further includes:
performing cloud model modeling on the historical load data to obtain first cloud model data;
carrying out cloud model modeling on the load estimation data to obtain second cloud model data;
the step of determining, as a training sample, data with the highest similarity to the environmental estimation data and the load estimation data in a preset proportion from the daily historical load data and the historical environmental data specifically includes:
and determining data with the highest similarity with the environmental estimation data and the load estimation data combined with the second cloud model data in a preset proportion from the historical load data and the historical environmental data combined with the first cloud model data as training samples.
Preferably, after the load data of the target day is predicted according to the extreme learning machine trained by the training sample, before the power plant is controlled to output according to the predicted load data on the target day, the power plant control method further includes:
taking the predicted load prediction data of the target day as the load prediction data;
judging whether the difference value of the load prediction data and the load prediction data is smaller than a preset threshold value or not;
if not, returning to the step: and determining data with the highest similarity with the environmental estimation data and the load estimation data combined with the second cloud model data in a preset proportion from the historical load data and the historical environmental data combined with the first cloud model data as training samples.
Preferably, the historical environmental data and the environmental forecast data include temperature, humidity, and barometric pressure.
Preferably, the preset period of time is three years.
Preferably, said similarity is highest, in particular weighted minkowski distance is lowest.
In order to solve the above technical problem, the present invention further provides a control apparatus for a power plant, including:
the first acquisition module is used for acquiring daily historical load data and historical environment data in a past preset time period;
the second acquisition module is used for acquiring environment estimation data and load estimation data of a target day to be predicted;
the determining module is used for determining data with the highest similarity with the environmental estimation data and the load estimation data in a preset proportion from the daily historical load data and the historical environmental data as training samples;
the prediction module is used for predicting the load data of the target day according to the extreme learning machine trained by the training sample;
and the control module is used for controlling the output of the power station on the target day according to the predicted load data.
In order to solve the above technical problem, the present invention further provides a control device of a power plant, including:
a memory for storing a computer program;
a processor for implementing the steps of the method of controlling a power plant as claimed in any one of the preceding claims when the computer program is executed.
The invention provides a control method of a power station, which can screen out a part of historical load data and historical environment data with highest similarity to two types of estimated data of a target day, wherein the part of data has the highest similarity to the estimated data of the target day, so that the accuracy of a prediction result can be improved by training a limit learning machine through the part of data, and a worker arranges the scheduling output of the power station of the target day according to the accurate prediction result, so that the supply and demand can be more balanced, the waste of electric energy is reduced, and the operation cost of a power grid is reduced.
The invention also provides a control device and equipment of the power station, and the control device and equipment have the same beneficial effects as the control method of the power station.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed in the prior art and the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a flow chart illustrating a method for controlling a power plant according to the present invention;
FIG. 2 is a schematic structural diagram of a control device of a power station provided by the invention;
fig. 3 is a schematic structural diagram of a control device of a power plant provided by the invention.
Detailed Description
The core of the invention is to provide a control method of a power station, so that the supply and demand of electric energy in a power grid are more balanced, the waste of the electric energy is reduced, and the operation cost of the power grid is reduced; the other core of the invention is to provide a control device and equipment of a power station, so that the supply and demand of electric energy in a power grid are more balanced, the waste of the electric energy is reduced, and the operation cost of the power grid is reduced.
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, 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 some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, fig. 1 is a schematic flow chart of a method for controlling a power plant according to the present invention, including:
step S1: acquiring historical load data and historical environment data of each day in a past preset time period;
specifically, the historical load data has a guiding effect on the prediction of the load data, and the historical environment data is also closely related to the load, for example, the air temperature on a certain day approaches 40 ℃, the usage rate of the air conditioner on the day will increase suddenly, and the corresponding load data in each time period on the day will be at a higher level.
Of course, besides the historical load data and the historical environment data, the acquired historical data may also include other types of data, and the embodiment of the present invention is not limited herein.
Specifically, the preset time period may be set autonomously, which may determine the size of the data sample obtained in this step, and the embodiment of the present invention is not limited herein.
Step S2: acquiring environment prediction data and load prediction data of a target day to be predicted;
specifically, the target day to be predicted may be any one of the natural days that do not come, for example, a tomorrow day, and the embodiment of the present invention is not limited herein.
The environmental estimated data are estimated values of various environmental indexes of the target day, the load estimated data can be load conditions of various time periods in the target day of initial rough estimation, and the two estimated data can be obtained to facilitate judgment of similarity between historical data and the estimated data of the target day in subsequent steps so as to determine the training sample.
The load estimation data may be load estimation data of a target day roughly estimated by a worker according to own experience, or may be data estimated by using other types of load data estimation methods, and the embodiment of the present invention is not limited herein.
Specifically, the environmental forecast data may be environmental forecast data of a target day acquired from a weather station, which may have higher accuracy.
And step S3: determining data with the highest similarity to environmental estimation data and load estimation data in a preset proportion from the daily historical load data and the historical environmental data as training samples;
specifically, considering that in the prior art, the limit learning machine is usually trained by using all the acquired historical load data and historical environment data, so as to predict the load data of the target day through the limit learning machine, but in the historical load data and the historical environment data of all the historical days, the similarity between the historical load data and the historical environment data of each day and the load predicted data and the environment predicted data of the target day is different, if the limit learning machine is trained by using all the historical load data and the historical environment data, the limit learning machine cannot have an accurate predicted result for the target day, and therefore, a worker can control a power plant according to inaccurate load data, the supply and demand matching degree of electric energy is not high, the cost is increased, and the user experience is reduced.
In the embodiment of the invention, historical environmental data and historical load data with higher similarity to the environmental estimated data and the load estimated data in a preset proportion can be determined as training samples, so that the trained extreme learning machine has higher estimation accuracy due to the higher similarity of the training samples and the two types of estimated data of the target day, and under the condition, a worker can well achieve the supply and demand balance of electric energy by controlling the power station according to the estimated load data, thereby reducing the cost and optimizing the user experience.
And step S4: predicting load data of a target day according to the extreme learning machine trained by the training sample;
specifically, after the extreme learning machine is trained by using the training samples, the extreme learning machine can predict the load data of the target day, wherein the load prediction data and the environment prediction data need to be input when the extreme learning machine is used for predicting the load data, and the training samples have high similarity with the two types of prediction data of the target day, so that the trained extreme learning machine can more accurately predict the load data of the target day.
Of course, other types of learning methods besides the extreme learning method may also be used, and the embodiment of the present invention is not limited herein.
Step S5: and controlling the output of the power station according to the predicted load data on the target day.
Specifically, after the load data of the target day is obtained, the worker can control the generator to generate the electric energy corresponding to the load data at the target day, for example, the estimated load data of a certain area at a certain hour is a, so that the generator is controlled to supply the electric energy corresponding to the load data a to the certain area at a certain hour, the supply and demand balance of the electric energy is achieved, the electric energy waste is reduced, and the user experience is also improved.
The invention provides a control method of a power station, which can screen out a part of historical load data and historical environment data with highest similarity to two types of estimated data of a target day, wherein the part of data has the highest similarity to the estimated data of the target day, so that the accuracy of a prediction result can be improved by training a limit learning machine through the part of data, and a worker arranges the scheduling output of the power station of the target day according to the accurate prediction result, so that the supply and demand can be more balanced, the waste of electric energy is reduced, and the operation cost of a power grid is reduced.
On the basis of the above-described embodiment:
as a preferred embodiment, after determining data with the highest similarity to the environmental estimation data and the load estimation data in a preset proportion as a training sample from the daily historical load data and the historical environmental data, and before predicting the load data of a target day according to the extreme learning machine trained by the training sample, the method for controlling the power station further includes:
load data and environmental data for a preset number of extreme days of largely inaccurate predictions are added to the training samples.
Specifically, in order to further enhance the prediction accuracy of the extreme learning machine, load data and environment data of an extreme day of the prediction of the large amplitude misalignment are added into the training samples, for example, in the historical load data prediction process, the load data of a certain day is predicted to be far from the real load data of the day, so that the day can be considered as belonging to the extreme day of the prediction of the large amplitude misalignment, and the load data and the environment data of the extreme day of the prediction of the large amplitude misalignment are added into the training samples, so that the load data and the environment data of the extreme day can be further combined in the prediction process, the situation of the large amplitude misalignment is prevented from reoccurring, and the prediction robustness is enhanced.
Specifically, the preset number can be set according to actual requirements, and the data volume of the common extreme days is far smaller than the data volume of the normal historical load data and the historical environment data.
As a preferred embodiment, after determining data with the highest similarity to the environmental estimation data and the load estimation data at a preset ratio in the daily historical load data and the historical environmental data as a training sample, and before predicting the load data of a target day according to the extreme learning machine trained by the training sample, the method for controlling the power plant further comprises:
load data and environmental data of a preset number of typical scenes corresponding to the target day are added to the training sample.
Specifically, in order to further enhance the accuracy of load data prediction, load data and environment data of a typical scene corresponding to a target day are added into a training sample, the typical scene refers to the typical scene corresponding to the target day, and the typical scene has many similar characteristics.
Specifically, the preset number may be set autonomously according to a requirement, and may be the same as the data size of the extreme day, and the embodiment of the present invention is not limited herein.
As a preferred embodiment, after acquiring the environmental forecast data and the load forecast data of the target day to be forecasted, before determining, as a training sample, data with the highest similarity to the environmental forecast data and the load forecast data in a preset proportion from the historical load data and the historical environmental data of each day, the method further includes:
carrying out cloud model modeling on the historical load data to obtain first cloud model data;
carrying out cloud model modeling on the load estimation data to obtain second cloud model data;
determining data with the highest similarity with the environmental estimation data and the load estimation data in a preset proportion from the daily historical load data and the historical environmental data as a training sample specifically comprises the following steps:
and determining the data with the highest similarity to the environment estimation data and the load estimation data combined with the second cloud model data in a preset proportion as a training sample from the historical load data and the historical environment data combined with the first cloud model data.
Specifically, the cloud model data can be used for more finely expressing the corresponding data, the cloud model data comprises index data of a plurality of cloud models, in this case, the number of the cloud models can deeply express the corresponding data, for example, the first cloud model data can deeply and accurately express historical load data, and the second cloud model data can deeply and accurately express load estimation data.
The type of the index data of the cloud model may be set autonomously, and the embodiment of the present invention is not limited herein.
As a preferred embodiment, after the load data of the target day is predicted according to the extreme learning machine trained by the training samples, before the power plant is controlled to output according to the predicted load data on the target day, the power plant control method further includes:
taking the predicted load prediction data of the target day as load prediction data;
judging whether the difference value of the load prediction data and the load prediction data is smaller than a preset threshold value or not;
if not, returning to the step: and determining the data with the highest similarity to the environment estimation data and the load estimation data combined with the second cloud model data in a preset proportion as a training sample from the historical load data and the historical environment data combined with the first cloud model data.
Specifically, although the accuracy of the predicted load data is higher than that of the load estimation data, the accuracy of the predicted load data is limited, so that the predicted result needs to be further iterated.
In addition, considering that the computing resources are limited, the method and the device can stop when the difference value between the load prediction data and the load prediction data is smaller than the preset threshold value, and finally obtained load prediction data can be used as the control basis of the power station.
The preset threshold may be set autonomously according to actual requirements, and the embodiment of the present invention is not limited herein.
As a preferred embodiment, the historical environmental data and environmental forecast data include temperature, humidity, and barometric pressure.
Specifically, the correlation between the temperature, the humidity and the air pressure and the load data is strong, so that the accuracy of the prediction result can be further improved by selecting several types of historical environmental data and environmental estimation data.
Of course, besides the above types, the historical environmental data and the environmental estimation data may also be of other types, and the embodiments of the present invention are not limited herein.
As a preferred embodiment, the preset period of time is three years.
Specifically, the preset time period is set to three years, so that on one hand, a sufficient sample can be ensured, and on the other hand, excessive calculation amount and low calculation speed caused by excessive samples are avoided.
Of course, the preset time period may be other specific values besides three years, and the embodiment of the present invention is not limited herein.
As a preferred embodiment, the highest degree of similarity is in particular the weighted minkowski distance minimum.
In particular, the weighted Minkowski distance can accurately describe the similarity between data, and can further improve the accuracy of the prediction result.
Of course, the similarity between the data may be described by other methods besides weighting the minkowski distance, and the embodiments of the present invention are not limited thereto.
Specifically, the embodiment of the present invention exemplifies a specific implementation manner, and the specific steps are as follows:
step A1, collecting daily hourly load and environmental data of places in last three years. Different combinations of the various types of environmental data represent various different environmental states. Various types of environmental data include air pressure, temperature, humidity data. Wherein:
Si＝{Si 1 ,Si 2 ,Si 3 ,P fi }
wherein Si represents individual data samples, and the combination of three kinds of data in one individual data sample represents an environmental state; si 1 Representing illumination intensity data; si 2 Representing temperature data; si 3 Representing humidity data, P fi Representing 24 hours of loading data on the day, i =1,2,3, \ 8230;, 24.
And step A2, collecting typical scenes in the last three years and load and environment data of each hour of extreme days with large-amplitude misalignment prediction, wherein the environment data comprises air pressure, temperature and humidity data. Typical scenes include a small random sampling of seasons, extreme weather such as typhoon, stormy weather scenes. Wherein:
Sr＝{Sr 1 ,Sr 2 ,Sr 3 ,P fi }
wherein, sr represents a data sample individual of a typical scene or an extreme day, sr 1 Representing illumination intensity data; sr 2 Representing temperature data; sr 3 Representing humidity data, P fi Representing 24 hours load data, i =1,2,3, \8230;, 24.
And step A3, carrying out cloud model modeling on the collected daily load data with granularity of one day as time within three years by using a reverse cloud generator. The cloud model numerical characteristics comprise an expectation Ex, an entropy En and a super entropy He, and the specific modeling is as follows:
step1: calculating the daily average load, wherein:
the expected Ex of the daily load cloud model is replaced by the daily average load.
Step2: calculating a daily load variance, wherein:
step3: computing an entropy En of the cloud model, wherein:
step4: and calculating the super entropy He according to the variance and entropy obtained by Step2 and Step3, wherein:
and A4, combining the three-year data sample cloud model with the environment data parameters, and calculating the weighted Minkowski distance with the environment estimated data of the target day, the load estimated data and the cloud model data of the load estimated data. The method comprises the following steps:
step1: combining the daily load cloud model data with the daily environment data to form a complete sample, wherein:
Si＝{Si 1 ,Si 2 ,Si 3 ,Ex,En,He,P fi }
Sr＝{Sr 1 ,Sr 2 ,Sr 3 ,Ex,En,He,P fi }
So＝{So 1 ,So 2 ,So 3 ,Ex,En,He,P fi }
in the above formula, si is a data composition of a common sample individual within three years, sr is an extreme sample, and So is a data sample individual of a target day.
Step2: three samples of Step1 were normalized as follows:
f imax ＝max(f i )
f imin ＝min(f i )
wherein f is i The method respectively represents 6 elements in the sample data set, and three normalized sample sets are obtained after normalization:
step3: the weighted Minkoff-based distance of Si and So is calculated as follows:
wherein the weight coefficient w i P is a norm indicator > 0, and if p =2, the weighted Minkoff-based distance is expressed as the squared value of the weighted two-norm sum, where p can be preset as desired. x is the number of i ,x j Elements in the common sample individual Si and the target day So respectively, wherein w i I =1-n, x i Including x i1 To x in And x is j Including x j1 To x jn 。
Step4: and (4) sorting the distances obtained in Step3 from small to large, selecting the samples with the distance sorting of the first 30% as similar day samples, and randomly extracting 5% of the sample capacity from the rest 70% of the samples to add into the extreme sample set.
Step A5, mixing and disordering the similar day sample and the extreme day sample obtained in the above steps to form a new sample S, wherein the sample S comprises a similar day sample Si and an extreme day sample Sr, and the proportion of the sample S is about 5:1.
step A6, sample
The cloud model parameters and the environment parameters are used as input, the predicted load is used as output, and an extreme learning machine is used for training, wherein the method specifically comprises the following steps:
step1: randomly dividing a training set and a testing set, and carrying out A5 on the samples obtained in the step
Randomly sampling, wherein 80% of samples are extracted as training samples without repeated sampling, and the rest 20% are used as test samples;
step2: ELMs (Extreme Learning machines) are created/trained. To be provided with
Taking the environmental parameters and the daily load cloud model parameters as input, taking the sample individuals as well>
Corresponding 24 hour daily load P i For output, i =1,2,3, \ 8230;, 24;
step3: performing ELM simulation test, namely predicting a test set by using the ELM trained by Step 2;
step4: and (4) evaluating the performance of the ELM, and evaluating the generalization capability of the model by calculating the error between the predicted value and the true value of the test set and indexes such as mean square error, decision coefficient, accuracy and the like.
And A7, predicting the load of the target day by using the extreme learning machine, establishing new cloud model data of the target day by using the predicted load, and forming a new target day data sample individual with the normalized value of the target day environmental parameter.
And step A8, repeating the steps A4 to A7 until the maximum difference of the predicted load of 24h of the previous iteration and the next iteration is not more than 2% of the load value in the time interval.
Referring to fig. 2, fig. 2 is a schematic structural diagram of a control device of a power station, which includes:
the system comprises a first acquisition module 1, a first storage module and a second acquisition module, wherein the first acquisition module is used for acquiring daily historical load data and historical environment data in a past preset time period;
the second acquisition module 2 is used for acquiring environment estimation data and load estimation data of a target day to be predicted;
the determining module 3 is used for determining data with the highest similarity to the environmental estimation data and the load estimation data in a preset proportion from the daily historical load data and the historical environmental data as training samples;
the prediction module 4 is used for predicting the load data of the target day according to the extreme learning machine trained by the training sample;
and the control module 5 is used for controlling the output of the power station according to the predicted load data on the target day.
For the description of the control device of the power plant provided by the embodiment of the present invention, reference is made to the foregoing embodiment of the control method of the power plant, and the embodiment of the present invention is not limited herein.
Referring to fig. 3, fig. 3 is a schematic structural diagram of a control device of a power plant provided in the present invention, including:
a memory 6 for storing a computer program;
a processor 7 for implementing the steps of the method of controlling a power plant as described above when executing a computer program.
For the introduction of the control device of the power plant provided by the embodiment of the present invention, refer to the foregoing embodiment of the control method of the power plant, and the embodiment of the present invention is not limited herein.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
It is further noted that, in the present specification, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrases "comprising a component of' 8230; \8230;" does not exclude the presence of additional like elements in a process, method, article, or apparatus that comprises the element.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (8)
1. A method of controlling a power plant, comprising:
acquiring historical load data and historical environment data of each day in a past preset time period;
acquiring environment prediction data and load prediction data of a target day to be predicted;
determining data with the highest similarity with the environmental estimation data and the load estimation data in a preset proportion from the daily historical load data and the historical environmental data as training samples;
predicting the load data of the target day according to the extreme learning machine trained by the training sample;
controlling the output of the power station on the target day according to the predicted load data;
after the environmental estimated data and the load estimated data of the target day to be predicted are obtained, before the data with the highest similarity with the environmental estimated data and the load estimated data in a preset proportion are determined to be used as training samples in the daily historical load data and the historical environmental data, the method further comprises the following steps:
performing cloud model modeling on the historical load data to obtain first cloud model data;
carrying out cloud model modeling on the load estimation data to obtain second cloud model data;
the step of determining, as a training sample, data with the highest similarity to the environmental estimation data and the load estimation data in a preset proportion from the daily historical load data and the historical environmental data specifically includes:
determining data with the highest similarity to the environmental estimation data combined with the second cloud model data and the load estimation data in a preset proportion from the historical load data combined with the first cloud model data and the historical environmental data as training samples;
after the load data of the target day is predicted according to the extreme learning machine trained by the training sample, and before the output of the power station is controlled according to the predicted load data on the target day, the control method of the power station further comprises the following steps:
taking the predicted load prediction data of the target day as the load prediction data;
judging whether the difference value of the load prediction data and the load prediction data is smaller than a preset threshold value or not;
if not, returning to the step: and determining data with the highest similarity with the environmental estimation data and the load estimation data combined with the second cloud model data in a preset proportion from the historical load data and the historical environmental data combined with the first cloud model data as training samples.
2. The method of controlling a power plant according to claim 1, wherein after determining a preset proportion of data with the highest similarity to the environmental estimation data and the load estimation data as a training sample from the daily historical load data and the historical environmental data, and before predicting the load data of the target day by using an extreme learning machine trained by the training sample, the method further comprises:
adding a preset number of load data and environmental data of extreme days of largely inaccurate prediction to the training sample.
3. The method for controlling a power plant according to claim 2, wherein the method further comprises, after determining a predetermined proportion of data with the highest similarity to the environmental estimation data and the load estimation data among the daily historical load data and the historical environmental data as training samples, and before predicting the load data of the target day based on the learning extreme machine trained using the training samples, the method comprising:
adding a preset number of load data and environment data of typical scenes corresponding to the target day into the training sample.
4. The method of controlling a power plant according to claim 1, characterized in that the historical environmental data and the environmental forecast data comprise temperature, humidity and air pressure.
5. The method for controlling a power plant according to claim 1, characterized in that the preset period of time is three years.
6. The method of controlling a power station according to any one of claims 1 to 5, characterized in that the degree of similarity is highest, in particular weighted Minkowski distance smallest.
7. A control apparatus for a power plant, characterized by comprising:
the first acquisition module is used for acquiring daily historical load data and historical environment data in a past preset time period;
the second acquisition module is used for acquiring environment estimation data and load estimation data of a target day to be predicted;
the determining module is used for determining data with the highest similarity with the environmental estimation data and the load estimation data in a preset proportion from the daily historical load data and the historical environmental data as training samples;
the prediction module is used for predicting the load data of the target day according to the extreme learning machine trained by the training sample;
the control module is used for controlling the output of the power station on the target day according to the predicted load data;
the control device of the power plant further comprises:
performing cloud model modeling on the historical load data to obtain first cloud model data;
carrying out cloud model modeling on the load estimation data to obtain second cloud model data;
the step of determining, as a training sample, data with the highest similarity to the environmental estimation data and the load estimation data in a preset proportion from the daily historical load data and the historical environmental data specifically includes:
determining data with the highest similarity between the environmental estimation data combined with the second cloud model data and the load estimation data in a preset proportion and the historical load data and the historical environmental data combined with the first cloud model data as training samples;
the control device of the power plant further comprises:
taking the predicted load prediction data of the target day as the load prediction data;
judging whether the difference value of the load prediction data and the load prediction data is smaller than a preset threshold value or not;
if not, returning to the step: and determining data with the highest similarity with the environmental estimation data and the load estimation data combined with the second cloud model data in a preset proportion from the historical load data and the historical environmental data combined with the first cloud model data as training samples.
8. A control apparatus of a power plant, characterized by comprising:
a memory for storing a computer program;
a processor for implementing the steps of the method of controlling a power plant according to any of claims 1 to 6 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910839181.XA CN110648248B (en) | 2019-09-05 | 2019-09-05 | Control method, device and equipment for power station |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910839181.XA CN110648248B (en) | 2019-09-05 | 2019-09-05 | Control method, device and equipment for power station |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110648248A CN110648248A (en) | 2020-01-03 |
| CN110648248B true CN110648248B (en) | 2023-04-07 |
Family
ID=68991632
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910839181.XA Active CN110648248B (en) | 2019-09-05 | 2019-09-05 | Control method, device and equipment for power station |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110648248B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111586127B (en) * | 2020-04-28 | 2023-04-28 | 深圳市爱路恩济能源技术有限公司 | Working condition data processing method based on Internet of things system |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103473601A (en) * | 2013-08-20 | 2013-12-25 | 上海海能信息科技有限公司 | System and method for intelligently forecasting short-term loads of power transmission lines |
| CN106096771A (en) * | 2016-06-07 | 2016-11-09 | 广州供电局有限公司 | The parameter selection method of support vector machine and system and load forecasting method and system |
| CN106447133A (en) * | 2016-11-03 | 2017-02-22 | 上海交通大学 | Short-term electric load prediction method based on deep self-encoding network |
| WO2018045642A1 (en) * | 2016-09-09 | 2018-03-15 | 国网山西省电力公司晋城供电公司 | A bus bar load forecasting method |
| CN108229754A (en) * | 2018-01-31 | 2018-06-29 | 杭州电子科技大学 | Short-term load forecasting method based on similar day segmentation and LM-BP networks |
| CN109034490A (en) * | 2018-08-13 | 2018-12-18 | 广东工业大学 | A kind of Methods of electric load forecasting, device, equipment and storage medium |
| CN109389238A (en) * | 2017-08-14 | 2019-02-26 | 中国电力科学研究院 | A kind of short-term load forecasting method and device based on ridge regression |
| CN110048420A (en) * | 2019-05-15 | 2019-07-23 | 广东电网有限责任公司 | A kind of method, apparatus and medium of the scheduling of power distribution network random optimization |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105337415A (en) * | 2015-11-18 | 2016-02-17 | 深圳合纵能源技术有限公司 | Regional power grid dispatching system and method based on prediction control |
| US20180284758A1 (en) * | 2016-05-09 | 2018-10-04 | StrongForce IoT Portfolio 2016, LLC | Methods and systems for industrial internet of things data collection for equipment analysis in an upstream oil and gas environment |
| WO2018015779A1 (en) * | 2016-07-20 | 2018-01-25 | Worldline | Multi-criteria adaptive scheduling for a market-oriented hybrid cloud infrastructure |
| US11120368B2 (en) * | 2017-09-27 | 2021-09-14 | Oracle International Corporation | Scalable and efficient distributed auto-tuning of machine learning and deep learning models |
| CN107944594B (en) * | 2017-09-30 | 2021-11-23 | 华南理工大学 | Short-term load prediction method based on spearman grade and RKELM microgrid |
| CN109726858A (en) * | 2018-12-21 | 2019-05-07 | 新奥数能科技有限公司 | Heat load prediction method and device based on dynamic time warping |
-
2019
- 2019-09-05 CN CN201910839181.XA patent/CN110648248B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103473601A (en) * | 2013-08-20 | 2013-12-25 | 上海海能信息科技有限公司 | System and method for intelligently forecasting short-term loads of power transmission lines |
| CN106096771A (en) * | 2016-06-07 | 2016-11-09 | 广州供电局有限公司 | The parameter selection method of support vector machine and system and load forecasting method and system |
| WO2018045642A1 (en) * | 2016-09-09 | 2018-03-15 | 国网山西省电力公司晋城供电公司 | A bus bar load forecasting method |
| CN106447133A (en) * | 2016-11-03 | 2017-02-22 | 上海交通大学 | Short-term electric load prediction method based on deep self-encoding network |
| CN109389238A (en) * | 2017-08-14 | 2019-02-26 | 中国电力科学研究院 | A kind of short-term load forecasting method and device based on ridge regression |
| CN108229754A (en) * | 2018-01-31 | 2018-06-29 | 杭州电子科技大学 | Short-term load forecasting method based on similar day segmentation and LM-BP networks |
| CN109034490A (en) * | 2018-08-13 | 2018-12-18 | 广东工业大学 | A kind of Methods of electric load forecasting, device, equipment and storage medium |
| CN110048420A (en) * | 2019-05-15 | 2019-07-23 | 广东电网有限责任公司 | A kind of method, apparatus and medium of the scheduling of power distribution network random optimization |
Non-Patent Citations (1)
| Title |
|---|
| 章国勇.风电功率预测区间评估及并网发电调度研究.《中国博士学位论文全文数据库》.2016,第1-74页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110648248A (en) | 2020-01-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104899431B (en) | Based on ant colony and swarm of particles into LSSVM fluctuating wind speed Forecasting Methodologies | |
| CN110232203B (en) | Knowledge distillation optimization RNN short-term power failure prediction method, storage medium and equipment | |
| CN112633316B (en) | Load prediction method and device based on boundary estimation theory | |
| CN109376772B (en) | Power load combination prediction method based on neural network model | |
| CN108429256B (en) | Power system operation optimization method and terminal equipment | |
| CN104951834A (en) | LSSVM (least squares support vector machine) wind speed forecasting method based on integration of GA (genetic algorithm) and PSO (particle swarm optimization) | |
| CN108171379B (en) | Power load prediction method | |
| CN104252649A (en) | Regional wind power output prediction method based on correlation between multiple wind power plants | |
| CN105809349B (en) | Dispatching method for step hydropower station group considering incoming water correlation | |
| CN112182720B (en) | Building energy consumption model evaluation method based on building energy management application scene | |
| CN105488592A (en) | Method for predicting generated energy of photovoltaic power station | |
| CN113222403B (en) | Big data-based power regulation method and device, storage medium and electronic equipment | |
| CN109492748A (en) | A kind of Mid-long term load forecasting method for establishing model of the electric system based on convolutional neural networks | |
| CN103838979A (en) | Statistics downscaling method based on SVM algorithm | |
| CN105224738A (en) | LSSVM non-gaussian fluctuating wind speed Forecasting Methodology | |
| CN110648248B (en) | Control method, device and equipment for power station | |
| CN115759467A (en) | Time-division integrated learning photovoltaic prediction method for error correction | |
| CN115099511A (en) | Photovoltaic power probability estimation method and system based on optimized copula | |
| CN110210677B (en) | Bus short-term daily load prediction method and device combining clustering and deep learning algorithm | |
| CN114266186B (en) | Air conditioner energy consumption data missing value filling method, equipment and medium | |
| CN114065646B (en) | Energy consumption prediction method based on hybrid optimization algorithm, cloud computing platform and system | |
| CN115358441A (en) | New energy cluster consumption intelligent control method and device based on federal learning | |
| CN115130764A (en) | Power distribution network situation prediction method and system based on state evaluation | |
| CN114021847A (en) | Short-term power load prediction method and system based on ISGU (inverse transformation and aggregation) hybrid model | |
| CN113051833B (en) | Flood-reservoir mapping relation simulation method for deep learning guided by physical mechanism |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |