CN113822714A - A method and system for industry electricity consumption forecasting considering price change factors - Google Patents
A method and system for industry electricity consumption forecasting considering price change factorsDownload PDFInfo
- Publication number
- CN113822714A CN113822714ACN202111113543.0ACN202111113543ACN113822714ACN 113822714 ACN113822714 ACN 113822714ACN 202111113543 ACN202111113543 ACN 202111113543ACN 113822714 ACN113822714 ACN 113822714A
- Authority
- CN
- China
- Prior art keywords
- sequence
- industry
- power consumption
- data
- price
- 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
Images
Classifications
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0206—Price or cost determination based on market factors
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- General Physics & Mathematics (AREA)
- Finance (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Primary Health Care (AREA)
- Tourism & Hospitality (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Translated fromChinese
Description
Technical Field
The invention relates to the field of power consumption prediction analysis of industries, in particular to a power consumption prediction method and a power consumption prediction system of industries considering price change factors.
Background
The demand prediction of the power consumption is an important basis for formulating power and electricity balance and energy layout planning, and a scientific prediction method and an accurate prediction result have very important significance for the economic operation of a power system, especially for the traditional high-energy-consumption industry. Meanwhile, the power consumption of the industry is influenced by economic factors, the economic factors which have important influence on the power consumption of the industry are analyzed, and a reasonable power consumption demand prediction model is constructed to provide decision basis for construction planning and operation of a power grid.
The summation autoregressive Moving Average model (ARIMA), abbreviated as ARIMA model in english, is a linear Regression model, is a prediction method for analyzing the correlation law in a time sequence, and can describe and predict a stationary random sequence and a non-stationary random sequence. The ARIMA model has wide application in time sequence analysis in the fields of economy, society and the like.
When the ARIMA model is used for carrying out prediction analysis on the power consumption, certain limitation exists. Actually, the sequence change rule is influenced by other sequences in many cases, for example, in the power consumption prediction problem, the industrial power consumption is greatly influenced by economic factors, the ARIMA model only considers the influence of the sequence change and cannot bring external influence factors such as economy into modeling, so that the ARIMA model is directly used for modeling and predicting the time sequence which is greatly influenced by the external factors, and the ARIMA model has certain defects in interpretability and prediction accuracy.
Disclosure of Invention
In order to solve the problems, the system and the method for predicting the industrial power consumption considering the price change factors are characterized in that industrial monthly power consumption data and industrial product monthly price data are obtained, data are preprocessed to obtain a power consumption sequence and a price change rate sequence, stability inspection and causal relationship inspection are carried out, ARIMAX models with different parameter combinations are constructed by using the power consumption sequence and the price change rate sequence, a prediction model is calculated and selected for predicting the future industrial monthly power consumption, the price change factors are introduced into an industrial power consumption prediction model, the accuracy of the prediction model for the industrial power consumption prediction is improved, and auxiliary decision can be provided for electric enterprise personnel to judge the industrial power consumption demand.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
an industry power consumption prediction method considering price change factors comprises the following steps:
step S1: acquiring monthly power consumption data of the industry;
step S2: acquiring monthly price data of industrial products;
step S3: preprocessing the data to obtain a power consumption sequence and a price change rate sequence;
step S4: respectively carrying out stability inspection on the power consumption sequence and the price change rate sequence, and if the power consumption sequence and the price change rate sequence do not pass the inspection, carrying out stabilization operation until the power consumption sequence and the price change rate sequence pass the inspection;
step S5: carrying out Granger causal test on the power consumption sequence and the price change rate sequence, and obtaining the month lag number of the price change rate sequence relative to the power consumption sequence;
step S6: constructing ARIMAX models with different parameter combinations by using the electricity consumption sequence and the price change rate sequence, and calculating and selecting a prediction model;
step S7: and predicting the monthly power consumption of the future industry by using the prediction model.
Furthermore, the mode of acquiring monthly price data of industrial products comprises a directional web crawler mode.
Further, the acquiring monthly price data of the industrial products comprises the following steps:
step S21: acquiring a main webpage website for publishing monthly price data of regional industry products;
step S22: constructing a regular expression matched with the characteristics of the sub-web character string containing monthly price data of the products in the regional industry;
step S23: determining monthly price data range and format of products in regional industry;
step S24: automatically acquiring and recording the link address of the sub-web page by adopting a computer program;
step S25: and automatically acquiring and storing the product price data of a certain area industry in the sub-web page by adopting a computer program.
Further, the preprocessing the data to obtain the power consumption sequence comprises the following steps:
step S31, selecting a power consumption reference value;
step S32, calculating the electricity consumption of each month based on the electricity reference value;
step S33: and attaching a month label to the electricity consumption of each month to obtain an electricity consumption sequence taking the month as the label.
