CN104636822A - Residential load prediction method of elman-based neural network - Google Patents
Residential load prediction method of elman-based neural network Download PDFInfo
- Publication number
- CN104636822A CN104636822A CN201510031651.1A CN201510031651A CN104636822A CN 104636822 A CN104636822 A CN 104636822A CN 201510031651 A CN201510031651 A CN 201510031651A CN 104636822 A CN104636822 A CN 104636822A
- Authority
- CN
- China
- Prior art keywords
- neural network
- data
- resident load
- day
- elman
- 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
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 103
- 238000000034 method Methods 0.000 title claims abstract description 32
- 230000001932 seasonal effect Effects 0.000 claims abstract description 36
- 210000002569 neuron Anatomy 0.000 claims abstract description 11
- 230000000153 supplemental effect Effects 0.000 claims description 27
- 238000012549 training Methods 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 4
- 210000004205 output neuron Anatomy 0.000 claims description 4
- 230000005611 electricity Effects 0.000 description 4
- 238000013507 mapping Methods 0.000 description 3
- 230000001537 neural effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
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
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a residential load prediction method of an elman-based neural network. The method comprises the following steps: acquiring residential load historical data of last year and corresponding historical weather parameter data; calculating a seasonal index of the residential load of each month; correcting the residential load historical data by using the seasonal index; determining input and output data of a neural network and determining an optimal hidden layer neuron number so as to establish the elman-based neural network; normalizing the corrected residential load historical data and the corresponding historical weather parameter data to further train the established neural network and control the prediction error in a preset range; predicting the residential load by using the trained neural network. The method has the capacity of adapting to time-variant characteristic and seasonal fluctuation of the residential load, the dynamic characteristic of the residential load can be directly predicted and reflected, the prediction accuracy is high, and the method can be widely applied to the charge prediction field of a power system.
Description
Technical field
The present invention relates to Load Prediction In Power Systems technical field, particularly relate to a kind of resident load Forecasting Methodology based on elman neural network.
Background technology
Electric power is as a kind of important energy source, all play a part very important in daily life and work, along with the fast development of national economy, Analyzing Total Electricity Consumption and each industry power consumption also steady growth, therefore the usage trend of power consumption not only affects production and management decision-making and the economic benefit of enterprises of managing electric wire netting, also can have influence on socioeconomic trend analysis.Reasonably carrying out load forecast is the precondition that electric system is dispatched electric power resource, planned.
Electric load generally can be divided into industrial load, Commercial Load, resident load etc., wherein industrial load and the Commercial Load proportion in electric load is higher, power grid enterprises always compare attention to the load prediction of this block, and have built up load control system and power information acquisition system successively to complete data acquisition to industry and commerce load and load prediction; The feature that resident's load disperses owing to distributing, scale is less than normal, what take is all the method concentrating prediction always, predict in units of Ji Yitai district or feeder load, the shortcoming of this Forecasting Methodology is exactly that precision is not high, especially along with the progressively popularization of the universal and electric automobile of the increasing year by year of resident's household electrical appliance, electric bicycle, the power load of resident presents steady-state growth trend and obvious seasonal fluctuation, is more manifested by the disadvantage of method to resident's load prediction of concentrated prediction.
Summary of the invention
In order to solve above-mentioned technical matters, the object of this invention is to provide a kind of resident load Forecasting Methodology based on elman neural network.
The technical solution adopted for the present invention to solve the technical problems is:
Based on a resident load Forecasting Methodology for elman neural network, comprising:
S1, the acquisition resident load historical data of previous year and the weather history supplemental characteristic of correspondence, carry out date type division to the effective number of days in this year simultaneously;
S2, according to the resident load historical data obtained, calculate the average same period of the resident load in each month, and then after the population mean calculating average all same periods, by each same period average and population mean to be divided by acquisition seasonal index number;
S3, employing seasonal index number are revised resident load historical data, by the resident load historical data in each month divided by after the seasonal index number of correspondence, obtain revised resident load historical data;
S4, determine the input and output data of neural network, and determine optimum hidden layer neuron number, thus set up the neural network based on elman;
S5, the weather history supplemental characteristic of revised resident load historical data and correspondence to be normalized, and then according to the data after normalized, the neural network set up is trained, the predicated error of neural network is controlled in preset range;
S6, the resident load historical data obtaining prediction the last week day, the weather parameters data of prediction day and date type are as the input of neural network, adopt the resident load of the neural network after training to prediction day to predict, and then obtain resident load predicted data after the predicted data of acquisition is multiplied by seasonal index number;
Described date type be divided into off-day and working day two type.
Further, preset range described in described step S5 is 5% ~ 10%.
Further, described resident load historical data comprises the resident load data of each hour, and described weather history supplemental characteristic comprises temperature, sunshine-duration and weather pattern.
Further, described step S4, comprising:
Resident load historical data, weather history supplemental characteristic and date type that S41, statistics obtain, using the output data of the resident load data of arbitrary day as neural network, simultaneously using the resident load data of each hour in the last week of this day and the weather parameters data of this day and the date type input data as neural network;
S42, initialization is carried out to neural network, determine input node unit vector, hidden layer node unit vector, feedback states vector sum output node vector according to input and output sequence, thus set up the neural network based on elman.
Further, the non-linear state space expression formula of the described neural network based on elman is:
Wherein, y (k) represents that m ties up output node vector, and l (k) represents that m ties up hidden layer node unit vector, and x (k) represents that u ties up input vector, and c (k) represents that n ties up feedback states vector, w
3represent the connection weights of hidden layer to output layer, w
2represent the connection weights of input layer to hidden layer, w
1represent and accept the connection weights of layer to hidden layer, g (*) represents the transport function of output neuron, and f (*) represents the transport function of hidden layer neuron.
Further, described step S5, comprising:
S51, to be normalized according to the weather history supplemental characteristic of following formula to revised resident load historical data and correspondence:
Wherein, x
krepresent the kth parameter value in resident load historical data sequence or weather history supplemental characteristic ordered series of numbers, k is natural number, x
maxrepresent x
kmaximal value in the data sequence of place, x
minrepresent x
kminimum value in the data sequence of place;
S52, according to the data after normalized, error calculation, right value update and threshold values are carried out to the neural network set up and upgrade, and then the predicated error of the neural network based on elman to be controlled in preset range.
Further, the described neural network based on elman adopts BP algorithm to carry out modified weight renewal, and adopts sum of squared errors function to carry out target function study, and the formula of described target function study is:
In above formula, E (x) represents target function,
represent target input vector.
Further, described step S6, comprising:
S61, the resident load historical data obtaining prediction the last week day, the weather parameters data of prediction day and date type are as the input of neural network, adopt the resident load of the neural network after training to prediction day to predict, and then obtain predicted data hourly on prediction same day day;
S62, the predicted data of acquisition is multiplied by seasonal index number after, obtain resident load predicted data hourly.
Further, further comprising the steps of after described step S62:
After the actual load data on same day day are predicted in S63, acquisition, calculate the error amount between resident load predicted data and actual load data obtained, and by error value back to neural network.
The invention has the beneficial effects as follows: a kind of resident load Forecasting Methodology based on elman neural network of the present invention, comprise: obtain the resident load historical data of previous year and the weather history supplemental characteristic of correspondence, date type division is carried out to the effective number of days in this year simultaneously; According to the resident load historical data obtained, calculate the average same period of the resident load in each month, and then after the population mean calculating average all same periods, by each same period average and population mean to be divided by acquisition seasonal index number; Adopt seasonal index number to revise resident load historical data, by the resident load historical data in each month divided by after the seasonal index number of correspondence, obtain revised resident load historical data; Determine the input and output data of neural network, and determine optimum hidden layer neuron number, thus set up the neural network based on elman; The weather history supplemental characteristic of revised resident load historical data and correspondence is normalized, and then according to the data after normalized, the neural network set up is trained, the predicated error of neural network is controlled in preset range; Obtain the input as neural network of the resident load historical data of prediction the last week day, the weather parameters data of prediction day and date type, adopt the resident load of the neural network after training to prediction day to predict, and then obtain resident load predicted data after the predicted data of acquisition is multiplied by seasonal index number.This method is by the structure based on the Elman neural network of seasonal index number, in conjunction with relevant regional resident load historical data and corresponding weather history supplemental characteristic, the measurable resident load data obtaining prediction day, and any Nonlinear Mapping can be approached with arbitrary precision, do not consider the impact of external noise, there is higher precision, and there is the ability of the seasonal fluctuations adapting to time-varying characteristics and resident load, directly can predict and reflect the dynamic perfromance of resident load, precision of prediction is higher.And this method adds seasonal index number feature, the comparatively large and historical data of the seasonal fluctuations that can overcome resident load utilizes the problems such as incomplete, effectively can improve the precision of predicted data and predict stability.
Accompanying drawing explanation
Below in conjunction with drawings and Examples, the invention will be further described.
Fig. 1 is the flow diagram of a kind of resident load Forecasting Methodology based on elman neural network of the present invention;
Fig. 2 is the structural representation of the neural network that a kind of resident load Forecasting Methodology based on elman neural network of the present invention is set up.
Embodiment
For the ease of following description, first provide following explanation of nouns:
Elman network: first J.L.Elman put forward for speech processing problems in nineteen ninety, it is a kind of typical local regression network (global feed for ward local recurrent).
BP algorithm: Error Back Propagation Algorithm, error backpropagation algorithm, is called for short BP algorithm.
With reference to Fig. 1, the invention provides a kind of resident load Forecasting Methodology based on elman neural network, comprising:
S1, the acquisition resident load historical data of previous year and the weather history supplemental characteristic of correspondence, carry out date type division to the effective number of days in this year simultaneously;
S2, according to the resident load historical data obtained, calculate the average same period of the resident load in each month, and then after the population mean calculating average all same periods, by each same period average and population mean to be divided by acquisition seasonal index number;
S3, employing seasonal index number are revised resident load historical data, by the resident load historical data in each month divided by after the seasonal index number of correspondence, obtain revised resident load historical data;
S4, determine the input and output data of neural network, and determine optimum hidden layer neuron number, thus set up the neural network based on elman;
S5, the weather history supplemental characteristic of revised resident load historical data and correspondence to be normalized, and then according to the data after normalized, the neural network set up is trained, the predicated error of neural network is controlled in preset range;
S6, the resident load historical data obtaining prediction the last week day, the weather parameters data of prediction day and date type are as the input of neural network, adopt the resident load of the neural network after training to prediction day to predict, and then obtain resident load predicted data after the predicted data of acquisition is multiplied by seasonal index number;
Described date type be divided into off-day and working day two type.
Be further used as preferred embodiment, preset range described in described step S5 is 5% ~ 10%.
Be further used as preferred embodiment, described resident load historical data comprises the resident load data of each hour, and described weather history supplemental characteristic comprises temperature, sunshine-duration and weather pattern.
Be further used as preferred embodiment, described step S4, comprising:
Resident load historical data, weather history supplemental characteristic and date type that S41, statistics obtain, using the output data of the resident load data of arbitrary day as neural network, simultaneously using the resident load data of each hour in the last week of this day and the weather parameters data of this day and the date type input data as neural network;
S42, initialization is carried out to neural network, determine input node unit vector, hidden layer node unit vector, feedback states vector sum output node vector according to input and output sequence, thus set up the neural network based on elman.
Be further used as preferred embodiment, the non-linear state space expression formula of the described neural network based on elman is:
Wherein, y (k) represents that m ties up output node vector, and l (k) represents that m ties up hidden layer node unit vector, and x (k) represents that u ties up input vector, and c (k) represents that n ties up feedback states vector, w
3represent the connection weights of hidden layer to output layer, w
2represent the connection weights of input layer to hidden layer, w
1represent and accept the connection weights of layer to hidden layer, g (*) represents the transport function of output neuron, and f (*) represents the transport function of hidden layer neuron.
Be further used as preferred embodiment, described step S5, comprising:
S51, to be normalized according to the weather history supplemental characteristic of following formula to revised resident load historical data and correspondence:
Wherein, x
krepresent the kth parameter value in resident load historical data sequence or weather history supplemental characteristic ordered series of numbers, k is natural number, x
maxrepresent x
kmaximal value in the data sequence of place, x
minrepresent x
kminimum value in the data sequence of place;
S52, according to the data after normalized, error calculation, right value update and threshold values are carried out to the neural network set up and upgrade, and then the predicated error of the neural network based on elman to be controlled in preset range.
Be further used as preferred embodiment, the described neural network based on elman adopts BP algorithm to carry out modified weight renewal, and adopts sum of squared errors function to carry out target function study, and the formula of described target function study is:
In above formula, E (x) represents target function,
represent target input vector.
Be further used as preferred embodiment, described step S6, comprising:
S61, the resident load historical data obtaining prediction the last week day, the weather parameters data of prediction day and date type are as the input of neural network, adopt the resident load of the neural network after training to prediction day to predict, and then obtain predicted data hourly on prediction same day day;
S62, the predicted data of acquisition is multiplied by seasonal index number after, obtain resident load predicted data hourly.
Be further used as preferred embodiment, further comprising the steps of after described step S62:
After the actual load data on same day day are predicted in S63, acquisition, calculate the error amount between resident load predicted data and actual load data obtained, and by error value back to neural network.
Below in conjunction with specific embodiment, the present invention will be further described.
With reference to Fig. 1, a kind of resident load Forecasting Methodology based on elman neural network, comprising:
S1, obtain the resident load historical data of previous year and the weather history supplemental characteristic of correspondence, date type division carried out to the effective number of days in this year simultaneously, the present embodiment date type is divided into off-day and working day two type;
S2, according to the resident load historical data obtained, calculate the average same period of the resident load in each month, and then after the population mean calculating average all same periods, by each same period average and population mean to be divided by acquisition seasonal index number;
S3, employing seasonal index number are revised resident load historical data, by the resident load historical data in each month divided by after the seasonal index number of correspondence, obtain revised resident load historical data;
S4, determine the input and output data of neural network, and determine optimum hidden layer neuron number, thus set up the neural network based on elman; Neural network based on elman comprises input layer, hidden layer, undertaking layer and output layer, accept layer for remembering the output valve of hidden layer previous moment and this output valve being returned to the input of hidden layer, add feedback, comparatively responsive to historical data, comparatively stable;
S5, the weather history supplemental characteristic of revised resident load historical data and correspondence to be normalized, and then according to the data after normalized, the neural network set up is trained, the predicated error of neural network is controlled in preset range; In the present embodiment, preset range is 5% ~ 10%; Carry out control errors, any Nonlinear Mapping can be approached with arbitrary accuracy;
S6, the resident load historical data obtaining prediction the last week day, the weather parameters data of prediction day and date type are as the input of neural network, adopt the resident load of the neural network after training to prediction day to predict, and then obtain resident load predicted data after the predicted data of acquisition is multiplied by seasonal index number;
Resident load historical data comprises the resident load data of each hour, and weather history supplemental characteristic comprises temperature, sunshine-duration and weather pattern.
Particularly, step S4 comprises step S41 ~ S42:
Resident load historical data, weather history supplemental characteristic and date type that S41, statistics obtain, using the output data of the resident load data of arbitrary day as neural network, simultaneously using the resident load data of each hour in the last week of this day and the weather parameters data of this day and the date type input data as neural network; In input data, resident load data amount to 168 load point, and the resident load data exporting data amount to 24 load point;
S42, initialization is carried out to neural network, according to input and output sequence (X, Y) determine that u dimension inputs node unit vector x, n ties up hidden layer node unit vector l, n and ties up feedback states vector c and m dimension output node vector y, thus the neural network set up based on elman, the neural metwork training model that the present embodiment is set up is as shown in Figure 2.Wherein, X1, X2Xu are the nodes of input layer, the weather parameters data of the prediction day of corresponding input, resident's load and date type after within upper one week, revising; Y1 is the node of output layer, the corresponding prediction day system resident load exported; L1, l2lN are the nodes of hidden layer, and wherein node in hidden layer n (namely optimum hidden layer neuron number) is by increasing progressively method of trial and error gradually, namely determines according to the way increasing exploration gradually; C1, C2CN accept the node of layer, are used for remembering the output valve of hidden layer unit previous moment and return to the input of hidden layer.
In the present embodiment, the non-linear state space expression formula based on the neural network of elman is:
Wherein, y (k) represents that m ties up output node vector, and l (k) represents that m ties up hidden layer node unit vector, and x (k) represents that u ties up input vector, and c (k) represents that n ties up feedback states vector, w
3represent the connection weights of hidden layer to output layer, w
2represent the connection weights of input layer to hidden layer, w
1represent and accept the connection weights of layer to hidden layer, g (*) represents the transport function of output neuron, and f (*) represents the transport function of hidden layer neuron, and f (*) generally adopts S function.
Particularly, step S5 comprises step S51 ~ S52:
S51, employing minimax method, be normalized according to the weather history supplemental characteristic of following formula to revised resident load historical data and correspondence:
Wherein, x
krepresent the kth parameter value in resident load historical data sequence or weather history supplemental characteristic ordered series of numbers, k is natural number, x
maxrepresent x
kmaximal value in the data sequence of place, x
minrepresent x
kminimum value in the data sequence of place;
S52, according to the data after normalized, error calculation, right value update and threshold values are carried out to the neural network set up and upgrade, and then the predicated error of the neural network based on elman to be controlled in preset range.
Due to the singularity of date type, will be labeled as 1 off-day here, working day is labeled as 0, meets minimax normalization principle equally, and being namely normalized date type to affect its occurrence.
In step S52, the neural network based on elman adopts BP algorithm to carry out modified weight renewal, and adopts sum of squared errors function to carry out target function study, and the formula of target function study is:
In above formula, E (x) represents target function,
represent target input vector.
Particularly, step S6 comprises step S61 ~ S63:
S61, the resident load historical data obtaining prediction the last week day, the weather parameters data of prediction day and date type are as the input of neural network, the resident load of the neural network after training to prediction day is adopted to predict, and then obtain predicted data hourly on prediction same day day, namely obtain 24 load point of prediction;
S62, the predicted data of acquisition is multiplied by seasonal index number after, obtain resident load predicted data hourly;
After the actual load data on same day day are predicted in S63, acquisition, calculate the error amount between resident load predicted data and actual load data obtained, and by error value back to neural network.After the actual load data obtaining prediction same day day, contrast with the load point of 24 in predicted data, calculate the error amount of 24 load point and actual load data respectively and feed back to neural network, the neural metwork training model set up can be adjusted, make it closer to actual conditions.
Resident load Forecasting Methodology of the present invention, by setting up the neural network prediction model of resident load, any Nonlinear Mapping can be approached with arbitrary precision, do not consider the impact of external noise, there is higher precision, and there is the ability of the seasonal fluctuations adapting to time-varying characteristics and resident load, directly can predict and reflect the dynamic perfromance of resident load, precision of prediction is higher.And this method adds seasonal index number feature, the comparatively large and historical data of the seasonal fluctuations that can overcome resident load utilizes the problems such as incomplete, effectively can improve the precision of predicted data and predict stability.
In addition, according to the residential electricity consumption historical data that residential electricity consumption demand, seasonal climate change are formed, predict its recent power consumption, will residential households using electricity wisely and slow down the urgency of electric power resource scarcity be consciously conducive to.Meanwhile, by resident load prediction is combined with the utilization of new forms of energy, more can accomplish such as sun power, making full use of of the various forms of new forms of energy such as wind energy, avoid the waste of the energy, for user creates larger economic benefit.
More than that better enforcement of the present invention is illustrated, but the invention is not limited to embodiment, those of ordinary skill in the art also can make all equivalent variations or replacement under the prerequisite without prejudice to spirit of the present invention, and these equivalent modification or replacement are all included in the application's claim limited range.
Claims (9)
1., based on a resident load Forecasting Methodology for elman neural network, it is characterized in that, comprising:
S1, the acquisition resident load historical data of previous year and the weather history supplemental characteristic of correspondence, carry out date type division to the effective number of days in this year simultaneously;
S2, according to the resident load historical data obtained, calculate the average same period of the resident load in each month, and then after the population mean calculating average all same periods, by each same period average and population mean to be divided by acquisition seasonal index number;
S3, employing seasonal index number are revised resident load historical data, by the resident load historical data in each month divided by after the seasonal index number of correspondence, obtain revised resident load historical data;
S4, determine the input and output data of neural network, and determine optimum hidden layer neuron number, thus set up the neural network based on elman;
S5, the weather history supplemental characteristic of revised resident load historical data and correspondence to be normalized, and then according to the data after normalized, the neural network set up is trained, the predicated error of neural network is controlled in preset range;
S6, the resident load historical data obtaining prediction the last week day, the weather parameters data of prediction day and date type are as the input of neural network, adopt the resident load of the neural network after training to prediction day to predict, and then obtain resident load predicted data after the predicted data of acquisition is multiplied by seasonal index number;
Described date type be divided into off-day and working day two type.
2. a kind of resident load Forecasting Methodology based on elman neural network according to claim 1, it is characterized in that, preset range described in described step S5 is 5% ~ 10%.
3. a kind of resident load Forecasting Methodology based on elman neural network according to claim 1, it is characterized in that, described resident load historical data comprises the resident load data of each hour, and described weather history supplemental characteristic comprises temperature, sunshine-duration and weather pattern.
4. a kind of resident load Forecasting Methodology based on elman neural network according to claim 3, it is characterized in that, described step S4, comprising:
Resident load historical data, weather history supplemental characteristic and date type that S41, statistics obtain, using the output data of the resident load data of arbitrary day as neural network, simultaneously using the resident load data of each hour in the last week of this day and the weather parameters data of this day and the date type input data as neural network;
S42, initialization is carried out to neural network, determine input node unit vector, hidden layer node unit vector, feedback states vector sum output node vector according to input and output sequence, thus set up the neural network based on elman.
5. a kind of resident load Forecasting Methodology based on elman neural network according to claim 4, is characterized in that, the non-linear state space expression formula of the described neural network based on elman is:
Wherein, y (k) represents that m ties up output node vector, and l (k) represents that m ties up hidden layer node unit vector, and x (k) represents that u ties up input vector, and c (k) represents that n ties up feedback states vector, w
3represent the connection weights of hidden layer to output layer, w
2represent the connection weights of input layer to hidden layer, w
1represent and accept the connection weights of layer to hidden layer, g (*) represents the transport function of output neuron, and f (*) represents the transport function of hidden layer neuron.
6. a kind of resident load Forecasting Methodology based on elman neural network according to claim 4, it is characterized in that, described step S5, comprising:
S51, to be normalized according to the weather history supplemental characteristic of following formula to revised resident load historical data and correspondence:
Wherein, x
krepresent the kth parameter value in resident load historical data sequence or weather history supplemental characteristic ordered series of numbers, k is natural number, x
maxrepresent x
kmaximal value in the data sequence of place, x
minrepresent x
kminimum value in the data sequence of place;
S52, according to the data after normalized, error calculation, right value update and threshold values are carried out to the neural network set up and upgrade, and then the predicated error of the neural network based on elman to be controlled in preset range.
7. a kind of resident load Forecasting Methodology based on elman neural network according to claim 6, it is characterized in that, the described neural network based on elman adopts BP algorithm to carry out modified weight renewal, and adopt sum of squared errors function to carry out target function study, the formula of described target function study is:
In above formula, E (x) represents target function,
represent target input vector.
8. a kind of resident load Forecasting Methodology based on elman neural network according to claim 7, it is characterized in that, described step S6, comprising:
S61, the resident load historical data obtaining prediction the last week day, the weather parameters data of prediction day and date type are as the input of neural network, adopt the resident load of the neural network after training to prediction day to predict, and then obtain predicted data hourly on prediction same day day;
S62, the predicted data of acquisition is multiplied by seasonal index number after, obtain resident load predicted data hourly.
9. a kind of resident load Forecasting Methodology based on elman neural network according to claim 8, is characterized in that, further comprising the steps of after described step S62:
After the actual load data on same day day are predicted in S63, acquisition, calculate the error amount between resident load predicted data and actual load data obtained, and by error value back to neural network.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510031651.1A CN104636822B (en) | 2015-01-21 | 2015-01-21 | A kind of resident load prediction technique based on elman neural networks |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510031651.1A CN104636822B (en) | 2015-01-21 | 2015-01-21 | A kind of resident load prediction technique based on elman neural networks |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104636822A true CN104636822A (en) | 2015-05-20 |
CN104636822B CN104636822B (en) | 2018-08-31 |
Family
ID=53215546
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510031651.1A Active CN104636822B (en) | 2015-01-21 | 2015-01-21 | A kind of resident load prediction technique based on elman neural networks |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104636822B (en) |
Cited By (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106845770A (en) * | 2016-12-16 | 2017-06-13 | 深圳市麦澜创新科技有限公司 | A kind of energy storage management system based on neutral net |
CN106971238A (en) * | 2017-03-10 | 2017-07-21 | 东华大学 | The Short-Term Load Forecasting Method of Elman neutral nets is obscured based on T S |
CN107038502A (en) * | 2017-04-18 | 2017-08-11 | 国网安徽省电力公司芜湖供电公司 | Consider the improvement wavelet packet electricity demand forecasting method of Seasonal Characteristics |
CN107590562A (en) * | 2017-09-05 | 2018-01-16 | 西安交通大学 | A kind of Short-Term Load Forecasting of Electric Power System based on changeable weight combination predicted method |
CN108038568A (en) * | 2017-12-05 | 2018-05-15 | 国家电网公司 | A kind of changeable weight combination Short-Term Load Forecasting of Electric Power System based on particle cluster algorithm |
CN108053070A (en) * | 2017-12-15 | 2018-05-18 | 国网江苏省电力有限公司淮安供电分公司 | The node load Forecasting Methodology of meter and high permeability distributed generation resource |
CN109376907A (en) * | 2018-09-25 | 2019-02-22 | 国网浙江省电力有限公司经济技术研究院 | Adapt to the high-voltage distribution network transformer substation load forecasting method of transmission and distribution network integration planning |
CN109996247A (en) * | 2019-03-27 | 2019-07-09 | 中国电子科技集团公司信息科学研究院 | Networked resource concocting method, device, equipment and storage medium |
CN110009136A (en) * | 2019-03-12 | 2019-07-12 | 国网江西省电力有限公司电力科学研究院 | A kind of load forecasting method of distribution transformer and distribution line |
CN110046765A (en) * | 2019-04-15 | 2019-07-23 | 国网甘肃省电力公司电力科学研究院 | A kind of power system stabilizer, PSS implementation method based on Elman neural network |
CN110569981A (en) * | 2019-07-22 | 2019-12-13 | 国网冀北电力有限公司电力科学研究院 | Method and device for determining fuel consumption of scattered coal |
CN110689208A (en) * | 2019-10-14 | 2020-01-14 | 国网山东省电力公司泰安供电公司 | Medium-and-long-term load prediction method based on uncertain load high permeability |
CN111080477A (en) * | 2019-12-06 | 2020-04-28 | 北京国电通网络技术有限公司 | Household power load prediction method and system |
CN111275571A (en) * | 2020-01-14 | 2020-06-12 | 河海大学 | Resident load probability prediction deep learning method considering microclimate and user mode |
CN111563827A (en) * | 2020-04-10 | 2020-08-21 | 杭州电子科技大学 | Load decomposition method based on electrical appliance physical characteristics and residential electricity consumption behaviors |
CN111898799A (en) * | 2020-06-24 | 2020-11-06 | 南京理工大学 | BFA-Elman-based power load prediction method |
CN112200384A (en) * | 2020-10-28 | 2021-01-08 | 宁波立新科技股份有限公司 | Power load short-time prediction method based on EWT neural network |
CN113139673A (en) * | 2020-11-04 | 2021-07-20 | 西安天和防务技术股份有限公司 | Method, device, terminal and storage medium for predicting air quality |
CN113191559A (en) * | 2021-05-08 | 2021-07-30 | 苏州瑞城电力科技有限公司 | Improved neural network short-term resident load prediction method based on autoregressive selection |
CN113300416A (en) * | 2021-07-07 | 2021-08-24 | 广东电网有限责任公司 | Power grid standby capacity configuration method, system, equipment and computer medium |
CN114997477A (en) * | 2022-05-20 | 2022-09-02 | 国网江苏省电力有限公司苏州供电分公司 | Low-voltage residential district load prediction method and system |
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 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101383023A (en) * | 2008-10-22 | 2009-03-11 | 西安交通大学 | Neural network short-term electric load prediction based on sample dynamic organization and temperature compensation |
CN103093285A (en) * | 2013-01-22 | 2013-05-08 | 清华大学 | Short-term load forecast method based on artificial neural network |
CN103793887A (en) * | 2014-02-17 | 2014-05-14 | 华北电力大学 | Short-term electrical load on-line predicting method based on self-adaptation enhancing algorithm |
CN103854068A (en) * | 2013-12-06 | 2014-06-11 | 国家电网公司 | Method for forecasting residential quarter short-term loads |
CN104268638A (en) * | 2014-09-11 | 2015-01-07 | 广州市香港科大霍英东研究院 | Photovoltaic power generation system power predicting method of elman-based neural network |
-
2015
- 2015-01-21 CN CN201510031651.1A patent/CN104636822B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101383023A (en) * | 2008-10-22 | 2009-03-11 | 西安交通大学 | Neural network short-term electric load prediction based on sample dynamic organization and temperature compensation |
CN103093285A (en) * | 2013-01-22 | 2013-05-08 | 清华大学 | Short-term load forecast method based on artificial neural network |
CN103854068A (en) * | 2013-12-06 | 2014-06-11 | 国家电网公司 | Method for forecasting residential quarter short-term loads |
CN103793887A (en) * | 2014-02-17 | 2014-05-14 | 华北电力大学 | Short-term electrical load on-line predicting method based on self-adaptation enhancing algorithm |
CN104268638A (en) * | 2014-09-11 | 2015-01-07 | 广州市香港科大霍英东研究院 | Photovoltaic power generation system power predicting method of elman-based neural network |
Cited By (28)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106845770A (en) * | 2016-12-16 | 2017-06-13 | 深圳市麦澜创新科技有限公司 | A kind of energy storage management system based on neutral net |
CN106971238A (en) * | 2017-03-10 | 2017-07-21 | 东华大学 | The Short-Term Load Forecasting Method of Elman neutral nets is obscured based on T S |
CN107038502A (en) * | 2017-04-18 | 2017-08-11 | 国网安徽省电力公司芜湖供电公司 | Consider the improvement wavelet packet electricity demand forecasting method of Seasonal Characteristics |
CN107590562A (en) * | 2017-09-05 | 2018-01-16 | 西安交通大学 | A kind of Short-Term Load Forecasting of Electric Power System based on changeable weight combination predicted method |
CN108038568A (en) * | 2017-12-05 | 2018-05-15 | 国家电网公司 | A kind of changeable weight combination Short-Term Load Forecasting of Electric Power System based on particle cluster algorithm |
CN108053070A (en) * | 2017-12-15 | 2018-05-18 | 国网江苏省电力有限公司淮安供电分公司 | The node load Forecasting Methodology of meter and high permeability distributed generation resource |
CN109376907B (en) * | 2018-09-25 | 2022-06-28 | 国网浙江省电力有限公司经济技术研究院 | High-voltage distribution network transformer substation load prediction method suitable for integrated planning of transmission and distribution network |
CN109376907A (en) * | 2018-09-25 | 2019-02-22 | 国网浙江省电力有限公司经济技术研究院 | Adapt to the high-voltage distribution network transformer substation load forecasting method of transmission and distribution network integration planning |
CN110009136A (en) * | 2019-03-12 | 2019-07-12 | 国网江西省电力有限公司电力科学研究院 | A kind of load forecasting method of distribution transformer and distribution line |
CN109996247A (en) * | 2019-03-27 | 2019-07-09 | 中国电子科技集团公司信息科学研究院 | Networked resource concocting method, device, equipment and storage medium |
CN109996247B (en) * | 2019-03-27 | 2022-03-11 | 中国电子科技集团公司信息科学研究院 | Networked resource allocation method, device, equipment and storage medium |
CN110046765A (en) * | 2019-04-15 | 2019-07-23 | 国网甘肃省电力公司电力科学研究院 | A kind of power system stabilizer, PSS implementation method based on Elman neural network |
CN110569981A (en) * | 2019-07-22 | 2019-12-13 | 国网冀北电力有限公司电力科学研究院 | Method and device for determining fuel consumption of scattered coal |
CN110689208A (en) * | 2019-10-14 | 2020-01-14 | 国网山东省电力公司泰安供电公司 | Medium-and-long-term load prediction method based on uncertain load high permeability |
CN111080477A (en) * | 2019-12-06 | 2020-04-28 | 北京国电通网络技术有限公司 | Household power load prediction method and system |
CN111275571A (en) * | 2020-01-14 | 2020-06-12 | 河海大学 | Resident load probability prediction deep learning method considering microclimate and user mode |
CN111563827A (en) * | 2020-04-10 | 2020-08-21 | 杭州电子科技大学 | Load decomposition method based on electrical appliance physical characteristics and residential electricity consumption behaviors |
CN111563827B (en) * | 2020-04-10 | 2023-08-04 | 杭州电子科技大学 | Load decomposition method based on physical characteristics of electrical appliance and electricity utilization behaviors of residents |
CN111898799A (en) * | 2020-06-24 | 2020-11-06 | 南京理工大学 | BFA-Elman-based power load prediction method |
CN111898799B (en) * | 2020-06-24 | 2022-09-27 | 南京理工大学 | BFA-Elman-based power load prediction method |
CN112200384A (en) * | 2020-10-28 | 2021-01-08 | 宁波立新科技股份有限公司 | Power load short-time prediction method based on EWT neural network |
CN112200384B (en) * | 2020-10-28 | 2024-05-17 | 宁波立新科技股份有限公司 | EWT neural network-based short-time prediction method for power load |
CN113139673A (en) * | 2020-11-04 | 2021-07-20 | 西安天和防务技术股份有限公司 | Method, device, terminal and storage medium for predicting air quality |
CN113191559A (en) * | 2021-05-08 | 2021-07-30 | 苏州瑞城电力科技有限公司 | Improved neural network short-term resident load prediction method based on autoregressive selection |
CN113191559B (en) * | 2021-05-08 | 2024-04-19 | 苏州瑞城电力科技有限公司 | Improved neural network short-term resident load prediction method based on autoregressive selection |
CN113300416A (en) * | 2021-07-07 | 2021-08-24 | 广东电网有限责任公司 | Power grid standby capacity configuration method, system, equipment and computer medium |
CN114997477A (en) * | 2022-05-20 | 2022-09-02 | 国网江苏省电力有限公司苏州供电分公司 | Low-voltage residential district load prediction method and system |
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 |
Also Published As
Publication number | Publication date |
---|---|
CN104636822B (en) | 2018-08-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104636822A (en) | Residential load prediction method of elman-based neural network | |
Ye et al. | Predicting electricity consumption in a building using an optimized back-propagation and Levenberg–Marquardt back-propagation neural network: Case study of a shopping mall in China | |
CN111353656B (en) | Steel enterprise oxygen load prediction method based on production plan | |
Hafiz et al. | Real-time stochastic optimization of energy storage management using deep learning-based forecasts for residential PV applications | |
Yan et al. | Reviews on uncertainty analysis of wind power forecasting | |
Shi et al. | Energy consumption prediction of office buildings based on echo state networks | |
US10873209B2 (en) | System and method for dynamic energy storage system control | |
He et al. | Forecasting China's total energy demand and its structure using ADL-MIDAS model | |
CN102426674B (en) | Power system load prediction method based on Markov chain | |
CN104598986A (en) | Big data based power load prediction method | |
CN102509173B (en) | A kind of based on markovian power system load Accurate Prediction method | |
Kumar et al. | Optimal sizing of stand alone hybrid renewable energy system with load shifting | |
WO2023070293A1 (en) | Long-term scheduling method for industrial byproduct gas system | |
CN101888087A (en) | Method for realizing distributed super-short-term area load forecasting in distribution network terminal | |
WO2020013884A1 (en) | Machine-learned prediction of network resources and margins | |
CN106355285A (en) | Electricity usage load predicting method based on parameter correction | |
Almalaq et al. | Deep learning application: Load forecasting in big data of smart grids | |
Wang et al. | Multi-criteria comprehensive study on predictive algorithm of heating energy consumption of district heating station based on timeseries processing | |
CN104134103A (en) | Method for predicting energy consumption of hot oil pipeline by using modified BP neural network model | |
CN115906408A (en) | Energy scheduling system and method based on building load prediction | |
CN117578534B (en) | Scheduling method, device, equipment and storage medium of photovoltaic energy storage system | |
Chen et al. | The predictive management in campus heating system based on deep reinforcement learning and probabilistic heat demands forecasting | |
Xie et al. | Integration of intermittent resources with price-responsive loads | |
KR102299698B1 (en) | Power Consumption Prediction System Using Long-Term and Short-Term Prediction method | |
Huang et al. | Wind prediction based on improved bp artificial neural network in wind farm |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |