Price prediction method and system based on interval search

Abstract

The invention provides a price prediction method and system based on interval search, and the method comprises the steps: carrying out the linear interpolation of an output unit price node of each unit, and generating a refined output unit price node set; according to the total electric quantity demand, the total price average growth rate and the output unit price node set, a unit generating capacity interval serial number is adjusted, a generating capacity interval assignment scheme is constructed, and the total electric quantity demand falls into a total generating capacity interval corresponding to the generating capacity interval assignment scheme; evaluating a generating capacity interval assignment scheme by using a price prediction quadratic programming model, and calculating a unit initial output prediction scheme and a unit initial prediction electricity price when a preset constraint condition is satisfied; and interval search and interval adjustment are carried out based on the relationship between the initial output prediction scheme of the unit and the corresponding interval endpoints, so that the initial predicted electricity price of the unit is iteratively optimized, the final predicted electricity price of the unit is obtained, and efficient and accurate price prediction under the complex constraint condition of the electric power spot market is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electricity spot market, and particularly relates to a price prediction method and system based on interval search. BACKGROUND

[0002] In the environment of electricity spot market, the operation of a power selling company needs to comprehensively consider various complex factors such as electricity quantity, weather, holidays, unit operation conditions, etc., and the price prediction of the electricity spot market is a key basis for the power selling company to formulate a reasonable trading strategy to realize market arbitrage.

[0003] However, the price prediction problem of the electricity spot market essentially belongs to a non-convex nonlinear mathematical programming problem, and the solving process is difficult. At present, the traditional electricity market price prediction method often depends on fixed costs and pre-set price strategies, which makes it difficult to guarantee the optimality of the prediction result and cannot meet the demand of the power selling company for accurate and optimal price prediction in market competition.

[0004] Therefore, how to efficiently and accurately predict the price of the electricity spot market has become a technical problem to be solved in the current operation of the electricity spot market. SUMMARY

[0005] The present application provides a price prediction method and system based on interval search to solve the defects of low accuracy and low efficiency of the price prediction of the electricity spot market in the prior art.

[0006] In one aspect, the present application provides a price prediction method based on interval search, comprising:

[0007] linearly interpolating the output unit price nodes of each unit to generate a refined set of output unit price nodes;

[0008] adjusting the unit generation quantity interval serial number according to the total electricity quantity demand, the total price average growth rate and the set of output unit price nodes, constructing a generation quantity interval assignment scheme, so that the total electricity quantity demand falls within the total generation quantity interval corresponding to the generation quantity interval assignment scheme;

[0009] using a price prediction quadratic programming model to evaluate the generation quantity interval assignment scheme, calculating the initial unit output prediction scheme and the initial predicted price of the unit when the preset constraint condition is satisfied;

[0010] based on the relationship between the initial unit output prediction scheme and the corresponding interval endpoints, performing interval search to adjust the interval, so as to iteratively optimize the initial predicted price of the unit, and obtain the final predicted price of the unit.

[0011] According to the price prediction method based on interval search provided by the application, the output unit price nodes of each unit are linearly interpolated to generate a refined output unit price node set, which comprises:

[0012] Based on the preset total price granularity parameter, the number of interval refinements between adjacent output unit price nodes is calculated, and an equal point is inserted between adjacent initial output unit price nodes to generate a refined output unit price node set.

[0013] The output unit price node comprises an initial power generation node and an initial unit power price node, and the absolute value of the difference between the product of the values of two adjacent initial power generation nodes in the output unit price node set and the values of the respective initial unit power price nodes is less than or equal to the total price granularity parameter.

[0014] According to the price prediction method based on interval search provided by the application, the unit power generation interval sequence number is adjusted according to the total power demand, the total price average growth rate and the output unit price node set to construct a power generation interval assignment scheme, which comprises:

[0015] Based on the power generation interval sequence number corresponding to the output unit price node set, an initial power generation interval assignment scheme is set, and the unit power generation interval sequence number is adjusted based on the initial power generation interval assignment scheme until the preset stop adjustment condition is met to obtain the power generation interval assignment scheme.

[0016] The unit power generation interval sequence number is adjusted until the preset stop adjustment condition is met, which comprises:

[0017] The current total power generation interval of the current power generation interval assignment scheme is calculated.

[0018] If the total power demand value is not in the current total power generation interval, the total price average growth rate of each unit from the currently allocated power generation interval to other power generation intervals that can reduce the gap between the current total power generation interval and the total power demand value is estimated.

[0019] The target unit with the minimum total price average growth rate and the new power generation interval corresponding to the target unit are selected, the currently allocated power generation interval of the target unit is updated, the step of calculating the current total power generation interval of the current power generation interval assignment scheme is re-executed until the total power demand value is in the current total power generation interval, and the adjustment is stopped.

[0020] According to the price prediction method based on interval search provided by the application, the estimation process of the total price average growth rate comprises:

[0021] The first product of the value of the lower limit endpoint of the currently allocated power generation interval and the corresponding unit power price is calculated.

[0022] calculating a second product of a value of a lower end point of the other power generation interval and a corresponding unit power price;

[0023] calculating a first difference value of the second product and the first product;

[0024] calculating a second difference value of the value of the lower end point of the other power generation interval and the value of the lower end point of the currently allocated power generation interval;

[0025] calculating a ratio of the first difference value and the second difference value as the total price average growth rate.

[0026] According to the interval search-based price prediction method provided by the application, the preset constraint conditions include total power constraint and power generation interval constraint; wherein the total power constraint requires that the sum of power generation of each unit is equal to the total power demand value, and the power generation interval constraint requires that the power generation of each unit is located in the allocated power generation interval;

[0027] According to the interval search-based price prediction method provided by the application, the objective function of the price prediction quadratic programming model is the average electricity price, which is the sum of the product of the unit power price of each unit and the power generation divided by the total power demand value;

[0028] The price prediction quadratic programming model is used to evaluate the power generation interval assignment scheme, and the unit output prediction scheme and the unit initial prediction price when the preset constraint condition is satisfied are calculated, including:

[0029] The optimal solution obtained by solving the quadratic programming model is taken as the unit output prediction scheme, and the unit initial prediction price is calculated by substituting the unit output prediction scheme into the objective function.

[0030] According to the interval search-based price prediction method provided by the application, the interval search is performed based on the relationship between the unit initial output prediction scheme and the corresponding interval end point, the interval is adjusted, the unit initial prediction price is iteratively optimized, and the unit final prediction price is obtained, including:

[0031] In each iteration process, a new power generation interval assignment scheme is constructed according to the relationship between the unit output prediction scheme at this time and the corresponding interval end point;

[0032] The power generation interval of part of the units in the new power generation interval assignment scheme is adjusted to ensure that the total power demand value is located in the total power generation interval of the power generation interval assignment scheme;

[0033] The unit output prediction new scheme of each unit in the power generation interval determined by the new power generation interval assignment scheme and the corresponding prediction new price are calculated;

[0034] If the predicted new electricity price is less than the unit predicted electricity price, the unit output prediction scheme is updated to a unit output prediction new scheme and the next iteration search is performed until the predicted new electricity price is greater than or equal to the unit predicted electricity price, the iteration is stopped, and the unit final predicted electricity price is obtained.

[0035] According to the price prediction method based on interval search provided by the application, a power generation interval assignment new scheme is constructed according to the relationship between the unit output prediction scheme and the corresponding interval end point, and the method comprises the following steps:

[0036] If the power generation of the unit in the unit output prediction scheme is located at the lower limit end point of the currently distributed power generation interval and the currently distributed power generation interval is not the first power generation interval, the serial number of the currently distributed power generation interval is reduced by one.

[0037] If the power generation of the unit is located at the upper limit end point of the currently distributed power generation interval and the currently distributed power generation interval is not the last power generation interval, the serial number of the currently distributed power generation interval is increased by one, thereby obtaining the power generation interval assignment new scheme.

[0038] According to the price prediction method based on interval search provided by the application, the method further comprises the following steps:

[0039] If the total power demand value is not in the total power generation interval of the power generation interval assignment new scheme, the currently distributed power generation interval of the unit whose power generation interval changes is restored until the total power generation interval of the power generation interval assignment new scheme contains the total power demand value.

[0040] In another aspect, the application further provides a price prediction system based on interval search, which comprises:

[0041] The extraction module is configured to extract features from the multi-modal data of the fan gearbox to obtain a plurality of feature data of the fan gearbox.

[0042] The interpolation module is configured to perform linear interpolation on the output unit price nodes of each unit to generate a refined output unit price node set.

[0043] The construction module is configured to adjust the power generation interval serial number of the unit according to the total power demand, the total price average growth rate and the output unit price node set, and construct a power generation interval assignment scheme so that the total power demand falls within the total power generation interval corresponding to the power generation interval assignment scheme.

[0044] The calculation module is configured to evaluate the power generation interval assignment scheme by using a price prediction quadratic programming model, and calculate a unit initial output prediction scheme and a unit initial predicted electricity price when a preset constraint condition is satisfied.

[0045] An optimization module is configured to perform interval search based on the relationship between the initial output prediction scheme of the generating unit and the corresponding interval end point, adjust the interval, and iteratively optimize the initial prediction price of the generating unit to obtain the final prediction price of the generating unit.

[0046] In another aspect, the present application also provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the interval search based price prediction method according to any one of the above aspects when executing the program.

[0047] In another aspect, the present application also provides a non-transitory computer readable storage medium having a computer program stored thereon, wherein the computer program is executable on a processor to implement the interval search based price prediction method according to any one of the above aspects.

[0048] In another aspect, the present application also provides a computer program product comprising a computer program, wherein the computer program is executable on a processor to implement the interval search based price prediction method according to any one of the above aspects.

[0049] The interval search based price prediction method and system provided by the present application can generate a refined output unit price node set by performing linear interpolation on the output unit price nodes of each generating unit, adjust the generating unit power generation interval serial number according to the total power demand, the average price growth rate, and the output unit price node set, construct a power generation interval assignment scheme to make the total power demand fall within the total power generation interval corresponding to the power generation interval assignment scheme, and evaluate the power generation interval assignment scheme by using a price prediction quadratic programming model to calculate the initial output prediction scheme of the generating unit and the initial prediction price of the generating unit under the condition of meeting the preset constraint condition. Then, the interval search is performed based on the relationship between the initial output prediction scheme of the generating unit and the corresponding interval end point, and the interval is adjusted to iteratively optimize the initial prediction price of the generating unit to obtain the final prediction price of the generating unit. In this way, the average market optimal configuration price can be predicted under the premise of meeting the power demand constraint without relying on specific artificial experience, and the effect is obviously better than that of the traditional price prediction method based on artificial experience or simple rule strategy. At the same time, the method has high efficiency and can realize real-time prediction of the market optimal configuration price, thereby adapting to the dynamic changes of the market. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort.

[0051] Figure 1is a flowchart of a price prediction method based on interval search provided by an embodiment of the present application.

[0052] Figure 2 is a schematic diagram of the principle of equal-interval linear interpolation refinement between adjacent initial output unit price nodes.

[0053] Figure 3 is a schematic diagram of a method example for estimating the average growth rate of the total price.

[0054] Figure 4 is a structural schematic diagram of a price prediction system based on interval search provided by an embodiment of the present application.

[0055] Figure 5 is a structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0056] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below with reference to the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0057] In the prior art, the price prediction method of the electricity spot market depends on the fixed cost model and the preset price strategy, and it is difficult to adapt to the dynamic changes of market supply and demand, which can easily lead to the deviation of the prediction result from the actual optimal configuration.

[0058] In order to solve the above problems, it is found through research that the key lies in how to quickly determine the reasonable interval range of the unit power generation, and on this basis, to carry out efficient optimization. By analyzing the nonlinear relationship between unit output and price, it is proposed to discretize continuous variables into interval combinations, to improve the model precision by using linear interpolation, and to reduce the computational complexity by using interval search strategy. The core of this idea is to build an adjustable power generation interval assignment scheme, and to use a quadratic programming model for local optimization, and finally to approach the global optimal solution through iterative search.

[0059] Therefore, based on the above idea, the present application provides the following technical solutions:

[0060] Figure 1 is a flowchart of a price prediction method based on interval search provided by an embodiment of the present application.

[0061] As shown in Figure 1 , the execution subject of the price prediction method based on interval search provided by an embodiment of the present application can be an electronic device, and the method mainly includes the following steps:

[0062] 101. Linearly interpolating each unit's output price node to generate a refined output price node set;

[0063] 102. Adjusting the unit generation interval sequence according to the total electricity demand, the total price average growth rate, and the output price node set, constructing a generation interval assignment scheme to make the total electricity demand fall within the total generation interval corresponding to the generation interval assignment scheme;

[0064] 103. Evaluating the generation interval assignment scheme using a price prediction quadratic programming model to calculate the unit initial output prediction scheme and the unit initial predicted price when the preset constraint condition is met;

[0065] 104. Based on the relationship between the unit initial output prediction scheme and the corresponding interval endpoint, performing interval search to adjust the interval for iterative optimization of the unit initial predicted price to obtain the unit final predicted price.

[0066] In a specific implementation process, linear interpolation refers to converting coarse-grained output price nodes into a fine-grained set by inserting equidistant points. Specifically, a preset total price granularity parameter can be used to control the interpolation density to ensure that the price change between adjacent nodes is controllable. Generation interval assignment scheme construction refers to dynamically adjusting the unit's corresponding generation interval to make the total electricity demand fall within the total generation interval corresponding to the scheme. Specifically, the total price average growth rate can be used as a basis for selecting the adjustment direction. The price prediction quadratic programming model refers to a mathematical programming model that takes the average electricity price as the optimization target and the total electricity constraint and the generation interval constraint as the conditions. Specifically, a standard quadratic programming algorithm can be used to solve the optimal generation distribution to obtain the unit initial output prediction scheme and further obtain the unit initial predicted price. Interval search iterative optimization refers to adjusting the unit generation interval sequence according to the distribution of the output prediction scheme at the corresponding interval endpoint through iterative search to generate a new generation interval assignment scheme and compare the prices until the new generation interval assignment scheme obtains a predicted price of the unit that is greater than or equal to the previous predicted price of the unit, and the previous predicted price of the unit is taken as the final predicted price of the unit.

[0067] Specifically, first, the original unit output price curve is linearly interpolated and refined to form a piecewise linear model containing more nodes. Based on the total electricity demand, the influence of adjusting the generation interval of each unit on the total price is calculated to gradually construct a feasible interval combination containing the target electricity. Under this interval constraint, a quadratic programming model is established to solve the optimal generation distribution scheme and the corresponding price that satisfies the electricity balance. When the unit output in the prediction scheme is at the interval endpoint, the interval boundary adjustment mechanism is triggered to generate a new scheme by expanding the adjacent interval and recalculate the price, and the cycle is iterated until the stable optimal solution is obtained.

[0068] Compared with the prior art, the application effectively reduces the calculation complexity by discretizing the continuous optimization problem into a search problem of finite interval combination. The traditional method needs to perform global search in the entire continuous space, while the application uses the strategy of combining interval assignment and local optimization to significantly improve the calculation efficiency under the premise of ensuring the quality of the solution. At the same time, the interval adjustment mechanism based on the average price growth rate can quickly locate the optimization direction that is conducive to reducing the overall electricity price, and avoid the waste of computing resources caused by blind search.

[0069] Through the above technical scheme, the application can quickly generate an optimal price prediction result that meets the total power demand in the electricity spot market environment. The method effectively balances the model precision and calculation efficiency by reasonably dividing the power generation interval, uses the quadratic programming model to ensure the feasibility of the local optimal solution, and combines the interval search strategy to continuously improve the global optimization effect, thereby providing real-time and accurate price prediction support for the power selling company.

[0070] In some embodiments, the application further proposes a method of linearly interpolating the output unit price nodes of each unit to generate a refined output unit price node set, which includes calculating the interval refinement number between adjacent output unit price nodes based on a preset total price granularity parameter, and inserting an equal point between adjacent initial output unit price nodes to generate a refined output unit price node set, wherein the output unit price node includes an initial power generation node and an initial unit power price node, and the absolute value of the difference between the product of the values of the adjacent two initial power generation nodes in the output unit price node set and the values of the respective initial unit power price nodes is less than or equal to the total price granularity parameter.

[0071] The total price granularity parameter refers to a threshold parameter for controlling the price difference between adjacent output unit price nodes, which can be realized by using a pre-set numerical value to balance the relationship between calculation accuracy and calculation efficiency. The interval refinement number refers to the number of equal points that need to be inserted between adjacent nodes, which can be realized by taking the integer greater than or equal to the ratio of the total price granularity parameter to the price difference between adjacent nodes, to ensure that the price change gradient between adjacent nodes is controllable. The equal point insertion refers to uniformly distributing new nodes between adjacent initial output unit price nodes, which can be realized by using a linear interpolation algorithm to calculate the power generation value and the corresponding unit power price value of the new nodes, so that the price change between the nodes presents a linear characteristic.

[0072] Specifically, in the implementation process, first, the original output unit price node data of the unit is obtained, including discrete power generation nodes and their corresponding unit power prices. The interval number that needs to be refined for each pair of adjacent nodes is calculated by the total price granularity parameter. This process ensures that the absolute value of the product difference of the power generation and price of any adjacent node does not exceed the total price granularity parameter, thereby maintaining the continuity of the price curve while avoiding excessive refinement.

[0073] Specifically, the number of units involved is denoted as m; each unit has a list of initial power generation nodes and a corresponding series of initial unit power price nodes, the values of the nodes are denoted as and where i is the unit number, is the number of unit price nodes of the ith unit. When the power generation of the ith unit is the value of the jth initial power generation node , it has the value of the initial unit power price node The values of the initial power generation nodes and the values of the initial unit power price nodes are monotonically increasing, and the specific formula is formula (1):

[0074]

[0075] In this embodiment, each unit output unit price node can be refined by linear interpolation according to the total price granularity parameter, where the total price granularity parameter is denoted as Δ, and the number of refined unit price nodes of the ith unit becomes n i , and the power generation node becomes The unit power price node becomes After linear interpolation, the absolute value of the difference between the product of the values of the two adjacent power generation nodes of the same unit and the corresponding unit power price nodes should be less than or equal to the total price granularity parameter, and the specific formula is formula (2):

[0076] |a i(j+1) c i(j+1) -a ij c ij |≤Δ (2)

[0077] where a i(j+1) represents the value of the j+1th power generation node, and c i(j+1) represents the value of the j+1th unit power price node.

[0078] Exemplarily, Figure 2 is a schematic diagram of the principle of equal interval linear interpolation refinement between adjacent initial output unit price nodes, as Figure 2 shown, the interval refinement number can be calculated from the value of each group of adjacent initial power generation nodes and the corresponding initial unit power price node, where the interval refinement number corresponding to the value of the jth and j+1th initial power generation nodes of the ith unit is denoted as k ij , and the specific formula is formula (3):

[0079]

[0080] The linear interpolation refinement can include interval refinement number equidivided points of values of each group of adjacent initial power generation nodes and unit power price nodes linearly interpolated at the equidivided points, wherein the values of the power generation nodes and the unit power price nodes are obtained by formula (4):

[0081]

[0082] Wherein, p represents the interval refinement number. Figure 2 p1 represents p = 1, p2 represents p = 2, and p(kij-1) represents p = k ij -1. The value corresponding to the horizontal coordinate of p1 is The value corresponding to the vertical coordinate is The value corresponding to the horizontal coordinate of p2 is The value corresponding to the vertical coordinate is The value corresponding to the horizontal coordinate of p(kij-1) is The value corresponding to the vertical coordinate is

[0083] Compared with the prior art, the traditional method using fixed interval interpolation can cause part of the section price gradient to suddenly change or the node to be redundant, and the interval refinement number is dynamically calculated in the present application, so that the interpolation density can be adaptively adjusted according to the actual price change amplitude of different sections. For example, the number of interpolation points is reduced in a section with a flat price change, and the number of interpolation points is increased in a section with a steep price change. This differential processing not only avoids the waste of computing resources caused by uniform interpolation, but also eliminates the linear approximation error of the key section, so that the price prediction model can more accurately reflect the real cost characteristics of the unit.

[0084] Through the above technical solution, the present application effectively solves the contradiction between precision and efficiency in the traditional linear interpolation method in power price prediction. The total price granularity parameter controls the upper limit of the economic cost difference of adjacent nodes, minimizes the number of calculation nodes on the premise of ensuring the smoothness of the price curve, so that the quadratic programming model can obtain high-precision input data when solving, and avoid the dimension disaster caused by too many nodes. This dynamic interpolation mechanism is particularly suitable for units with nonlinear cost characteristics, and can provide basic data support for subsequent optimization calculation with consideration of accuracy and feasibility.

[0085] In some embodiments, the application further proposes to construct a power generation interval assignment scheme according to total power demand, total price average growth rate and power output unit price node set, which includes setting an initial power generation interval assignment scheme based on the corresponding power generation interval sequence number of the power output unit price node set, and adjusting the power generation interval sequence number of the unit on the basis of the initial scheme until the preset stop adjustment condition is met, to obtain the power generation interval assignment scheme. The adjustment process includes: calculating the current total power generation interval, if the total power demand value is not in the current total power generation interval, estimating the total price average growth rate of each unit after adjusting to other power generation intervals, selecting the target unit with the minimum growth rate and its new interval for updating until the total power demand falls into the current total power generation interval. The other power generation interval is the interval that can reduce the gap between the current total power generation interval and the total power demand value.

[0086] Wherein, the total price average growth rate refers to the price growth rate per unit of power generation after adjusting the power generation interval of the unit, which can be realized by calculating the ratio of the product difference of the lower limit endpoint of power generation before and after adjustment and the corresponding price to the power generation difference. This parameter is used to measure the economic difference of different adjustment schemes. The preset stop adjustment condition refers to the state judgment standard that the total power demand value falls into the current total power generation interval, which can be realized by interval endpoint comparison. This condition ensures the feasibility of the constructed assignment scheme.

[0087] Specifically, the initial power generation interval assignment scheme can be set as all units selecting the first power generation interval, and then the upper and lower limits of the total power generation interval corresponding to the current scheme are calculated. When the total power demand does not fall into this interval, the total price average growth rate of each unit after adjusting its currently allocated power generation interval to the adjacent power generation interval is calculated. For example, the lower limit of the current interval of a unit is 100 MW, corresponding to a price of 50 yuan, and the lower limit of the adjusted interval is 150 MW, corresponding to a price of 55 yuan, then the growth rate is (55*150-50*100) / (150-100)=65 yuan / MW. After obtaining the total price average growth rate of multiple units after adjusting to other power generation intervals, the adjustment scheme with the minimum growth rate among all units can be selected, the interval of the unit is updated, and the total power generation interval is recalculated, and the cycle is executed until the total power demand is included.

[0088] Compared with the prior art, the traditional method uses fixed interval division or manual experience adjustment, which is easy to cause the absence of feasible solution or low calculation efficiency. The present application adjusts the interval sequence number dynamically and selects the optimal adjustment path based on the economic index, which not only ensures the existence of solution, but also significantly reduces the calculation amount through gradient descent iteration.

[0089] By the technical scheme, the application effectively solves the problem of difficult initial solution construction in the traditional price prediction method, quickly generates a feasible solution meeting the power supply and demand balance through a systematic interval adjustment mechanism, provides a high-quality initial solution basis for subsequent price optimization, and avoids subjective bias caused by manual intervention.

[0090] In some embodiments, the application further proposes an estimation process of the total price average growth rate, which includes: calculating a first product of a value of a lower limit endpoint of a currently assigned power generation interval and a corresponding unit power price; calculating a second product of a value of a lower limit endpoint of another power generation interval and a corresponding unit power price; calculating a first difference value of the second product and the first product; calculating a second difference value of the value of the lower limit endpoint of the another power generation interval and the value of the lower limit endpoint of the currently assigned power generation interval; and calculating a ratio of the first difference value and the second difference value as the total price average growth rate.

[0091] The total price average growth rate refers to a cost change rate corresponding to an increase in unit power generation when the power generation interval is adjusted, and can be specifically realized by calculating a ratio of a cost increment at an adjacent interval endpoint and a power generation increment. The index is used to measure the economic difference of different interval adjustment schemes and provide a quantitative basis for selecting an optimal adjustment path. The first product and the second product respectively represent the benchmark cost of the current interval and the candidate interval, and the cost change trend can be accurately reflected through difference calculation. The ratio of the first difference value and the second difference value forms a standardized index, which eliminates the influence of different unit power generation bases.

[0092] Specifically, in the process of constructing the power generation interval assignment scheme, when the unit power generation interval needs to be adjusted to meet the total power demand, the economic efficiency of different adjustment schemes needs to be evaluated. Taking a unit currently assigned power generation interval lower limit of 100 MW and a corresponding price of 50 yuan / MWh as an example, if the lower limit of the candidate adjustment interval is 150 MW and the corresponding price is 55 yuan / MWh, the first product is 100*50=5000 yuan, the second product is 150*55=8250 yuan. The first difference value is 8250-5000=3250 yuan, the second difference value is 150-100=50 MW, and the total price average growth rate is 3250 / 50=65 yuan / MW. By comparing the total price average growth rates of different units in different adjustment directions, the adjustment scheme with the smallest growth rate can be preferentially selected to achieve the goal of meeting the total power demand with the smallest cost increase.

[0093] Compared with the prior art, the traditional method is often based on experience judgment or simple sorting when adjusting the power generation interval, and lacks quantitative evaluation index. The method can convert the economic evaluation into a comparable numerical index by establishing an accurate growth rate calculation model, so that the interval adjustment decision has a clear mathematical basis. The quantitative evaluation mechanism effectively avoids the deviation caused by subjective judgment, and improves the adjustment efficiency through the standardized calculation process.

[0094] Through the above technical solutions, the economic difference of the adjustment scheme of each unit in different power generation intervals can be accurately identified, and the adjustment path with the lowest cost increase rate is preferentially selected under the premise of meeting the total power demand constraint. The decision mechanism based on quantitative indicators significantly improves the economic rationality of the interval assignment scheme, and reduces manual intervention through the standardized calculation process, enhancing the applicability and stability of the algorithm in complex scenarios.

[0095] In a specific implementation process, the power generation interval assignment scheme is a non-negative integer sequence, denoted as J = [J(1), J(2),..., J(m)], where the value of the ith integer represents the power generation interval sequence number of the ith unit, denoted as J(i), and the jth power generation interval of the ith unit is the closed interval [a ij ,a i(j+1) ].

[0096] In a specific implementation process, first, the power generation interval assignment scheme is initialized, which can be [1, 1,..., 1], indicating that the assigned power generation interval of each unit is the first interval, i.e., based on the power generation interval sequence number corresponding to the output unit price node set, the initial power generation interval assignment scheme is set; the total power generation interval of the current power generation interval assignment scheme J is calculated, denoted as [l J ,u J ], and the specific formula is formula (5):

[0097]

[0098] If the total power demand value is not in the total power generation interval [l J ,u J ] of the current power generation interval assignment scheme J, the total price average growth rate of the current allocated power generation interval of the ith unit to other power generation intervals that can reduce the gap between the current total power generation interval and the total power demand value is estimated, denoted as r i J(i)→j Then, the minimum one is selected from all r i J(i)→j , denoted as And the in the current power generation interval assignment scheme is updated to The process is repeated until the total electricity demand value is within the total generation interval of the current generation interval assignment scheme, and the adjustment is stopped.

[0099] Exemplary, Figure 3 is an example schematic diagram of a method for estimating the average growth rate of total price, as Figure 3 constructed, the total price average growth rate of the generation interval J(i) currently allocated to the i th unit to the generation interval j can be calculated from the left end point of the two intervals and the corresponding unit electricity price, the specific formula is formula (6):

[0100] r i J(i)→j = (c ij a ij -c iJ(i) a iJ(i) ) / (a ij -a iJ(i) ) (6)

[0101] Where a iJ(i) is the lower limit of the i th unit in the J(i) generation interval, c iJ(i) is the corresponding unit electricity price of a iJ(i) , a ij is the lower limit of the i th unit in the j th generation interval, c ij is the corresponding unit electricity price of a ij .

[0102] In some embodiments, the present application further proposes that the preset constraints include total electricity constraints and generation interval constraints, the total electricity constraints require the sum of the generation of each unit to be equal to the total electricity demand value, and the generation interval constraints require the generation of each unit to be located in the respective allocated generation interval.

[0103] Wherein, the total electricity constraint refers to the sum of the generation of each unit must be strictly equal to the total electricity demand value required by the system, which can be realized by setting an equality constraint in the quadratic programming model, which ensures the balance between supply and demand, and avoids insufficient or excessive generation. The generation interval constraint refers to the generation of each unit must be within the range of the generation interval assigned to it, which can be realized by setting upper and lower inequality constraints, which ensures that the unit operates within its feasible operating interval and meets the actual operating conditions.

[0104] Specifically, in constructing the price prediction quadratic programming model, the total power constraint is summed up by mathematical expression to establish an equation relationship between the power generation variables of each unit and the total demand value, ensuring that the optimization result meets the overall demand of the power system. The power generation interval constraint limits the unit output within the pre-divided interval range by setting the lower and upper limit parameters of the power generation of each unit. For example, when a unit is assigned to the second power generation interval, its power generation variable will be constrained between the minimum and maximum values corresponding to the interval. By applying these two types of constraints simultaneously, the optimization model can find the optimal power generation scheme that meets the actual operating capacity of each unit while meeting the demand of the power system.

[0105] Compared with the prior art, the traditional price prediction method often only considers the total power balance and ignores the unit operating interval constraint, or uses a simplified linear model to handle the unit constraint, resulting in a predicted result deviating from the actual operating condition. Although some methods in the prior art consider the unit constraint, they do not optimize the interval constraint and the total power constraint simultaneously, which can easily lead to a suboptimal solution.

[0106] Through the above technical solutions, the present application effectively solves the problem of collaborative optimization of unit operating constraints and system demand balance in power spot market price prediction, ensures that the predicted result meets both the total power demand and the actual operating capacity of each unit, and improves the practicality and operability of the price prediction result, providing a decision basis closer to the actual market conditions for power selling companies.

[0107] In some embodiments, the present application further proposes that the objective function of the price prediction quadratic programming model is the average electricity price, which is the sum of the product of the unit electricity price of each unit and the power generation divided by the total power demand value; the price prediction quadratic programming model is used to evaluate the power generation interval assignment scheme, calculate the predicted power generation scheme of the power generation unit and the initial predicted electricity price of the power generation unit under the preset constraint condition, including: obtaining the optimal solution as the predicted power generation scheme of the power generation unit by solving the quadratic programming model, and substituting the predicted power generation scheme of the power generation unit into the objective function to calculate the initial predicted electricity price of the power generation unit.

[0108] The average electricity price is the ratio of the total generation cost of each unit to the total power generation, which can be realized by summing the product of the unit electricity price of each unit and the power generation and then dividing by the total power demand value. This definition can accurately reflect the comprehensive cost level under market supply and demand balance, providing a quantitative index for the optimization target. The quadratic programming model is a mathematical programming model with the average electricity price as the optimization target and containing linear constraint conditions, which can be solved by the interior point method or the efficient set method. The model finds the optimal power generation scheme under the conditions of meeting the power supply and demand balance and the unit operating constraints through mathematical optimization methods.

[0109] Specifically, after constructing the generation interval assignment scheme, a quadratic programming model containing total power constraint and generation interval constraint is established. The total power constraint requires the sum of generation of each unit to be strictly equal to the total demand value, ensuring supply-demand balance; the generation interval constraint limits the output of each unit within the range of its assigned interval, consistent with the unit operation characteristics. The objective function is set to minimize the average electricity price, and the optimal generation distribution scheme that satisfies all constraints is calculated by the quadratic programming solver. Substituting the solution into the objective function expression, the corresponding initial predicted electricity price can be calculated, providing a benchmark value for subsequent iterative optimization.

[0110] Compared with the prior art, the traditional power price prediction method mostly uses linear programming or fixed cost allocation mode, which is difficult to accurately reflect the nonlinear characteristics of unit generation cost. The present application can more accurately describe the relationship between unit output and cost through the quadratic programming model combined with interval constraints. The existing technology uses total cost minimization as the objective function, which cannot directly reflect the market value of unit power, while the present application uses average electricity price as the optimization target, which is more in line with the market pricing mechanism of power commodity.

[0111] Through the above technical solutions, the present application can effectively solve the problem that the traditional prediction method cannot balance economy and calculation efficiency. The optimization solution based on the quadratic programming model can quickly obtain the optimal generation scheme under the premise of ensuring power supply-demand balance, providing a high-quality initial solution for subsequent interval search. This method establishes an objective function directly related to the market pricing mechanism, making the prediction result closer to the real trading environment and providing a reliable basis for the transaction strategy of the power selling company.

[0112] In one specific implementation process, the price prediction quadratic programming model contains a set of continuous variables, denoted as x i , x i represents the generation of the i-th unit. The total power constraint is where S represents the total power demand value. The generation interval constraint is a iJ(i) ≤ x i ≤ a i(J(i)+1) . The objective function of the price prediction quadratic programming model is formula (7):

[0113]

[0114] In some embodiments, the application further proposes to search the interval based on the relationship between the initial output prediction scheme of the generating set and the corresponding interval endpoints, adjust the interval, and iteratively optimize the initial predicted electricity price of the generating set to obtain the final predicted electricity price of the generating set, which comprises: in each iteration process, constructing a new power generation interval assignment scheme according to the relationship between the current generating set output prediction scheme and the corresponding interval endpoints; adjusting the power generation interval of part of the generating set in the new power generation interval assignment scheme to ensure that the total power demand value is located in the total power generation interval of the new power generation interval assignment scheme; calculating the new generating set output prediction scheme and the corresponding new predicted electricity price of each generating set power generation in the interval determined by the new power generation interval assignment scheme; if the new predicted electricity price is less than the predicted electricity price of the generating set, updating the generating set output prediction scheme to the new generating set output prediction scheme and performing the next iteration search, until the new predicted electricity price is greater than or equal to the predicted electricity price of the generating set, stopping iteration, and obtaining the final predicted electricity price of the generating set.

[0115] wherein the new power generation interval assignment scheme refers to an interval allocation strategy dynamically adjusted according to the positional relationship between the power generation of each generating set in the current generating set output prediction scheme and the interval endpoints of the allocated power generation interval, which can be specifically implemented by reducing one from the interval serial number of the generating set located at the lower limit of the interval and not in the first interval, or adding one to the interval serial number of the generating set located at the upper limit of the interval and not in the last interval. This feature enables the algorithm to explore better solutions in the direction of price decline, avoiding falling into local optimum.

[0116] wherein the total power demand value located in the total power generation interval refers to adjusting the power generation interval of part of the generating set, so that the sum of the lower limits of all generating set power generation intervals is less than or equal to the total demand, and the sum of the upper limits is greater than or equal to the total demand. This constraint ensures the feasibility of the power generation scheme and avoids the situation that the total demand cannot be met.

[0117] wherein the new predicted electricity price refers to the average electricity price calculated based on the quadratic programming model under the new power generation interval assignment scheme, which is specifically obtained by solving the optimization problem with the objective function being the sum of the product of the power generation of each generating set and the unit price divided by the total demand. This step quantifies the economic differences of different interval allocation schemes, providing a clear comparison benchmark for iterative optimization.

[0118] Specifically, in each iteration process, first, a new interval allocation scheme is generated according to the relationship between the power generation of each unit in the current output prediction scheme and the position of the end point of the allocated interval. For example, if the power generation of the unit in the current unit output prediction scheme is located at the lower limit end point of the allocated power generation interval and the allocated power generation interval is not the first power generation interval, the serial number of the allocated power generation interval is reduced by one; if the power generation of the unit is located at the upper limit end point of the allocated power generation interval and the allocated power generation interval is not the last power generation interval, the serial number of the allocated power generation interval is increased by one, thereby obtaining the new power generation interval assignment scheme.

[0119] That is, when the power generation of a certain unit is located at the lower limit of the current allocated interval and the interval is not the first interval, the serial number of the allocated interval is reduced by one, so that the lower interval cost parameter is used in the next round of optimization; when it is located at the upper limit and is not the last interval, it is increased by one. For example, a certain unit is currently allocated to the third interval and its predicted power generation is equal to the minimum value 200MW of the interval, and the system adjusts it to the second interval; if another unit is allocated to the fifth interval and the predicted power generation reaches the maximum value 500MW of the interval, it is adjusted to the sixth interval. This adjustment method triggers interval switching through boundary conditions and can systematically explore the optimization possibility of adjacent intervals.

[0120] Compared with the prior art, the traditional method usually uses random exploration or global traversal method when adjusting the interval, resulting in low calculation efficiency. The present application establishes the association rule between the output prediction result and the interval boundary, and only triggers adjustment when the prediction value reaches the interval boundary, effectively reducing the invalid interval switching operation. The prior art does not consider the dynamic relationship between the prediction result and the interval boundary, which easily produces a large amount of redundant calculation, and the present application realizes directional adjustment through the boundary triggering mechanism, so that the search process is always carried out in the direction that may produce price optimization.

[0121] Through the above technical scheme, the present application can intelligently adjust the interval allocation based on the boundary state of the unit output prediction result, significantly reduce the invalid calculation steps in the iterative optimization process under the premise of ensuring the total power demand, and establish the causal relationship between the prediction result and the interval adjustment, so that each interval switching has a clear optimization orientation, thereby accelerating the convergence speed and improving the accuracy of price prediction. At the same time, the boundary triggering mechanism avoids the solution space shock problem caused by blind adjustment of the interval in the traditional method, and ensures that the optimization process stably approaches the optimal solution.

[0122] After obtaining the new generation capacity interval assignment scheme, it can be checked whether the new generation capacity interval assignment scheme satisfies that the total power demand is within the total generation capacity interval, and if not, the original interval of part of the units is selectively restored, that is, if the total power demand value is not in the total generation capacity interval of the new generation capacity interval assignment scheme, the assigned generation capacity interval of part of the units whose generation capacity interval has changed is restored until the total generation capacity interval of the new generation capacity interval assignment scheme contains the total power demand value.

[0123] Among them, restoring part of the units whose generation capacity interval has changed means selectively reverting the interval adjustment operation of part of the units when the adjusted interval assignment scheme cannot cover the total power demand. Specifically, a priority sorting mechanism can be used to realize it, for example, the units with higher average growth rate of total price are preferentially restored, or the units are restored in reverse order according to the unit number. This operation ensures that the inclusiveness of the total generation capacity interval is restored while maintaining as many effective adjustments as possible.

[0124] Specifically, after constructing the new generation capacity interval assignment scheme, first, the upper and lower limits of the total generation capacity interval corresponding to the new scheme are calculated. If the total power demand value exceeds the range of the interval, then all the units whose generation capacity interval has been adjusted are traversed, and the unit interval is restored to the state before adjustment in a preset restoration order. The total generation capacity interval is recalculated after each restoration operation until the total power demand value falls within the updated total generation capacity interval. This process restores the feasibility of the scheme by gradually reverting part of the adjustment operation while preserving the effective interval changes.

[0125] In a specific implementation process, when multiple units need to be restored, a binary search method can also be used to determine the minimum number of restorations. For example, half of the adjusted units are first attempted to be restored, and if the total power demand is still not satisfied, the other half of the remaining units are continued to be restored, and the process is repeated until the minimum restoration set that meets the condition is found. This way maximizes the preservation of optimization adjustment results while ensuring the effectiveness of the scheme.

[0126] Compared with the prior art, the traditional method usually directly gives up the current optimization path or performs global reset when the interval adjustment fails, resulting in waste of computing resources and reduction of convergence speed. The present application quickly restores the feasibility on the basis of maintaining part of the effective adjustment through the local restoration mechanism, which avoids invalid iteration and preserves the effective information in the optimization process.

[0127] Through the above technical scheme, the present application effectively solves the problem of invalidation of the total generation capacity interval caused by local adjustment in the interval search process. This technical means ensures that the interval assignment scheme generated in each iteration meets the total power demand constraint, avoids the waste of computing resources caused by solving the quadratic programming model to obtain an invalid scheme, and at the same time, maintains the continuity of the optimization process by minimizing the restoration operation, significantly improves the iteration efficiency and result reliability of the electricity price prediction.

[0128] After determining whether the new scheme satisfies that the total electricity demand is within the total generation interval, the optimal generation distribution under the new scheme can be solved by a quadratic programming model, and the corresponding average electricity price is calculated. If the new electricity price is better than the current optimal value, the scheme is updated and the iteration is continued, otherwise the search is terminated. This mechanism adjusts the interval distribution in a targeted manner, gradually approaching the global optimal solution under the premise of ensuring feasibility.

[0129] Compared with the prior art, the traditional method usually adopts a fixed interval or a single optimization strategy, which is easy to fall into local optimum and cannot dynamically respond to market changes. The present application can break through the limitation of local optimum by establishing an iterative search mechanism to actively adjust the generation interval distribution after each optimization, while avoiding the generation of invalid search paths by constraining the feasibility of the total electricity interval, significantly improving the accuracy and efficiency of price prediction.

[0130] Through the above technical solutions, the present application effectively solves the problems of local optimum and poor dynamic adaptability existing in traditional power market price prediction methods. By constructing a dynamic interval assignment scheme and performing iterative optimization, a better price allocation scheme can be systematically explored under the premise of ensuring power supply and demand balance. This scheme not only improves the global optimality of the prediction result, but also reduces the amount of invalid calculation through the constraint-driven interval adjustment mechanism, achieving a double improvement in calculation efficiency and prediction accuracy, providing reliable support for real-time trading decisions in the electricity market.

[0131] Based on the same overall inventive concept, the present application also protects a price prediction system based on interval search. The price prediction system based on interval search provided by the present application is described below, and the price prediction system based on interval search described below can be mutually corresponding and referred to the price prediction method based on interval search described above.

[0132] Figure 4 is a structural schematic diagram of the price prediction system based on interval search provided by an embodiment of the present application, as Figure 4 shown, the price prediction system based on interval search of the present embodiment includes an interpolation module 41, a construction module 42, a calculation module 43 and an optimization module 44.

[0133] The interpolation module 41 is used for linear interpolation of the output unit price nodes of each unit to generate a refined set of output unit price nodes.

[0134] The construction module 42 is used for adjusting the unit generation interval serial number according to the total electricity demand, the total price average growth rate and the set of output unit price nodes, constructing a generation interval assignment scheme, so that the total electricity demand falls within the total generation interval corresponding to the generation interval assignment scheme.

[0135] The computing module 43 is configured to evaluate the generation capacity interval assignment scheme by using a price prediction quadratic programming model, and calculate a unit initial output prediction scheme and a unit initial predicted electricity price when a preset constraint condition is met.

[0136] The optimization module 44 is configured to perform interval search based on a relationship between the unit initial output prediction scheme and a corresponding interval end point, and adjust the interval to iteratively optimize the unit initial predicted electricity price, so as to obtain a unit final predicted electricity price.

[0137] Figure 5 FIG. 1 is a structural schematic diagram of an electronic device provided by an embodiment of the present application. The electronic device can include a processor 510, a communications interface 520, a memory 530, and a communications bus 540. The processor 510, the communications interface 520, and the memory 530 can communicate with each other through the communications bus 540. The processor 510 can invoke a logical instruction in the memory 530 to execute a price prediction method based on interval search.

[0138] In addition, the logical instruction in the memory 530 described above can be implemented in the form of a software function unit and sold or used as an independent product. When used, the logical instruction can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or partly with respect to the prior art or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0139] On the other hand, the present application also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer readable storage medium. When the computer program is executed by a processor, a computer can execute the price prediction method based on interval search provided by the above-mentioned methods.

[0140] In yet another aspect, the present application also provides a non-transitory computer readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the price prediction method based on interval search provided by the above-mentioned methods is implemented.

[0141] It should be noted that the related information involved in each embodiment of the present application is strictly in accordance with the requirements of laws and regulations, and follows the principles of legality, legitimacy and necessity, and is based on the reasonable purpose of the business scene, and processes the information provided by the user in the process of using the product / service or generated due to the use of the product / service, and authorized by the user.

[0142] The related information processed by the present application will be different due to the specific product / service scene, and the specific scene of the user using the product / service should be used as the reference, which may involve the user's account information, device information or other related information. The present application will treat the related information and its processing with high diligence.

[0143] The present application attaches great importance to the security of related information, and has taken reasonable and feasible security protection measures to protect the related information in accordance with industry standards, to prevent unauthorized access, public disclosure, use, modification, damage or loss of related information.

[0144] The device embodiments described above are only schematic, and the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e. they can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present embodiment scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0145] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be realized by means of software and necessary general hardware platform, and of course it can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of software product, which can be stored in computer readable storage medium such as ROM / RAM, magnetic disk, optical disk, etc., including a plurality of instructions to make a computer device (which can be a personal computer, server, or network device, etc.) execute the method described in each embodiment or some part of the embodiment.

[0146] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of each embodiment of the present application.

Claims

1. A price prediction method based on interval search, characterized by, The method comprises the following steps: linear interpolation is performed on the output unit price nodes of each unit to generate a refined output unit price node set; based on the total electricity demand, the average total price growth rate and the output unit price node set, the unit generation interval sequence number is adjusted to construct a generation interval assignment scheme, so that the total electricity demand falls within the total generation interval corresponding to the generation interval assignment scheme; a price prediction quadratic programming model is used to evaluate the generation interval assignment scheme, and the initial output prediction scheme of the unit and the initial predicted price of the unit are calculated when the preset constraint condition is satisfied; based on the relationship between the initial output prediction scheme of the unit and the interval endpoint, interval search is performed to adjust the interval, so that the initial predicted price of the unit is iteratively optimized to obtain the final predicted price of the unit.

2. The interval search-based price forecasting method according to claim 1, characterized by, The method comprises the following steps: linear interpolation is performed on the output unit price nodes of each unit to generate a refined output unit price node set, comprising: based on the preset total price granularity parameter, the number of interval refinements between adjacent output unit price nodes is calculated, and equally divided points are inserted between adjacent initial output unit price nodes to generate a refined output unit price node set; 3. The interval search-based price forecasting method according to claim 1, characterized by, wherein the output unit price nodes include initial generation capacity nodes and initial unit electricity price nodes, and the absolute value of the difference between the product of the values of two adjacent initial generation capacity nodes and the values of their corresponding initial unit electricity price nodes is less than or equal to the total price granularity parameter. based on the total electricity demand, the average total price growth rate and the output unit price node set, the unit generation interval sequence number is adjusted to construct a generation interval assignment scheme, comprising: based on the generation interval sequence number corresponding to the output unit price node set, an initial generation interval assignment scheme is set, and based on the initial generation interval assignment scheme, the unit generation interval sequence number is adjusted until the preset stop adjustment condition is met to obtain the generation interval assignment scheme; wherein the unit generation interval sequence number is adjusted until the preset stop adjustment condition is met, comprising: calculating the current total generation interval of the current generation interval assignment scheme; if the total electricity demand value is not in the current total generation interval, estimating the total price average growth rate of each unit from the currently allocated generation interval to other generation intervals that can reduce the gap between the current total generation interval and the total electricity demand value; 4. The interval search-based price forecasting method according to claim 3, characterized by, selecting the target unit with the minimum total price average growth rate and the new generation interval corresponding to the target unit, updating the currently allocated generation interval of the target unit, and re-executing the step of calculating the current total generation interval of the current generation interval assignment scheme until the total electricity demand value is in the current total generation interval, and stopping adjustment. The estimation process of the total price average growth rate comprises: calculating the first product of the value of the lower limit endpoint of the currently allocated generation interval and the corresponding unit electricity price; calculating the second product of the value of the lower limit endpoint of the other generation interval and the corresponding unit electricity price; calculating the first difference between the second product and the first product; calculating the second difference between the value of the lower limit endpoint of the other generation interval and the value of the lower limit endpoint of the currently allocated generation interval; calculating a ratio of the first difference value and the second difference value as the total price average growth rate.

5. The interval search-based price forecasting method according to claim 1, characterized by, The preset constraint conditions include a total power constraint and a power generation interval constraint; the total power constraint requires that a sum of power generations of the units is equal to a total power demand value, and the power generation interval constraint requires that the power generation of each unit is located in an individually assigned power generation interval.

6. The interval search-based price forecasting method according to claim 5, characterized by, The objective function of the price prediction quadratic programming model is an average price, which is a sum of products of unit power prices of the units and power generations divided by the total power demand value; The price prediction quadratic programming model is used to evaluate the power generation interval assignment scheme, calculate a unit output prediction scheme and an initial prediction price of the units when the preset constraint conditions are satisfied, and includes the following steps: An optimal solution obtained by solving the quadratic programming model is taken as the unit output prediction scheme, and the unit output prediction scheme is substituted into the objective function to calculate the initial prediction price of the units.

7. The interval search-based price forecasting method according to claim 1, characterized by, Interval searching is performed based on a relationship between the initial unit output prediction scheme and corresponding interval endpoints, the interval is adjusted, and the initial prediction price of the units is iteratively optimized to obtain a final prediction price of the units, and includes the following steps: In each iteration process, a new power generation interval assignment scheme is constructed according to a relationship between a unit output prediction scheme at this time and corresponding interval endpoints; The power generation interval of part of the units in the new power generation interval assignment scheme is adjusted to ensure that the total power demand value is located in a total power generation interval of the power generation interval assignment scheme; A new unit output prediction scheme in the power generation interval determined by the new power generation interval assignment scheme and a corresponding new prediction price are calculated; If the new prediction price is less than the prediction price of the units, the unit output prediction scheme is updated to the new unit output prediction scheme, and the next iteration searching is performed until the new prediction price is greater than or equal to the prediction price of the units, the iteration is stopped, and the final prediction price of the units is obtained.

8. The interval search-based price forecasting method according to claim 7, characterized by, The new power generation interval assignment scheme is constructed according to the relationship between the unit output prediction scheme at this time and the corresponding interval endpoints, and includes the following steps: If the power generation of the unit in the unit output prediction scheme at this time is located at a lower limit endpoint of the currently assigned power generation interval and the currently assigned power generation interval is not the first power generation interval, the serial number of the currently assigned power generation interval is reduced by one; If the power generation of the unit is located at an upper limit endpoint of the currently assigned power generation interval and the currently assigned power generation interval is not the last power generation interval, the serial number of the currently assigned power generation interval is increased by one, and thus the new power generation interval assignment scheme is obtained.

9. The interval search-based price forecasting method according to claim 8, characterized in that, Further including: If the total power demand value is not located in the total power generation interval of the new power generation interval assignment scheme, the currently assigned power generation interval of the unit whose power generation interval is changed is restored until the total power generation interval of the new power generation interval assignment scheme contains the total power demand value.

10. A price prediction system based on interval search, characterized in that, Further including: an interpolation module configured to perform linear interpolation on the output unit price nodes of each unit to generate a refined output unit price node set; A construction module is configured to adjust a generating capacity interval sequence number according to a total electricity demand, an average price growth rate and the set of unit price nodes, and to construct a generating capacity interval assignment scheme so that the total electricity demand falls within a total generating capacity interval corresponding to the generating capacity interval assignment scheme; A calculation module is configured to evaluate the generating capacity interval assignment scheme by using a price prediction quadratic programming model, and to calculate a unit initial output prediction scheme and a unit initial predicted price when a preset constraint condition is satisfied; An optimization module is configured to perform interval search based on a relationship between the unit initial output prediction scheme and interval endpoints, to iteratively optimize the unit initial predicted price based on the interval, and to obtain a unit final predicted price.

Images