CN117829380A - Method, system, equipment and medium for long-term prediction of power use - Google Patents
Method, system, equipment and medium for long-term prediction of power use Download PDFInfo
- Publication number
- CN117829380A CN117829380A CN202410245372.4A CN202410245372A CN117829380A CN 117829380 A CN117829380 A CN 117829380A CN 202410245372 A CN202410245372 A CN 202410245372A CN 117829380 A CN117829380 A CN 117829380A
- Authority
- CN
- China
- Prior art keywords
- power
- model
- prediction
- amount
- source 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 31
- 230000007774 longterm Effects 0.000 title claims abstract description 29
- 238000007667 floating Methods 0.000 claims abstract description 27
- 238000013145 classification model Methods 0.000 claims abstract description 17
- 238000005265 energy consumption Methods 0.000 claims abstract description 11
- 238000005457 optimization Methods 0.000 claims abstract description 5
- 230000005611 electricity Effects 0.000 claims description 30
- 230000001932 seasonal effect Effects 0.000 claims description 29
- 238000010586 diagram Methods 0.000 claims description 11
- 238000010606 normalization Methods 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 4
- 238000012549 training Methods 0.000 claims description 4
- 230000010354 integration Effects 0.000 claims description 3
- 238000007637 random forest analysis Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 238000004891 communication Methods 0.000 claims description 2
- YHXISWVBGDMDLQ-UHFFFAOYSA-N moclobemide Chemical compound C1=CC(Cl)=CC=C1C(=O)NCCN1CCOCC1 YHXISWVBGDMDLQ-UHFFFAOYSA-N 0.000 description 7
- 238000005311 autocorrelation function Methods 0.000 description 4
- 239000012530 fluid Substances 0.000 description 4
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 description 3
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 description 3
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 2
- 229910052799 carbon Inorganic materials 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000000737 periodic effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013467 fragmentation Methods 0.000 description 1
- 238000006062 fragmentation reaction Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000012447 hatching Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the technical field of power prediction, and provides a method, a system, equipment and a medium for predicting long-term power use, which are used for solving the problem of inaccurate power prediction at present, wherein a SARIMAX model is used for respectively predicting multi-source data related to power use to obtain a floating point value of corresponding power prediction amount, an energy consumption type is obtained according to a classification model, a limited space range of corresponding power prediction use amount is obtained based on the energy consumption type, the SARIMAX model is optimally adjusted based on the comparison of the limited space range of the power prediction use amount and the obtained floating point value of the power prediction use amount, and the power prediction is performed based on the model after the optimization adjustment; not only the multi-source data related to the power use is fully utilized, but also the model is optimized and adjusted based on the classification result obtained by the classification model, so that the power prediction of the subsequent model is more accurate.
Description
Technical Field
The invention belongs to the technical field of power prediction, and particularly relates to a method, a system, equipment and a medium for long-term prediction of power use.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
Long-term prediction of power usage is critical to ensure energy supply and demand balance, optimize grid construction, and reduce carbon emissions. In the conventional long-term prediction, the following problems mainly exist: 1. multi-source heterogeneous data fusion: predictions often involve a variety of data sources, such as historical power usage data, meteorological data (images), economic data, local GDP, industrial structures, and so forth. These data may come from different systems, with different formats and granularity, it is critical how to completely extract the useful information inside; 2. data completion type: the method is limited by various factors of insufficient physical environment and computer informatization, so that the acquired data is insufficient in integrity, poor in timeliness and serious in fault; 3. characteristics of the time series: the power data is typically time series data, which has seasonal, trending, and periodic characteristics. Therefore, how to accurately predict the long-term use of power using power-related data is a problem that needs to be solved at present.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention provides a method, a system, equipment and a medium for predicting the power use in a long term.
To achieve the above object, a first aspect of the present invention provides a power use long-term prediction method including:
acquiring historical multi-source data related to power use;
predicting based on a classification model according to the acquired historical multi-source data to obtain an electricity consumption type, and determining a limited space range of electricity prediction using amount according to the predicted electricity consumption type;
according to the type of the acquired historical multi-source data, respectively predicting based on the corresponding SARIMAX model to obtain a plurality of floating point values of the power prediction using amount;
comparing the limited space range of the power prediction using amount with a plurality of floating point values of the power prediction using amount, and if the floating point value of the power prediction using amount is not in the limited space range of the power prediction using amount, optimizing and adjusting the corresponding SARIMAX model to obtain an adjusted SARIMAX model;
and predicting the acquired power to be predicted by using related multi-source data based on the adjusted SARIMAX model to obtain a prediction result.
A second aspect of the present invention provides a long-term prediction system for electric power use, comprising:
the acquisition module is used for: acquiring historical multi-source data related to power use;
and a classification module: predicting based on a classification model according to the acquired historical multi-source data to obtain an electricity consumption type, and determining a limited space range of electricity prediction using amount according to the predicted electricity consumption type;
and (3) a regression module: according to the type of the acquired historical multi-source data, respectively predicting based on the corresponding SARIMAX model to obtain a plurality of floating point values of the power prediction using amount;
and an optimization module: comparing the limited space range of the power prediction using amount with a plurality of floating point values of the power prediction using amount, and if the floating point value of the power prediction using amount is not in the limited space range of the power prediction using amount, optimizing and adjusting the corresponding SARIMAX model to obtain an adjusted SARIMAX model;
and a prediction module: and predicting the acquired power to be predicted by using related multi-source data based on the adjusted SARIMAX model to obtain a prediction result.
A third aspect of the present invention provides a computer apparatus comprising: a processor, a memory, and a bus, the memory storing machine-readable instructions executable by the processor, the processor in communication with the memory via the bus when the computer device is running, the machine-readable instructions when executed by the processor performing a power use long-term prediction method.
A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs a method of long-term prediction of power usage.
The one or more of the above technical solutions have the following beneficial effects:
in the invention, a seasonal autoregressive comprehensive moving average model with exogenous variables, namely an SARIMAX model, predicts multi-source data related to power use respectively to obtain a floating point value of corresponding power prediction, an energy consumption type is obtained according to a classification model, a limited space range of corresponding power prediction use amount is obtained based on the energy consumption type, the SARIMAX model is optimally adjusted based on comparison of the limited space range of the power prediction use amount and the floating point value of the power prediction use amount obtained by the SARIMAX model, and the power prediction is performed based on the model after optimization adjustment.
Additional aspects of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention.
FIG. 1 is a flowchart of a method for long-term prediction of power usage based on the SARIMAX model in accordance with an embodiment of the present invention;
FIG. 2 is a flow chart of the construction of the SARIMAX regression flow model according to the first embodiment of the present invention;
FIG. 3 is a predicted value of the long-term month-to-month power consumption of the middle-aged person according to the embodiment of the invention;
FIG. 4 is a schematic diagram of an electric power in accordance with an embodiment of the present invention;
FIG. 5 is an autocorrelation diagram in accordance with a first embodiment of the present invention;
fig. 6 is a partial autocorrelation diagram in accordance with a first embodiment of the present invention.
Detailed Description
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the invention. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the present invention.
Embodiments of the invention and features of the embodiments may be combined with each other without conflict.
Example 1
The embodiment discloses a long-term prediction method for power use, comprising the following steps:
acquiring historical multi-source data related to power use;
predicting based on a classification model according to the acquired historical multi-source data to obtain an electricity consumption type, and determining a limited space range of electricity prediction using amount according to the predicted electricity consumption type;
according to the type of the acquired historical multi-source data, respectively predicting based on the corresponding SARIMAX model to obtain a plurality of floating point values of the power prediction using amount;
comparing the limited space range of the power prediction using amount with a plurality of floating point values of the power prediction using amount, and if the floating point value of the power prediction using amount is not in the limited space range of the power prediction using amount, optimizing and adjusting the corresponding SARIMAX model to obtain an adjusted SARIMAX model;
and predicting the acquired power to be predicted by using related multi-source data based on the adjusted SARIMAX model to obtain a prediction result.
In the embodiment, a dual-fluid model is adopted, namely, a model of a classified flow auxiliary regression flow is adopted, a classified flow model is used for constructing a limited space, a regression flow is used for calculating the power use condition, meanwhile, in order to ensure the accuracy of predicting the power trend of the regression flow, a seasonal autoregressive comprehensive moving average model (Seasonal Autoregressive Integrated Moving Average with Exogenous Variables, SARIMAX) with exogenous variables is particularly adopted for fitting the power data distribution condition of a time sequence, the traditional power use long-term prediction generally adopts only a single regression flow model for calculation, in the embodiment, a classified flow is introduced on the basis of using the regression flow for firstly constructing a result limited space, and in addition, the time sequence of a field to be predicted is also added into the model on the basis of a common regression flow, so that the model is ensured to fully fit the data distribution condition.
The method mainly adopts a SARIMAX model and SVM model double-fluid structure, the SVM model adjusts the output result as a limited control unit, and then the SARIMAX model is used for calculating the current power usage amount obtained by calculating the data of the industrial structure, population, past power usage condition, meteorological data and the like. The power utilization condition of the next month or the next year can be determined in advance, the power utilization condition is regulated in advance, planning is performed in advance, and power faults are avoided.
The scheme of this embodiment is described in detail with reference to fig. 1, and specifically includes:
step 1: and acquiring a power usage prediction related parameter table to be predicted.
And filling in parameter values to be predicted according to the power business related parameter template file.
In the present embodiment, the analysis of the parameters related to the power usage mainly refers to analyzing which field is urban population, GDP, industrial structure, user type, etc. in prediction, and performing corresponding outlier processing, normalization processing, discretization processing, etc.
Step 2: building power use prediction classification model based on SVM model
Discretization of long-term usage of electric power: the measured power related data are used, the power long-term use amount in the training data is discretized, and the power is packed into six types, which are respectively: the rated voltage of the electric equipment is lower than 220V, and the electricity consumption of the electric equipment is lower than 500 kilowatt-hours; the rated voltage of the electric equipment is lower than 220V, and the electricity consumption is higher than 500 kilowatt-hours, and the low-voltage electric load has high energy consumption; the rated voltage of the electric equipment is 1 kV-35 kV, and the annual electricity consumption is lower than the medium-voltage electricity consumption load low energy consumption type of 5000 kilowatt hours; the rated voltage of the electric equipment is 1 kV-35 kV, and the annual electricity consumption is higher than the high energy consumption type of medium-voltage electricity load with 5000 kilowatt hours; the rated voltage of the electric equipment is higher than 100kV, and the electricity consumption is lower than one hundred million kilowatt-hours; the rated voltage of the electric equipment is higher than 100kV in the year, and the electricity consumption is higher than one hundred million kilowatt-hour.
And converting the power usage related parameters into classification model prediction related values, and inputting the classification model prediction related values into a classification model to calculate what energy consumption type the classification model prediction related values are.
Power related data normalization: and screening out parameters related to the power consumption according to the Pearson correlation coefficient and the random forest integration model, and carrying out standardization processing on the selected parameters.
（1）
（2）
Wherein:: values after the current normalization; />: current values before normalization; />: the average value of the column in which the current parameter is located; />: standard deviation of the column in which the current parameter is located.
Parameters related to the amount of electricity usageThe method specifically comprises the following steps: GDP, industry structure, population, user type, electricity used in the last period, area, weather.
Step 3: based on SARIMAX model, constructing power use prediction model
The SARIMAX model is used to predict specific floating point values based on the power usage related parameters.
The seasonal autoregressive moving average model (Seasonal Autoregressive Integrated Moving Average with Exogenous Variables, SARIMAX) with exogenous variables is an extension of the ARIMA model, a time series model for seasonal data. Seasonal data is a periodic data, such as the same month of the year, different days of the same week, etc. The SARIMAX model can effectively model and predict seasonal data by taking into account seasonal variations in the time series.
A specific expression of the ARIMA model is ARIMA (p, d, q), where AR refers to the autoregressive process, MA: moving average, p: autoregressive term number, q: the number of terms, I, is the single integer, and d is the number of times the time series needs to be differentiated from non-stationary to stationary. When d=0, the ARIMA model becomes ARIMA. Taking into account most of the non-stationarity of the carbon emission time series, the general expression formula of the ARIMA (p, d, q) model is as follows by introducing a difference term d and a hysteresis operator l:
wherein,for the value of the previous cycle, +.>For error (S)>The model is first-order difference with difference coefficient, < ->Is an autocorrelation coefficient>As a constant term, the ARIMA model involves 3 parameters: the 3 parameters p, d, q reflect the trend of the sequence, but do not characterize the periodicity of the sequence. Thus SARIMAX increases the periodicity parameter: p, D, Q, S are denoted as sarmax (P, D, Q) S, and are specifically as follows:
wherein: s is a seasonal period;for seasonal shift operator, < ->As a seasonal regression coefficient, this parameter is seasonal in the model [ 3 ]; />Seasonal difference coefficient, which is a first order difference of [ 1 ] in the model, i.e. representingThe method comprises the steps of carrying out a first treatment on the surface of the P, D, Q the seasonal autoregressive order, the seasonal moving average order, and the seasonal differential order, respectively; />) And->A regression average coefficient polynomial and a moving average coefficient polynomial, respectively.
The idea of the SARIMAX model is to treat the data sequence formed by the predicted object over time as a random sequence, i.e. the time sequence is a set of random variables that depend on time t. By mathematically modeling the autocorrelation of the set of random variables, future values can be predicted from past and present time series values. The correlation between data can be reduced by taking into account seasonal differences, seasonal variations in the time series by taking into account seasonal autoregressions, and irregular seasonal fluctuations by taking seasonal moving averages before the time series prediction is performed. This allows seasonal variation factors to be incorporated into the features of the model to more accurately predict future seasonal data trends.
Step 3-1: normalization of the power usage: the analyzed power usage presents long tail distribution, so that small samples or low frequency samples can negatively affect the data quality, and simultaneously affect the prediction accuracy of the constructed model. Here, the amounts used in this scheme are normally distributed.
（3）
Wherein,for the amount of electricity usage, +.>The power consumption is normalized.
Step 3-2: according to the pearson correlation coefficient and the random forest integration model, parameters related to the receiving and releasing time are screened out (GDP, industrial structure, population, user type, upper period used electric quantity, area and weather), the selected parameters are standardized, and an electric power usage prediction model, namely SARIMAX, is constructed, specifically comprising the following steps:
1. and reading the data, preprocessing the data, setting the time as an index, and converting the data into a time sequence format.
2. As shown in fig. 4, a time chart of the original data is drawn, and the change trend and periodicity of the data are observed.
3. And checking the stability of the data, and carrying out differential processing on the unstable data to eliminate the trend.
The first order difference is used to form a new data set by subtracting the previous data from the next data conversion for the power usage.
4. And selecting exogenous factors (X), determining key factors influencing the use of electric power according to a scatter diagram and a correlation analysis method, and screening seven exogenous factors including GDP, an industrial structure, population, user type, upper-period electric quantity, region and weather.
5. The order of the SARIMAX model is determined and the values of p, q, d, P, D, Q are determined from the autocorrelation diagrams and the partial autocorrelation diagrams.
As shown in FIG. 5, p, q, d are filtered out as (2, 1, 2), the second one of FIG. 5 begins with shading, and 1 is first order differenced according to a timing diagram.
As shown in fig. 6, P, D, Q parameters are: in fig. 6, the hatching part from the beginning, 3 is the quaternary expression in the amount of electric power usage (1, 0, 3).
6. Model parameters are trained with seasonal and non-seasonal orders and exogenous variables and the SARIMAX model is fitted.
7. And predicting the power consumption of each month by using the fitted SARIMAX model, and evaluating the accuracy of the prediction result.
8. If the fitting result of the SARIMAX model to the historical data is St, reconstructing a residual sequence by using the residual sequence Qt of St.
As shown in fig. 2, the model training specifically includes the following steps:
and (3) data observation: and (5) detecting data stability and observing data time sequence.
And (3) the difference is stable: and carrying out first-order difference on the data to be predicted so as to stabilize the data.
Data normalization: the data to be predicted is normalized.
Determining parameters: autocorrelation function (ACF), partial autocorrelation function (PACF).
And (3) constructing a model: the model is tentative according to ACF and PACF values.
And (3) model checking: and evaluating the model effect according to MAPE.
Model preservation: the trained model is output as a PKL file.
The model prediction comprises the following specific steps:
loading data: and loading the data file to be predicted.
Loading a model: and loading the SARIMAX model PKL file of the corresponding industry.
Data normalization: and normalizing the exogenous characteristic data to be predicted.
Model prediction: and carrying out model prediction according to the historical trend of model learning and exogenous characteristic data.
The predicted outcome is reversed: the prediction result is normalized and inverted (the training data is the data after normalization, and the data to be tested needs to be normalized).
Step 4: and (3) constructing a limited space according to the energy consumption type of the related parameters obtained by the classification model, taking the limited space as a regression model 3 theta normalization boundary, and optimizing a floating point value of the output power consumption. I.e., the range of values obtained by classifying the model, it is considered that the regression model is accurate if it is stabilized at its 3 theta normalization, i.e., the value is in the 99.73% range.
And according to the predicted result of the double-fluid model, performing eight-value coding processing on the floating point value. The last decimal part of the power consumption is standardized, and the power values are generally 1.125, 1.25, 1.375 and the like.
The result of 1-MAPE is used as the accuracy of the model in this example, the mean absolute percentage error (Mean Absolute Percentage Error, MAPE) is a relative error metric that uses absolute values to avoid positive and negative errors from canceling each other, and relative errors can be used to compare the accuracy of predictions for various time series models. Theoretically, the smaller the MAPE value, the better the prediction model fitting effect is, and the better the accuracy is. The following is disclosed:
where n is the number of samples,is the actual value +.>Is a model predictive value.
The invention solves the problems of multi-source heterogeneous data, data fragmentation and data time sequence in the long-term prediction of the power consumption, and particularly the comparison of traditional models mainly comprising other single regression flow models, and the model can be more fully fitted with a time sequence model.
The invention uses a method for calculating the power consumption based on a double-fluid model, firstly uses a classification flow model to construct a limited space, then uses a regression flow model to accurately predict the power consumption, and finally combines the space of the classification flow prediction to check the result of the regression flow.
For the regression flow part, the characteristic data of the traditional prediction for using the power is used in the invention, and the requirements of time sequence type characteristics are ensured. The model is characterized quarterly, if the current 4 months of power usage is predicted, the model will fit 1 month, 2 months, 3 months of power usage.
Example two
It is an object of the present embodiment to provide a power usage long-term prediction system including:
the acquisition module is used for: acquiring historical multi-source data related to power use;
and a classification module: predicting based on a classification model according to the acquired historical multi-source data to obtain an electricity consumption type, and determining a limited space range of electricity prediction using amount according to the predicted electricity consumption type;
and (3) a regression module: according to the type of the acquired historical multi-source data, respectively predicting based on the corresponding SARIMAX model to obtain a plurality of floating point values of the power prediction using amount;
and an optimization module: comparing the limited space range of the power prediction using amount with a plurality of floating point values of the power prediction using amount, and if the floating point value of the power prediction using amount is not in the limited space range of the power prediction using amount, optimizing and adjusting the corresponding SARIMAX model to obtain an adjusted SARIMAX model;
and a prediction module: and predicting the acquired power to be predicted by using related multi-source data based on the adjusted SARIMAX model to obtain a prediction result.
Example III
It is an object of the present embodiment to provide a computing device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, which processor implements the steps of the method described above when executing the program.
Example IV
An object of the present embodiment is to provide a computer-readable storage medium.
A computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of the above method.
The steps involved in the devices of the second, third and fourth embodiments correspond to those of the first embodiment of the method, and the detailed description of the embodiments can be found in the related description section of the first embodiment. The term "computer-readable storage medium" should be taken to include a single medium or multiple media including one or more sets of instructions; it should also be understood to include any medium capable of storing, encoding or carrying a set of instructions for execution by a processor and that cause the processor to perform any one of the methods of the present invention.
It will be appreciated by those skilled in the art that the modules or steps of the invention described above may be implemented by general-purpose computer means, alternatively they may be implemented by program code executable by computing means, whereby they may be stored in storage means for execution by computing means, or they may be made into individual integrated circuit modules separately, or a plurality of modules or steps in them may be made into a single integrated circuit module. The present invention is not limited to any specific combination of hardware and software.
While the foregoing description of the embodiments of the present invention has been presented in conjunction with the drawings, it should be understood that it is not intended to limit the scope of the invention, but rather, it is intended to cover all modifications or variations within the scope of the invention as defined by the claims of the present invention.
Claims (10)
1. A method for long-term prediction of power usage, comprising:
acquiring historical multi-source data related to power use;
predicting based on a classification model according to the acquired historical multi-source data to obtain an electricity consumption type, and determining a limited space range of electricity prediction using amount according to the predicted electricity consumption type;
according to the type of the acquired historical multi-source data, respectively predicting based on the corresponding SARIMAX model to obtain a plurality of floating point values of the power prediction using amount;
comparing the limited space range of the power prediction using amount with a plurality of floating point values of the power prediction using amount, and if the floating point value of the power prediction using amount is not in the limited space range of the power prediction using amount, optimizing and adjusting the corresponding SARIMAX model to obtain an adjusted SARIMAX model;
and predicting the acquired power to be predicted by using related multi-source data based on the adjusted SARIMAX model to obtain a prediction result.
2. A method of long-term prediction of power usage as claimed in claim 1, wherein the types of multi-source data include: GDP, industry structure, population, user type, last used electricity, area, and weather.
3. The method for long-term prediction of power usage as defined in claim 1, wherein the prediction is performed based on the corresponding sarmax model according to the type of the obtained historical multi-source data, so as to obtain a plurality of floating point values of the predicted usage of power, specifically:
drawing a power usage time sequence diagram, and performing differential processing on unstable data in the power usage time sequence diagram;
determining a type affecting the power usage as an exogenous variable of the SARIMAX model;
determining parameters of the SARIMAX model according to the autocorrelation diagrams and the partial autocorrelation diagrams;
training SARIMAX model parameters by utilizing seasonal and non-seasonal orders and exogenous variables, and performing fitting treatment;
and predicting the future power usage based on the SARIMAX model after the fitting process.
4. A method for long-term prediction of power usage as defined in claim 1, further comprising: and screening the multi-source data types related to the power consumption according to the Pearson correlation coefficient and the random forest integration model, and carrying out standardization processing on the screened multi-source data related to the power consumption.
5. The method of claim 1, wherein a limited spatial range of power usage prediction determined from a classification model is used as a 3 theta normalization boundary for the sarmax model, and floating point values obtained from the sarmax model are optimized.
6. A method for long-term prediction of power usage as defined in claim 1, wherein the prediction is based on a classification model based on the obtained historical multi-source data to obtain corresponding energy consumption types, each energy consumption type corresponding to a range of values of power usage.
7. A method of long-term prediction of power usage as claimed in claim 1, wherein the accuracy of the sarmax model is assessed using average absolute percentage error.
8. A long-term prediction system for power usage, comprising:
the acquisition module is used for: acquiring historical multi-source data related to power use;
and a classification module: predicting based on a classification model according to the acquired historical multi-source data to obtain an electricity consumption type, and determining a limited space range of electricity prediction using amount according to the predicted electricity consumption type;
and (3) a regression module: according to the type of the acquired historical multi-source data, respectively predicting based on the corresponding SARIMAX model to obtain a plurality of floating point values of the power prediction using amount;
and an optimization module: comparing the limited space range of the power prediction using amount with a plurality of floating point values of the power prediction using amount, and if the floating point value of the power prediction using amount is not in the limited space range of the power prediction using amount, optimizing and adjusting the corresponding SARIMAX model to obtain an adjusted SARIMAX model;
and a prediction module: and predicting the acquired power to be predicted by using related multi-source data based on the adjusted SARIMAX model to obtain a prediction result.
9. A computer device, comprising: a processor, a memory and a bus, the memory storing machine-readable instructions executable by the processor, the processor and the memory in communication via the bus when the computer device is running, the machine-readable instructions when executed by the processor performing a power use long-term prediction method according to any one of claims 1 to 7.
10. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, performs a power usage long-term prediction method according to any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410245372.4A CN117829380B (en) | 2024-03-05 | 2024-03-05 | Method, system, equipment and medium for long-term prediction of power use |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410245372.4A CN117829380B (en) | 2024-03-05 | 2024-03-05 | Method, system, equipment and medium for long-term prediction of power use |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117829380A true CN117829380A (en) | 2024-04-05 |
| CN117829380B CN117829380B (en) | 2024-05-28 |
Family
ID=90524254
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410245372.4A Active CN117829380B (en) | 2024-03-05 | 2024-03-05 | Method, system, equipment and medium for long-term prediction of power use |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117829380B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110264001A (en) * | 2019-06-20 | 2019-09-20 | 国网上海市电力公司 | Electro-load forecast method based on multiple timings |
| CN110414719A (en) * | 2019-07-05 | 2019-11-05 | 电子科技大学 | A kind of vehicle flowrate prediction technique based on Multi-variable Grey Model time series |
| CN112200377A (en) * | 2020-10-16 | 2021-01-08 | 国能日新科技股份有限公司 | Photovoltaic medium-long term power generation capacity forecasting method and device based on SARIMAX model |
| US20210224700A1 (en) * | 2019-03-15 | 2021-07-22 | State Street Corporation | Techniques to forecast financial data using deep learning |
| CN114943383A (en) * | 2022-06-07 | 2022-08-26 | 平安科技（深圳）有限公司 | Prediction method and device based on time series, computer equipment and storage medium |
| CN115470682A (en) * | 2022-10-20 | 2022-12-13 | 江苏大学 | Pump station unit hydraulic component wear and service life prediction method based on digital twinning |
| CN115764870A (en) * | 2022-11-14 | 2023-03-07 | 国网冀北电力有限公司计量中心 | Multivariable photovoltaic power generation power prediction method and device based on automatic machine learning |
| US20230229911A1 (en) * | 2022-01-14 | 2023-07-20 | Tata Consultancy Services Limited | System and method for online time-series forecasting using spiking reservoir |
| CN117634693A (en) * | 2023-12-06 | 2024-03-01 | 国网天津市电力公司电力科学研究院 | Carbon emission prediction method based on energy data |
-
2024
- 2024-03-05 CN CN202410245372.4A patent/CN117829380B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210224700A1 (en) * | 2019-03-15 | 2021-07-22 | State Street Corporation | Techniques to forecast financial data using deep learning |
| CN110264001A (en) * | 2019-06-20 | 2019-09-20 | 国网上海市电力公司 | Electro-load forecast method based on multiple timings |
| CN110414719A (en) * | 2019-07-05 | 2019-11-05 | 电子科技大学 | A kind of vehicle flowrate prediction technique based on Multi-variable Grey Model time series |
| CN112200377A (en) * | 2020-10-16 | 2021-01-08 | 国能日新科技股份有限公司 | Photovoltaic medium-long term power generation capacity forecasting method and device based on SARIMAX model |
| US20230229911A1 (en) * | 2022-01-14 | 2023-07-20 | Tata Consultancy Services Limited | System and method for online time-series forecasting using spiking reservoir |
| CN114943383A (en) * | 2022-06-07 | 2022-08-26 | 平安科技（深圳）有限公司 | Prediction method and device based on time series, computer equipment and storage medium |
| CN115470682A (en) * | 2022-10-20 | 2022-12-13 | 江苏大学 | Pump station unit hydraulic component wear and service life prediction method based on digital twinning |
| CN115764870A (en) * | 2022-11-14 | 2023-03-07 | 国网冀北电力有限公司计量中心 | Multivariable photovoltaic power generation power prediction method and device based on automatic machine learning |
| CN117634693A (en) * | 2023-12-06 | 2024-03-01 | 国网天津市电力公司电力科学研究院 | Carbon emission prediction method based on energy data |
Non-Patent Citations (1)
| Title |
|---|
| 徐闻;王斌;: "基于高阶累积量的核Logistic回归调制分类算法", 太赫兹科学与电子信息学报, no. 02, 25 April 2013 (2013-04-25) * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117829380B (en) | 2024-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chai et al. | Coal consumption in China: how to bend down the curve? | |
| Zheng et al. | Addressing the conditional and correlated wind power forecast errors in unit commitment by distributionally robust optimization | |
| CN109193748B (en) | Evaluation method and computing device for photovoltaic absorption capacity | |
| CN106529704A (en) | Monthly maximum power load forecasting method and apparatus | |
| CN109840633B (en) | Photovoltaic output power prediction method, system and storage medium | |
| CN111680841B (en) | Short-term load prediction method, system and terminal equipment based on principal component analysis | |
| CN115169703A (en) | Short-term power load prediction method based on long-term and short-term memory network combination | |
| CN114372360A (en) | Method, terminal and storage medium for power load prediction | |
| CN111625947A (en) | Distributed energy development scale prediction method, equipment and medium | |
| CN115860797A (en) | Electric quantity demand prediction method suitable for new electricity price reform situation | |
| CN114638425B (en) | Historical data-based large-user monthly electricity consumption prediction method and system | |
| CN113177366A (en) | Comprehensive energy system planning method and device and terminal equipment | |
| CN112348287A (en) | Electric power system short-term load probability density prediction method based on LSTM quantile regression | |
| CN116128315A (en) | Long-short term energy storage planning method, system, medium and equipment | |
| CN111626543A (en) | Method and device for processing power related data | |
| Wang et al. | Multi‐stage stochastic wind‐thermal generation expansion planning with probabilistic reliability criteria | |
| CN113869556A (en) | Power consumption prediction method, device and equipment | |
| CN104077231B (en) | Transformer maintenance optimization method based on symbol dynamics and LS-SVM | |
| CN117277312A (en) | Gray correlation analysis-based power load influence factor method and equipment | |
| CN117829380B (en) | Method, system, equipment and medium for long-term prediction of power use | |
| CN109034458A (en) | It is a kind of to guarantee to permit taking in acquired electric grid investment planing method | |
| CN115130924A (en) | Microgrid power equipment asset evaluation method and system under source grid storage background | |
| Song et al. | Power supply, coal price, and economic growth in China | |
| CN117154723B (en) | Platform short-term load prediction method and system based on multi-source data and model fusion | |
| Yang et al. | Research on Financing Efficiency of Photovoltaic Industry Based on Three-stage DEA Model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |