CN111340525A

CN111340525A - Method and system for determining time-sharing electricity price peak electricity price

Info

Publication number: CN111340525A
Application number: CN202010079007.2A
Authority: CN
Inventors: 周颖; 韩凝辉; 李德智; 卜凡鹏; 潘明明; 卢毓东; 刘周斌; 谢祥颖; 骆欣
Original assignee: State Grid Corp of China SGCC; China Electric Power Research Institute Co Ltd CEPRI; Electric Power Research Institute of State Grid Zhejiang Electric Power Co Ltd; State Grid E Commerce Co Ltd
Current assignee: State Grid Corp of China SGCC; China Electric Power Research Institute Co Ltd CEPRI; Electric Power Research Institute of State Grid Zhejiang Electric Power Co Ltd; State Grid E Commerce Co Ltd
Priority date: 2020-02-03
Filing date: 2020-02-03
Publication date: 2020-06-26

Abstract

The invention provides a method and a system for determining a time-sharing electricity price peak electricity price, which comprise the following steps: determining a plurality of days of data most similar to the predicted day from historical data of the scheduling automation system; predicting the load power of a prediction day based on data of a plurality of days to obtain a prediction power curve; and determining the electricity price in the peak time period of the forecast day according to the forecast power curve. The method and the device can measure and calculate the data of the predicted daily load power curve, provide a core algorithm for accurately applying the peak load electricity price and guide the scientificity of demand side management time-sharing pricing.

Description

Method and system for determining time-sharing electricity price peak electricity price

Technical Field

The invention belongs to the technical field of power markets, and particularly relates to a method and a system for determining time-sharing electricity price peak electricity prices.

Background

Peak current rating (CPP): the CPP is a dynamic electricity rate mechanism developed on the basis of time-of-use electricity rates, that is, formed by superimposing a peak rate on the time-of-use electricity rates. The CPP implements a time period setting standard (e.g., system emergency or peak period) for pre-publishing the peak event and a corresponding peak rate, and notifies the user (usually within 1 d) in advance, so that the user can make a corresponding power plan adjustment. Because the CPP rate is also determined in advance, the CPP rate can be up and down floated in each implementation, and the short-term power supply cost in the system peak period is reflected. For example, in 8 months of 2015, according to a weather forecast issued by a central television station, the temperature continues to be high for 5 consecutive days in the next day of south kyo, and the highest temperature meets the implementation condition peak electricity price policy, so that the electricity charge is added by 0.1 yuan in the peak period of 5 days, the province executes a seasonal peak electricity price policy on 5.6 million-family large-scale industrial enterprises, and the electricity charge is 18.9 kilowatt hours in the peak period, wherein the peak electricity charge for effective adding is 3.2 million kilowatt hours, the load of the power grid is reduced by about 10 kilowatts in the early peak, the load of the waist peak is reduced by about 80 kilowatts, the peak load is transferred and reduced by about 90 kilowatts in the late peak, the peak shifting effect is significant, and the income of the peak electricity price is increased by 3200 trillions.

The electricity price of the user is the electricity generation price and the network loss price plus the power grid price plus the electricity selling cost plus the auxiliary service cost.

The short-term power supply cost of the peak period of the system is reflected, the electricity price of the peak period is calculated, the maximum amplitude of the load of the peak period is measured and calculated, and at present, the electricity price of the peak period is added according to a fixed price, and the pertinence is lacked.

Disclosure of Invention

In order to overcome the defects of the prior art, the invention provides a method for determining a time-sharing electricity price peak electricity price, which is improved by comprising the following steps:

determining a plurality of days of data most similar to the predicted day from historical data of the scheduling automation system;

predicting the load power of a prediction day based on the data of the plurality of days to obtain a prediction power curve;

and determining the electricity price in the peak time period of the forecast day according to the forecast power curve.

Preferably, the determining, from historical data of the scheduling automation system, a number of days of data that are most similar to the predicted day includes:

selecting a date with the same type as the predicted daily weather from historical weather data according to the type of the predicted daily weather;

screening out a plurality of dates which are most similar to the predicted day in a preset number by utilizing the minimum Euclidean distance between weather data in the same day as the weather type of the predicted day;

acquiring historical load curves of a plurality of days from historical data of the dispatching automation system as data of the plurality of days according to the screened dates;

wherein the weather data comprises: weather type, maximum air temperature, and minimum air temperature; the weather types include: sunny, partly cloudy, cloudy and rainy.

Preferably, the predicting the load power of the prediction day based on the data of the plurality of days to obtain a predicted power curve includes:

decomposing the historical load curve of each day by respectively adopting a collective empirical mode decomposition method for a plurality of days which are most similar to the predicted day to obtain a plurality of inherent modal function components of the historical load curve of each day;

predicting the value of each inherent modal function component at each moment by adopting a least square method support vector machine;

and obtaining a predicted power curve of a predicted day according to the predicted value of each inherent mode function component at each moment.

Preferably, the obtaining a predicted power curve of a prediction day according to the predicted value of each eigenmode function component at each time includes:

for each moment, carrying out equivalent weighted summation on the value of each inherent modal function component at the moment to obtain the value of the predicted daily load power at each moment;

and obtaining a predicted power curve of the predicted day according to the value of the predicted daily load power at each moment.

Preferably, the determining the peak period electricity price according to the predicted power curve comprises:

obtaining a load amplitude at a peak period according to the predicted power curve;

and determining the electricity price in the peak time period according to the load amplitude in the peak time period.

Preferably, the calculation formula of the electricity price in the peak period is as follows:

the peak period electricity price is the electricity generation price and the grid loss price + the power grid price + the electricity selling cost + the auxiliary service cost/the peak load amplitude.

Preferably, the prediction day is the next day.

Based on the same inventive concept, the invention also provides a system for determining the peak time-share electricity price, and the improvement comprises: the device comprises a data acquisition module, a prediction curve module and a peak electricity price module;

the data acquisition module is used for determining a plurality of days of data which are most similar to the predicted days from historical data of the dispatching automation system;

the prediction curve module is used for predicting the load power of a prediction day based on the data of the plurality of days to obtain a prediction power curve;

and the peak electricity price module is used for determining the electricity price in the peak time period on the forecast day according to the forecast power curve.

Preferably, the data acquisition module includes: the system comprises a type screening unit, a weather screening unit and a curve acquisition unit;

the type screening unit is used for selecting the date which is the same as the predicted day weather type from the historical weather data according to the predicted day weather type;

the weather screening unit is used for screening a plurality of dates which are the most similar to the predicted day in a preset number by using the minimum Euclidean distance between weather data in the same date with the weather type of the predicted day;

the curve acquisition unit is used for acquiring historical load curves of a plurality of days from historical data of the dispatching automation system as data of the plurality of days according to the screened dates;

Preferably, the prediction curve module comprises: the device comprises a curve decomposition unit, a numerical value prediction unit and a prediction curve unit;

the curve decomposition unit is used for decomposing the historical load curve of each day by respectively adopting an ensemble empirical mode decomposition method aiming at a plurality of days most similar to the predicted day to obtain a plurality of inherent mode function components of the historical load curve of each day;

the numerical value prediction unit is used for predicting the value of each inherent modal function component at each moment by adopting a least square method support vector machine;

and the prediction curve unit is used for obtaining a prediction power curve of a prediction day according to the prediction value of each inherent mode function component at each moment.

Compared with the closest prior art, the invention has the following beneficial effects:

Drawings

Fig. 1 is a schematic flow chart of a method for determining a peak electricity price of a time-sharing electricity price according to the present invention;

fig. 2 is a schematic flow chart of an embodiment of a method for determining a peak time-of-use electricity price according to the present invention;

FIG. 3 is a schematic diagram of model training provided by the present invention;

FIG. 4 is a schematic diagram of a load prediction method according to the present invention;

FIG. 5 is a one day load graph relating to the present invention;

FIG. 6 is a schematic diagram of a basic structure of a system for determining a peak electricity price of a time-sharing electricity price according to the present invention;

fig. 7 is a detailed structural diagram of a system for determining a peak time-of-use electricity price according to the present invention.

Detailed Description

The following describes embodiments of the present invention in further detail with reference to the accompanying drawings.

At present, the electric power system transformation gradually develops towards the direction of cleanness, low carbon, safety and high efficiency. The proposal of the concept of three-type two-network raises the importance of demand side management to a new height. It is necessary to further advance peak-valley time-of-use electricity prices and peak electricity prices on the basis of implemented step electricity prices according to the characteristics of electricity loads in different time periods, so that the purposes of adjusting the peak and valley loads of a power grid, improving the curve shape of system loads, cutting peaks and filling valleys, relieving power tension, improving load rate, improving the safety and economy of the power grid, enhancing the operation efficiency and stability of a power system and the like are achieved. At present, a system for demand response of users is being constructed, and a pointed peak electricity price measuring and calculating method and means are lacked. Therefore, research on the peak electricity price core algorithm and means is urgently needed.

The invention provides a method and a system for determining time-sharing electricity price peak electricity price, which are used for calculating a next-day load power curve by using historical load power curve data of a scheduling system, thereby determining the time period adopted by the peak electricity price and the short-time power supply cost, and being used as a relevant policy suggestion for formulating the peak electricity price in a region.

Example 1:

the flow diagram of the method for determining the peak electricity price of the time-sharing electricity price provided by the invention is shown in fig. 1, and the method comprises the following steps:

step 1: determining a plurality of days of data most similar to the predicted day from historical data of the scheduling automation system;

step 2: predicting the load power of a prediction day based on data of a plurality of days to obtain a prediction power curve;

and step 3: and determining the electricity price in the peak time period of the forecast day according to the forecast power curve.

An embodiment of a method for determining a peak time-of-use electricity rate is shown in fig. 2.

Wherein, step 1 includes:

step 101, based on historical data of a dispatching automation system, classifying the data into 5 types including sunny days, local cloudy days, cloudy days and rainy days. In the subsequent steps, in the load power curve data sets in the five types, 96-point load power curve data which are most similar to the next day for 5 days and 24 hours every day every 15 minutes are selected by using the Euclidean distance, and are predicted by using the similar day, so that the influence of the environment is avoided, and the prediction precision is high. Through weather classification, 96-point load power curve data of every 15 minutes in 24 hours of the next day can be measured and calculated under the condition that environmental influence factors change, and the prediction effect is more accurate.

Step 102, utilizing Euclidean distance d_ikCalculating to obtain a Euclidean distance set { d) from the historical data k of the weather of the same type as the next day i to the weather data of the next day_i1,d_i2,…,d_ik,…,d_inAnd sorting the Euclidean distance sets from small to large, and selecting the 5-day data most similar to the next day.

The step 2 specifically comprises the following steps:

step 103, an Ensemble Empirical Mode Decomposition (EEMD) method is a novel adaptive signal time-frequency processing method, and is particularly suitable for analyzing and processing nonlinear and non-stationary signals. And decomposing the original load power curve data value according to the ensemble empirical mode decomposition EEMD to realize the analysis and processing of the nonlinear non-stationary signal.

And step 104, predicting each moment t of each IMF component by adopting a Least Square Support Vector Machine (LSSVM), carrying out equivalent weighting on the IMF components, and carrying out inverse normalization processing to finally obtain a prediction result of the next-day load power curve.

The basic idea of the LSSVM regression algorithm is to map the input vector to a high dimensional feature space by a pre-selected non-linear mapping, in which space the optimal decision function is constructed. When the optimal decision function is constructed, the principle of minimizing the structural risk is utilized, and the kernel function of the original space is skillfully utilized to replace the dot product operation in the high-dimensional feature space. For a given training data set: s ═ x₁,y₁),(x₂,y₂),…,(x_L,y_L))∈Rⁿ× R by first mapping ψ (x) to (Φ (x) with a non-linearity₁),φ(x₂),…,φ(x_L)). An optimal decision function y (x) ω · Φ (x) + b is constructed in this high-dimensional feature space. This translates the non-linear estimation function into a linear estimation function in a high dimensional feature space. By using the principle of minimizing the structural risk, the search for omega and b is to minimize

Wherein |²Controlling the complexity of the model; c is a normalization parameter; r_empThe method is characterized in that the method is an error control function, namely an epsilon loss function and a huber loss function, different loss functions are selected, different forms of support vector machines can be constructed, and the loss function of the least square support vector machine in the process of optimizing a target is the error ξ_iSo the optimization problem is:

s.t:y_i＝φ(x_i).ω+b+ξ_ii＝1,…l

where c is a penalty factor, achieving a compromise between empirical risk and confidence range ξ_i-relaxing the variables, facilitating finding the coefficients. The Lagrange method is used for solving the formula:

in the formula, a_i(i ═ 1, …, l), is the lagrange multiplier. According to the optimization conditions:

the following can be obtained:

a_i＝cξ_i

ω·φ(x_i)+b+ξ_i-y_i＝0

defining a kernel function K (x)_i,y_i)＝φ(x_i)φ(x_j)，K(x_i,y_i) Is a symmetric function satisfying the Mercer condition, and commonly used kernel functions include:

linearityKernel function K (x, x)_i)＝x^Tx_i；

Polynomial kernel function K (x, x)_i)＝(γ₁x^Tx_i+r)^p,γ₁>0；

Radial Basis (RBF) kernel function: k (x, x)_i)＝exp(-γ₂‖x-xi‖²),γ₂>0

According to the above equation, the optimization problem is transformed to solve a linear equation:

finally, a nonlinear equation is obtained:

the unknown points can be predicted from the regression function.

The step 3 comprises the following steps:

step 105: and solving the load power curve data of 96 points of the next day, calculating the short-time power supply cost according to the peak load maximum amplitude, and determining the peak time period electricity price.

Example 2:

a specific example of the time-of-use electricity rate spike electricity rate determination method is given below.

Step 201: load power curve data normalization processing

Because the used data have different dimensions and have large value ranges, the data from different sources can be unified into a reference range through data normalization, thereby improving the prediction efficiency and precision.

In the formula, P_m-load curve power value at mth time point (1. ltoreq. m. ltoreq.96); pmin and Pmax-the minimum and maximum values of the load power in the sequence;

-sequence values after normalization.

Step 202: using Euclidean distance d_ikAnd calculating to obtain a Euclidean distance set { d) of the next day i_i1,d_i2,…,d_ik,…,d_inSorting the Euclidean distance sets from small to large, selecting the 5-day data most similar to the next day,

in the formula, X_i1，X_i2，X_i3The day highest temperature and the day lowest temperature of the next day i; x_k1，X_k2，X_k3The day highest temperature, the day lowest temperature and the date in the kth day data in the historical weather data sample D.

Step 203: the temperatures were normalized, where the temperatures included the highest daily temperature and the lowest daily temperature.

Wherein, T is the temperature value to be normalized; tmin and Tmax-the minimum and maximum values in the temperature series; t-normalized results.

Step 204: the date is normalized, and is normalized as the earth revolves around the sun to generate four seasons change:

in the formula, n is the chronological order.

Step 205: decomposing the normalized data by using EEMD, and decomposing the load power curve data according to EEMD decomposition termination conditions to obtain IMF1, IMF2, IMF3, IMF4, IMF5 … and r_nAnd (4) components. And finally classifying the data according to the serial numbers of the components. The specific decomposition process is as follows:

1) determining all local extreme points of the power time sequence x (t), and then connecting all the extreme points and all the minimum points by using a sample curve respectively to obtain the upper envelope line and the lower envelope line of x (t). The mean value of the upper and lower envelope lines is recorded as m (t).

2) Subtracting the mean value m (t) of the envelope curve from the original power time sequence x (t) to obtain h₁X (t) -m (t), detecting h₁(t) whether two conditions for IMF are satisfied. If not, let h₁(t) as data to be processed, repeating the first step until h₁(t) is a fundamental mode component, denoted c₁(t)＝h₁(t)。

3) Decomposing a first fundamental mode component c from the original time series x (t)₁After (t), subtracting c from x (t)₁(t) obtaining a residue sequence x₁(t)＝x(t)-c₁(t) of (d). B is x₁(t) repeating the above steps as a new "original sequence", sequentially extracting the 2 nd, 3 rd, … … th and nth fundamental mode components until the residual signal satisfies a predetermined termination criterion, such as the residual signal r obtained by decomposition_n(t) sufficiently small or r_n(t) becomes a monotonic function, the whole decomposition process is terminated. Thus, the load power data value x (t) can be expressed as n fundamental mode components c_i(t) and a remainder r_n(t) the sum of (a).

In the above formula, the fundamental mode component c_i(t) may be IMF1, IMF2, IMF3, IMF4, IMF5 …

Because each IMF component represents a data sequence of a characteristic scale, the process of "screening" is actually a superposition process that decomposes the load power curve data values into various characteristic fluctuation sequences.

Similarly, the 5-day load power curve data most similar to the next day is decomposed into IMF1, IMF2, IMF3, IMF4, IMF5 … and r_nAnd (4) components. If the number of decompositions is insufficient, such as historical data for a certain day only decomposing to IMF4, IMF5 is filled with zeros.

Step 206: and (5) training a model.

Respectively applying LSSVM to each decomposed IMF component for prediction, adopting a multi-input and single-output prediction method, taking the time t of an IMF1 component of a sunny type as an example, an IMF1 prediction model LSSVM1(t) at the time t is used for predicting the value IMF1(t) of the IMF1 component at the time t, and the training process of the LSSVM1 is shown in FIG. 3.

The method comprises the following steps that IMFi (1,1, t), IMFi (1,2, t), IMFi (1,3, t), IMFi (1,4, t) and IMFi (1,5, t) sequentially represent IMF1 component values of clear day data and t moments of similar days 1-5 d of a next day i, the highest day temperature, the lowest day temperature and the date of the next day i form input of an LSSVM1(t), the output is IMF1i (t), namely the IMF1 component of the next day i at the t moment is represented, wherein t is more than or equal to 1 and less than or equal to 96, and for a given training data set: s ═ x₁,y₁),(x₂,y₂),…,(x_L,y_L))∈Rⁿ× R by first mapping ψ (x) to (Φ (x) with a non-linearity₁),φ(x₂),…,φ(x_L)). An optimal decision function y (x) ω · Φ (x) + b is constructed in this high-dimensional feature space. This translates the non-linear estimation function into a linear estimation function in a high dimensional feature space. By using the principle of minimizing the structural risk, the search for omega and b is to minimize

s.t:y_i＝φ(x_i).ω+b+ξ_ii＝1,…l (2)

where c is a penalty factor, achieving a compromise between empirical risk and confidence range ξ_i-the value of the relaxation variable,and the coefficient is easy to find. The Lagrange method is used for solving the formula:

the following can be obtained:

a_i＝cξi (6)

ω·φ(x_i)+b+ξ_i-y_i＝0 (7)

linear kernel function K (x, x)_i)＝x^Tx_i；

Polynomial kernel function K (x, x)_i)＝(γ₁x^Tx_i+r)^p,γ₁>0；

According to the equations (4) to (7), the optimization problem is converted into solving a linear equation:

finally, a nonlinear equation is obtained:

the unknown points can be predicted from the regression function.

Step 207: through training of multiple groups of data, an IMF1 prediction model LSSVM1(t) at the t moment under a sunny type can be obtained. Similarly, training resulted in IMF2, IMF3, IMF4, IMF5 … and r_nLSSVM1(t) to LSSVMn (t).

In step 207: and selecting LSSVM1(t) -LSSVM 6(t) with the same weather type as the day to be measured, predicting values of IMF 1-IMF 6 components at the time t, namely IMF1(t) -IMF 6(t), carrying out equivalent weighting on the values, and carrying out inverse normalization processing to finally obtain a prediction result of the next-day load power curve.

IMF (1, β, t) -IMF (6, β, t) represent data sets of IMF 1-IMF 6 components of 5 similar days at time t, where 1 ≦ β ≦ 5, representing 5 similar days, as shown in FIG. 4.

Step 208: and solving the load power curve data of 96 points of the next day, calculating the short-time power supply cost according to the peak load maximum amplitude, and determining the peak time period added value.

As shown in fig. 5, the predicted 20: 00-22: 30, the load ratio is higher than usual, and the power supply cost is increased, and the electricity price in the peak period is calculated.

Example 3:

based on the same inventive concept, the invention also provides a system for determining the time-of-use electricity price peak electricity price.

The basic structure of the system is shown in fig. 6, and comprises: the device comprises a data acquisition module, a prediction curve module and a peak electricity price module;

the prediction curve module is used for predicting the load power of a prediction day based on data of a plurality of days to obtain a prediction power curve;

The detailed structure of the time-of-use electricity price peak electricity price determination system is shown in fig. 7.

Wherein, the data acquisition module includes: the system comprises a type screening unit, a weather screening unit and a curve acquisition unit;

the weather screening unit is used for screening a plurality of dates which are the most similar to the predicted days in a preset number by utilizing the minimum Euclidean distance between weather data in the days with the same weather types as the predicted days;

the curve acquisition unit is used for acquiring historical load curves of a plurality of days from the historical data of the dispatching automation system as data of the plurality of days according to the screened dates;

wherein the weather data includes: weather type, maximum air temperature, and minimum air temperature; the weather types include: sunny, partly cloudy, cloudy and rainy.

Wherein the prediction curve module comprises: the device comprises a curve decomposition unit, a numerical value prediction unit and a prediction curve unit;

the curve decomposition unit is used for decomposing the historical load curve of each day by respectively adopting a set empirical mode decomposition method aiming at a plurality of days most similar to the predicted day to obtain a plurality of inherent modal function components of the historical load curve of each day;

Wherein the prediction curve unit includes: a predictor subunit and a curve subunit;

the predicted value subunit is used for carrying out equivalent weighted summation on the value of each inherent modal function component at each moment to obtain the value of the predicted daily load power at each moment;

and the curve subunit is used for obtaining a predicted power curve of the predicted day according to the value of the predicted daily load power at each moment.

Wherein the spike price module includes: a peak load unit and a peak electricity price unit;

the peak load unit is used for obtaining a load amplitude value in a peak period according to the predicted power curve;

and the peak electricity price unit is used for determining the electricity price in the peak time period according to the load amplitude in the peak time period.

As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

It should be noted that the above-mentioned embodiments are only for illustrating the technical solutions of the present application and not for limiting the scope of protection thereof, and although the present application is described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that after reading the present application, they can make various changes, modifications or equivalents to the specific embodiments of the application, but these changes, modifications or equivalents are all within the scope of protection of the claims to be filed.

Claims

1. A method for determining a peak time-of-use electricity price, comprising:

2. The method of claim 1, wherein determining the number of days of data that are most similar to the predicted day from historical data of the scheduling automation system comprises:

3. The method of claim 2, wherein predicting the load power for a predicted day based on the data for the number of days to obtain a predicted power curve comprises:

4. The method of claim 3, wherein obtaining the predicted power curve for the predicted day based on the predicted values of the respective components of the eigenmode function at the respective time instants comprises:

5. The method of claim 1, wherein determining a spike period electricity price from the predicted power curve comprises:

6. The method of claim 5, wherein the peak period electricity prices are calculated as follows:

7. The method of claim 1, wherein the predicted day is the next day.

8. A time-of-use electricity price peak electricity price determination system, comprising: the device comprises a data acquisition module, a prediction curve module and a peak electricity price module;

9. The system of claim 8, wherein the data acquisition module comprises: the system comprises a type screening unit, a weather screening unit and a curve acquisition unit;

10. The system of claim 9, wherein the prediction curve module comprises: the device comprises a curve decomposition unit, a numerical value prediction unit and a prediction curve unit;