Time-of-use electricity price short-term load prediction method based on demand response
By establishing a demand response model and utility theory, and combining seasonal load characteristics and dynamic electricity prices, a radial basis function neural network model is adopted to solve the problems of insufficient short-term load dispatch and power reliability in existing power forecasting systems, thus achieving higher accuracy and more stable load forecasting.
Patent Information
- Application Number
- CN202411097186.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-12
- Publication Date
- 2024-11-22
AI Technical Summary
The lack of research on reasonable scheduling of short-term loads and power reliability in existing power forecasting systems leads to insufficient accuracy and stability in short-term load forecasting.
A demand response-based time-of-use pricing method for short-term load forecasting is adopted. By establishing a demand response model and determining the elasticity coefficient, and combining seasonal load characteristics and dynamic peak-valley pricing, load forecasting is performed using utility theory and a radial basis function neural network model, taking into account the influence of various factors on user electricity consumption behavior.
It has improved the accuracy and stability of short-term load forecasting, provided more reasonable references for electricity pricing and electricity consumption behavior, and optimized the market structure.
Smart Images

Figure CN119005613A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of power forecasting, and in particular, relates to a time-of-use electricity price short-term load forecasting method based on demand response. Background Art
[0002] With the development of the power industry, it is proposed to vigorously improve the elasticity of power load, strengthen the construction of power demand side response capabilities, integrate decentralized demand response resources, guide users to optimize power storage and use modes, and release the elasticity of residential and general industrial and commercial power loads at a high proportion.
[0003] In the traditional electricity market, demand-side management is dominated by government departments, while demand-side response focuses more on active user participation and is more market-oriented, which can be regarded as the mature stage of demand-side management. The introduction of demand-side response and the rapid development of smart grids have further solved the interaction problem between the demand side and the supply side through the active and spontaneous behavior of major participants. At present, there are many research angles on the economic dispatch of demand response to complete mathematical modeling and load forecasting. While implementing demand response, it alleviates the contradiction between the supply side and the demand side and reasonably optimizes the market structure, but it also brings challenges to the reasonable dispatch of short-term loads and the reliability of electricity use. Summary of the invention
[0004] In view of the problem that the existing power forecasting system lacks consideration of reasonable scheduling of short-term loads and power reliability, the present invention provides a time-of-use electricity price short-term load forecasting method based on demand response.
[0005] In order to achieve the above technical objectives, the technical solution adopted by the present invention is as follows:
[0006] A time-of-use electricity price short-term load forecasting method based on demand response, comprising the steps of:
[0007] S1. Collect users’ electricity demand at different time periods and establish a demand response model;
[0008] S2, demand response type selection;
[0009] S3. Establish the basic model of utility theory;
[0010] S4. Determine the elasticity coefficient of the demand response model;
[0011] S5. Determine the dynamic peak and valley electricity price based on seasonal characteristics of load;
[0012] S6. Conduct a detailed analysis of the implementation plan for the user's electricity consumption behavior;
[0013] S7. Demand response time-of-use electricity price short-term load forecast.
[0014] Furthermore, demand response types include time-of-use electricity prices and tiered electricity prices.
[0015] Furthermore, the basic model of utility theory is:
[0016]
[0017] In the model: U represents the utility function of electricity consumers, and ρ represents the elasticity coefficient.
[0018] Furthermore, the price demand elasticity matrix is used to describe the relationship between user electricity consumption and electricity price changes: the elasticity coefficient e ii It represents the relationship between the change of electricity price at time i and the change of electricity consumption of user i, and the cross elasticity coefficient e i j represents the relationship between the change of j electricity price and the change of i user's electricity consumption, and the self-elasticity coefficient e ii and the cross elastic coefficient e ij The expressions are:
[0019]
[0020] In the model:
[0021] e ii ——autoelastic coefficient;
[0022] ΔQ i ——The change in the user's electricity consumption at time i;
[0023] Q i ——The user’s original electricity consumption at time i;
[0024] ΔP i ——The change in electricity price at time i;
[0025] P i ——The original electricity price at time i;
[0026] e i j——cross elastic coefficient;
[0027] ΔP j ——The change in electricity price at time j;
[0028] P j ——The original electricity price at time j.
[0029] Furthermore, the expression of the price demand elasticity matrix is:
[0030]
[0031] In the model: n is the number of time periods for electricity consumption data. This paper selects n=96, takes 15 minutes as the sampling interval, and collects 24 hours of user data.
[0032] According to the price demand elasticity matrix, the power consumption after response is obtained, and its expression is:
[0033]
[0034] In the model: Q' n ——The electricity consumption after the user implements demand response at time n;
[0035]
[0036] In the model: η n Indicates the user's responsiveness at time n.
[0037] Furthermore, according to the actual electricity load of the enterprise, the load consumed by the user during peak hours, normal hours and valley hours before the implementation of demand response is recorded as D f0 , D p0 , D g0 After implementing demand response, the load of users in the corresponding time periods is recorded as D f , D p , D g ; Before implementing demand response, the basic load of the enterprise during peak, flat and valley periods is: D fd , D pd , D gd ; The current basic electricity prices for peak, flat and valley periods are: P f0 , P p0 , P g0 ; The electricity price after implementing demand response is: P f , P p , P g ; The expected ratio of enterprise production optimization target is: q pf ,q g f, q gp ;
[0038] The budget constraint before implementing demand response is:
[0039] C 0 =D f0 ×P f0 +D p0 ×P p0 +D g0 ×P g0
[0040] The budget constraint after implementing demand response is:
[0041] C=D fd ×P f +D pd ×P p +(D f0 -D fd)×(q pf ×P p +q gf ×P g )+D gd ×P g +(D p0 -D pd )×q gp ×P g
[0042] Final budget change:
[0043] ΔC=CC 0
[0044] Additional power consumption:
[0045] Q fx =ΔC / P f
[0046] Q px =ΔC / P p
[0047] Q gx =ΔC / P g
[0048] Load transfer situation:
[0049] Q f =D fd -D f0
[0050] Q p =D pd +(D f0 -D fd )×q pf -D p0
[0051] Q g =D gd +(D f0 -D fd )×q gf +(D p0 -D pd )×q gp -D g0
[0052] Load in peak, flat and valley periods after demand response:
[0053] D f =D f0 +Q f
[0054] D p =D p0 +Qp .
[0055] Compared with the prior art, the present invention has the following beneficial effects:
[0056] In the current short-term load forecasting research, there are still few studies that consider the impact of time-of-use electricity prices and seasonal characteristics of loads. In the existing short-term load forecasting evaluation, the research methods and contents are relatively simple. This invention proves through research that the forecasting model taking into account demand response has high forecasting accuracy and good stability. By using utility theory and simulating the actual demand response curve, the influence of multiple factors is introduced to construct a radial basis function neural network short-term load forecasting model taking into account demand response signals, and establish a reasonable and effective forecasting method to provide a reference for users' future electricity price formulation and electricity consumption behavior. BRIEF DESCRIPTION OF THE DRAWINGS
[0057] Figure 1 The present invention is an overall flow chart of a method for short-term load forecasting based on time-of-use electricity prices and demand response in an embodiment of the present invention. DETAILED DESCRIPTION
[0058] In order to facilitate the understanding of those skilled in the art, the present invention is further described below in conjunction with embodiments and drawings. The contents mentioned in the implementation modes are not intended to limit the present invention.
[0059] like Figure 1 As shown, this embodiment provides a time-of-use electricity price short-term load forecasting method based on demand response, including the steps of:
[0060] S1. Collect users’ electricity demand at different time periods and establish a demand response model;
[0061] S2, demand response type selection;
[0062] S3. Establish the basic model of utility theory;
[0063] S4. Determine the elasticity coefficient of the demand response model;
[0064] S5. Determine the dynamic peak and valley electricity price based on seasonal characteristics of load;
[0065] S6. Conduct a detailed analysis of the implementation plan for the user's electricity consumption behavior;
[0066] S7. Demand response time-of-use electricity price short-term load forecast.
[0067] Demand response types include time-of-use electricity prices and tiered electricity prices. Capacity electricity prices are not involved, and load transfer is considered in the process of user behavior transfer, but load reduction is not considered.
[0068] By analyzing the literature, starting from the two dimensions of high reliability and economy, the research scope of load forecasting based on demand response dynamic electricity price is preliminarily defined.
[0069] The basic model of utility theory is:
[0070]
[0071] In the model: U represents the utility function of electricity consumers, and ρ represents the elasticity coefficient. In the existing economic research theory, it is shown that if the price of a certain item rises, the demand for another item that is a substitute for it will increase; if the price of a certain item falls, the demand for another item that is a substitute for it will decrease. Combining the utility theory with consumers' electricity purchasing behavior, electricity products at different time periods are regarded as interchangeable commodities. The commodity selection process is the process of consumers shifting from using electricity during high-priced periods to using electricity during low-priced periods.
[0072] Furthermore, the price demand elasticity matrix is used to describe the relationship between user electricity consumption and electricity price changes: the elasticity coefficient e ii It represents the relationship between the change of electricity price at time i and the change of electricity consumption of user i, and the cross elasticity coefficient e i j represents the relationship between the change of j electricity price and the change of i user's electricity consumption, and the self-elasticity coefficient e ii and the cross elastic coefficient e ij The expressions are:
[0073]
[0074] In the model:
[0075] e ii ——autoelastic coefficient;
[0076] ΔQ i ——The change in the user's electricity consumption at time i;
[0077] Q i ——The user’s original electricity consumption at time i;
[0078] ΔP i ——The change in electricity price at time i;
[0079] P i ——The original electricity price at time i;
[0080] e i j——cross elastic coefficient;
[0081] ΔP j ——The change in electricity price at time j;
[0082] P j——The original electricity price at time j.
[0083] Different electricity prices have different impacts on users and lead to different reactions. When the electricity price fluctuation is too small, it is not enough to change the user's electricity usage, and the user will not react. However, when the electricity price fluctuates to a certain range, it will have a significant impact on the user's behavior, and the user will react and change his or her electricity usage.
[0084] The expression of price demand elasticity matrix is:
[0085]
[0086] In the model: n is the number of time periods for electricity consumption data. This paper selects n=96, takes 15 minutes as the sampling interval, and collects 24 hours of user data.
[0087] According to the price demand elasticity matrix, the power consumption after response is obtained, and its expression is:
[0088]
[0089] In the model: Q' n ——The electricity consumption after the user implements demand response at time n;
[0090]
[0091] In the model: η n Indicates the user's responsiveness at time n.
[0092] Furthermore, according to the actual electricity load of the enterprise, the load consumed by the user during peak hours, normal hours and valley hours before the implementation of demand response is recorded as D f0 , D p0 , D g0 After implementing demand response, the load of users in the corresponding time periods is recorded as D f , D p , D g ; Before implementing demand response, the basic load of the enterprise during peak, flat and valley periods is: D fd , D pd , D gd ; The current basic electricity prices for peak, flat and valley periods are: P f0 , P p0 , P g0 ; The electricity price after implementing demand response is: P f , P p , P g ; The expected ratio of enterprise production optimization target is: q pf ,q g f, q gp ;
[0093] The specific indicator details and corresponding expressions are shown in Table 1 below:
[0094] Table 1 Index list
[0095]
[0096] The budget constraint before implementing demand response is:
[0097] C 0 =D f0 ×P f0 +D p0 ×P p0 +D g0 ×P g0
[0098] The budget constraint after implementing demand response is:
[0099] C=D fd ×P f +D pd ×P p +(D f0 -D fd )×(q pf ×P p +q gf ×P g )+D gd ×P g +(D p0 -D pd )×q gp ×P g
[0100] Final budget change:
[0101] ΔC=CC 0
[0102] Additional power consumption:
[0103] Q fx =ΔC / P f
[0104] Q px =ΔC / P p
[0105] Q gx =ΔC / P g
[0106] Load transfer situation:
[0107] Q f =D fd -D f0
[0108] Q p =D pd +(D f0 -D fd )×q pf -D p0
[0109] Q g =D gd +(D f0 -D fd )×q gf +(D p0 -D pd )×q gp -D g0
[0110] Load in peak, flat and valley periods after demand response:
[0111] D f =D f0 +Q f
[0112] D p =D p0 +Q p .
[0113] Compared with the prior art, the present invention has the following beneficial effects:
[0114] In the current short-term load forecasting research, there are still few studies that consider the impact of time-of-use electricity prices and seasonal characteristics of loads. In the existing short-term load forecasting evaluation, the research methods and contents are relatively simple. The present invention proves through research that the forecasting model taking into account demand response has high forecasting accuracy and good stability. Therefore, the present invention studies and utilizes utility theory, adopts a method of simulating the actual demand response curve, introduces the influence of multiple factors, constructs a radial basis function neural network short-term load forecasting model taking into account demand response signals, and establishes a reasonable and effective forecasting method to provide a reference for users' future electricity price formulation and electricity consumption behavior.
[0115] The above is a detailed introduction to a time-of-use electricity price short-term load forecasting method based on demand response provided by the present application. The description of the specific embodiment is only used to help understand the method and its core idea of the present application. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of the present application, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the scope of protection of the claims of the present application.
Claims
1. A time-of-use electricity price short-term load forecasting method based on demand response, characterized in that: Includes steps: S1. Collect users’ electricity demand at different time periods and establish a demand response model; S2, demand response type selection; S3. Establish the basic model of utility theory; S4. Determine the elasticity coefficient of the demand response model; S5. Determine the dynamic peak and valley electricity price based on seasonal characteristics of load; S6. Conduct a detailed analysis of the implementation plan for the user's electricity consumption behavior; S7. Demand response time-of-use electricity price short-term load forecast.
2. A time-of-use electricity price short-term load forecasting method based on demand response according to claim 1, characterized in that: Demand response types include time-of-use pricing and tiered pricing.
3. A time-of-use electricity price short-term load forecasting method based on demand response according to claim 2, characterized in that: The basic model of utility theory is: In the model: U represents the utility function of electricity consumers, and ρ represents the elasticity coefficient.
4. A time-of-use electricity price short-term load forecasting method based on demand response according to claim 3, characterized in that: The price demand elasticity matrix is used to describe the relationship between user electricity consumption and electricity price changes: the elasticity coefficient e ii It represents the relationship between the change of electricity price at time i and the change of electricity consumption of user i, and the cross elasticity coefficient e ij It represents the relationship between the change of electricity price j and the change of electricity consumption of user i, and the elasticity coefficient e ii and the cross elastic coefficient e i The expressions of j are: In the model: e ii ——autoelastic coefficient; ΔQ i ——The change in the user's electricity consumption at time i; Q i ——The user’s original power consumption at time i; ΔP i ——The change in electricity price at time i; P i ——The original electricity price at time i; e i j——cross elastic coefficient; ΔP j ——The change in electricity price at time j; P j ——The original electricity price at time j.
5. A time-of-use electricity price short-term load forecasting method based on demand response according to claim 4, characterized in that: The expression of price demand elasticity matrix is: In the model: n is the number of time periods for electricity consumption data, n=96 is selected, with 15 minutes as the sampling interval, and 24 hours of user data is collected.
6. A time-of-use electricity price short-term load forecasting method based on demand response according to claim 5, characterized in that: According to the actual situation of the enterprise's electricity load, the load consumed by the user during peak hours, normal hours and valley hours before the implementation of demand response is recorded as D f0 , D p0 , D g0 After implementing demand response, the load of users in the corresponding time periods is recorded as D f , D p , D g ; Before implementing demand response, the basic load of the enterprise during peak, flat and valley periods is: D fd , D pd , D gd ; The current basic electricity prices for peak, flat and valley periods are: P f0 , P p0 , P g0 ; The electricity price after implementing demand response is: P f , P p , P g ; The expected ratio of enterprise production optimization target is: q pf ,q gf ,q gp ; The budget constraint before implementing demand response is: C0=D f0 ×P f0 +D p0 ×P p0 +D g0 ×P g0 The budget constraint after implementing demand response is: C=D fd ×P f +D pd ×P p +(D f0 -D fd )×(q pf ×P p +q gf ×P g )+D gd ×P g +(D p0 -D pd )×q gp ×P g Final budget change: ΔC=C-C0 Additional power consumption: Q fx =ΔC / P f Q px =ΔC / P p Q gx =ΔC / P g Load transfer situation: Q f =D fd -D f0 Q p =D pd +(D f0 -D fd )×q pf -D p0 Q g =D gd +(D f0 -D fd )×q gf +(D p0 -D pd )×q gp -D g0 Loads during peak, flat and valley periods after demand response: D f =D f0 +Q f D p =D p0 +Q p 。
