Method for predicting spot market deviation settlement price through bidding space

Abstract

The invention provides a method for predicting a spot market deviation settlement price through a bidding space, and the method comprises the steps: determining the bidding space, carrying out the monthly feature analysis, and dividing the bidding space into four spot time periods; and obtaining a monthly spot goods deviation settlement price by adopting a spot goods deviation settlement price prediction model and comprehensively considering a monthly medium and long term price, a monthly spot goods price, a monthly new energy signing curve and a monthly output characteristic curve. According to the method, the concept of bidding space is introduced, the conditions of different time periods of supply and demand tension and supply and demand looseness in the spot market are effectively reflected, a good and simple spot market time period division mode is established, and the mode is well linked with a medium-and-long-term market time period division mode. On the basis of medium and long term market price prediction and calculation, spot market deviation settlement price is introduced, a more comprehensive and objective new energy expected electricity price income calculation result is obtained, and important support is provided for new energy project full life cycle income level calculation and investment decision.

Description

Technical Field

[0001] The present invention belongs to the technical field of price prediction, and in particular relates to a method for predicting a deviation settlement price in a spot market through a bidding space. Background Art

[0002] Under the guidance of accelerating the construction of a unified power market system, the power market has basically formed a unified power market "1+N" basic rule system with medium- and long-term, spot, and auxiliary service rules as the main trunk, and information disclosure, access registration, and metering and settlement rules as the support. Under this rule system, the transaction price of electricity in the power market is mainly composed of medium- and long-term transaction prices and spot transaction prices, which is also the core price of most concern in the cost-benefit analysis of new energy projects. At present, the power spot market in most parts of the country has entered the trial operation stage. On the basis of medium- and long-term transactions, the new energy electricity needs to be finally settled in the spot market. The settlement method is the deviation settlement between the medium- and long-term and spot, that is, the part of the medium- and long-term transaction transaction volume is still settled at the medium- and long-term contract price. In the intraday spot market, the actual transaction volume and the deviation of the medium- and long-term transaction agreed volume are settled at the spot price, and the final settlement price of the market transaction is obtained.

[0003] The spot deviation settlement price is an important part that affects the revenue of new energy projects. However, there is no mature model to carry out the long-term electricity market spot deviation settlement price forecasting work. The difficulty lies in the difficulty of judging the spot price level at each moment and the degree of deviation from the medium and long-term price from a long-term perspective. The existing electricity price forecasting models often only calculate the spot market price within a short-term range of 1 to 3 days through sequence analysis and trend forecasting to provide decision support for day-ahead electricity market transactions, and cannot effectively support the spot deviation settlement price forecast from a long-term perspective. Summary of the invention

[0004] In view of the defects of the prior art, the present invention provides a method for predicting the deviation settlement price of the spot market through bidding space, which can effectively solve the above problems.

[0005] The technical solution adopted by the present invention is as follows:

[0006] The present invention provides a method for predicting a spot market deviation settlement price through a bidding space, comprising the following steps:

[0007] Step S1, based on the characteristic load curve, new energy output characteristic curve and adjustable power installed capacity of the target area planned in the future years, obtain the bidding space j(t) time series at each time t in the future years;

[0008] Step S2, taking months as units, analyze the monthly characteristics of the bidding space j(t) time series at each moment t in the future years, and divide the H moments in a day into four spot periods, namely: spot peak period T1, spot peak period T2, spot normal period T3 and spot valley period T4, each spot period corresponds to the intraday time range;

[0009] Step S3, pre-set the spot price coefficients of the spot peak period T1, the spot peak period T2, the spot normal period T3 and the spot valley period T4, which correspond to the spot price coefficients k1, k2, k3 and k4 respectively, so as to obtain the monthly spot price coefficient matrix k of each month i i =[k i (h)],h=1,2,…,H; where k i (h) represents the spot price coefficient at time h of each month i;

[0010] Step S4, determine the monthly medium- and long-term prices of each month i; according to the monthly spot price coefficient matrix k of each month i i =[k i (h)], h=1,2,…,H, considering the relationship with the monthly medium- and long-term prices of various periods of each month i, the monthly spot price is obtained;

[0011] Step S5, using the spot deviation settlement price prediction model, comprehensively considering the monthly medium- and long-term prices, monthly spot prices, the monthly new energy contract curve of the future medium- and long-term market, and the monthly output characteristic curve of each month i, predicts the monthly spot deviation settlement price.

