CN109886445A - A kind of tomorrow requirement prediction technique based on material requirements property quantification - Google Patents
A kind of tomorrow requirement prediction technique based on material requirements property quantification Download PDFInfo
- Publication number
- CN109886445A CN109886445A CN201811526760.0A CN201811526760A CN109886445A CN 109886445 A CN109886445 A CN 109886445A CN 201811526760 A CN201811526760 A CN 201811526760A CN 109886445 A CN109886445 A CN 109886445A
- Authority
- CN
- China
- Prior art keywords
- materials
- value
- goods
- modeling
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to power grid operation maintenance technology fields, and in particular to a kind of tomorrow requirement prediction technique based on material requirements property quantification, comprising the following steps: A) establish goods and materials historic demand tables of data；B goods and materials) are divided into continuity demand characteristic goods and materials and discontinuity demand characteristic goods and materials；C linear prediction model modeling) is carried out to the historic demand data of continuity demand characteristic goods and materials；D exponential smoothing modeling and Grey Prediction Modeling) are carried out respectively to the historic demand data of discontinuity demand characteristic goods and materials；E it) rolls and obtains nearest material requirements data, repeat step A-E and roll update prediction.Substantial effect of the invention is: it is effective to improve material requirements prediction accuracy by the way that goods and materials are carried out classification prediction, enterprise procurement cost is saved, market synthesized competitiveness is promoted.
Description
Technical field
The present invention relates to power grid operation maintenance technology fields, and in particular to a kind of future based on material requirements property quantification
Needing forecasting method.
Background technique
In the construction and maintenance of power grid, need to consume a large amount of goods and materials and equipment.Different goods and materials have different
Consumption needs to prepare enough goods and materials for the normal operation for maintaining power grid.And the scale of power grid constantly expands at present, goods and materials
Consumption is increasing, in the case where lacking amount of materials consumed prediction, be easy to cause goods and materials short.Part goods and materials shortage can shadow
It rings normal maintenance and updates construction speed.If goods and materials prepare excessively, it will cause the wastes of storage capacity.Thus need to matching
The demand in various goods and materials futures is predicted in net.
Chinese patent CN102831489B, publication date on March 9th, 2016, a kind of power matching network construction material requirements prediction
Method various is gone through comprising steps of obtaining the parameter and various history item goods and materials usage amounts of history item preset attribute by described
History project goods and materials usage amount is normalized into preset range；According to the parameter of the history item preset attribute, standardized each
Kind history item goods and materials usage amount and default implicit number of nodes, construct prediction model using extreme learning machine, according to the prediction
Model determines implicit node weights parameter matrix；The parameter for obtaining project preset attribute to be measured, according to the implicit node weights
The parameter of parameter matrix and the project preset attribute to be measured determines that corresponding project goods and materials to be measured use using the prediction model
The predicted value of amount restores the predicted value in corresponding ratio is standardized, determines corresponding project goods and materials usage amount to be measured.Its scheme is real
Now model is simple to be predicted to a variety of goods and materials dosages simultaneously, takes into account relevance.But it relies on a large amount of data and is trained, sample
The accuracy of the quality of data most prediction result is affected, and does not distinguish and treats to different types of goods and materials, thus
Prediction accuracy is not high enough.
Summary of the invention
The technical problem to be solved by the present invention is the technology that the method for the prediction power grid consumption of materials is lack of pertinence at present is asked
Topic.It is pre- to propose a kind of higher tomorrow requirement based on material requirements property quantification of the accuracy for predicting power grid materials and equipment classification
Survey method.
In order to solve the above technical problems, the technical solution used in the present invention are as follows: one kind is based on material requirements property quantification
Tomorrow requirement prediction technique, comprising the following steps: A) obtain goods and materials history receive data, establish goods and materials historic demand data
Table；B goods and materials) are divided into continuity demand characteristic goods and materials and discontinuity demand characteristic goods and materials；C) to continuity demand characteristic goods and materials
Historic demand data carry out stationary test, if by carrying out linear prediction model modeling if stationary test, conversely, right
The historic demand data of continuity demand characteristic goods and materials carry out linear prediction model modeling after carrying out single exponential smoothing processing；D)
Exponential smoothing modeling and Grey Prediction Modeling, correlation index are carried out respectively to the historic demand data of discontinuity demand characteristic goods and materials
The accuracy of smooth modeling and Grey Prediction Modeling, the higher model of accuracy of selection as discontinuity demand characteristic goods and materials not
Carry out Demand Forecast Model；E it) rolls and obtains nearest material requirements data, repeat step A-E and roll update prediction.
Preferably, the field of the goods and materials historic demand tables of data include project name, material, material description, unit,
Plan gets the time, actually gets time, the plan amount of getting and the practical amount of getting.
Preferably, the method for the stationary test are as follows: calculate the historic demand data of continuity demand characteristic goods and materials
P value, if P value be less than significance 0.05 if stationary test pass through.
Preferably, the linear prediction model is AR model or arma modeling, when going through for continuity demand characteristic goods and materials
The absolute value of the high-order autocorrelation value of history demand data is less than 0.2 and when the absolute value of its partial autocorrelation value is less than 0.3, the line
Property prediction model be AR model, conversely, the linear prediction model be arma modeling.
Preferably, the recursive calculation formula of the exponential smoothing modeling are as follows:
WhereinFor value to be predicted,For t moment obtain the predicted value to subsequent time, i.e.,AndThe t+1 for being
The value at moment, ytFor the actual value of t moment,It is the t-1 moment to the predicted value of subsequent time, the t moment as predicted
Value, α are weighted value, α value interval be [0.65,1).
Preferably, the exponential smoothing is modeled as double smoothing modeling, the side of the double smoothing modeling
Method are as follows: after carrying out exponential smoothing modeling to the historic demand data of discontinuity demand characteristic goods and materials, each moment is obtained
Predicted value is repeated once exponential smoothing modeling, as final model as new initial data.
Substantial effect of the invention is: it is effective to improve material requirements prediction standard by the way that goods and materials are carried out classification prediction
Exactness saves enterprise procurement cost, promotes market synthesized competitiveness.
Detailed description of the invention
Fig. 1 is one needing forecasting method flow diagram of embodiment.
Fig. 2 is using exponential smoothing modeling and forecasting result figure.
Fig. 3 is the prediction result figure using grey method.
Specific embodiment
Below by specific embodiment, and in conjunction with attached drawing, a specific embodiment of the invention is further described in detail.
Embodiment one:
A kind of tomorrow requirement prediction technique based on material requirements property quantification, as shown in Figure 1, being one requirement forecasting side of embodiment
Method flow diagram, the present embodiment is the following steps are included: A) obtain goods and materials history receive data, establish goods and materials historic demand data
Table；B goods and materials) are divided into continuity demand characteristic goods and materials and discontinuity demand characteristic goods and materials；C) to continuity demand characteristic goods and materials
Historic demand data carry out stationary test, if by carrying out linear prediction model modeling if stationary test, conversely, right
The historic demand data of continuity demand characteristic goods and materials carry out linear prediction model modeling after carrying out single exponential smoothing processing；D)
Exponential smoothing modeling and Grey Prediction Modeling, correlation index are carried out respectively to the historic demand data of discontinuity demand characteristic goods and materials
The accuracy of smooth modeling and Grey Prediction Modeling, the higher model of accuracy of selection as discontinuity demand characteristic goods and materials not
Carry out Demand Forecast Model；E it) rolls and obtains nearest material requirements data, repeat step A-E and roll update prediction.
High-tension fuse is discontinuity demand characteristic goods and materials, as shown in Fig. 2, for using exponential smoothing modeling and forecasting result
Scheme, indicate to model high-tension fuse goods and materials data 2016-2017 in figure, abscissa indicates month in Fig. 2, indulges and sits
Mark indicates of that month quantity required, and carry out Demand Forecast as a result, dotted line indicates prediction result in figure, as shown in figure 3, to adopt
With the prediction result figure of grey method, dotted line indicates prediction result in figure, respectively will height between prediction result and 2016-2017
Pressure fuse goods and materials data compare, and abscissa indicates month in Fig. 3, and ordinate indicates of that month quantity required, obtains index
The prediction accuracy of smooth modeling and grey method, exponential smoothing modeling error rate are 16.62%, the model of grey method
Error rate is 25.88%, thus selects the prediction result of exponential smoothing modeling, demand as the following high-tension fuse it is pre-
Survey result.
The field of goods and materials historic demand tables of data include project name, material, material description, unit, plan get the time,
Actually get time, the plan amount of getting and the practical amount of getting.The method of stationary test are as follows: calculate continuity demand characteristic goods and materials
Historic demand data P value, if P value be less than significance 0.05 if stationary test pass through.
Linear prediction model is AR model or arma modeling, when the height of the historic demand data of continuity demand characteristic goods and materials
For the absolute value of rank autocorrelation value less than 0.2 and when the absolute value of its partial autocorrelation value is less than 0.3, linear prediction model is AR mould
Type, conversely, linear prediction model is arma modeling.
The recursive calculation formula of exponential smoothing modeling are as follows:WhereinFor to pre-
Measured value,For t moment obtain the predicted value to subsequent time, i.e.,AndThe value at the t+1 moment for being, ytFor t
The actual value at moment,It is the t-1 moment to the predicted value of subsequent time, the value for the t moment as predicted, α is weighted value, α
[0.65,1) value interval is.
Exponential smoothing is modeled as double smoothing modeling, the method for double smoothing modeling are as follows: to discontinuity demand
After the historic demand data of feature goods and materials carry out exponential smoothing modeling, using predicted value obtained of each moment as newly original
Data are repeated once exponential smoothing modeling, as final model.
Above-mentioned embodiment is only a preferred solution of the present invention, not the present invention is made in any form
Limitation, there are also other variations and modifications on the premise of not exceeding the technical scheme recorded in the claims.
Claims (8)
1. a kind of tomorrow requirement prediction technique based on material requirements property quantification, which is characterized in that
The following steps are included:
A it) obtains goods and materials history and receives data, establish goods and materials historic demand tables of data；
B goods and materials) are divided into continuity demand characteristic goods and materials and discontinuity demand characteristic goods and materials；
C stationary test) is carried out to the historic demand data of continuity demand characteristic goods and materials, is carried out if through stationary test
Linear prediction model modeling, conversely, then being carried out at single exponential smoothing to the historic demand data of continuity demand characteristic goods and materials
Linear prediction model modeling is carried out after reason；
D exponential smoothing modeling and Grey Prediction Modeling) are carried out respectively to the historic demand data of discontinuity demand characteristic goods and materials, it is right
Ratio index smoothly models and the accuracy of Grey Prediction Modeling, and the higher model of accuracy of selection is as discontinuity demand characteristic object
The tomorrow requirement prediction model of money；
E it) rolls and obtains nearest material requirements data, repeat step A-E and roll update prediction.
2. a kind of tomorrow requirement prediction technique based on material requirements property quantification according to claim 1, feature exist
In, the field of the goods and materials historic demand tables of data include project name, material, material description, unit, plan get the time,
Actually get time, the plan amount of getting and the practical amount of getting.
3. a kind of tomorrow requirement prediction technique based on material requirements property quantification according to claim 1, feature exist
In the method for the stationary test are as follows: the P value for calculating the historic demand data of continuity demand characteristic goods and materials, if P value is less than
Then stationary test passes through significance 0.05.
4. a kind of tomorrow requirement prediction technique based on material requirements property quantification according to claim 1 or 2, feature
It is, the linear prediction model is AR model or arma modeling, when the historic demand data of continuity demand characteristic goods and materials
Less than 0.2 and when the absolute value of its partial autocorrelation value is less than 0.3, the linear prediction model is the absolute value of high-order autocorrelation value
AR model, conversely, the linear prediction model is arma modeling.
5. a kind of tomorrow requirement prediction technique based on material requirements property quantification according to claim 3, feature exist
In the linear prediction model is AR model or arma modeling, when the height of the historic demand data of continuity demand characteristic goods and materials
For the absolute value of rank autocorrelation value less than 0.2 and when the absolute value of its partial autocorrelation value is less than 0.3, the linear prediction model is AR
Model, conversely, the linear prediction model is arma modeling.
6. a kind of tomorrow requirement prediction technique based on material requirements property quantification according to claim 1 or 2, feature
It is, the recursive calculation formula of the exponential smoothing modeling are as follows:WhereinFor to
Predicted value,For t moment obtain the predicted value to subsequent time, i.e.,AndThe value at the t+1 moment for being, yt
For the actual value of t moment,It is the t-1 moment to the predicted value of subsequent time, the value for the t moment as predicted, α is weight
Value, α value interval be [0.65,1).
7. a kind of tomorrow requirement prediction technique based on material requirements property quantification according to claim 3, feature exist
In the recursive calculation formula of the exponential smoothing modeling are as follows:WhereinFor to pre-
Measured value,For t moment obtain the predicted value to subsequent time, i.e.,AndThe value at the t+1 moment for being, ytFor t
The actual value at moment,It is the t-1 moment to the predicted value of subsequent time, the value for the t moment as predicted, α is weighted value, α
[0.65,1) value interval is.
8. a kind of tomorrow requirement prediction technique based on material requirements property quantification according to claim 6, feature exist
In the exponential smoothing is modeled as double smoothing modeling, the method for the double smoothing modeling are as follows: need to discontinuity
After asking the historic demand data of feature goods and materials to carry out exponential smoothing modeling, using predicted value obtained of each moment as new original
Beginning data are repeated once exponential smoothing modeling, as final model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811526760.0A CN109886445A (en) | 2018-12-13 | 2018-12-13 | A kind of tomorrow requirement prediction technique based on material requirements property quantification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811526760.0A CN109886445A (en) | 2018-12-13 | 2018-12-13 | A kind of tomorrow requirement prediction technique based on material requirements property quantification |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109886445A true CN109886445A (en) | 2019-06-14 |
Family
ID=66925024
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811526760.0A Pending CN109886445A (en) | 2018-12-13 | 2018-12-13 | A kind of tomorrow requirement prediction technique based on material requirements property quantification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109886445A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110866656A (en) * | 2019-11-26 | 2020-03-06 | 广州供电局有限公司 | Power material demand prediction method and device, computer equipment and storage medium |
| CN111192083A (en) * | 2019-12-26 | 2020-05-22 | 浙江大学 | Method for predicting intermittent demand |
| CN113283924A (en) * | 2020-02-20 | 2021-08-20 | 纬创资通股份有限公司 | Demand forecasting method and demand forecasting device |
| CN113361745A (en) * | 2021-05-07 | 2021-09-07 | 云南电网有限责任公司曲靖供电局 | Power distribution network material demand prediction method and system |
| CN113902188A (en) * | 2021-10-08 | 2022-01-07 | 国网江苏省电力有限公司镇江供电分公司 | Electric energy metering material demand prediction method |
| CN113919687A (en) * | 2021-10-08 | 2022-01-11 | 国网江苏省电力有限公司镇江供电分公司 | Electric energy metering material inventory distribution method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2016208556A (en) * | 2013-10-03 | 2016-12-08 | 三菱電機株式会社 | Demand prediction device and demand prediction method |
| CN106600032A (en) * | 2016-10-28 | 2017-04-26 | 北京国电通网络技术有限公司 | Inventory material demand predicting method and device |
| CN107993033A (en) * | 2017-11-14 | 2018-05-04 | 广东电网有限责任公司物流服务中心 | A kind of Power Material Forecasting Methodology |
-
2018
- 2018-12-13 CN CN201811526760.0A patent/CN109886445A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2016208556A (en) * | 2013-10-03 | 2016-12-08 | 三菱電機株式会社 | Demand prediction device and demand prediction method |
| CN106600032A (en) * | 2016-10-28 | 2017-04-26 | 北京国电通网络技术有限公司 | Inventory material demand predicting method and device |
| CN107993033A (en) * | 2017-11-14 | 2018-05-04 | 广东电网有限责任公司物流服务中心 | A kind of Power Material Forecasting Methodology |
Non-Patent Citations (2)
| Title |
|---|
| 夏丽: "基于ARIMA模型及回归分析的区域用电量预测方法研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
| 龚巍: "基于需求特性分类的电力物资库存控制与需求预测方法研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110866656A (en) * | 2019-11-26 | 2020-03-06 | 广州供电局有限公司 | Power material demand prediction method and device, computer equipment and storage medium |
| CN111192083A (en) * | 2019-12-26 | 2020-05-22 | 浙江大学 | Method for predicting intermittent demand |
| CN113283924A (en) * | 2020-02-20 | 2021-08-20 | 纬创资通股份有限公司 | Demand forecasting method and demand forecasting device |
| CN113283924B (en) * | 2020-02-20 | 2024-05-07 | 纬创资通股份有限公司 | Demand prediction method and demand prediction device |
| CN113361745A (en) * | 2021-05-07 | 2021-09-07 | 云南电网有限责任公司曲靖供电局 | Power distribution network material demand prediction method and system |
| CN113902188A (en) * | 2021-10-08 | 2022-01-07 | 国网江苏省电力有限公司镇江供电分公司 | Electric energy metering material demand prediction method |
| CN113919687A (en) * | 2021-10-08 | 2022-01-11 | 国网江苏省电力有限公司镇江供电分公司 | Electric energy metering material inventory distribution method |
| CN113902188B (en) * | 2021-10-08 | 2023-12-22 | 国网江苏省电力有限公司镇江供电分公司 | Electric energy metering material demand prediction method |
| CN113919687B (en) * | 2021-10-08 | 2024-03-08 | 国网江苏省电力有限公司镇江供电分公司 | Electric energy metering material stock distribution method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109886445A (en) | A kind of tomorrow requirement prediction technique based on material requirements property quantification | |
| Levén et al. | Inventory control with a modified Croston procedure and Erlang distribution | |
| Zhao et al. | Combining forecasts of electricity consumption in China with time-varying weights updated by a high-order Markov chain model | |
| CN108364085A (en) | A kind of take-away distribution time prediction technique and device | |
| CN109615226A (en) | A kind of operation indicator method for monitoring abnormality | |
| CN109214948A (en) | A kind of method and apparatus of electric system heat load prediction | |
| Behboudian et al. | A long-term prediction of domestic water demand using preprocessing in artificial neural network | |
| CN109165809B (en) | Power grid planning project investment sequencing assessment method under new electricity-to-electricity environment | |
| CN107506845A (en) | A kind of electricity sales amount Forecasting Methodology and its system based on multi-model fusion | |
| CN107993088A (en) | A kind of Buying Cycle Forecasting Methodology and device, electronic equipment | |
| CN111178585A (en) | Fault reporting amount prediction method based on multi-algorithm model fusion | |
| CN108830417A (en) | A kind of residential energy consumption prediction technique and system based on ARMA and regression analysis | |
| CN110019420A (en) | A kind of data sequence prediction technique and calculate equipment | |
| CN105956716A (en) | Total social electricity consumption prediction method based on industry economy and electricity relationship | |
| CN112288172A (en) | Prediction method and device for line loss rate of transformer area | |
| CN111178957B (en) | Method for early warning sudden increase of electric quantity of electricity consumption customer | |
| Roach | Reconciled boosted models for GEFCom2017 hierarchical probabilistic load forecasting | |
| Zhao et al. | Short-term microgrid load probability density forecasting method based on k-means-deep learning quantile regression | |
| CN108932562A (en) | Integrated energy system comprehensive benefit assessment model method for building up | |
| Žmuk et al. | Forecasting labour productivity in the European Union member states: is labour productivity changing as expected? | |
| CN107679666A (en) | A kind of power consumption prediction method based on Shapley values and economic development | |
| CN114757416A (en) | Water demand prediction method and system based on rating method | |
| CN110991768A (en) | Electric quantity prediction method and system | |
| CN105488595A (en) | Leading index construction method for monthly power consumption prediction model | |
| CN111967684A (en) | Metering asset active distribution method based on big data analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190614 |