Further, the preprocessing the data to obtain the price change rate sequence includes the following steps:
step S34: calculating monthly prices of industrial products;
step S35: calculating monthly price sequences of the industrial products;
step S36: and calculating a price change rate sequence.
Further, the stationarity test adopts a unit root test method.
Further, the unit root inspection method includes an ADF method.
Further, the smoothing operation includes performing a number of differential operations on the data sequence.
Further, the ARIMAX model of different parameter combinations is constructed by the electricity consumption sequence and the price change rate sequence, and the selection prediction model is calculated, and the method comprises the following steps:
step S61: constructing an ARIMAX (p, d, q) model by taking the power consumption sequence as a response variable and taking the price change rate sequence lagging by a plurality of month parts as an external variable, wherein p and q are limited to be within the ranges of 0,1 and 2, and d is the difference frequency of the stabilizing operation of the power consumption sequence;
step S61: and (3) determining a prediction model after data simulation and white noise test of the ARIMAX (p, d, q) model with the minimum AIC value.
An industry power consumption prediction system considering price change factors is used for implementing the industry power consumption prediction method considering price change factors, and comprises a data collection module, a data preprocessing module, a sequence stabilization module, a Granger causal test module, a model construction module, a model selection module and a model prediction module; the data preprocessing module is respectively connected with the data collecting module and the sequence stabilizing module so as to preprocess the monthly price data of the industrial products and the monthly electricity consumption data of the industry collected by the data collecting module to obtain an electricity consumption sequence and a price change rate sequence, and transmit the electricity consumption sequence and the price change rate sequence to the sequence stabilizing module for stability inspection; the Granger causal test module is respectively connected with the sequence stabilization module and the model construction module, and transmits the power consumption sequence and the price change rate sequence subjected to stability test to the model construction module after causal relationship test to construct an ARIMAX (p, d, q) model; the model selection module is respectively connected with the model construction module and the model prediction module, and after the constructed ARIMAX (p, d, q) model is calculated and checked, the optimal ARIMAX model is selected as a prediction model to be transmitted to the model prediction module for predicting the power consumption of the industry.
According to the method, the system firstly adopts technologies such as directional web crawlers and the like to obtain monthly power consumption data of the industry and monthly price data of industry products, preprocesses the data to obtain a power consumption sequence and a price change rate sequence, then respectively carries out stability inspection on the power consumption sequence and the price change rate sequence, uses Granger causal inspection to inspect the causal relationship between the power consumption sequence and the price change rate sequence and obtain a delayed monthly number, and then uses the power consumption sequence and the price change rate sequence to construct ARIMAX models with different parameter combinations and calculates and selects a prediction model for predicting the monthly power consumption of the future industry. The method can collect the historical price data of main products of regional industries and the historical power consumption data of the regional industries, introduces the price change factor into the power consumption prediction model of the industries, improves the accuracy of the prediction model on the power consumption prediction of the industries, and can provide an auxiliary decision for the electric enterprise personnel to judge the power consumption requirements of the industries.
Drawings
FIG. 1 is a flow chart of an industry power consumption prediction method considering price variation factors;
FIG. 2 is a flow chart for obtaining monthly price data for an industrial product;
FIG. 3 is a graph of predicted monthly power usage versus actual power usage with and without consideration of price factors;
FIG. 4 is a schematic diagram of a system for forecasting industry power consumption in consideration of price variation.
Detailed Description
The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.
The embodiments of the present disclosure are described below with specific examples, and other advantages and effects of the present disclosure will be readily apparent to those skilled in the art from the disclosure in the specification. It is to be understood that the described embodiments are merely illustrative of some, and not restrictive, of the embodiments of the disclosure. The disclosure may be embodied or carried out in various other specific embodiments, and various modifications and changes may be made in the details within the description without departing from the spirit of the disclosure. It is to be noted that the features in the following embodiments and examples may be combined with each other without conflict. All other embodiments, which can be derived by a person skilled in the art from the embodiments disclosed herein without making any creative effort, shall fall within the protection scope of the present disclosure.
Example one
Fig. 1 is a flowchart of an industry power consumption prediction method considering price change factors in an embodiment of the present invention, where a certain area steel industry is selected in the embodiment, the method includes the following steps:
step S1: acquiring monthly power consumption data of the industry;
step S2: acquiring monthly price data of industrial products;
step S3: preprocessing the data to obtain a power consumption sequence and a price change rate sequence;
step S4: respectively carrying out stability inspection on the power consumption sequence and the price change rate sequence, and if the power consumption sequence and the price change rate sequence do not pass the inspection, carrying out stabilization operation until the power consumption sequence and the price change rate sequence pass the inspection;
step S5: testing the causal relationship between the power consumption sequence and the price change rate sequence by using a Granger causal test;
step S6: using the power consumption sequence and the price change rate sequence to construct ARIMAX models with different parameter combinations and calculate and select a prediction model;
step S7: and predicting monthly power consumption of the future industry by using the prediction model.
In a specific implementation, the step S1 of obtaining the monthly electricity consumption data of the industry is to obtain the electricity consumption data of the steel industry from a regional power grid company, and the time span of the electricity consumption data is 42 months from 1 month in 2018 to 6 months in 2021.
In a specific implementation, as shown in fig. 2, the acquiring monthly price data of the industry products in step S2 is monthly price data of main products in the steel industry in a corresponding area, and the acquiring mode includes using a directional web crawler technology, and includes the following steps:
step S21: acquiring a main webpage website for publishing monthly price data of regional industry products; in the embodiment of the invention, the main webpage for publishing monthly price data of regional industry products is an 'regional price' column webpage under an official website of a certain iron and steel industry association.
Step S22: constructing a regular expression matched with the characteristics of the sub-web character string containing monthly price data of the products in the regional industry; in the embodiment of the invention, a sub-webpage containing monthly price data of regional industry products is accessed as an 'regional price' column webpage under an official website of a certain iron and steel industry association, the names of character strings linked by the sub-webpage are determined to be 'main market price in XXXXXX year X month at week X' or 'regional price in XXXX year X month at week X', and a regular expression capable of matching the character strings beginning with 'XXXXX year X month' is constructed.
Step S23: determining monthly price data range and format of products in regional industry; in the embodiment of the invention, the data range determined and obtained by visiting the sub-web page is 6 main steels which are respectively: a high-speed wire HPB300, a deformed steel HRB400, angle steel Q235, a medium plate Q235, a hot plate coil Q235 and a cold thin plate SPCC; and the regional steel price data is determined to be displayed in the form of pictures in part of the sub-webpages and displayed in the form of tables in the rest of the sub-webpages, so that the data in two different formats can be respectively processed in the subsequent steps. The time span of the acquired monthly electricity consumption data of the steel industry in a certain area is from 1 month in 2018 to 6 months in 2021, and in order to ensure a certain elastic space, the time span of the acquired main steel price data in the certain area is from 1 month in 2017 to 6 months in 2021.
Step S24: automatically acquiring and recording the link address of the sub-web page by using a computer program according to the constructed regular expression; in the embodiment of the present invention, the following functions are implemented by Python program codes: controlling an open source webpage automatic testing tool to access an 'area price' page of an official website of a certain iron and steel industry association, acquiring all HTML elements in a current page, traversing character strings contained in all HTML elements, if the character strings can be matched with the regular expression constructed in the step S22, representing that the HTML elements are linked to a sub-web page containing area steel price data, and recording the link address of the sub-web page in a text document; and if all HTML elements of the current page are traversed and the data of the required time period is not completely obtained, controlling an open-source webpage automatic testing tool to click the next page in the page, then re-acquiring the HTML elements of the page, and repeating the traversing, matching and recording operations until link addresses of all sub-pages of steel data in a certain area from 2017, 1 month to 2021, 6 months are published and recorded.
Step S25: automatically acquiring and storing the product price data of a certain area industry in the sub-web page by using a computer program; in the embodiment of the invention, the following functions are realized through Python program codes according to the recorded subnet page link address and the product price data range and format of the regional industry: traversing the link addresses recorded in the step S24, controlling the open-source web page automation test tool to sequentially access the sub-web pages pointed by the link addresses, acquiring all HTML elements of the sub-web pages, delivering the HTML elements to the HTML data extraction tool, and downloading and storing the pictures in the local computer when the HTML elements of the sub-web pages contain the pictures; and when the table exists in the HTML element of the sub-webpage, all character strings in the table are stored in the Excel table of the local computer.
In a specific implementation, the preprocessing the data to obtain the power consumption sequence in step S3 includes the following steps:
step S31: selecting a certain area and a certain industry at a certain month as a reference month, wherein the actual power consumption of the industry in the month is a power consumption reference value, the reference month generally adopts the first month of a certain year, and the power consumption of the steel industry in the certain area at 1 month of 2011 can be selected as the reference value in the embodiment;
step S32: acquiring the actual power consumption of each month in the industry within a certain period of time after the reference month, and setting the power consumption of each month based on the power consumption reference value as the ratio of the actual power consumption of each month to the power consumption reference value; in the embodiment, a total of 42 historical months from 1 month in 2018 to 6 months in 2021 are selected to obtain the electricity consumption of the 42 months;
step S33: the method includes the steps that a month label is attached to electricity consumption of each month, a monthly electricity consumption sequence with the month as the label and contents as the industry is obtained, and the time sequence of the electricity consumption of the steel industry in a certain area is obtained in the embodiment.
In a specific implementation, the preprocessing the data to obtain the price change rate sequence in step S3 includes the following steps:
step S34: and (3) calculating monthly prices of industrial products, wherein the calculation formula is as follows:
wherein, ctThe monthly price of the industrial product is shown, and in the embodiment, the monthly price of steel in a certain area is shown; w represents the week of the industry product price in the month, and in the embodiment, represents the week of the main steel price published by certain iron and steel industry association in the month; n represents the total number of the types of industrial products, in this embodiment, the total number of the types of main steel materials published by a certain iron and steel industry association; e.g. of the typeijThe price of the jth industry product in the ith week is represented, and in the embodiment, the average value of the prices of each main steel in each week of steel monitored by a monitoring point in a certain area in the month is represented.
Step S35: calculating monthly price sequences of the industrial products; the calculation of the step S31 is repeated for the main steel price data of the monitoring network point in a certain region from 1 month in 2017 to 6 months in 2021, and a steel monthly price sequence in a certain region with the time span from 1 month in 2017 to 6 months in 2021 is obtained.
Step S36: calculating a price change rate sequence, wherein the monthly price change rate is calculated according to the following formula:
where t represents month, xtRepresenting monthly price change rate, ctRepresenting the price of the product in the month of the year, ct-1And the product price of the industry in the previous month is represented.
Repeating the calculation on the industry product prices of each month in a period of time to obtain a price change rate sequence. In this example, the above calculation is repeated for the steel monthly price sequence in a certain region from 1 month in 2017 to 6 months in 2021, and a steel price change rate sequence in a certain region with a time span from 2 months in 2017 to 6 months in 2021 is obtained.
In a specific implementation, in step S4, the smoothness check is performed on the power consumption sequence and the price change rate sequence, and if the smoothness check does not pass, the smoothing operation is performed until the smoothness check passes. The stationarity test comprises a unit root test method, and the theoretical basis of the unit root test method is as follows: if the data sequence is stationary, then all feature roots of the data sequence should be within the unit circle; if the data sequence has a characteristic root on or off the unit circle, the data sequence is not stationary. The unit root inspection method includes an ADF method, which is a common unit root inspection method that can perform smoothness inspection on a data sequence of an arbitrary time period. The smoothing operation includes 1 difference operation on the data sequence.
In this embodiment, the ADF method is used to perform stationarity check on the price rate sequence, the original hypothesis is checked to be non-stationary in sequence, the alternative hypothesis is sequence stationary, and if the significance level is 0.05, the original hypothesis may be rejected and the alternative hypothesis may be selected if the P value is less than 0.05. Firstly, using an ADF method to carry out stability inspection on the price change rate sequence, wherein the inspection result shows that the price change rate sequence is stable; and then continuing to perform stability test on the power consumption sequence, wherein the result shows that the power consumption sequence does not pass the stability test under the condition that the significance level is 0.05, the power consumption sequence needs to be subjected to the stabilization operation, namely 1 difference operation is performed on the power consumption sequence, the obtained power consumption sequence passes the stability test, and the difference number d is 1. The results of the stationarity test of each data sequence are shown in table 1:
TABLE 1
In a specific implementation, the Granger causal test described in step S5 has the original hypothesis that there is no causal relationship between the two sequences, and the alternative hypothesis has a causal relationship between the two sequences, and if the significance level is 0.1, the original hypothesis may be rejected and the alternative hypothesis may be selected if the P value is less than 0.1. In the embodiment of the present invention, Granger causal test is performed using a power consumption sequence with a difference number of times of 1 and a price change rate sequence, the Granger causal test results of the power consumption sequence and the price change rate sequence are shown in table 2, it can be considered that a price change rate sequence lagging by 2 months or 3 months is the Granger cause of the power consumption sequence with a difference number of times of 1, and a minimum P value is taken, that is, an optimal lagging number of months m is 2.
TABLE 2
In a specific implementation, the step S6 of constructing an ARIMAX model with different parameter combinations by using the power consumption sequence and the price change rate sequence and calculating and selecting a prediction model refers to constructing an ARIMAX (p, d, q) model by using the power consumption sequence as a response variable and the price change rate sequence lagging by several months as an external variable, where p and q are limited to ranges of 0,1, and 2, and d is a difference number of times of smoothing operation on the power consumption sequence; and selecting a model with the minimum AIC value, and determining a prediction model after data simulation and white noise test.
In the embodiment of the invention, an ARIMAX (p, d, q) model is constructed by using a power consumption sequence as a response variable and a price change rate sequence lagging by 2 months as an external variable, the response variable and the external variable are in one-to-one correspondence, namely, a monthly power consumption sequence of a certain area of steel industry from 2018, 1 month to 2021, 6 months is used as the response variable, and a monthly price change rate sequence of a certain area from 2017, 11 months to 2021, 4 months is used as the external variable. Limiting p and q to be within the ranges of 0,1 and 2, and as known from step S4, if d is 1, constructing all ARIMAX (p, d, q) models, that is, ARIMAX (0,1,2), ARIMAX (1,1,2), ARIMAX (2,1,0), ARIMAX (2,1,2), ARIMAX (1,1,1), ARIMAX (2,1,1), ARIMAX (1,1,0), ARIMAX (0,1,1) and ARIMAX (0,1,0), wherein the ARIMAX (1,1,0) are 9 models in total, and the constructed ARIMAX models are evaluated by AIC values for 9 different parameter combinations. The AIC value is calculated as follows:
AIC＝2k-2l n(L)
wherein k is the number of unknown parameters of the model, and L is the maximum likelihood function value of the model.
The larger the maximum likelihood function of the model is, the better the model fitting effect is; the more the unknown parameters of the model are, the more the estimation difficulty of the unknown parameters is; therefore, a good fitting model should be a comprehensive optimal configuration of fitting accuracy and the number of unknown parameters, and the AIC value is provided based on the idea, so that the model which minimizes the AIC value is the optimal model. The AIC values of the constructed ARIMAX (p, d, q) model of 9 parameters are shown in table 3:
TABLE 3
| Model (model) | AIC value |
| ARIMAX(0,1,2) | -61.845934 |
| ARIMAX(1,1,2) | -60.850102 |
| ARIMAX(2,1,0) | -60.682772 |
| ARIMAX(2,1,2) | -59.622291 |
| ARIMAX(1,1,1) | -59.457178 |
| ARIMAX(2,1,1) | -58.735592 |
| ARIMAX(1,1,0) | -56.631244 |
| ARIMAX(0,1,1) | -53.156610 |
| ARIMAX(0,1,0) | -29.906928 |
ARIMAX (0,1,2) with the smallest AIC value was preliminarily selected as the prediction model from table 3. Generating a monthly electricity consumption fitting value of the steel industry in a certain region from 3 months in 2021 to 6 months in 2021 by using a model ARIMAX (0,1,2) considering price change factors; and (3) constructing a power consumption prediction model ARIMA (0,1,2) of the steel industry in a certain area without considering price change factors by using the same parameters and data, and generating a monthly power consumption fitting value of the steel industry in the certain area from 3 months in 2021 to 6 months in 2021. The percentage error between the fitted value and the actual value of the two models is shown in table 4, and the results in table 4 show that the ARIMAX (0,1,2) model considering the price change factor is reduced in the percentage of the prediction error compared with the ARIMAX (0,1,2) model not considering the price change factor, that is, the prediction accuracy is improved by considering the price change factor.
TABLE 4
And performing white noise test on a residual sequence of the ARIMAX (0,1,2) model, wherein the original assumption of the white noise test is that the sequence is a white noise sequence, the alternative assumption is that the sequence is not a white noise sequence, and under the condition that the significance level is 0.05, if the P value is less than 0.05, the original assumption can be rejected, and the white noise sequence is also called a pure random sequence, which indicates that the sequence has no available information extraction. The white noise test result of the residual sequence of the ARIMAX (0,1,2) model shows that the residual sequence of the ARIMAX (0,1,2) model cannot reject the original hypothesis under the condition of 0.05 of significance level, namely the residual sequence of the ARIMAX (0,1,2) model is a white noise sequence, which proves that the residual of the model has no available information, namely the model completely extracts effective information in the used construction sequence, and the test result is shown in table 5:
TABLE 5
| Number of delay steps | P value |
| 1 | 0.776223 |
| 2 | 0.868665 |
| 3 | 0.962285 |
| 4 | 0.990391 |
| 5 | 0.794768 |
| 6 | 0.805762 |
From the above results, it was determined that using ARIMAX (0,1,2) model as the prediction model, the effect of applying this model to monthly electricity usage prediction (fitting) of the steel industry in a certain area from 1 month in 2018 to 6 months in 2021 is shown in fig. 3. In fig. 3, a fitted value sequence of monthly electricity consumption of the steel industry in a certain area from 1 month to 2021 in 2018 to 6 months is generated by using an ARIMAX (0,1,2) model considering the price change factor and an ARIMAX (0,1,2) model not considering the price change factor, and compared with an actual value sequence of monthly electricity consumption of the steel industry in a certain area from 1 month to 2021 in 2018 to 6 months, and the three sequence comparison broken lines are presented in the figure, wherein the broken line represents the fitted value sequence of the ARIMAX (0,1,2) model considering the price change factor, the dotted broken line represents the fitted value sequence of the ARIMAX (0,1,2) model not considering the price change factor, and the solid line represents the actual value sequence. As can be seen from the prediction (fitting) result of fig. 3, in the time period from 7 months in 2019 to 12 months in 2019, from 1 month in 2021 to 6 months in 2021, and the like, the fitting value obtained by considering the price change factor model is closer to the actual value than the fitting value obtained by not considering the price change factor model, which indicates that by using the method described in this patent, the price change factor is considered in the model building process, and the accuracy of the industry power consumption prediction can be effectively improved.
In the embodiment of the present invention, step S7 illustrates a specific usage method of the prediction model by the following example. Suppose that a user needs to predict the monthly electricity consumption of the steel industry in a certain region from 6 months to 7 months in 2021 for 2 months in total, namely, a monthly electricity consumption sequence of the steel industry in a certain region from 7 months to 8 months in 2021 is obtained through a prediction model
Then it will need steel month in a certain region from 5 months in 2021 to 6 months in 2021The price change rate sequence is recorded as a sequence
As an external variable input of the model, the price change rate sequence may be obtained by the web crawler targeting technique in step S2 and by the calculation method in step S3.
Example two
Fig. 4 shows an industry power consumption prediction system considering price variation, which is used to implement the industry power consumption prediction method considering price variation in the above embodiments. The system comprises a data collection module, a data preprocessing module, a sequence stabilization module, a Granger causal test module, a model construction module, a model selection module and a model prediction module.
The data preprocessing module is respectively connected with the data collecting module and the sequence stabilizing module so as to preprocess the monthly price data of the products of a certain area industry and the monthly electricity consumption data of the industry collected by the data collecting module, generate a month time index, obtain an electricity consumption sequence and construct a monthly price sequence of the products of the area industry, calculate to obtain a price change rate sequence, and transmit the price change rate sequence to the sequence stabilizing module for stability inspection; and if the stationarity test does not pass, the sequence stationarity module performs differential operation on the corresponding sequence until the sequence is stable, wherein the differential times are recorded as d.
The Granger causal test module is respectively connected with the sequence stabilization module and the model construction module, carries out causal relationship test on a power consumption sequence and a price change rate sequence subjected to stability test, and then transmits the power consumption sequence and the price change rate sequence to the model construction module to construct an ARIMAX (p, d, q) model, wherein under the condition of single causal relationship, the optimal delayed monthly number of the price change rate sequence to the power consumption sequence is determined and recorded as m, and if no causal relationship exists, m is set as 0.
The model selection module is respectively connected with the model construction module and the model prediction module, the model construction module uses a power consumption sequence as a response variable, generates a price change rate sequence lagging for m months as an external variable, uses the difference times d, limits the parameters p and q to 0,1 and 2, and constructs ARIMAX (p, d, q) models with 9 different parameter combinations in total. And the model selection module compares the AIC values of the ARIMAX (p, d, q) models with the 9 different parameter combinations on the basis of an AIC criterion, selects the ARIMAX (p, d, q) model with the minimum AIC value, and transmits the ARIMAX (p, d, q) model as a prediction model after calculation and white noise test to the model prediction module for predicting the power consumption of the industry.
The system comprises a data collection module, a data preprocessing module, a sequence smoothing module, a Granger causal test module, a model construction module, a model selection module and a model prediction module, and the system firstly adopts the technologies of directional web crawlers and the like to obtain monthly power consumption data of industry products and monthly price data of the industry products according to the method, preprocesses the data to obtain a power consumption sequence and a price change rate sequence, respectively carries out stability test on the power consumption sequence and the price change rate sequence, uses the Granger causal test to test the causal relationship between the power consumption sequence and the price change rate sequence and obtain a number of delayed months, then uses the power consumption sequence and the price change rate sequence to construct an ARIMAX model with different parameter combinations and calculates the number of delayed months, And selecting a prediction model for predicting the monthly power consumption of the future industry. The method can collect the historical price data of main products of regional industries and the historical power consumption data of the regional industries, introduces the price change factor into the power consumption prediction model of the industries, improves the accuracy of the prediction model on the power consumption prediction of the industries, and can provide an auxiliary decision for the electric enterprise personnel to judge the power consumption requirements of the industries.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
In the present invention, unless otherwise expressly stated or limited, the terms "mounted," "connected," "secured," and the like are to be construed broadly and can, for example, be fixedly connected, detachably connected, or integrally formed; may be mechanically coupled, may be electrically coupled or may be in communication with each other; they may be directly connected or indirectly connected through intervening media, or they may be connected internally or in any other suitable relationship, unless expressly stated otherwise. The specific meanings of the above terms in the present invention can be understood by those skilled in the art according to specific situations.
The above description is for the purpose of illustrating embodiments of the invention and is not intended to limit the invention, and it will be apparent to those skilled in the art that any modification, equivalent replacement, or improvement made without departing from the spirit and principle of the invention shall fall within the protection scope of the invention.
Claims (10)
1. An industry power consumption prediction method considering price change factors is characterized by comprising the following steps:
step S1: acquiring monthly power consumption data of the industry;
step S2: acquiring monthly price data of industrial products;
step S3: preprocessing the data to obtain a power consumption sequence and a price change rate sequence;
step S4: respectively carrying out stability inspection on the power consumption sequence and the price change rate sequence, and if the power consumption sequence and the price change rate sequence do not pass the inspection, carrying out stabilization operation until the power consumption sequence and the price change rate sequence pass the inspection;
step S5: carrying out Granger causal test on the power consumption sequence and the price change rate sequence, and obtaining the month lag number of the price change rate sequence relative to the power consumption sequence;
step S6: constructing ARIMAX models with different parameter combinations by using the electricity consumption sequence and the price change rate sequence, and calculating and selecting a prediction model;
step S7: and predicting the monthly power consumption of the future industry by using the prediction model.
2. The method of predicting industry power usage in view of price changing factors as claimed in claim 1, wherein said manner of obtaining monthly price data for industry products includes a directed web crawler manner.
3. The method of predicting industry power usage in view of price variability of claim 2, wherein said obtaining industry product monthly price data comprises the steps of:
step S21: acquiring a main webpage website for publishing monthly price data of regional industry products;
step S22: constructing a regular expression matched with the characteristics of the sub-web character string containing monthly price data of the products in the regional industry;
step S23: determining monthly price data range and format of products in regional industry;
step S24: automatically acquiring and recording the link address of the sub-web page by adopting a computer program;
step S25: and automatically acquiring and storing the product price data of a certain area industry in the sub-web page by adopting a computer program.
4. The method for predicting power consumption of industry in consideration of price change factors as claimed in claim 1, wherein the step of preprocessing the data to obtain the power consumption sequence comprises the following steps:
step S31, selecting a power consumption reference value;
step S32, calculating the electricity consumption of each month based on the electricity reference value;
step S33: and attaching a month label to the electricity consumption of each month to obtain an electricity consumption sequence taking the month as the label.
5. The method for forecasting electricity consumption in industry considering price change factors according to claim 1, wherein the step of preprocessing the data to obtain a price change rate sequence comprises the following steps:
step S34: calculating monthly prices of industrial products;
step S35: calculating monthly price sequences of the industrial products;
step S36: and calculating a price change rate sequence.
6. The method of predicting business electricity usage in view of price change factors as in claim 1, wherein the stationarity test employs a unit root test method.
7. The method of predicting business power usage in view of price changing factors of claim 6, wherein said unit root inspection method comprises an ADF method.
8. The method of any one of claims 1 to 7, wherein the smoothing operation comprises performing a number of differential operations on the data sequence.
9. The method for predicting power consumption of industry with price change factors considered according to any one of claims 1 to 7, wherein the ARIMAX model of different parameter combinations is constructed according to the power consumption sequence and the price change rate sequence, and the calculation and selection prediction model comprises the following steps:
step S61: constructing an ARIMAX (p, d, q) model by taking the power consumption sequence as a response variable and taking the price change rate sequence lagging by a plurality of month parts as an external variable, wherein p and q are limited to be within the ranges of 0,1 and 2, and d is the difference frequency of the stabilizing operation of the power consumption sequence;
step S61: and (3) determining a prediction model after data simulation and white noise test of the ARIMAX (p, d, q) model with the minimum AIC value.
10. An industry power consumption prediction system considering price change factors is characterized in that the industry power consumption prediction system is used for implementing the industry power consumption prediction method considering price change factors in any one of claims 1 to 9, and comprises a data collection module, a data preprocessing module, a sequence smoothing module, a Granger causal test module, a model construction module, a model selection module and a model prediction module;
the data preprocessing module is respectively connected with the data collecting module and the sequence stabilizing module so as to preprocess the monthly price data of the industrial products and the monthly electricity consumption data of the industry collected by the data collecting module to obtain an electricity consumption sequence and a price change rate sequence, and transmit the electricity consumption sequence and the price change rate sequence to the sequence stabilizing module for stability inspection;
the Granger causal test module is respectively connected with the sequence stabilization module and the model construction module, and transmits the power consumption sequence and the price change rate sequence subjected to stability test to the model construction module after causal relationship test to construct an ARIMAX (p, d, q) model;
the model selection module is respectively connected with the model construction module and the model prediction module, and after the constructed ARIMAX (p, d, q) model is calculated and checked, the optimal ARIMAX model is selected as a prediction model to be transmitted to the model prediction module for predicting the power consumption of the industry.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111113543.0ACN113822714A (en) | 2021-09-23 | 2021-09-23 | A method and system for industry electricity consumption forecasting considering price change factors |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111113543.0ACN113822714A (en) | 2021-09-23 | 2021-09-23 | A method and system for industry electricity consumption forecasting considering price change factors |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113822714Atrue CN113822714A (en) | 2021-12-21 |
Family
ID=78915211
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111113543.0APendingCN113822714A (en) | 2021-09-23 | 2021-09-23 | A method and system for industry electricity consumption forecasting considering price change factors |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113822714A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114493705A (en)* | 2022-01-27 | 2022-05-13 | 广东工业大学 | A monitoring system for influencing factors of product sales for the whole life cycle |
| CN115271796A (en)* | 2022-07-13 | 2022-11-01 | 广东电网有限责任公司 | An early warning method, device, storage medium and system for electricity price |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104123600A (en)* | 2014-08-14 | 2014-10-29 | 国家电网公司 | Electrical manager's index forecasting method for typical industry big data |
| US20150046221A1 (en)* | 2011-09-17 | 2015-02-12 | Autogrid Inc. | Load forecasting from individual customer to system level based on price |
| CN104657788A (en)* | 2015-02-04 | 2015-05-27 | 国家电网公司 | Method for predicting key industrial electricity consumption based on industrial condition index |
| CN109063892A (en)* | 2018-06-25 | 2018-12-21 | 华北电力大学 | Industry watt-hour meter prediction technique based on BP-LSSVM combination optimization model |
- 2021
- 2021-09-23CNCN202111113543.0Apatent/CN113822714A/enactivePending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150046221A1 (en)* | 2011-09-17 | 2015-02-12 | Autogrid Inc. | Load forecasting from individual customer to system level based on price |
| CN104123600A (en)* | 2014-08-14 | 2014-10-29 | 国家电网公司 | Electrical manager's index forecasting method for typical industry big data |
| CN104657788A (en)* | 2015-02-04 | 2015-05-27 | 国家电网公司 | Method for predicting key industrial electricity consumption based on industrial condition index |
| CN109063892A (en)* | 2018-06-25 | 2018-12-21 | 华北电力大学 | Industry watt-hour meter prediction technique based on BP-LSSVM combination optimization model |
Non-Patent Citations (1)
| Title |
|---|
| 崔和瑞 等: "基于ARIMAX 模型的夏季短期电力负荷预测", 电力系统保护与控制, vol. 43, no. 4, pages 108 - 114* |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114493705A (en)* | 2022-01-27 | 2022-05-13 | 广东工业大学 | A monitoring system for influencing factors of product sales for the whole life cycle |
| CN115271796A (en)* | 2022-07-13 | 2022-11-01 | 广东电网有限责任公司 | An early warning method, device, storage medium and system for electricity price |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Rai et al. | Analysis of classical and machine learning based short-term and mid-term load forecasting for smart grid | |
| Sun et al. | Deep learning: spatiotemporal impact of digital economy on energy productivity | |
| US20020194042A1 (en) | Method of business analysis | |
| Balcilar et al. | Long memory, economic policy uncertainty and forecasting US inflation: a Bayesian VARFIMA approach | |
| Jeon et al. | Measuring efficiency of total productive maintenance (TPM): A three-stage data envelopment analysis (DEA) approach | |
| Lam et al. | An artificial neural network and entropy model for residential property price forecasting in Hong Kong | |
| CN114048436A (en) | Construction method and construction device for forecasting enterprise financial data model | |
| Dzobo et al. | Multi-dimensional customer segmentation model for power system reliability-worth analysis | |
| CN113822714A (en) | A method and system for industry electricity consumption forecasting considering price change factors | |
| Velasco et al. | Next day electric load forecasting using Artificial Neural Networks | |
| El Maghraoui et al. | Smart energy management system: A comparative study of energy consumption prediction algorithms for a hotel building | |
| Wu et al. | Data mining algorithms and statistical analysis for sales data forecast | |
| CN117151276A (en) | Intelligent management system of electricity selling platform | |
| Youssef et al. | Exploring the correlation between energy intensity and specific energy consumption in food and kindred industry for the Midwest states | |
| CN115130924A (en) | Microgrid power equipment asset evaluation method and system under source grid storage background | |
| Bareth et al. | Load forecasting model using LSTM for Indian state load dispatch centre | |
| Xia et al. | Correlation analysis of different vegetable categories based on Spearman rank correlation coefficient | |
| CN105741018A (en) | Asset life cycle cost calculation analysis method and system | |
| Prihandi et al. | Comparative Analysis of ARIMA and LSTM Algorithms for Predicting Chilli Prices in Bali | |
| Bai et al. | Foreign trade survey data: Do they help in forecasting exports and imports? | |
| Shahraeini et al. | Towards a Centrality Control Center: Innovations in Complex Power Network Operation and Control | |
| CN113935524A (en) | Prediction method of work order quantity based on long and short-term memory network | |
| Muralidharan et al. | Stochastic hourly load forecasting for smart grids in korea using narx model | |
| Bao et al. | Dynamic XGBoost-based Quantile Predictor for Real-time Electricity Price Forecasting | |
| Gallagher et al. | Utilising the Cross Industry Standard Process for Data Mining to reduce uncertainty in the Measurement and Verification of energy savings |
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 | Application publication date:20211221 | |
| RJ01 | Rejection of invention patent application after publication |