[0012] Preferably, in step S1, the characteristic load curve of the target area planned in the next year is c=f(t), the new energy output characteristic curve is p=g(t), and the adjustable power supply installed capacity is P T , using formula (1), we can get the bidding space j(t) at each time t in the future years:

[0013]

[0014] Preferably, step S2 specifically comprises:

[0015] Step S2.1, divide the future year into 12 months by month; based on the time series of the bidding space j(t) at each moment t in the future year, for each month i, i = 1, 2, ..., 12, extract the bidding space at each moment h of each day d to obtain the bidding space monthly matrix J i ={J i (d,h)}, d=1,2,...,D, h=1,2,...,H, where D is the total number of days in each month i, H is the total number of hours in a day, and J i(d,h) represents the bidding space value of month i at time h on day d;

[0016] Step S2.2: For each month i, the monthly matrix J of the bidding space i ={J i (d,h)}, and obtain the bidding space sequence {J i (1,h),J i (2,h),...,J i (d,h),...,J i (D,h)}, and extract the median of the bidding space from the bidding space sequence as the monthly mean of the bidding space at time h in the day jm i (h), the monthly average of the bidding space at each intraday time h jm i (h) Arrange in chronological order to obtain the monthly mean vector jm of the bidding space i =[jm i (1),jm i (2),...,jm i (h),...,jm i (H)];

[0017] Step S2.3: For each month i, the monthly mean vector jm of the bidding space i =[jm i (1),jm i (2),...,jm i (h),...,jm i The distribution of the monthly averages of each bidding space in (H)] is analyzed, and the mapping relationship between the monthly average level of the bidding space and the level of the spot time period is established, so as to divide the H moments in the day into four spot time periods, namely: spot peak time period T1, spot peak time period T2, spot normal time period T3 and spot valley time period T4, and each spot time period corresponds to the intraday time range.

[0018] Preferably, step S2.3 is specifically:

[0019] Step S2.3.1, using formula (2), for each month i, the monthly mean vector jm of the bidding space i =[jm i (1),jm i (2),...,jm i (h),...,jm i The monthly averages of each bidding space in (H)] are arranged in descending order to obtain the descending vector jmd of the monthly average of the bidding space shown in formula (3): i :

[0020] jmd i=descent[jm i (1),jm i (2),...,jm i (h),...,jm i (H)] (2)

[0021] Among them: descent[] represents the descending sort operation;

[0022] jmd i =[jmd i (h),jmd loc,i (h)],h=1,2,...,H (3)

[0023] Among them: jmd loc,i (h) represents the bid space value jmd after descending order i (h) The corresponding intraday time h.

[0024] Step S2.3.2, dividing the H bidding space monthly averages arranged in descending order into four sets from high to low, namely: the first bidding space monthly average set, the second bidding space monthly average set, the third bidding space monthly average set and the fourth bidding space monthly average set;

[0025] The time range corresponding to the first set of bidding space monthly average sets, the time range corresponding to the second set of bidding space monthly average sets, the time range corresponding to the third set of bidding space monthly average sets and the time range corresponding to the fourth set of bidding space monthly average sets are respectively used as the spot peak period T1, the spot peak period T2, the spot normal period T3 and the spot valley period T4, thereby obtaining four spot periods divided by H time points within the day, and the intraday time range corresponding to each spot period.

[0026] Preferably, step S4 is specifically:

[0027] Step S4.1, determine the monthly medium and long-term prices for each month i:

[0028] Step S4.1.1, obtain the monthly medium- and long-term normal time period forecast price pz for each month i in the medium- and long-term market i ,i=1,2,…,12;

[0029] Step S4.1.2, obtain the monthly mid- to long-term peak trading price pz of each month i in the mid- to long-term market i1 and the monthly medium- and long-term normal time period forecast price pz i The ratio s1, the monthly mid- and long-term peak trading price pz i2 and the monthly medium- and long-term normal time period forecast price pz iThe ratio s2 and the monthly mid- to long-term valley trading price pz i3 and the monthly medium- and long-term normal time period forecast price pz i The ratio s3;

[0030] Therefore, according to formula (4), the monthly mid- to long-term peak trading price pz of each month i in the mid- to long-term market is obtained: i1 、Monthly mid- to long-term peak trading price pz i2 and the monthly mid- to long-term valley trading price pz i3 ;

[0031]

[0032] Step S4.1.3, according to the pre-agreed correspondence between the medium- and long-term peak period, the medium- and long-term peak period, the medium- and long-term normal period, the medium- and long-term valley period and the intraday period, the monthly medium- and long-term peak period price pz of each month i in the medium- and long-term market is calculated. i1 , monthly mid- to long-term peak period price pz i2 、Medium and long term normal time period forecast price pz i and the monthly mid- to long-term valley price pz i3 , corresponding to each intraday period, so as to obtain the monthly medium- and long-term price vector r_pz for each month i i =[r_pz i (h)], h = 1, 2, ..., H, where r_pz i (h) represents the medium- and long-term price at time h during the day of month i;

[0033] Step S4.2, determine the monthly spot price for each month i:

[0034] Step S4.2.1, according to formula (5), the spot prices of the spot peak period T1, the spot peak period T2, the spot normal period T3 and the spot valley period T4 are obtained respectively, expressed as: spot peak period spot price px i1 , spot price during peak period px i2 , spot price during normal hours px i3 and spot price px during the spot valley period i4 :

[0035]

[0036] Step S4.2.2, according to the intraday time range corresponding to each spot time period determined in step S2, the spot prices of various spot time periods are mapped to each intraday time period, so as to obtain the monthly spot price vector r_px of each month i i =[r_px i (h)], h=1,2,…,H.

[0037] Preferably, step S5 is specifically:

[0038] Step S5.1: The monthly medium- and long-term prices r_pz at time h on each day of each month i are i (h) is extended to the same time h of each day in month i, so as to obtain the monthly medium- and long-term price series r_pz at each time t of each month i i (1),r_pz i (2),...,r_pz i (N i ),N i is the number of moments t in month i;

[0039] The monthly spot price r_px at time h on each day of month i i (h) is extended to the same time h of each day in month i, thus obtaining the monthly spot price series r_px at each time t of each month i i (1),r_px i (2),...,r_px i (N i );

[0040] Step S5.2, obtain the monthly i new energy contract curve c of the future mid- to long-term market 1,i (N t ),N t =1,2,…,N i and monthly output characteristic curve c 2,i (N t ),N t =1,2,…,N i , using the spot deviation settlement price prediction model of formula (6), we get the monthly spot deviation settlement price pp i :

[0041]

[0042] This step is finished.

[0043] The method provided by the present invention for predicting the deviation settlement price of the spot market through the bidding space has the following advantages:

[0044] By introducing the concept of bidding space, the present invention can effectively reflect the different periods of tight supply and demand and loose supply and demand in the spot market, thereby establishing a good and simple spot market period division method, which can be well connected with the medium- and long-term market period division method, and further realize the price calculation of the deviation settlement in the spot market in the medium- and long-term market. On the basis of the medium- and long-term market price forecast and calculation, by introducing the spot market deviation settlement price calculation results, a more comprehensive and objective new energy expected electricity price income calculation result can be obtained, which provides important support for the calculation of the income level and investment decision-making of the new energy project throughout its life cycle. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A flowchart of a method for predicting spot market deviation settlement prices through bidding space provided by the present invention. DETAILED DESCRIPTION

[0046] In order to make the technical problems, technical solutions and beneficial effects solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0047] With the continuous development of the power market and the continuous improvement of power market rules, the connection and coupling between the medium- and long-term trading market and the spot market is the core link of power market transactions. The combination of medium- and long-term market transaction prices and spot deviation settlement prices forms the source of income for new energy projects in the power market. In order to reasonably carry out investment decision-making analysis and cost-benefit analysis of new energy projects, it is crucial to predict and judge the power market price throughout the life cycle, especially the spot deviation settlement price. The spot deviation settlement price is affected by the real-time load demand and power output level, has strong volatility and irregularity, and is difficult to predict. At present, the time period of the more mature spot price prediction model is often 1 to 3 days, which has a more effective support role for the real-time operation of the power market and the spot market revenue decision-making, but it cannot support the long-term electricity price revenue analysis and forecast, and it is difficult to serve the measurement of revenue analysis in the process of new energy project investment decision-making. Therefore, on the basis of fully considering the spot market price and the real-time supply and demand relationship, the concept of bidding space is proposed to model the spot price coefficients at different times. The bidding space parameters of future years are calculated by combining the hourly characteristic curve of load in future years, the hourly characteristic curve of new energy and the installed capacity of regulating power supply. The spot price level at different times in future years and the deviation from the medium- and long-term predicted transaction price are predicted to obtain the spot deviation settlement price, which provides important information support for the calculation of the income level and investment decision-making of new energy projects throughout their life cycle.

[0048] The present invention proposes a method for predicting the spot market deviation settlement price through bidding space. Taking the target area power market as the object, the hourly bidding space parameters of the target area are calculated through the hourly characteristic curve of future annual load, the hourly characteristic curve of new energy and the installed capacity of regulating power supply. Within each monthly time range, the median of the bidding space for 24 hours a day is calculated respectively, and sorted from high to low and divided into four spot price periods of peak, peak, flat and valley, thereby calculating the monthly average daily spot price forecast results for 12 months a year. Combined with the medium- and long-term monthly transaction average price forecast data of the long-term power market, the new energy electricity contract curve and the power generation characteristic curve, the long-term spot deviation settlement price forecast results are obtained.

[0049] See also Figure 1 The present invention provides a method for predicting the deviation settlement price of the spot market through the bidding space, comprising the following steps:

[0050] Step S1, based on the characteristic load curve, new energy output characteristic curve and adjustable power installed capacity of the target area planned in the future years, obtain the bidding space j(t) time series at each time t in the future years;

[0051] When considering the settlement price of electricity supply in the spot market, the decisive factor is the relationship between the power supply capacity and the expected load demand at that moment, and the marginal price of marginal units such as thermal power or hydropower is used as the market clearing price. When the output of new energy is small and a large number of regulating power sources are required to provide power support, the bidding space of regulating power sources in the province in the spot market is sufficient, and the marginal price will be correspondingly higher, resulting in a higher market clearing price; on the contrary, when new energy is developed, the output demand of regulating power sources is small, the bidding space in the spot market is compressed, and the marginal price is correspondingly reduced, resulting in a lower spot clearing price. Therefore, this method proposes the bidding space parameters at each moment in the spot market to reflect the high and low capacity of the marginal price in the spot market.

[0052] Specifically, the characteristic load curve of the target area in the next year is c = f (t), the new energy output characteristic curve is p = g (t), and the adjustable power installed capacity is P T , using formula (1), we can get the bidding space j(t) at each time t in the future years:

[0053]

[0054] The time t may be hourly.

[0055] Step S2, taking months as units, analyze the monthly characteristics of the bidding space j(t) time series at each moment t in the future years, and divide the H moments in a day into four spot periods, namely: spot peak period T1, spot peak period T2, spot normal period T3 and spot valley period T4, each spot period corresponds to the intraday time range;

[0056] Step S2 is specifically as follows:

[0057] Step S2.1, divide the future year into 12 months by month; based on the time series of the bidding space j(t) at each moment t in the future year, for each month i, i = 1, 2, ..., 12, extract the bidding space at each moment h of each day d to obtain the bidding space monthly matrix J i ={J i (d,h)}, d=1,2,...,D, h=1,2,...,H, where D is the total number of days in each month i, H is the total number of hours in a day, and when the time is in hours, H is 24, representing 24 hours in a day; J i (d,h) represents the bidding space value of month i at time h on day d;

[0058] Specifically, in order to effectively connect the medium- and long-term electricity market with the spot market and to settle the deviation of the medium- and long-term transaction prices in the spot market transaction prices, this method simplifies the price curve of the spot market for 8760 hours a year into four main spot time periods: peak, peak, flat, and valley. The spot time periods are divided based on the size of the bidding space.

[0059] Divide a year into 12 months by month, extract the bidding space values ​​of 24 hours a day at each moment in each month, and form the bidding space monthly matrix J i ={J i (d,h)}, the formula can be expressed as:

[0060] J i (d,h)=reshape(j(t i-1 :t i ),d,h),i=1,2,…,12

[0061] Where: t i is the hour at the end of month i, t i-1 is the hour at the beginning of month i; j(t i-1 :t i ) represents the sequence range data in month i in the bidding space 8760 sequence; reshape(j(t i-1 :t i ),d,h) function represents the j(t i-1 :t i) sequence is rearranged into a matrix, which is a matrix with D rows and H columns in the present invention.

[0062] Step S2.2: For each month i, the monthly matrix J of the bidding space i ={J i (d,h)}, and obtain the bidding space sequence {J i (1,h),J i (2,h),...,J i (d,h),...,J i (D,h)}, and extract the median of the bidding space from the bidding space sequence as the monthly mean of the bidding space at time h in the day jm i (h), the expression is:

[0063] jm i (h)=median(J i (d,h)),i=1,2,…,12

[0064] The monthly average of the bidding space at each intraday time h is jm i (h) Arrange in chronological order to obtain the monthly mean vector jm of the bidding space i =[jm i (1),jm i (2),...,jm i (h),...,jm i (H)];

[0065] Step S2.3: For each month i, the monthly mean vector jm of the bidding space i =[jm i (1),jm i (2),...,jm i (h),...,jm i The distribution of the monthly averages of each bidding space in (H)] is analyzed, and the mapping relationship between the monthly average level of the bidding space and the level of the spot time period is established, so as to divide the H moments in the day into four spot time periods, namely: spot peak time period T1, spot peak time period T2, spot normal time period T3 and spot valley time period T4, and each spot time period corresponds to the intraday time range.

[0066] Step S2.3 is specifically as follows:

[0067] Step S2.3.1, using formula (2), for each month i, the monthly mean vector jm of the bidding space i =[jm i (1),jm i (2),...,jm i (h),...,jmi The monthly averages of each bidding space in (H)] are arranged in descending order to obtain the descending vector jmd of the monthly average of the bidding space shown in formula (3): i :

[0068] jmd i =descent[jm i (1),jm i (2),...,jm i (h),...,jm i (H)] (2)

[0069] Among them: descent[] represents the descending sort operation;

[0070] jmd i =[jmd i (h),jmd loc,i (h)],h=1,2,...,H (3)

[0071] Among them: jmd loc,i (h) represents the bid space value jmd after descending order i (h) The corresponding intraday time h.

[0072] Step S2.3.2, dividing the H bidding space monthly averages arranged in descending order into four sets from high to low, namely: the first bidding space monthly average set, the second bidding space monthly average set, the third bidding space monthly average set and the fourth bidding space monthly average set;

[0073] The time range corresponding to the first set of bidding space monthly average sets, the time range corresponding to the second set of bidding space monthly average sets, the time range corresponding to the third set of bidding space monthly average sets and the time range corresponding to the fourth set of bidding space monthly average sets are respectively used as the spot peak period T1, the spot peak period T2, the spot normal period T3 and the spot valley period T4, thereby obtaining four spot periods divided by H time points within the day, and the intraday time range corresponding to each spot period.

[0074] Therefore, when H is 24 intraday hours, as an embodiment, the H descending monthly averages of the bidding space are divided into 4 categories, the 6 hours with the highest bidding space parameter values ​​correspond to the spot peak period T1, the 6 hours with the second highest parameter values ​​correspond to the spot peak period T2, the 6 hours with the third highest parameter values ​​correspond to the spot normal period T3, and the 6 hours with the lowest parameter values ​​correspond to the spot valley period T4.

[0075] Step S3, pre-set the spot price coefficients of the spot peak period T1, the spot peak period T2, the spot normal period T3 and the spot valley period T4, which correspond to the spot price coefficients k1, k2, k3 and k4 respectively, so as to obtain the monthly spot price coefficient matrix k of each month i i =[k i (h)],h=1,2,…,H; where k i (h) represents the spot price coefficient at time h of each month i;

[0076] As an embodiment, k1, k2≥1, k4≤1, k3=1.

[0077] Step S4, determine the monthly medium- and long-term prices of each month i; according to the monthly spot price coefficient matrix k of each month i i =[k i (h)], h=1,2,…,H, considering the relationship with the monthly medium- and long-term prices of various periods of each month i, the monthly spot price is obtained;

[0078] Step S4 is specifically as follows:

[0079] Step S4.1, determine the monthly medium and long-term prices for each month i:

[0080] Step S4.1.1, obtain the monthly medium- and long-term normal period forecast price pz for each month i in the medium- and long-term market i ,i=1,2,…,12, usually agreed upon in the medium and long term markets.

[0081] Step S4.1.2, obtain the monthly mid- to long-term peak trading price pz of each month i in the mid- to long-term market i1 and the monthly medium- and long-term normal time period forecast price pz i The ratio s1, the monthly mid- and long-term peak trading price pz i2 and the monthly medium- and long-term normal time period forecast price pz i The ratio s2 and the monthly mid- and long-term valley trading price pz i3 and the monthly medium- and long-term normal time period forecast price pz i The ratio s3; each ratio is also usually agreed upon by the medium and long-term markets.

[0082] Therefore, according to formula (4), the monthly mid- to long-term peak trading price pz of each month i in the mid- to long-term market is obtained: i1 、Monthly mid- to long-term peak trading price pz i2 and the monthly mid- to long-term valley trading price pz i3 ;

[0083]

[0084] Step S4.1.3, according to the pre-agreed correspondence between the medium- and long-term peak period, the medium- and long-term peak period, the medium- and long-term normal period, the medium- and long-term valley period and the intraday period, the monthly medium- and long-term peak period price pz of each month i in the medium- and long-term market is calculated. i1 , monthly mid- to long-term peak period price pz i2 、Medium and long term normal time period forecast price pz i and the monthly mid- to long-term valley price pz i3 , corresponding to each intraday period, so as to obtain the monthly medium- and long-term price vector r_pz for each month i i =[r_pz i (h)], h = 1, 2, ..., H, where r_pz i (h) represents the medium- and long-term price at time h during the day of month i;

[0085] It should be noted that the intraday time period h corresponding to various medium and long-term time periods is not the same as the intraday time period h corresponding to various spot time periods.

[0086] Step S4.2, determine the monthly spot price for each month i:

[0087] Step S4.2.1, according to formula (5), the spot prices of the spot peak period T1, the spot peak period T2, the spot normal period T3 and the spot valley period T4 are obtained respectively, expressed as: spot peak period spot price px i1 , spot price during peak period px i2 , spot price during normal hours px i3 and spot price px during the spot valley period i4 :

[0088]

[0089] Step S4.2.2, according to the intraday time range corresponding to each spot time period determined in step S2, the spot prices of various spot time periods are mapped to each intraday time period, so as to obtain the monthly spot price vector r_px of each month i i =[r_px i (h)], h=1,2,…,H.

[0090] Step S5, using the spot deviation settlement price prediction model, comprehensively considering the monthly medium- and long-term prices, monthly spot prices, the monthly new energy contract curve of the future medium- and long-term market, and the monthly output characteristic curve of each month i, predicts the monthly spot deviation settlement price.

[0091] Step S5 is specifically as follows:

[0092] Step S5.1: The monthly medium- and long-term prices r_pz at time h on each day of each month i are i (h) is extended to the same time h of each day in month i, so as to obtain the monthly medium- and long-term price series r_pz at each time t of each month i i (1),r_pz i (2),...,r_pz i (N i ),N i is the number of moments t in month i;

[0093] The monthly spot price r_px at time h on each day of month i i (h) is extended to the same time h of each day in month i, thus obtaining the monthly spot price series r_px at each time t of each month i i (1),r_px i (2),...,r_px i (N i );

[0094] Step S5.2, obtain the monthly i new energy contract curve c of the future mid- to long-term market 1,i (N t ),N t =1,2,…,N i and monthly output characteristic curve c 2,i (N t ),N t =1,2,…,N i , using the spot deviation settlement price prediction model of formula (6), we get the monthly spot deviation settlement price pp i :

[0095]

[0096] This step is finished.

[0097] As an embodiment, the present invention provides a method for predicting a spot market deviation settlement price through a bidding space, comprising the following steps:

[0098] Step 1: Based on the planned provincial characteristic load curve, new energy output characteristic curve, and regulating power installed capacity for future years, propose the concept of bidding space and calculate the bidding space to obtain the bidding space parameter values ​​for each hour within a year.

[0099] Step 2: Divide a year into 12 months on a monthly basis, extract the bidding space values ​​at each moment of the 24 hours of the day for each month, and form a bidding space monthly matrix, for a total of 12 bidding space monthly matrices.

[0100] In the bidding space monthly matrix, the median of the bidding space parameters at the same time of each day is taken as the monthly mean of the bidding space at that time, thereby obtaining a bidding space monthly mean vector consisting of 24 hours.

[0101] Step 3: Sort the monthly mean values ​​of each bidding space in the monthly mean vector of the bidding space of each month. Each bidding space monthly mean vector obtains 24 monthly mean values ​​of the bidding space from high to low, which are divided into 4 categories. The 6 hours with the highest monthly mean value of the bidding space correspond to the peak period of the spot market, the 6 hours with the second highest parameter value correspond to the peak period of the spot market, the 6 hours with the third highest parameter value correspond to the normal period of the spot market, and the 6 hours with the lowest parameter value correspond to the valley period of the spot market. The results of the spot market period division for each of the 12 months are obtained and included in the monthly spot price coefficient matrix.

[0102] Step 4: By defining the spot price coefficients of different time periods in the spot market, the spot price in normal time periods is equal to the medium- and long-term normal time prices (k3=1), the spot price in peak and peak time periods is equal to k1 times and k2 times the medium- and long-term peak and peak time prices (k1, k2>1), and the spot price in valley time periods is equal to k4 times the medium- and long-term valley price (k4<1). Combined with the new energy contracting curve and output characteristic curve of the medium- and long-term market, the forecast results of the spot deviation settlement prices in the future years and months are obtained.

[0103] Currently, the more mature spot price forecasting models often have a time period of 1 to 3 days, which can effectively support the real-time operation of the power market and the profit decision-making of the spot market. However, since they do not effectively consider the matching degree between the long-term electricity load demand and the output capacity of new energy, as well as the impact of bidding space factors on the spot market price, they cannot reasonably calculate the long-term spot market electricity price level in the next few months to several years and the deviation settlement profit level in the spot market in the medium and long term. They cannot support the important profit analysis in the process of project investment decision-making during the period of rapid development of new energy.

[0104] The present invention takes the provincial power market as the object, and on the basis of fully considering the spot market price and the real-time supply and demand relationship, proposes the concept of bidding space to model the spot price coefficients at different times, and combines the hourly characteristic curve of the load in the future year, the hourly characteristic curve of the new energy and the installed capacity of the regulating power supply to calculate the bidding space parameters in the future year, and uses this to predict the deviation between the spot price level at different times in the future year and the medium- and long-term predicted transaction price, and obtains the spot deviation settlement price.

[0105] Unlike other methods, the present invention can effectively reflect the different periods of tight supply and demand and loose supply and demand in the spot market by introducing the concept of bidding space, thereby establishing a good and simple spot market period division method, which can be well connected with the medium- and long-term market period division method, and further realize the price calculation of deviation settlement in the spot market in the medium- and long-term market. On the basis of the medium- and long-term market price forecast and calculation, by introducing the spot market deviation settlement price calculation results, a more comprehensive and objective new energy expected electricity price benefit calculation result can be obtained, which provides important support for the calculation of the benefit level and investment decision-making of the whole life cycle of new energy projects.

[0106] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be considered as the scope of protection of the present invention.

Claims

1. A method for predicting the deviation settlement price of the spot market through the bidding space, characterized in that: The following steps are involved: Step S1, based on the characteristic load curve, new energy output characteristic curve and adjustable power installed capacity of the target area planned in the future years, obtain the bidding space j(t) time series at each time t in the future years; Step S2, taking months as units, analyze the monthly characteristics of the bidding space j(t) time series at each moment t in the future years, and divide the H moments in a day into four spot periods, namely: spot peak period T1, spot peak period T2, spot normal period T3 and spot valley period T4, each spot period corresponds to the intraday time range; Step S3, pre-set the spot price coefficients of the spot peak period T1, the spot peak period T2, the spot normal period T3 and the spot valley period T4, which correspond to the spot price coefficients k1, k2, k3 and k4 respectively, so as to obtain the monthly spot price coefficient matrix k of each month i i =[k i (h)],h=1,2,…,H; where k i (h) represents the spot price coefficient at time h of each month i; Step S4, determine the monthly medium- and long-term prices of each month i; according to the monthly spot price coefficient matrix k of each month i i =[k i (h)], h=1,2,…,H, considering the relationship with the monthly medium- and long-term prices of various periods of each month i, the monthly spot price is obtained; Step S5, using the spot deviation settlement price prediction model, comprehensively considering the monthly medium- and long-term prices, monthly spot prices, the monthly new energy contract curve of the future medium- and long-term market, and the monthly output characteristic curve of each month i, predicts the monthly spot deviation settlement price.

2. A method for predicting spot market deviation settlement price through bidding space according to claim 1, characterized in that: In step S1, the characteristic load curve of the target area in the future is c=f(t), the new energy output characteristic curve is p=g(t), and the adjustable power supply installed capacity is P T , using formula (1), we can get the bidding space j(t) at each time t in the future years:

3. The method for predicting the deviation settlement price of the spot market through the bidding space according to claim 1, characterized in that: Step S2 is specifically as follows: Step S2.1, divide the future year into 12 months by month; based on the time series of the bidding space j(t) at each moment t in the future year, for each month i, i = 1, 2, ..., 12, extract the bidding space at each moment h of each day d to obtain the bidding space monthly matrix J i ={J i (d,h)}, d=1,2,...,D, h=1,2,...,H, where D is the total number of days in each month i, H is the total number of hours in a day, and J i (d,h) represents the bidding space value of month i at time h on day d; Step S2.2: For each month i, the monthly matrix J of the bidding space i ={J i (d,h)}, and obtain the bidding space sequence {J i (1,h),J i (2,h),...,J i (d,h),...,J i (D,h)}, and extract the median of the bidding space from the bidding space sequence as the monthly mean of the bidding space at time h in the day jm i (h), the monthly average of the bidding space at each intraday time h jm i (h) Arrange in chronological order to obtain the monthly mean vector jm of the bidding space i =[jm i (1),jm i (2),...,jm i (h),...,jm i (H)]; Step S2.3: calculate the monthly mean vector jm of the bidding space for each month i i =[jm i (1),jm i (2),...,jm i (h),...,jm i The distribution of the monthly averages of each bidding space in (H)] is analyzed, and the mapping relationship between the monthly average level of the bidding space and the level of the spot time period is established, so as to divide the H moments in the day into four spot time periods, namely: spot peak time period T1, spot peak time period T2, spot normal time period T3 and spot valley time period T4, and each spot time period corresponds to the intraday time range.

4. The method for predicting the deviation settlement price of the spot market through the bidding space according to claim 3, characterized in that: Step S2.3 is specifically as follows: Step S2.3.1, using formula (2), for each month i, the monthly mean vector jm of the bidding space i =[jm i (1),jm i (2),...,jm i (h),...,jm i The monthly averages of each bidding space in (H)] are arranged in descending order to obtain the descending vector jmd of the monthly average of the bidding space shown in formula (3): i : jmd i =descent[jm i (1),jm i (2),...,jm i (h),...,jm i (H)] (2) Among them: descent[] represents the descending sort operation; someone i =[someone i (h),someone loc,i (h)],h=1,2,...,H (3) Among them: jmd loc,i (h) represents the bid space value jmd after descending order i (h) The corresponding intraday time h. Step S2.3.2, dividing the H bidding space monthly averages arranged in descending order into four sets from high to low, namely: the first bidding space monthly average set, the second bidding space monthly average set, the third bidding space monthly average set and the fourth bidding space monthly average set; The time range corresponding to the first set of bidding space monthly average sets, the time range corresponding to the second set of bidding space monthly average sets, the time range corresponding to the third set of bidding space monthly average sets and the time range corresponding to the fourth set of bidding space monthly average sets are respectively used as the spot peak period T1, the spot peak period T2, the spot normal period T3 and the spot valley period T4, thereby obtaining four spot periods divided by H time points within the day, and the intraday time range corresponding to each spot period.

5. The method for predicting the deviation settlement price of the spot market through the bidding space according to claim 1, characterized in that: Step S4 is specifically as follows: Step S4.1, determine the monthly medium and long-term prices for each month i: Step S4.1.1, obtain the monthly medium- and long-term normal time period forecast price pz for each month i in the medium- and long-term market i ,i=1,2,…,12; Step S4.1.2, obtain the monthly mid- to long-term peak trading price pz of each month i in the mid- to long-term market i1 and the monthly medium- and long-term normal time period forecast price pz i The ratio s1, the monthly mid- and long-term peak trading price pz i2 and the monthly medium- and long-term normal time period forecast price pz i The ratio s2 and the monthly mid- and long-term valley trading price pz i3 and the monthly medium- and long-term normal time period forecast price pz i The ratio s3; Therefore, according to formula (4), the monthly mid- to long-term peak trading price pz of each month i in the mid- to long-term market is obtained: i1 , monthly mid- to long-term peak trading price pz i2 and the monthly mid- to long-term valley trading price pz i3 ; Step S4.1.3, according to the pre-agreed correspondence between the medium- and long-term peak period, the medium- and long-term peak period, the medium- and long-term normal period, the medium- and long-term valley period and the intraday period, the monthly medium- and long-term peak period price pz of each month i in the medium- and long-term market is calculated. i1 , monthly mid- to long-term peak period price pz i2 、Medium and long term normal time period forecast price pz i and the monthly mid- to long-term valley price pz i3 , corresponding to each intraday period, so as to obtain the monthly medium- and long-term price vector r_pz for each month i i =[r_pz i (h)], h = 1, 2, ..., H, where r_pz i (h) represents the medium- and long-term price at time h during the day of month i; Step S4.2, determine the monthly spot price for each month i: Step S4.2.1, according to formula (5), the spot prices of the spot peak period T1, the spot peak period T2, the spot normal period T3 and the spot valley period T4 are obtained respectively, expressed as: spot peak period spot price px i1 , spot price during peak period px i2 , spot price during normal hours px i3 and spot price px during the spot valley period i4 : Step S4.2.2, according to the intraday time range corresponding to each spot time period determined in step S2, the spot prices of various spot time periods are mapped to each intraday time period, so as to obtain the monthly spot price vector r_px of each month i i =[r_px i (h)], h=1,2,…,H.

6. A method for predicting spot market deviation settlement price through bidding space according to claim 5, characterized in that: Step S5 is specifically as follows: Step S5.1: The monthly medium- and long-term prices r_pz at time h on each day of each month i are i (h) is extended to the same time h of each day in month i, so as to obtain the monthly medium- and long-term price series r_pz at each time t of each month i i (1),r_pz i (2),...,r_pz i (N i ),N i is the number of moments t in month i; The monthly spot price r_px at time h on each day of month i i (h) is extended to the same time h of each day in month i, so as to obtain the monthly spot price series r_px at each time t of each month i i (1),r_px i (2),...,r_px i (N i ); Step S5.2, obtain the monthly i new energy contract curve c of the future mid- to long-term market 1,i (N t ),N t =1,2,…,N i and monthly output characteristic curve c 2,i (N t ),N t =1,2,…,N i , using the spot deviation settlement price prediction model of formula (6), we get the monthly spot deviation settlement price pp i : This step is finished.

Images