CN112070268A - Power load prediction method and device based on hotel demand side response - Google Patents
Power load prediction method and device based on hotel demand side response Download PDFInfo
- Publication number
- CN112070268A CN112070268A CN202010759919.4A CN202010759919A CN112070268A CN 112070268 A CN112070268 A CN 112070268A CN 202010759919 A CN202010759919 A CN 202010759919A CN 112070268 A CN112070268 A CN 112070268A
- Authority
- CN
- China
- Prior art keywords
- day
- load
- power
- prediction
- power load
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 230000004044 response Effects 0.000 title claims abstract description 23
- 238000012545 processing Methods 0.000 claims abstract description 55
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 25
- 238000012806 monitoring device Methods 0.000 claims abstract description 15
- 238000004364 calculation method Methods 0.000 claims abstract description 13
- 230000005611 electricity Effects 0.000 claims description 40
- 238000011049 filling Methods 0.000 claims description 20
- 238000007781 pre-processing Methods 0.000 claims description 13
- 238000013519 translation Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 7
- 230000009466 transformation Effects 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 5
- 238000012544 monitoring process Methods 0.000 claims description 3
- 238000013277 forecasting method Methods 0.000 claims 2
- 230000006870 function Effects 0.000 description 27
- 230000008859 change Effects 0.000 description 6
- 238000010438 heat treatment Methods 0.000 description 6
- 238000001556 precipitation Methods 0.000 description 5
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000001816 cooling Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000004083 survival effect Effects 0.000 description 2
- 238000009423 ventilation Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000011068 loading method Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification 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
- G06Q10/04—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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- 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—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/12—Hotels or restaurants
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/003—Load forecast, e.g. methods or systems for forecasting future load demand
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- 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
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Tourism & Hospitality (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Entrepreneurship & Innovation (AREA)
- Power Engineering (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention provides a power load prediction method and a power load prediction device based on hotel demand side response, which comprise the following steps of; generating a day eigenvector according to factors influencing the hotel power load; calculating the distance between the day characteristic vector of the predicted day and the day characteristic vector of the historical day, and selecting the historical similar day of the predicted day according to the calculation result; generating a power load curve by a load monitoring device deployed in a hotel; performing wavelet decomposition on the power load curve to obtain a sub-load prediction result corresponding to each power utilization unit in the hotel; and carrying out error processing on the sub-load prediction results, and adding the sub-load prediction results subjected to error processing to obtain a total load prediction result. Aiming at the particularity of complex types of the power utilization units of the hotel, the prediction result obtained by wavelet decomposition is subjected to further error processing, and compared with the method that the prediction loads of all the power utilization units of the hotel are simply added, more accurate adjustment is realized through error processing, and the prediction precision is improved.
Description
Technical Field
The invention belongs to the field of load prediction, and particularly relates to a power load prediction method and device based on hotel demand side response.
Background
The demand side response means that a series of incentive mechanisms are implemented, and the power price is adjusted to guide a user to actively carry out matching management on the operation of the power distribution network, so that the problem that the traditional power distribution network cannot actively generate power according to the demand of the user is solved, and the stability of the power supply demand is improved.
The power load prediction has an important role in the demand side response, and a power supply party can adjust the demand side response in time according to the result of the power load prediction to balance the power supply load of the power distribution network. The current ultra-short term load prediction mainly takes artificial intelligence and wavelet analysis as main components, although a short prediction period can be realized, the prediction precision is low, and particularly for power load prediction in hotel application scenes, the prediction accuracy requirement cannot be met due to complex interference factors.
Disclosure of Invention
In order to solve the defects in the prior art, the invention provides a power load prediction method based on hotel demand side response, which comprises the following steps:
selecting day characteristics according to factors influencing the hotel power load, and generating a day characteristic vector;
calculating the distance between the day characteristic vector of the predicted day and the day characteristic vector of the historical day, and selecting the historical similar day of the predicted day according to the calculation result;
recording real-time power loads of historical similar days by a load monitoring device deployed in a hotel to generate a power load curve;
performing wavelet decomposition on the power load curve to obtain a sub-load prediction result corresponding to each power utilization unit in the hotel;
and carrying out error processing on the sub-load prediction results, and adding the sub-load prediction results subjected to error processing to obtain a total load prediction result.
Optionally, the selecting a day feature according to factors affecting the hotel power load and generating a day feature vector includes:
determining corresponding day characteristics according to factors influencing the hotel power load;
and analyzing the closeness degree of the historical day and the prediction day to the features of each day, and generating a day feature vector according to the analysis result.
Optionally, the calculating a distance between the day feature vector of the predicted day and the day feature vector of the historical day, and selecting the historical similar day of the predicted day according to the calculation result includes:
calculating the Euclidean distance between the day characteristic vector of the prediction day and the day characteristic vector of the historical day;
and if the Euclidean distance is smaller than a preset similarity threshold value, selecting the historical date corresponding to the Euclidean distance as the historical similar date of the prediction date.
Optionally, the performing wavelet decomposition on the power load curve corresponding to the historical similar day to obtain the sub-load prediction result corresponding to each power consumption unit in the hotel includes:
respectively carrying out wavelet decomposition on the power load curve of each power utilization unit on historical similar days through a discrete wavelet transformation functionComprises the following steps:
wherein a is a scaling factor in a discrete wavelet transform function, b is a translation factor in the discrete wavelet transform function, and a and b are both manually set; Δ t is the sampling interval to the power load curve;a power load curve for wavelet decomposition; n is the total number of samples, k is the sample number; a. the value ranges of b, delta t and t are positive numbers, and the value ranges of k and n are positive integers;
according to different values of a and b, load prediction results decomposed by the power utilization unit i based on different historical similar days t are obtained
Will belong to the same electricity consuming unitAdding to obtain the sub-load prediction result x of the electricity utilization unit i corresponding to the prediction dayi。
Further, the performing error processing on the sub-load prediction results, and adding the sub-load prediction results after the error processing to obtain a total load prediction result includes:
performing error processing on each sub-power load by an error processing function f, wherein the error processing function is as follows:
αiis the weight, x 'of the electricity unit i in the predicted day'iThe actual sub-load of the electricity utilization unit i at the predicted time of day; alpha is alphai、xiAnd x'iThe value ranges of i and t are positive integers;
and f, acquiring the minimum value of f, and obtaining the total load prediction result of each power utilization unit in the prediction time of day after error processing.
Specifically, the error processing function further comprises the weight alpha of the power utilization unit i in the prediction dayiThe calculation method of (2) comprises:
wherein the absolute errorn is the number of days of a historical similar day,is the predicted load x 'of the pre-stored power consumption unit i on the historical similar day j'i,jThe actual load of the electricity utilization unit i on the historical similar day j; said Eit、And x'it,jThe value ranges of (a) and (b) are positive integers.
Optionally, the power load prediction method further includes performing a data preprocessing process on power load curves of historical similar days, where the data preprocessing process includes:
analyzing the data missing condition of the power load curve, and if the missing data does not exceed a preset threshold, performing data filling according to a filling formula, wherein the filling formula is as follows:
Dn+jfor missing data, it is the (n + j) th data in the power load curve, DnFor the nth data in the power data set, Dn+iThe data is the n + i th data in the power load curve, and the value of n is manually selected according to experience; the value ranges of i, j and n are positive integers, and i is>j,Dn、Dn+jAnd Dn+iThe value range of (a) is positive.
The invention also provides a prediction device of power load based on hotel demand side response based on the same invention thought, which comprises:
a vector generation unit: the system is used for selecting day characteristics according to factors influencing the hotel power load and generating a day characteristic vector;
the similar day selecting unit: the system is used for calculating the distance between the day characteristic vector of the predicted day and the day characteristic vector of the historical day and selecting the historical similar day of the predicted day according to the calculation result;
a load monitoring unit: the system comprises a load monitoring device, a power load curve generation device and a power load monitoring device, wherein the load monitoring device is deployed in a hotel and is used for recording real-time power loads of historical similar days and generating the power load curve;
a prediction unit: the power load curve is subjected to wavelet decomposition to obtain a sub-load prediction result corresponding to each power utilization unit in the hotel;
an error unit: and the load prediction unit is used for carrying out error processing on the sub-load prediction results and adding the sub-load prediction results after error processing to obtain a total load prediction result.
Optionally, the prediction unit is specifically configured to:
respectively carrying out wavelet decomposition on the power load curve of each power utilization unit on historical similar days through a discrete wavelet transformation functionComprises the following steps:
wherein a is a scaling factor in a discrete wavelet transform function, b is a translation factor in the discrete wavelet transform function, and a and b are both manually set; Δ t is the sampling interval to the power load curve;a power load curve for wavelet decomposition; n is the total number of samples, k is the sample number; a. the value ranges of b, delta t and t are positive numbers, and the value ranges of k and n are positive integers;
according to different values of a and b, load prediction results decomposed by the power utilization unit i based on different historical similar days t are obtained
Will belong to the same electricity consuming unitAdding to obtain the sub-load prediction result x of the electricity utilization unit i corresponding to the prediction dayi。
Further, the error unit is specifically configured to:
performing error processing on each sub-power load by an error processing function f, wherein the error processing function is as follows:
αiis the weight, x 'of the electricity unit i in the predicted day'iThe actual sub-load of the electricity utilization unit i at the predicted time of day; alpha is alphai、xiAnd x'iThe value ranges of i and t are positive integers;
and f, acquiring the minimum value of f, and obtaining the total load prediction result of each power utilization unit in the prediction time of day after error processing.
The technical scheme provided by the invention has the beneficial effects that:
on the basis of the traditional short-term power load prediction, aiming at the complex particularity of the types of the electric units of the hotel, the prediction result obtained by wavelet decomposition is subjected to further error processing. Compared with the method that the predicted loads of all power utilization units of a hotel are simply added, more accurate adjustment is achieved through error processing, and prediction accuracy is improved.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced 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 based on the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a power load prediction method based on hotel demand side response according to the present invention;
fig. 2 is a block diagram of a power load prediction device based on hotel demand side response according to the present invention.
Detailed Description
To make the structure and advantages of the present invention clearer, the structure of the present invention will be further described with reference to the accompanying drawings.
Example one
As shown in fig. 1, the present invention provides a power load prediction method based on hotel demand side response, including:
s1: and selecting day characteristics according to factors influencing the hotel power load, and generating a day characteristic vector.
According to the factors influencing the hotel power load, determining the corresponding day characteristics, if the day characteristics are working days, air temperatures, rainfall conditions and the like, in the embodiment, the setting of the day characteristics comprises the following steps: according to the date type, the forecast day and the historical day are classified into a weekday type, and the Monday to Saturday type are classified into another type, and the check-in rate of the hotel on the weekdays and other times is judged to have obvious difference according to experience, and the check-in rate indirectly influences the change of the power load of the hotel. The maximum air temperature and the minimum air temperature of the prediction day and the historical day are divided into four categories according to the temperature of less than 0 ℃, 0-10 ℃, 10-20 ℃ and 20-30 ℃, and the air temperature can influence the working conditions of heating and cooling systems of a hotel, such as air conditioners, heating systems and other equipment, so that the power load is changed. Classifying the precipitation conditions of the forecast day and the historical day into no precipitation, light rain, medium rain and heavy rain based on the precipitation amountRain is in four categories, and the rainfall condition influences the travel condition of people, so that the survival rate of the hotel is influenced, and the change of the power load of the hotel is indirectly influenced. Based on the above-mentioned division of the day features, a day feature vector X ═ (Date, T) is generatedmax,TminRain), wherein Date, Tmax、TminAnd Rain sequentially represents the membership degrees of different categories of the date type, the highest temperature of the day, the lowest temperature of the day and the precipitation condition of the forecast day and the historical day. The membership degree is a matrix, is calculated based on a fuzzy C-means algorithm, and represents the degree of the prediction day and the historical day belonging to each category on a certain day characteristic.
Compared with a hard clustering method, the fuzzy clustering idea provides a more scientific classification method, and a more objective classification result can be obtained by calculating the membership degree, so that the generated daily feature vector is more accurate.
S2: and calculating the distance between the day characteristic vector of the predicted day and the day characteristic vector of the historical day, and selecting the historical similar day of the predicted day according to the calculation result.
And analyzing the distance between the day feature vector of the predicted day and the day feature vector of the historical day, namely calculating the Euclidean distance between the day feature vector of the predicted day and the day feature vector of the historical day. If the Euclidean distance is smaller than a preset similarity threshold, the fact that the day characteristic type corresponding to the historical day is similar to the predicted day is shown, the power load predicted by taking the power load curve of the historical day as a sample is supposed to be the actual power load closest to the predicted day, and therefore the historical day is judged to be the historical similar day of the predicted day.
In many historical days, the accuracy of the prediction result may be reduced due to the fact that day characteristics of some historical days are greatly different from those of the prediction days, so that history similar days which are closer to each other are screened out, and the power load prediction accuracy of the prediction days is improved.
S3: a power load curve is generated by recording real-time power loads of historical similar days through a load monitoring device deployed in a hotel.
The method comprises the steps of deploying corresponding load monitoring devices such as power load control terminals and the like in each power utilization loop of the hotel building in advance, recording real-time power loads of historical similar days through the load monitoring devices, and generating a power load curve according to recorded real-time data, wherein the abscissa is a time axis, and the ordinate is a load value of the historical similar days at a certain time point.
S4: and performing wavelet decomposition on the power load curve to obtain a sub-load prediction result corresponding to each power utilization unit in the hotel.
Since there is generally more than one history similar day obtained by S2, there should be several power load curves corresponding to the history similar days, and in this case, it is difficult to decide which curve should be used for prediction at the same time point, so that wavelet decomposition is adopted, the power load curves corresponding to each history similar day are decomposed by a discrete wavelet transform function to obtain a more minute time period, and frequency conversion processing is performed to decompose a low-frequency curve and a high-frequency curve. The low-frequency curve mainly represents the basic load and has strong regularity and periodicity, for example, when the previous prediction day and the months and weeks of the historical day can be aligned, the low-frequency curves of the previous prediction day and the historical day are close to each other; the high-frequency curve mainly represents sudden load, when the forecast day and the historical day are in the aspects of weather, time interval air temperature and the like, the uncertainty is large, and the part of the high-frequency curve with sudden change is relatively close.
wherein a is a scaling factor in a discrete wavelet transform function, b is a translation factor in the discrete wavelet transform function, and a and b are both manually set; Δ t isSampling intervals for the power load curve;a power load curve for wavelet decomposition; n is the total number of samples, k is the sample number; a. the value ranges of b, delta t and t are positive numbers, and the value ranges of k and n are positive integers;
according to different values of a and b, load prediction results decomposed by the power utilization unit i based on different historical similar days t are obtainedAnd c, selecting the expansion degree of the curve by changing the value of a, wherein the contracted width corresponds to high frequency, the expanded width corresponds to low frequency, and the curve is subjected to translation processing by changing the value of b. Will belong to the same electricity consuming unitAdding to obtain the sub-load prediction result x of the electricity utilization unit i corresponding to the prediction dayi。
In practical application, not all the characteristics of the selected historical similar days are similar to the predicted days, for example, although the date type of one historical similar day is the same as that of the predicted day, the rainfall condition of the historical similar day is different from that of the predicted day, so that a low-frequency part can be decomposed by a discrete wavelet transform function in load prediction; at the same time, although the date type of the other historical similar day is different from that of the forecast day, the rainfall condition of the historical similar day is the same as that of the forecast day, so that the high-frequency part can be decomposed by a discrete wavelet transform function in load forecast, and the low-frequency part and the high-frequency part are added to restore a forecast curve, namely the power load curve x of the forecast dayiThe abscissa is a time axis, and the ordinate is a predicted value of the power load.
S5: and carrying out error processing on the sub-load prediction results, and adding the sub-load prediction results subjected to error processing to obtain a total load prediction result.
Due to the wide variety of hotel power loads, for the electricity consumption particularity of the building, electricity consumption metering can be performed in each regional loop with different functions of the hotel according to different branch tables, for example, electricity consumption metering devices are respectively arranged on electricity consumption units such as a heating ventilation system, an air conditioner power supply system, a kitchen power supply system and an illumination electricity consumption loop of the hotel, and a plurality of sub-power load prediction curves corresponding to the electricity consumption units are obtained through the wavelet transformation function described in S4. Theoretically, the final total load prediction result can be obtained by adding the obtained plurality of power load prediction curves, but actually, the sum of the total load and the sub-load has a deviation, so that further correction is needed to perform error processing on the sub-load prediction result.
The error processing is to perform error processing on each sub-power load through an error processing function f, where the error processing function is:
αiis the weight, x 'of the electricity unit i in the predicted day'iFor actual subpooling of electricity utilization unit i at predicted time of day
Loading; alpha is alphai、xiAnd x'iThe value ranges of i and t are positive integers;
and f, acquiring the minimum value of f, and obtaining the total load prediction result of each power utilization unit in the prediction time of day after error processing.
Wherein the weight alpha of the power utilization unit i in the prediction day is also includediThe calculation method of (2) comprises:
wherein the absolute errorn is the number of days of a historical similar day,is the predicted load x 'of the pre-stored power consumption unit i on the historical similar day j'i,jThe actual load of the electricity utilization unit i on the historical similar day j; said Eit、And x'it,jThe value ranges of (a) and (b) are positive integers.
The power load prediction method further comprises the step of preprocessing data of power load curves of historical similar days, wherein the data in the power load curves of the historical similar days are acquired by power consumption metering devices deployed in each power consumption unit of a hotel, so that the data are influenced by the devices, and the data may have large errors, so that the data are preprocessed during the power load prediction through wavelet decomposition, and the data preprocessing step comprises the following steps:
analyzing the data missing condition of the power load curve, and if the missing data does not exceed a preset threshold, performing data filling according to a filling formula, wherein the filling formula is as follows:
Dn+jfor missing data, it is the (n + j) th data in the power load curve, DnFor the nth data in the power data set, Dn+iThe data is the n + i th data in the power load curve, and the value of n is manually selected according to experience; the value ranges of i, j and n are positive integers, and i is>j,Dn、Dn+jAnd Dn+iThe value range of (a) is positive.
For example, there is a set of power load curves containing data (1, 3, 5, x, 9, 11), a fourth data T4If only one of the 7 data is missing, n is preset to 1, i is preset to 6, i is preset to 3, and then j is calculated according to the filling formula, assuming that the preset threshold is not exceeded, i.e. the missing data is not largeI.e., 5 fills at x, is closer to the expected 7. However, if the missing data exceeds the preset threshold, or the proportion of the missing data to all the data in the power load curve exceeds the preset proportion, that is, the missing data is excessive, at this time, if the data filling is still performed by using the filling formula, the power load curve on the historical similar day may be inaccurate, and a large error may be generated in the subsequent prediction, so the power load curve with the excessive missing data is regarded as being unsatisfactory and is not used as an input sample of the power load prediction.
Meanwhile, the data preprocessing process also comprises the step of filtering out obviously wrong power load curves in history similar days by presetting a certain value range, so that the influence of wrong data contained in the power load curves on the prediction precision of the power load is avoided.
By carrying out data preprocessing on the power load curves of historical similar days and carrying out appropriate data filling in the range where the error can be received, the power load curves serving as input samples can be further improved, and the accuracy of subsequent prediction results can be improved.
Example two
As shown in fig. 2, the present invention further provides an electric load prediction apparatus 5 based on hotel demand side response, including:
the vector generation unit 51: and the method is used for selecting the day characteristics according to the factors influencing the hotel power load and generating the day characteristic vector.
According to the factors influencing the hotel power load, determining the corresponding day characteristics, if the day characteristics are working days, air temperatures, rainfall conditions and the like, in the embodiment, the setting of the day characteristics comprises the following steps: according to the date type, the forecast day and the historical day are classified into a weekday type, and the Monday to Saturday type are classified into another type, and the check-in rate of the hotel on the weekdays and other times is judged to have obvious difference according to experience, and the check-in rate indirectly influences the change of the power load of the hotel. The highest temperature and the lowest temperature of the prediction day and the historical day are divided into four categories according to the temperature of less than 0 ℃, 0-10 ℃, 10-20 ℃ and 20-30 ℃, and the temperature can influence the hotelThe operating conditions of heating and cooling systems, such as air conditioners, heating systems, etc., further cause changes in electrical loads. The rainfall conditions of the prediction day and the historical day are divided into four categories of no rainfall, light rain, medium rain and heavy rain based on the rainfall amount, and the falling rainfall conditions influence the trip conditions of people, so that the survival rate of the hotel is influenced, and the change of the power load of the hotel is further influenced indirectly. Based on the above-mentioned division of the day features, a day feature vector X ═ (Date, T) is generatedmax,TminRain), wherein Date, Tmax、TminAnd Rain sequentially represents the membership degrees of different categories of the date type, the highest temperature of the day, the lowest temperature of the day and the precipitation condition of the forecast day and the historical day. The membership degree is a matrix, is calculated based on a fuzzy C-means algorithm, and represents the degree of the prediction day and the historical day belonging to each category on a certain day characteristic.
Compared with a hard clustering method, the fuzzy clustering idea provides a more scientific classification method, and a more objective classification result can be obtained by calculating the membership degree, so that the generated daily feature vector is more accurate.
Similar day selecting unit 52: and the distance between the day characteristic vector of the predicted day and the day characteristic vector of the historical day is calculated, and the historical similar day of the predicted day is selected according to the calculation result.
And analyzing the distance between the day feature vector of the predicted day and the day feature vector of the historical day, namely calculating the Euclidean distance between the day feature vector of the predicted day and the day feature vector of the historical day. If the Euclidean distance is smaller than a preset similarity threshold, the fact that the day characteristic type corresponding to the historical day is similar to the predicted day is shown, the power load predicted by taking the power load curve of the historical day as a sample is supposed to be the actual power load closest to the predicted day, and therefore the historical day is judged to be the historical similar day of the predicted day.
In many historical days, the accuracy of the prediction result may be reduced due to the fact that day characteristics of some historical days are greatly different from those of the prediction days, so that history similar days which are closer to each other are screened out, and the power load prediction accuracy of the prediction days is improved.
The load monitoring unit 53: the method is used for recording real-time power loads of historical similar days through a load monitoring device deployed in a hotel and generating a power load curve.
The method comprises the steps of deploying corresponding load monitoring devices such as power load control terminals and the like in each power utilization loop of the hotel building in advance, recording real-time power loads of historical similar days through the load monitoring devices, and generating a power load curve according to recorded real-time data, wherein the abscissa is a time axis, and the ordinate is a load value of the historical similar days at a certain time point.
The prediction unit 54: the method is used for performing wavelet decomposition on the power load curves corresponding to historical similar days to obtain the sub-load prediction results corresponding to each power utilization unit in the hotel.
Because more than one historical similar day is obtained by the similar day selecting unit 52, the power load curves corresponding to the historical similar days also have a plurality of curves, and in this case, the curve which is used as the basis for predicting the same time point is difficult to decide, so that wavelet decomposition is adopted, the power load curves corresponding to each historical similar day are decomposed by a discrete wavelet transform function respectively to obtain a more tiny time period, and frequency conversion processing is carried out to decompose a low-frequency curve and a high-frequency curve. The low-frequency curve mainly represents the basic load and has strong regularity and periodicity, for example, when the previous prediction day and the months and weeks of the historical day can be aligned, the low-frequency curves of the previous prediction day and the historical day are close to each other; the high-frequency curve mainly represents sudden load, when the forecast day and the historical day are in the aspects of weather, time interval air temperature and the like, the uncertainty is large, and the part of the high-frequency curve with sudden change is relatively close.
wherein a is a scaling factor in a discrete wavelet transform function, b is a translation factor in the discrete wavelet transform function, and a and b are both manually set; Δ t is the sampling interval to the power load curve;a power load curve for wavelet decomposition; n is the total number of samples, k is the sample number; a. the value ranges of b, delta t and t are positive numbers, and the value ranges of k and n are positive integers;
according to different values of a and b, load prediction results decomposed by the power utilization unit i based on different historical similar days t are obtainedAnd c, selecting the expansion degree of the curve by changing the value of a, wherein the contracted width corresponds to high frequency, the expanded width corresponds to low frequency, and the curve is subjected to translation processing by changing the value of b. Will belong to the same electricity consuming unitAdding to obtain the sub-load prediction result x of the electricity utilization unit i corresponding to the prediction dayi。
In practical application, not all the characteristics of the selected historical similar days are similar to the predicted days, for example, although the date type of one historical similar day is the same as that of the predicted day, the rainfall condition of the historical similar day is different from that of the predicted day, so that a low-frequency part can be decomposed by a discrete wavelet transform function in load prediction; at the same time, although the date type of the other historical similar day is different from that of the forecast day, the rainfall condition of the historical similar day is the same as that of the forecast day, so that the high-frequency part can be decomposed by a discrete wavelet transform function in load forecast, and the low-frequency part and the high-frequency part are added to restore a forecast curve, namely the power load curve x of the forecast dayiWith the abscissa being the time axis, ordinateLabeled as the predicted value of the power load.
Error unit 55: and the load prediction unit is used for carrying out error processing on the sub-load prediction results and adding the sub-load prediction results after error processing to obtain a total load prediction result.
Due to the wide variety of hotel power loads, for the electricity consumption particularity of the building, electricity consumption metering can be performed in each regional loop with different functions of the hotel according to different branch tables, for example, electricity consumption metering devices are respectively arranged on electricity consumption units such as a heating ventilation system, an air conditioner power supply system, a kitchen power supply system and an illumination electricity consumption loop of the hotel, and a plurality of sub-power load prediction curves corresponding to the electricity consumption units are obtained through a wavelet transformation function described in the prediction unit 54. Theoretically, the final total load prediction result can be obtained by adding the obtained plurality of power load prediction curves, but actually, the sum of the total load and the sub-load has a deviation, so that further correction is needed to perform error processing on the sub-load prediction result.
The error processing is to perform error processing on each sub-power load through an error processing function f, where the error processing function is:
αiis the weight, x 'of the electricity unit i in the predicted day'iThe actual sub-load of the electricity utilization unit i at the predicted time of day; alpha is alphai、xiAnd x'iThe value ranges of i and t are positive integers;
and f, acquiring the minimum value of f, and obtaining the total load prediction result of each power utilization unit in the prediction time of day after error processing.
Wherein the weight alpha of the power utilization unit i in the prediction day is also includediThe calculation method of (2) comprises:
wherein the absolute errorn is the number of days of a historical similar day,is the predicted load x 'of the pre-stored power consumption unit i on the historical similar day j'i,jThe actual load of the electricity utilization unit i on the historical similar day j; said Eit、And x'it,jThe value ranges of (a) and (b) are positive integers.
The power load prediction device 5 further includes a preprocessing unit for preprocessing the power load curves of the historical similar days, and since the data in the power load curves of the historical similar days are collected by the power consumption metering devices deployed in the power consumption units of the hotel, the data are influenced by the device itself and have a large error, the data are preprocessed during the power load prediction by wavelet decomposition, and the data preprocessing process includes:
analyzing the data missing condition of the power load curve, and if the missing data does not exceed a preset threshold, performing data filling according to a filling formula, wherein the filling formula is as follows:
Dn+jfor missing data, it is the (n + j) th data in the power load curve, DnFor the nth data in the power data set, Dn+iThe data is the n + i th data in the power load curve, and the value of n is manually selected according to experience; the value ranges of i, j and n are positive integers, and i is>j,Dn、Dn+jAnd Dn+iThe value range of (a) is positive.
For example, there is a groupData (1, 3, 5, x, 9, 11) contained in the power load curve, and fourth data T4If only one of the 7 data is missing, n is preset to 1, i is preset to 6, i is preset to 3, and then j is calculated according to the filling formula, assuming that the preset threshold is not exceeded, i.e. the missing data is not largeI.e., 5 fills at x, is closer to the expected 7. However, if the missing data exceeds the preset threshold, or the proportion of the missing data to all the data in the power load curve exceeds the preset proportion, that is, the missing data is excessive, at this time, if the data filling is still performed by using the filling formula, the power load curve on the historical similar day may be inaccurate, and a large error may be generated in the subsequent prediction, so the power load curve with the excessive missing data is regarded as being unsatisfactory and is not used as an input sample of the power load prediction.
Meanwhile, the data preprocessing process also comprises the step of filtering out obviously wrong power load curves in history similar days by presetting a certain value range, so that the influence of wrong data contained in the power load curves on the prediction precision of the power load is avoided.
By carrying out data preprocessing on the power load curves of historical similar days and carrying out appropriate data filling in the range where the error can be received, the power load curves serving as input samples can be further improved, and the accuracy of subsequent prediction results can be improved.
The sequence numbers in the above embodiments are merely for description, and do not represent the sequence of the assembly or the use of the components.
The above description is only exemplary of the present invention and should not be taken as limiting the invention, as any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. The power load prediction method based on hotel demand side response is characterized by comprising the following steps:
selecting day characteristics according to factors influencing the hotel power load, and generating a day characteristic vector;
calculating the distance between the day characteristic vector of the predicted day and the day characteristic vector of the historical day, and selecting the historical similar day of the predicted day according to the calculation result;
recording real-time power loads of historical similar days by a load monitoring device deployed in a hotel to generate a power load curve;
performing wavelet decomposition on the power load curve to obtain a sub-load prediction result corresponding to each power utilization unit in the hotel;
and carrying out error processing on the sub-load prediction results, and adding the sub-load prediction results subjected to error processing to obtain a total load prediction result.
2. The hotel demand side response-based power load prediction method of claim 1, wherein selecting a day feature according to factors affecting hotel power load and generating a day feature vector comprises:
determining corresponding day characteristics according to factors influencing the hotel power load;
and generating a day feature vector according to the day features of the historical day and the predicted day.
3. The hotel demand side response-based power load prediction method according to claim 1, wherein the step of calculating the distance between the day eigenvector of the predicted day and the day eigenvector of the historical day, and selecting the historical similar day of the predicted day according to the calculation result comprises the steps of:
calculating the Euclidean distance between the day characteristic vector of the prediction day and the day characteristic vector of the historical day;
and if the Euclidean distance is smaller than a preset similarity threshold value, selecting the historical date corresponding to the Euclidean distance as the historical similar date of the prediction date.
4. The hotel demand side response-based power load prediction method according to claim 1, wherein the performing wavelet decomposition on the power load curve to obtain a sub-load prediction result corresponding to each power unit in the hotel comprises:
respectively carrying out wavelet decomposition on the power load curve of each power utilization unit on historical similar days through a discrete wavelet transformation functionComprises the following steps:
wherein a is a scaling factor in a discrete wavelet transform function, b is a translation factor in the discrete wavelet transform function, and a and b are both manually set; Δ t is the sampling interval to the power load curve;a power load curve for wavelet decomposition; n is the total number of samples, k is the sample number; a. the value ranges of b, delta t and t are positive numbers, and the value ranges of k and n are positive integers;
according to different values of a and b, load prediction results decomposed by the power utilization unit i based on different historical similar days t are obtained
5. The hotel demand side response-based power load prediction method according to claim 1, wherein the error processing is performed on the sub-load prediction results, and the sub-load prediction results after the error processing are added to obtain a total load prediction result, and the method comprises the following steps:
performing error processing on each sub-power load by an error processing function f, wherein the error processing function is as follows:
αiis the weight of the power utilization unit i in the prediction day, xiIs the sub-load prediction result, x ', of the electricity unit i on the prediction day'iActual sub-loads of the electricity utilization unit i on the forecast day; alpha is alphai、xiAnd x'iThe value ranges of i and t are positive integers;
and f, acquiring the minimum value of f, and obtaining the total load prediction result of each power utilization unit in the prediction time of day after error processing.
6. The hotel demand side response based power load forecasting method of claim 5, wherein the error handling function further comprises a weight α of the power utilization unit i in the forecast dayiThe calculation method of (2) comprises:
wherein the absolute errorn is the number of days of a historical similar day,is the predicted load x 'of the electric unit i acquired in advance on the historical similar day j'i,jFor electricity unit i in calendarActual load of days j with similar history; said Eit、And x'i,jThe value ranges of (a) and (b) are positive integers.
7. The hotel demand side response based power load forecasting method according to claim 1, further comprising a data preprocessing process for power load curves of historical similar days, the data preprocessing process comprising:
analyzing the data missing condition of the power load curve, and if the missing data does not exceed a preset threshold, performing data filling according to a filling formula, wherein the filling formula is as follows:
Dn+jfor missing data, it is the (n + j) th data in the power load curve, DnFor the nth data in the power data set, Dn+iThe data is the n + i th data in the power load curve, and the value of n is manually selected according to experience; the value ranges of i, j and n are positive integers, and i is>j,Dn、Dn+jAnd Dn+iThe value range of (a) is positive.
8. An electrical load prediction device based on hotel demand side response, the electrical load prediction device comprising:
a vector generation unit: the system is used for selecting day characteristics according to factors influencing the hotel power load and generating a day characteristic vector;
the similar day selecting unit: the system is used for calculating the distance between the day characteristic vector of the predicted day and the day characteristic vector of the historical day and selecting the historical similar day of the predicted day according to the calculation result;
a load monitoring unit: the system comprises a load monitoring device, a power load curve generation device and a power load monitoring device, wherein the load monitoring device is deployed in a hotel and is used for recording real-time power loads of historical similar days and generating the power load curve;
a prediction unit: the power load curve is subjected to wavelet decomposition to obtain a sub-load prediction result corresponding to each power utilization unit in the hotel;
an error unit: and the load prediction unit is used for carrying out error processing on the sub-load prediction results and adding the sub-load prediction results after error processing to obtain a total load prediction result.
9. The hotel demand side response-based power load prediction device of claim 8, wherein the prediction unit is specifically configured to:
respectively carrying out wavelet decomposition on the power load curve of each power utilization unit on historical similar days through a discrete wavelet transformation functionComprises the following steps:
wherein a is a scaling factor in a discrete wavelet transform function, b is a translation factor in the discrete wavelet transform function, and a and b are both manually set; Δ t is the sampling interval to the power load curve;a power load curve for wavelet decomposition; n is the total number of samples, k is the sample number; a. the value ranges of b, delta t and t are positive numbers, and the value ranges of k and n are positive integers;
according to different values of a and b, load forecast decomposed by the power utilization unit i based on different historical similar days t is obtainedMeasurement results
10. The hotel demand side response-based power load prediction device of claim 8, wherein the error unit is specifically configured to:
performing error processing on each sub-power load by an error processing function f, wherein the error processing function is as follows:
αiis the weight of the power utilization unit i in the prediction day, xiIs the sub-load prediction result, x ', of the electricity unit i on the prediction day'iActual sub-loads of the electricity utilization unit i on the forecast day; alpha is alphai、xiAnd x'iThe value ranges of i and t are positive integers;
and f, acquiring the minimum value of f, and obtaining the total load prediction result of each power utilization unit in the prediction time of day after error processing.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010759919.4A CN112070268B (en) | 2020-07-31 | 2020-07-31 | Power load prediction method and device based on hotel demand side response |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010759919.4A CN112070268B (en) | 2020-07-31 | 2020-07-31 | Power load prediction method and device based on hotel demand side response |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112070268A true CN112070268A (en) | 2020-12-11 |
CN112070268B CN112070268B (en) | 2024-01-12 |
Family
ID=73657283
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010759919.4A Active CN112070268B (en) | 2020-07-31 | 2020-07-31 | Power load prediction method and device based on hotel demand side response |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112070268B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112990597A (en) * | 2021-03-31 | 2021-06-18 | 国家电网有限公司 | Ultra-short-term prediction method for industrial park factory electrical load |
CN113919600A (en) * | 2021-12-08 | 2022-01-11 | 国网湖北省电力有限公司经济技术研究院 | Resident load ultra-short term prediction method |
CN115829152A (en) * | 2022-12-21 | 2023-03-21 | 杭州易龙电安科技有限公司 | Power load prediction method, device and medium based on machine learning algorithm |
CN117272121A (en) * | 2023-11-21 | 2023-12-22 | 江苏米特物联网科技有限公司 | Hotel load influence factor quantitative analysis method based on Deep SHAP |
CN117370770A (en) * | 2023-12-08 | 2024-01-09 | 江苏米特物联网科技有限公司 | Hotel load comprehensive prediction method based on shape-XGboost |
CN117951555A (en) * | 2024-03-27 | 2024-04-30 | 山东德源电力科技股份有限公司 | Strategy formulation method for load control installation testing device |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101923139A (en) * | 2010-04-19 | 2010-12-22 | 昆明理工大学 | Intelligent method for single-ended traveling wave fault location of power transmission line |
CN104820876A (en) * | 2015-05-21 | 2015-08-05 | 国家电网公司 | Short-term load forecasting method and system |
RU2600099C1 (en) * | 2015-03-23 | 2016-10-20 | Юрий Анатольевич Кропотов | Method of neural network forecasting of change of values of function with its complementary wavelet processing and device for its implementation |
CN106203683A (en) * | 2016-06-29 | 2016-12-07 | 西安美林数据技术股份有限公司 | A kind of modeling method of power customer electro-load forecast system |
CN110991815A (en) * | 2019-11-12 | 2020-04-10 | 国网电动汽车服务有限公司 | Distribution room power energy scheduling method and system |
-
2020
- 2020-07-31 CN CN202010759919.4A patent/CN112070268B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101923139A (en) * | 2010-04-19 | 2010-12-22 | 昆明理工大学 | Intelligent method for single-ended traveling wave fault location of power transmission line |
RU2600099C1 (en) * | 2015-03-23 | 2016-10-20 | Юрий Анатольевич Кропотов | Method of neural network forecasting of change of values of function with its complementary wavelet processing and device for its implementation |
CN104820876A (en) * | 2015-05-21 | 2015-08-05 | 国家电网公司 | Short-term load forecasting method and system |
CN106203683A (en) * | 2016-06-29 | 2016-12-07 | 西安美林数据技术股份有限公司 | A kind of modeling method of power customer electro-load forecast system |
CN110991815A (en) * | 2019-11-12 | 2020-04-10 | 国网电动汽车服务有限公司 | Distribution room power energy scheduling method and system |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112990597A (en) * | 2021-03-31 | 2021-06-18 | 国家电网有限公司 | Ultra-short-term prediction method for industrial park factory electrical load |
CN112990597B (en) * | 2021-03-31 | 2024-02-27 | 国家电网有限公司 | Ultra-short-term prediction method for industrial park power consumption load |
CN113919600A (en) * | 2021-12-08 | 2022-01-11 | 国网湖北省电力有限公司经济技术研究院 | Resident load ultra-short term prediction method |
CN113919600B (en) * | 2021-12-08 | 2022-04-01 | 国网湖北省电力有限公司经济技术研究院 | Resident load ultra-short term prediction method |
CN115829152A (en) * | 2022-12-21 | 2023-03-21 | 杭州易龙电安科技有限公司 | Power load prediction method, device and medium based on machine learning algorithm |
CN115829152B (en) * | 2022-12-21 | 2023-07-07 | 杭州易龙电安科技有限公司 | Power load prediction method, device and medium based on machine learning algorithm |
CN117272121A (en) * | 2023-11-21 | 2023-12-22 | 江苏米特物联网科技有限公司 | Hotel load influence factor quantitative analysis method based on Deep SHAP |
CN117272121B (en) * | 2023-11-21 | 2024-03-12 | 江苏米特物联网科技有限公司 | Hotel load influence factor quantitative analysis method based on Deep SHAP |
CN117370770A (en) * | 2023-12-08 | 2024-01-09 | 江苏米特物联网科技有限公司 | Hotel load comprehensive prediction method based on shape-XGboost |
CN117370770B (en) * | 2023-12-08 | 2024-02-13 | 江苏米特物联网科技有限公司 | Hotel load comprehensive prediction method based on shape-XGboost |
CN117951555A (en) * | 2024-03-27 | 2024-04-30 | 山东德源电力科技股份有限公司 | Strategy formulation method for load control installation testing device |
Also Published As
Publication number | Publication date |
---|---|
CN112070268B (en) | 2024-01-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112070268A (en) | Power load prediction method and device based on hotel demand side response | |
Dash et al. | A real-time short-term load forecasting system using functional link network | |
CN110610280A (en) | Short-term prediction method, model, device and system for power load | |
CN109934386B (en) | Heat load prediction method for combined heat and power system | |
CN110033134A (en) | A kind of short-term load forecasting algorithm of segmentation day by day considering meteorologic factor | |
CN107423836B (en) | Short-term load prediction method based on body sensing temperature | |
CN110782094B (en) | Subentry load prediction method for fine-grained electricity consumption behaviors of residential users | |
CN110610121B (en) | Small-scale source load power abnormal data identification and restoration method based on curve clustering | |
Yu et al. | Load forecasting based on smart meter data and gradient boosting decision tree | |
CN112257962B (en) | Method and device for predicting line loss of transformer area | |
CN114372360A (en) | Method, terminal and storage medium for power load prediction | |
CN110991815A (en) | Distribution room power energy scheduling method and system | |
CN116227637A (en) | Active power distribution network oriented refined load prediction method and system | |
Ku et al. | Building electric energy prediction modeling for BEMS using easily obtainable weather factors with Kriging model and data mining | |
CN110489893B (en) | Variable weight-based bus load prediction method and system | |
CN116307111A (en) | Reactive load prediction method based on K-means clustering and random forest algorithm | |
de Vilmarest et al. | Adaptive probabilistic forecasting of electricity (net-) load | |
JPH0830581A (en) | Method for predicting quantity of demand | |
CN114997470A (en) | Short-term power load prediction method based on LSTM neural network | |
Imani | Deep learning based electricity demand forecasting in different domains | |
CN112561153A (en) | Scenic spot crowd gathering prediction method based on model integration | |
Chatterjee et al. | Prediction of Household-level Heat-Consumption using PSO enhanced SVR Model | |
CN112488418A (en) | Full-topology load prediction method and device and computer equipment | |
CN116233312B (en) | Regression-decomposition-based power grid customer service traffic prediction method | |
CN117955094B (en) | Power load prediction method and system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |