CN108596405A - The prediction technique and system of grey library yield - Google Patents
The prediction technique and system of grey library yield Download PDFInfo
- Publication number
- CN108596405A CN108596405A CN201810440144.7A CN201810440144A CN108596405A CN 108596405 A CN108596405 A CN 108596405A CN 201810440144 A CN201810440144 A CN 201810440144A CN 108596405 A CN108596405 A CN 108596405A
- Authority
- CN
- China
- Prior art keywords
- prediction
- predicted
- prediction result
- grey library
- result
- 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 72
- 238000004519 manufacturing process Methods 0.000 claims abstract description 84
- 238000013528 artificial neural network Methods 0.000 claims abstract description 25
- 239000003245 coal Substances 0.000 claims description 43
- 238000003062 neural network model Methods 0.000 claims description 16
- NINIDFKCEFEMDL-UHFFFAOYSA-N Sulfur Chemical group [S] NINIDFKCEFEMDL-UHFFFAOYSA-N 0.000 claims description 14
- 229910052717 sulfur Inorganic materials 0.000 claims description 14
- 239000011593 sulfur Substances 0.000 claims description 14
- 238000012545 processing Methods 0.000 claims description 13
- 238000004458 analytical method Methods 0.000 claims description 12
- 238000013459 approach Methods 0.000 claims description 10
- 230000007613 environmental effect Effects 0.000 claims description 7
- 210000002364 input neuron Anatomy 0.000 claims description 7
- 210000002569 neuron Anatomy 0.000 claims description 5
- 238000010248 power generation Methods 0.000 claims description 3
- 241000208340 Araliaceae Species 0.000 claims 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims 2
- 235000003140 Panax quinquefolius Nutrition 0.000 claims 2
- 235000008434 ginseng Nutrition 0.000 claims 2
- 239000002956 ash Substances 0.000 description 11
- 238000010586 diagram Methods 0.000 description 10
- 238000000611 regression analysis Methods 0.000 description 5
- 238000007619 statistical method Methods 0.000 description 3
- 239000010883 coal ash Substances 0.000 description 2
- 238000010835 comparative analysis Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000004445 quantitative analysis Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 239000006227 byproduct Substances 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 238000006467 substitution reaction Methods 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 prediction meanss and method of a kind of ash library yield, including：First prediction module determines the first prediction result according to the historical information of grey library yield and grey library production information to be predicted；Second prediction module builds a neural network prediction model, according to historical information and grey library production information to be predicted, determines the second prediction result；First prediction result and the second prediction result are carried out convergent operation by computing module, determine the final prediction result of grey library production information to be predicted；Update module replaces second prediction result, to update the neural network prediction model using the final prediction result.Historical information and to be predicted grey library production information of the present invention according to grey library yield, determine the first prediction result, neural network prediction model is applied in the prediction of grey library yield again, determine the second prediction result, most latter two prediction result carries out convergent, constantly update neural network prediction model simultaneously so that final prediction result is more accurate.
Description
Technical field
The present invention relates to neural network and power generation byproduct field more particularly to a kind of grey library yield based on neural network
Prediction technique and system.
Background technology
Currently, the conventional model in prediction technique uses regression analysis model (Quantitative Analysis Model).Wherein mathematical statistics
The attention of related field is constantly subjected to as a kind of effective knowledge discovering technologies, it is important one of the pillar of Knowledge Discovery.
Among these there are many regular method between variable of finding, and regression analysis is a kind of with very extensive statistical analysis side
Method has very extensive purposes in Knowledge Discovery.
So-called regression analysis is exactly to be established using mathematical statistics method because becoming on the basis of grasping a large amount of observation data
Regression relation function expression (being known as regression equation) between amount and independent variable.
In practical problem, often there is such situation, although there is certain relationship between variable Y and variable X, this relationship
Different from common functional relation, the value of variable Y cannot be determined by the value of X completely.That is, the relationship between variable can
To be divided into deterministic relationship and uncertainty relationship.For the variable of correlativity, although specific between cannot obtaining variable
Accurate function expression, but by largely observing data, using statistical method, it can be found that existing between them
Certain statistical regularity.When X obtains any probable value, Y correspondingly obeys certain probability distribution.
Invention content
(1) technical problems to be solved
The purpose of the present invention is to provide a kind of prediction meanss and method of grey library yield, to solve at least one above-mentioned
Technical problem.
(2) technical solution
An aspect of of the present present invention provides a kind of prediction meanss of grey library yield, including：
First prediction module, for the historical information and ash library to be predicted production information according to grey library yield, determination waits for pre-
Survey the first prediction result of grey library production information；
Second prediction module, for building a neural network prediction model, and according to the historical information and ash to be predicted
Library production information determines the second prediction result of the grey library production information to be predicted；
Computing module, for first prediction result and second prediction result to be carried out convergent operation, determining should
The final prediction result of ash library to be predicted production information；And
Update module replaces second prediction result, to update the neural network using the final prediction result
Prediction model.
In some embodiments of the invention, the historical information of the grey library yield includes at least following history parameters：It goes through
History exports blower pressure, history generation schedule, history generated energy, history coal card numerical value, history coal sulfur-bearing numerical value and/or goes through
History coal moisture content；
The grey library production information to be predicted includes at least following parameter：Export blower pressure, generation schedule, generated energy,
Coal card numerical value, coal sulfur-bearing numerical value and/or coal moisture content.
In some embodiments of the invention, second prediction module includes：
Method determination unit, the prediction technique for determining the grey library production information to be predicted；And
Pattern determining unit, for according to the prediction technique and historical information, combination processing sale pattern to determine
Second prediction result, the processing sale pattern include pattern, the free roping pattern undersold, and unit price sale mould
Formula.
In some embodiments of the invention, the method determination unit includes：
Judging unit, for judging that the parameter whether there is successively, if each parameter in the presence of if judging result Aa be 1,
Judging result Aa is -1 if being not present, and the judging result Aa is denoted as to six input neuron Xa of neural network model
Threshold values, 1≤a≤6；
Converting unit, for the prediction technique is denoted as two intrerneurons Xb, 7≤b of neural network model≤
8；
Determination unit determines the threshold values Fb, each defeated of each intrerneuron for the historical information according to the grey library yield
Enter the weights Wab between neuron and the intrerneuron, wherein the value of Fb is 0 and ± 1；And
Selecting unit, the weighted sum for determining each intrerneuronAnd according to Pb
Value, select prediction technique：If Pb is more than 1, historical forecast analytic approach is selected；If Pb is less than 1, average price is selected to predict
Method；If Pb is equal to 1, convergence analysis method is selected.
In some embodiments of the invention, the first prediction module includes：
Database, the historical information of the grey library yield for storing extraneous input；
Query unit, for inquire extraneous input grey library production information to be predicted and historical information in the database it
Between matching result；
Predicting unit, for according to the matching result, combining environmental factor determines the of grey library production information to be predicted
One prediction result.
In some embodiments of the invention, the convergent operation of the arithmetic element is the arithmetic element to described first
Prediction result seeks arithmetic mean of instantaneous value with second prediction result.
Another aspect of the present invention additionally provides a kind of prediction technique of grey library yield, is applied to any description above
Device, this method include：
According to the historical information of grey library yield and grey library production information to be predicted, the of grey library production information to be predicted is determined
One prediction result；
A neural network prediction model is built, and according to the historical information and grey library production information to be predicted, determining should
Second prediction result of ash library to be predicted production information；
First prediction result and second prediction result are subjected to convergent operation, determine the grey library yield to be predicted
The final prediction result of information；And
Second prediction result is replaced using the final prediction result, to update the neural network prediction model.
In some embodiments of the invention, the historical information of the grey library yield includes at least following history parameters：It goes through
History exports blower pressure, history generation schedule, history generated energy, history coal card numerical value, history coal sulfur-bearing numerical value and/or goes through
History coal moisture content；The grey library production information to be predicted includes at least following parameter：Export blower pressure, generation schedule, hair
Electricity, coal card numerical value, coal sulfur-bearing numerical value and/or coal moisture content.
In some embodiments of the invention, the second prediction result of the grey library production information to be predicted, including step are determined
Suddenly：
Determine the prediction technique of the grey library production information to be predicted；And
According to the prediction technique and historical information, combination processing sale pattern determines the second prediction result, described
Processing sale pattern includes pattern of underselling, free roping pattern and monovalent sale pattern.
In some embodiments of the invention, the prediction technique of the grey library production information to be predicted, including step are determined：
Judge that the parameter whether there is successively, is 1 if the judging result Aa of each parameter in the presence of if, judges if being not present
As a result Aa is -1, and the judging result Aa is denoted as to the threshold values of six of neural network model input neuron Xa, 1≤a≤
6；
The prediction technique is denoted as to two intrerneuron Xb of neural network model, 7≤b≤8；
According to the historical information of the grey library yield, the threshold values Fb, each input neuron and institute of each intrerneuron are determined
State the weights Wab between intrerneuron, wherein the value of Fb is 0 and ± 1；And
Determine the weighted sum of each intrerneuronAnd according to the value of Pb, selection is pre-
Survey method：If Pb is more than 1, historical forecast analytic approach is selected；If Pb is less than 1, average price predicted method is selected；If Pb is equal to 1,
Select convergence analysis method.
In some embodiments of the invention, the first prediction result of grey library production information to be predicted, including step are determined：
Establish the database of the historical information of the grey library yield for storing extraneous input；
Matching result in the grey library production information to be predicted and the database of the extraneous input of inquiry between historical information；
And
According to the matching result, combining environmental factor determines the first prediction result of grey library production information to be predicted.
In some embodiments of the invention, the historical forecast analytic approach includes step：The database is inquired, with institute
Parameter comparison is stated, the history parameters in database have at least five kinds of identical parameter informations with the parameter at this time, using institute
State the corresponding price of history parameters；
The average price predicted method includes step：The database is inquired, and the parameter comparison, going through in database at this time
There are four kinds of identical parameter informations with the parameter for history parameter, are started to the average price of current time using of that month No. 1；
The convergence analysis method includes step：The database is inquired, and the parameter comparison, going through in database at this time
At most there are three kinds of identical parameter informations with the parameter for history parameter, judge becoming when the historical data information in the last week
Gesture is set price.
In some embodiments of the invention, the first prediction result and second prediction result are subjected to convergent operation,
Specially：Calculate the arithmetic mean of instantaneous value of first prediction result and second prediction result.
(3) advantageous effect
The prediction meanss and method of the grey library yield of the present invention have at least the following advantages compared to the prior art：
1, according to the historical information of grey library yield and grey library production information to be predicted, the first prediction result is determined, then will be refreshing
It is applied to through Network Prediction Model in the prediction of grey library yield, determines that the second prediction result, most latter two prediction result become
Together, more accurately final prediction result is obtained, while the neural network prediction model can also be constantly updated so that final prediction
As a result more precisely.
2, compared with traditional regression analysis model, when being analyzed without limiting model, especially when data variable is deposited
It can be checked automatically in interaction, enhance real-time.
3, relating to parameters of the neural network model with input information itself, and these parameters are calculated by using study
Method is trained neural network model, therefore the present invention is with stronger adaptive inferential.
4, the present invention also in the prediction model of grey library yield by select prediction technique, in conjunction with processing sale mould
Formula and the first prediction result obtain the final prediction result of grey library production information to be predicted, have given full play to neural network model
Calculation features, while also enhancing the accuracy of final prediction result.
Description of the drawings
Fig. 1 is the structural schematic diagram of the prediction meanss of the grey library yield of the embodiment of the present invention.
Fig. 2 is the module diagram of the first prediction module of the embodiment of the present invention.
Fig. 3 is the module diagram of the second prediction module of the embodiment of the present invention.
Fig. 4 is the module diagram of the method determination unit of the embodiment of the present invention.
Fig. 5 is the schematic diagram of the neural network prediction model of the embodiment of the present invention.
Fig. 6 is the step schematic diagram of the prediction technique of the grey library yield of the embodiment of the present invention.
Fig. 7 is the sub-step signal of the first prediction result of the determination of the embodiment of the present invention grey library production information to be predicted
Figure.
Fig. 8 is the specific steps signal of the prediction result of the determination of the embodiment of the present invention grey library production information to be predicted
Figure.
Fig. 9 is the sub-step schematic diagram of the step S21 of the embodiment of the present invention.
Specific implementation mode
In the prior art, the conventional model in prediction technique uses regression analysis model (Quantitative Analysis Model), needs certainly
It is dynamic to limit model, and when data variable when interaction there are being difficult to check automatically, in view of this, the present invention provides one kind
The prediction meanss and method of grey library yield determine first according to the historical information of grey library yield and grey library production information to be predicted
Prediction result, then neural network prediction model is applied in the prediction of grey library yield, determines the second prediction result, most latter two
Prediction result carries out convergent, obtains more accurately final prediction result, while can also constantly update the neural network prediction mould
Type so that final prediction result is more accurate.
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with specific embodiment, and reference
Attached drawing, the present invention is described in more detail.
The one side of the embodiment of the present invention provides a kind of prediction meanss of grey library yield, and Fig. 1 is the embodiment of the present invention
The structural representation of the prediction meanss of grey library yield, as shown in Figure 1, the prediction meanss may include：
First prediction module 1 determines ash to be predicted according to the historical information of grey library yield and grey library production information to be predicted
First prediction result of library production information；
Second prediction module 2, for building a neural network prediction model, and according to the historical information and ash to be predicted
Library production information determines the second prediction result of the grey library production information to be predicted；
Computing module 3, for first prediction result and second prediction result to be carried out convergent operation, determining should
The final prediction result of ash library to be predicted production information；And
Update module 4, for replacing second prediction result using the final prediction result, to described in update
Neural network prediction model.
In general, the historical information of the grey library yield includes at least following history parameters：History outlet blower pressure,
History generation schedule, history generated energy, history coal card numerical value, history coal sulfur-bearing numerical value and/or history coal moisture content；Institute
It states grey library production information to be predicted and includes at least following parameter：Export blower pressure, generation schedule, generated energy, coal card numerical value,
Coal sulfur-bearing numerical value and/or coal moisture content.
In some embodiments, in order to more easily inquire historical information, so that it is determined that the first prediction result, such as
Shown in Fig. 2, which may include：
Database 11, the historical information of the grey library yield for storing extraneous input；
Query unit 12, the grey library production information to be predicted for inquiring extraneous input and historical information in the database
Between matching result；And
Predicting unit 13, for according to the matching result, combining environmental factor to determine grey library production information to be predicted
First prediction result.
In some embodiments, in order to rapidly and accurately combine Neural Network model predictive, as shown in figure 3, this second
Prediction module 2 may include：
Method determination unit 21, the prediction technique for determining the grey library production information to be predicted；And
Pattern determining unit 22, the prediction technique for being determined according to the method determination unit 21, combination processing sale
Operating mode determines that the second prediction result, the processing sale pattern include underselling pattern, free roping pattern,
With monovalent sale pattern.
Wherein, pattern, free roping pattern and monovalent sale pattern are undersold, this is three kinds of treated coal ash sale
Operating mode can be selected according to actual conditions.
In general, it when grey library ash position height, can not sell away, the normal production of power generation of trying hard to keep starts free roping pattern；
Since weather environment factor causes the market demand less, start sale at low prices pattern；Normal relation between supply and demand sets unit price sale.
More specifically, as shown in Figure 4 and Figure 5, this method determination unit 21 may include with lower unit：
Judging unit 211, for judging that the parameter whether there is successively, if in the presence of if the judging result Aa of each parameter be
1, judging result Aa is -1 if being not present, and the judging result Aa is denoted as to six input neurons of neural network model
The threshold values of Xa, 1≤a≤6；
Converting unit 212, two intrerneurons Xb, 7≤b for the prediction technique to be denoted as to neural network model
≤8；
Determination unit 213, for the historical information according to the grey library yield, determine each intrerneuron threshold values Fb,
Weights Wab between each input neuron and the intrerneuron, wherein the value of Fb is 0 and ± 1；And
Selecting unit 214, the weighted sum for determining each intrerneuronAnd according to
The value of Pb selects prediction technique：If Pb is more than 1, historical forecast analytic approach, and historical forecast analytic approach operating mode are selected
For enquiry of historical data library information, information comparative analysis obtains prediction numerical value, and correction data information is as follows：History exports wind turbine pressure
Power, history generation schedule, history generated energy, history coal card numerical value, history coal sulfur-bearing numerical value and/or history coal are aqueous
Point, the data that each history parameters of the historical information of grey library yield are mainly recorded by computer obtain, and are believed by comparison data
If breath wherein five kinds of data above information are identical, result is using historical data information result at that time.
If Pb is less than 1, average price predicted method is selected, i.e., is wherein four kinds of data informations are identical by comparison data information
System calculates of that month No. 1 and starts to the average average price data information of current time automatically).
If Pb is equal to 1, convergence analysis method is selected, i.e., by comparison data information, wherein there are three types of only or less
Identical data then starts convergence analysis.Convergence analysis pattern, generally when data information Trend judgement in the last week, to pass through letter
Single statistical analysis calculating can be obtained data information.
Another aspect of the present invention additionally provides a kind of prediction technique of grey library yield, and Fig. 6 is the ash of the embodiment of the present invention
The step schematic diagram of the prediction technique of library yield, as shown in fig. 6, the prediction technique includes the following steps：
S1, the historical information according to grey library yield and grey library production information to be predicted, determine grey library production information to be predicted
The first prediction result；
S2, one neural network prediction model of structure, and according to the historical information and grey library production information to be predicted, determine
Second prediction result of the grey library production information to be predicted；
S3, first prediction result and second prediction result are subjected to convergent operation, determine the grey library to be predicted
The final prediction result of production information；And
S4, second prediction result is replaced using the final prediction result, to pre- to update the neural network
Survey model.
Each step is described in detail in conjunction with Fig. 7 to Fig. 9 with that.
S1, the historical information according to grey library yield and grey library production information to be predicted, determine grey library production information to be predicted
The first prediction result.
Wherein, the historical information of the grey library yield may include following history parameters：History exports blower pressure, history
Generation schedule, history generated energy, history coal card numerical value, history coal sulfur-bearing numerical value and/or history coal moisture content, grey library production
The data that each history parameters of the historical information of amount are mainly recorded by computer obtain；The grey library production information to be predicted can
To include following parameter：Export blower pressure, generation schedule, generated energy, coal card numerical value, coal sulfur-bearing numerical value and/or coal
Each parameter of moisture content, ash library to be predicted production information mainly judges according to actual conditions.
Fig. 7 is the sub-step signal of the first prediction result of the determination of the embodiment of the present invention grey library production information to be predicted
Figure specifically includes following sub-step as shown in fig. 7, determining the first prediction result of grey library production information to be predicted in step S1：
The database of the historical information of the grey library yield of S11, foundation for storing extraneous input.
Matching in S12, the grey library production information to be predicted of the extraneous input of inquiry and the database between historical information
As a result.
S13, according to the matching result, combining environmental factor determines the first prediction knot of grey library production information to be predicted
Fruit.That is, according to the matching degree of historical information and grey library production information to be predicted in the database, in conjunction with such as temperature
With the environmental factors such as season, you can determine the first prediction result.For example, there are historical information and grey library to be predicted in database
The matching degree of production information is 95%, and is that winter, temperature are low at this time, then coal amount demand is big, therefore, the first prediction result at this time
More than 95% of the grey library yield corresponding to the historical information, increased amplitude can be selected as the case may be.
S2, one neural network prediction model of structure, and according to the historical information and grey library production information to be predicted, determine
Second prediction result of the grey library production information to be predicted.
Fig. 8 is the specific steps signal of the prediction result of the determination of the embodiment of the present invention grey library production information to be predicted
Figure, as shown in figure 8, the specific following sub-step of the step：
S21, the prediction technique for determining the grey library production information to be predicted；
S22, the prediction technique for determining the grey library production information to be predicted；And
According to the prediction technique and historical information, combination processing sale pattern determines the second prediction result, described
Processing sale pattern includes pattern of underselling, free roping pattern and monovalent sale pattern.
Pattern, free roping pattern and monovalent sale pattern are undersold, this is three kinds of work of treated coal ash sale
Pattern can be selected according to actual conditions.
Fig. 9 is the sub-step schematic diagram of the step S21 of the embodiment of the present invention, and Fig. 9 is the neural network of the embodiment of the present invention
The schematic diagram of prediction model, as shown in figs. 5 and 9, step S21 specifically include following sub-step：
S211, judge that the parameter whether there is successively, if each parameter in the presence of if judging result Aa be 1, if being not present
Then judging result Aa is -1, and the judging result Aa is denoted as to six threshold values for inputting neuron Xa of neural network model, 1
≤a≤6；
S212, two intrerneuron Xb that the prediction technique is denoted as to neural network model, 7≤b≤8；
S213, according to the historical information of the grey library yield, determine the threshold values Fb of each intrerneuron, respectively input neuron
Weights Wab between the intrerneuron, wherein the value of Fb is 0 and ± 1；I.e. according to the historical information of grey library yield
Database, referring again to the parameters of grey library production information to be predicted, you can obtain Fb and Wab；
S214, the weighted sum for determining each intrerneuronAnd according to the value of Pb,
Select prediction technique：
If Pb is more than 1, historical forecast analytic approach is selected, wherein historical forecast analytic approach operating mode is query history
Database information, information comparative analysis obtain prediction numerical value, and correction data information is as follows：History exports blower pressure, history hair
Electric plan, history generated energy, history coal card numerical value, history coal sulfur-bearing numerical value and/or history coal moisture content, grey library yield
The data that are mainly recorded by computer of each history parameters of historical information obtain, by wherein five kinds of comparison data information with
If upper data information is identical, result is using historical data information result at that time.
If Pb is less than 1, average price predicted method is selected, i.e., is wherein four kinds of data informations are identical by comparison data information
System calculates of that month No. 1 and starts to the average average price data information for currently subtracting one automatically.
If Pb is equal to 1, convergence analysis method is selected, i.e., by comparison data information, wherein there are three types of only or less
Identical data then starts convergence analysis.Convergence analysis pattern, when data information Trend judgement in the last week, to obtain data letter
Breath.
Convergent operation is carried out for step S3, by the first prediction result and second prediction result, specially：Calculate institute
The arithmetic mean of instantaneous value for stating the first prediction result and second prediction result, can correct the first prediction result so that most in this way
Actual value of the whole prediction result closer to grey library yield.
To sum up, historical information and to be predicted ash of the prediction meanss and method of grey library yield of the invention according to grey library yield
Library production information determines the first prediction result, and the second prediction result is determined in conjunction with neural network prediction model, finally by two
Prediction result carries out convergent, obtains the more accurately final prediction result of grey library yield, while can also constantly update the nerve
Network Prediction Model so that final prediction result is more accurate, improves accuracy.
Furthermore "comprising" does not exclude the presence of element or step not listed in the claims." one " before element
Or "one" does not exclude the presence of multiple such elements.
The word of specification and ordinal number such as " first ", " second ", " third " etc. used in claim, with modification
Corresponding element, itself is not meant to that the element has any ordinal number, does not also represent the suitable of a certain element and another element
Sequence in sequence or manufacturing method, the use of those ordinal numbers are only used for enabling the element with certain name and another tool
There is the element of identical name that can make clear differentiation.
Particular embodiments described above has carried out further in detail the purpose of the present invention, technical solution and advantageous effect
It describes in detail bright, it should be understood that the above is only a specific embodiment of the present invention, is not intended to restrict the invention, it is all
Within the spirit and principles in the present invention, any modification, equivalent substitution, improvement and etc. done should be included in the guarantor of the present invention
Within the scope of shield.
Claims (13)
1. a kind of prediction meanss of ash library yield, including：
First prediction module determines ash to be predicted for the historical information and ash library to be predicted production information according to grey library yield
First prediction result of library production information；
Second prediction module, for building a neural network prediction model, and according to the historical information and grey library production to be predicted
Information is measured, determines the second prediction result of the grey library production information to be predicted；
It is pre- to determine that this is waited for for first prediction result and second prediction result to be carried out convergent operation for computing module
Survey the final prediction result of grey library production information；And
Update module replaces second prediction result, to update the neural network prediction using the final prediction result
Model.
2. the apparatus according to claim 1, wherein
The historical information of ash library yield includes at least following history parameters：History export blower pressure, history generation schedule,
History generated energy, history coal card numerical value, history coal sulfur-bearing numerical value and/or history coal moisture content；
The grey library production information to be predicted includes at least following parameter：Export blower pressure, generation schedule, generated energy, coal
Card numerical value, coal sulfur-bearing numerical value and/or coal moisture content.
3. the apparatus of claim 2, wherein second prediction module includes：
Method determination unit, the prediction technique for determining the grey library production information to be predicted；And
Pattern determining unit, for according to the prediction technique and historical information, combination processing sale pattern to determine second
Prediction result, the processing sale pattern include pattern, the free roping pattern undersold, and monovalent sale pattern.
4. device according to claim 3, wherein the method determination unit includes：
Judging unit, for judging that the parameter whether there is successively, if each parameter in the presence of if judging result Aa be 1, if not
It is -1 in the presence of then judging result Aa, and the judging result Aa is denoted as to six valves for inputting neuron Xa of neural network model
Value, 1≤a≤6；
Converting unit, two intrerneuron Xb for the prediction technique to be denoted as to neural network model, 7≤b≤8；
Determination unit determines the threshold values Fb of each intrerneuron, each input god for the historical information according to the grey library yield
Through the weights Wab between member and the intrerneuron, wherein the value of Fb is 0 and ± 1；And
Selecting unit, the weighted sum for determining each intrerneuronAnd taking according to Pb
Value selects prediction technique：If Pb is more than 1, historical forecast analytic approach is selected；If Pb is less than 1, average price predicted method is selected；If
Pb is equal to 1, then selects convergence analysis method.
5. the apparatus according to claim 1, wherein the first prediction module includes：
Database, the historical information of the grey library yield for storing extraneous input；
Query unit, for inquiring in the extraneous grey library production information to be predicted inputted and the database between historical information
Matching result；And
Predicting unit, for according to the matching result, combining environmental factor to determine that the first of grey library production information to be predicted is pre-
Survey result.
6. the apparatus according to claim 1, wherein the convergent operation of the arithmetic element is the arithmetic element to described
First prediction result seeks arithmetic mean of instantaneous value with second prediction result.
7. a kind of prediction technique of ash library yield, is applied to the device as described in any in claim 1 to 6, this method includes：
According to the historical information of grey library yield and grey library production information to be predicted, determine that the first of grey library production information to be predicted is pre-
Survey result；
A neural network prediction model is built, and according to the historical information and grey library production information to be predicted, it is pre- to determine that this is waited for
Survey the second prediction result of grey library production information；
First prediction result and second prediction result are subjected to convergent operation, determine the grey library production information to be predicted
Final prediction result；And
Second prediction result is replaced using the final prediction result, to update the neural network prediction model.
8. according to the method described in claim 7, wherein, the historical information of ash library yield is joined including at least following history
Number：History exports blower pressure, history generation schedule, history generated energy, history coal card numerical value, history coal sulfur-bearing numerical value
And/or history coal moisture content；The grey library production information to be predicted includes at least following parameter：Export blower pressure, power generation
Plan, generated energy, coal card numerical value, coal sulfur-bearing numerical value and/or coal moisture content.
9. according to the method described in claim 8, wherein it is determined that the second prediction result of the grey library production information to be predicted, packet
Include step：
Determine the prediction technique of the grey library production information to be predicted；And
According to the prediction technique and historical information, combination processing sale pattern determines the second prediction result, the processing
Sale pattern includes pattern of underselling, free roping pattern and monovalent sale pattern.
10. according to the method described in claim 9, wherein it is determined that the prediction technique of the grey library production information to be predicted, including step
Suddenly：
Judge that the parameter whether there is successively, if each parameter in the presence of if judging result Aa be 1, if be not present judging result
Aa is -1, and the judging result Aa is denoted as to six threshold values for inputting neuron Xa of neural network model, 1≤a≤6；
The prediction technique is denoted as to two intrerneuron Xb of neural network model, 7≤b≤8；
According to the historical information of the grey library yield, determine the threshold values Fb of each intrerneuron, each input neuron with it is described in
Between weights Wab between neuron, wherein the value of Fb is 0 and ± 1；And
Determine the weighted sum of each intrerneuronAnd according to the value of Pb, select prediction side
Method：If Pb is more than 1, historical forecast analytic approach is selected；If Pb is less than 1, average price predicted method is selected；If Pb is equal to 1, select
Convergence analysis method.
11. according to the method described in claim 10, wherein it is determined that the first prediction result of ash library to be predicted production information, packet
Include step：
Establish the database of the historical information of the grey library yield for storing extraneous input；
Matching result in the grey library production information to be predicted and the database of the extraneous input of inquiry between historical information；And
According to the matching result, combining environmental factor determines the first prediction result of grey library production information to be predicted.
12. the method according to claim 11, wherein：
The historical forecast analytic approach includes step：The database is inquired, and the parameter comparison, going through in database at this time
There are at least five kinds of identical parameter informations in history parameter, with the parameter using the corresponding price of the history parameters；
The average price predicted method includes step：The database is inquired, and the parameter comparison, the history ginseng in database at this time
There are four kinds of identical parameter informations with the parameter for number, are started to the average price of current time using of that month No. 1；
The convergence analysis method includes step：The database is inquired, and the parameter comparison, the history ginseng in database at this time
At most there are three kinds of identical parameter informations with the parameter for number, judge the trend when the historical data information in the last week, really
Price lattice.
13. according to the method described in claim 7, wherein, the first prediction result and second prediction result are carried out convergent
Operation, specially：Calculate the arithmetic mean of instantaneous value of first prediction result and second prediction result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810440144.7A CN108596405B (en) | 2018-05-09 | 2018-05-09 | Prediction method and system for ash warehouse yield |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810440144.7A CN108596405B (en) | 2018-05-09 | 2018-05-09 | Prediction method and system for ash warehouse yield |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108596405A true CN108596405A (en) | 2018-09-28 |
| CN108596405B CN108596405B (en) | 2020-03-06 |
Family
ID=63636240
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810440144.7A Active CN108596405B (en) | 2018-05-09 | 2018-05-09 | Prediction method and system for ash warehouse yield |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108596405B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109979033A (en) * | 2019-03-26 | 2019-07-05 | 广东艾胜物联网科技有限公司 | A kind of industrial production yield monitoring method based on width learning algorithm |
| CN113610657A (en) * | 2021-10-10 | 2021-11-05 | 江苏四方精密钢管有限公司 | Method and system for popularizing and selling steel pipe products |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN105512758A (en) * | 2015-12-01 | 2016-04-20 | 中国石油天然气股份有限公司 | Oil-gas well yield prediction method and device |
-
2018
- 2018-05-09 CN CN201810440144.7A patent/CN108596405B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN105512758A (en) * | 2015-12-01 | 2016-04-20 | 中国石油天然气股份有限公司 | Oil-gas well yield prediction method and device |
Non-Patent Citations (2)
| Title |
|---|
| 张祺: "中国石油生产量的组合预测模型及应用", 《经济问题》 * |
| 戴枫: "燃煤发电固体排放物销售项目优化决策模型研究", 《中国博士学位论文全文数据库》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109979033A (en) * | 2019-03-26 | 2019-07-05 | 广东艾胜物联网科技有限公司 | A kind of industrial production yield monitoring method based on width learning algorithm |
| CN109979033B (en) * | 2019-03-26 | 2021-08-20 | 广东艾胜物联网科技有限公司 | Industrial production yield monitoring method based on width learning algorithm |
| CN113610657A (en) * | 2021-10-10 | 2021-11-05 | 江苏四方精密钢管有限公司 | Method and system for popularizing and selling steel pipe products |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108596405B (en) | 2020-03-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Deng et al. | Simulation-based evaluation of defuzzification-based approaches to fuzzy multiattribute decision making | |
| Hu et al. | Research and application of a hybrid model based on Meta learning strategy for wind power deterministic and probabilistic forecasting | |
| CN112100745B (en) | Automobile girder steel mechanical property prediction method based on LDA theory | |
| CN106067088A (en) | E-bank accesses detection method and the device of behavior | |
| CN105844371A (en) | Electricity customer short-term load demand forecasting method and device | |
| CN108320053A (en) | A kind of region electricity demand forecasting method, apparatus and system | |
| CN113627735B (en) | Early warning method and system for engineering construction project security risk | |
| CN111091241A (en) | BP neural network-based drug sales prediction and decision method and system | |
| CN113408869A (en) | Power distribution network construction target risk assessment method | |
| CN111582548A (en) | Power load prediction method based on multivariate user behavior portrait | |
| CN112950067A (en) | Power consumer electricity utilization efficiency assessment method based on fuzzy comprehensive evaluation | |
| CN105844353A (en) | Aquatic product price prediction method and device | |
| CN108596405A (en) | The prediction technique and system of grey library yield | |
| CN112884359A (en) | Electric power spot market risk assessment method | |
| CN112818525A (en) | Hydroelectric generating set state evaluation method and system | |
| CN108764528A (en) | A kind of Daily treatment cost interval prediction method of allowed for influencing factors analysis | |
| Son et al. | Prediction of government-owned building energy consumption based on an RReliefF and support vector machine model | |
| Yang et al. | Short-term wind power prediction based on multi-parameters similarity wind process matching and weighed-voting-based deep learning model selection | |
| CN108830405B (en) | Real-time power load prediction system and method based on multi-index dynamic matching | |
| CN113361776A (en) | Power load probability prediction method based on user power consumption behavior clustering | |
| Azadeh et al. | An integrated PCA DEA framework for assessment and ranking of manufacturing systems based on equipment performance | |
| CN110142803B (en) | Method and device for detecting working state of mobile welding robot system | |
| CN107705160A (en) | A kind of combination econometrics and heuristic intelligent automobile Method for Sales Forecast method and system | |
| CN110084400A (en) | Information forecasting method, device, computer equipment and storage medium | |
| CN110533213A (en) | Transmission line of electricity defect Risk Modeling and its prediction technique based on support vector machines |
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 | ||
| CB03 | Change of inventor or designer information |
Inventor after: Zhang Guofeng Inventor after: Liu Huigen Inventor after: Xu Hongfei Inventor after: Dai Xiaorui Inventor after: Zhou Cheng Inventor after: Tan Wei Inventor before: Zhang Guofeng Inventor before: Liu Huigen Inventor before: Xu Hongfei Inventor before: Dai Xiaorui Inventor before: Zhou Cheng |
|
| CB03 | Change of inventor or designer information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |