A method and device for optimizing unit output based on maximizing the interests of power generators

Abstract

The present invention discloses a method and device for optimizing the output of a generator set based on maximizing the interests of a power generator, the method comprising: constructing a double-layer optimization model based on maximizing the interests of a power generator; converting the double-layer optimization model into a single-layer optimization model based on a value function approximation method; solving the single-layer optimization model under the segmented range constraints of the output of the generator set, the range constraints of the output of the generator set, the segmented range constraints of the load output, the line transmission power range constraints, and the range constraints of the node phase angle, and generating the output of the generator set when the interests of the power generator are maximized; regulating the generator set according to the output of the generator set. The present invention converts the double-layer optimization model into a single-layer optimization model based on the value function approximation method, and obtains the corresponding output of the generator set when the interests of the power generator are maximized by solving the problem, so that the present invention can automatically obtain the output of the generator set, and the total operating cost of the power generator set corresponding to the output of the generator set can be minimized, which is more in line with the needs of analyzing the actual power market.

Description

Technical Field

[0001] The present invention relates to the field of computer processing technology, and in particular to a method and device for optimizing unit output based on maximizing the benefits of a power generator. Background Art

[0002] The pricing mechanism of the electricity spot market is one of the key issues in market design, and it interacts with the trading behavior of power generators. The design of the pricing mechanism needs to take into account the possible trading behavior of power generators, and the bidding strategies of power generators under different pricing mechanisms are different. The existing technology for the calculation and acquisition of power generator bids and power generation unit output under the pricing mechanism is still subject to the constraints of personnel experience and human subjective factors. When calculating the unit output, if there are multiple optimal unit outputs that meet the maximum profit of the power generator, the result of manual calculation cannot guarantee that the final selected unit output can meet the maximum profit of the power generator and reduce the cost, so that the total operating cost of the power unit corresponding to the final determined unit output cannot reach the minimum. Summary of the invention

[0003] The embodiments of the present invention provide a method and device for optimizing unit output based on maximizing the interests of power generators, which can effectively solve the problem in the prior art that the results of manual calculations cannot guarantee that the output of the unit finally selected can meet the maximum interests of the power generator and reduce costs, so that the total operating cost of the power unit corresponding to the finally determined unit output cannot reach the lowest.

[0004] An embodiment of the present invention provides a method for optimizing unit output based on maximizing the interests of power generators, including:

[0005] Obtain data of power generators, market clearing data and power network data, and construct a two-layer optimization model to maximize the interests of power generators; wherein the two-layer optimization model includes segment range constraints of generator set output, range constraints of generator set output, segment range constraints of load output, line transmission power range constraints and range constraints of node phase angles;

[0006] Converting the two-layer optimization model into a single-layer optimization model based on a value function approximation method;

[0007] Under the constraints of the segmented range of generator set output, the range of generator set output, the segmented range of load output, the range of line transmission power and the range of node phase angle, the single-layer optimization model is solved to generate the unit output when the power generator benefits the most;

[0008] The generator set is regulated according to the output of the set.

[0009] Preferably, the two-layer optimization model is specifically:

[0010]

[0011] in, is the revenue of generator j in the day-ahead electricity market, is the revenue of generator j in the real-time electric energy market, represents the income of renewable energy generator j in the green certificate market, represents the income of traditional power producer j in the carbon market, C j is the total unit operating cost of generator j;

[0012] represents the marginal electricity price of node n in the day-ahead electricity market in time period t, T represents a set of multiple time periods t (such as 1-24h), Ω j represents the set of units of generator j; n Represents a collection of nodes. represents the winning bid capacity of generator i in time period t and capacity segment b in the day-ahead electricity market, W represents the set composed of various scenarios ω, π ω represents the probability of scenario ω occurring, represents the marginal electricity price of node n in the real-time electricity market in scenario ω and time period t, represents the winning capacity of generator i in the real-time electric energy market in scenario ω, time period t, and capacity segment b, Indicates the price of the green certificate, represents the number of green certificates obtained by renewable energy generator j under scenario ω, λ CT represents the carbon trading price in the carbon market, σ j represents the carbon quota decomposition factor of power producer j, represents the carbon quota of power producer j, represents the carbon emission factor of generator set i, a i represents the quadratic coefficient of the operating cost of generator set i, b i represents the first-order coefficient of the operating cost of generator set i, c i The constant term representing the operating cost of generator set i, Ω RE represents the set of renewable energy generators, Ω T represents the set of traditional power generators, Used to mark whether power generator j participates in the green certificate market, with a value of 0 or 1;

[0013] The two-level optimization model satisfies the following constraints:

[0014]

[0015] α i,b,trepresents the slope of the bth segment of the bidding curve of generator set i in period t, Indicates the minimum slope of the bth segment of the bidding curve of generator set i, Indicates the maximum slope of the bth segment of the bidding curve of generator set i, β i,t represents the intercept of the quotation curve of generator set i in period t, It represents the maximum intercept value of the quotation curve of generator set i, Indicates the minimum intercept value of the quotation curve of generator set i, Ω j Represents the set of generator sets of generator j.

[0016] Preferably, the converting the double-layer optimization model into a single-layer optimization model based on the value function approximation method comprises:

[0017] Fit the output variables of the lower model in the two-level optimization model to a linear combination of the polynomial basis functions of the input variables of the lower model:

[0018]

[0019] Among them, V represents the social welfare value in the lower model, α represents the slope and intercept of the bid curve of each power producer, is the mapping function, Representation value function The approximate value of represents the set of polynomial basis functions, φ l (α) represents the lth polynomial basis function, d l represents the coefficients of the lth polynomial to be solved.

[0020] Preferably, after fitting the output variables of the lower model in the two-layer optimization model to the linear combination of the polynomial basis functions of the input variables of the lower model, the method further comprises:

[0021] The input variables are transformed into high-dimensional value functions by Convert multiple input variables into low-dimensional value functions

[0022] At the preset point α C The pair value function Expanding it, we get the following formula:

[0023]

[0024] Among them, the high-dimensional variable α={α1,α2,...,α n}, Representation value function For the k-th dimension α of variable α kFind the j-order partial derivative, R represents the expansion term of the next dimension;

[0025] The value function Decompose into multiple input variables into low-dimensional value functions As shown in the following formula:

[0026]

[0027] n=n1r

[0028] in, represents a low-dimensional variable-valued function, express At the expansion point α C The specific value of n1 is α m The dimension of .

[0029] Preferably, the lower model of the two-layer optimization model includes: a day-ahead electricity energy market clearing sub-model;

[0030] When converting the two-layer optimization model into a single-layer optimization model based on the value function approximation method, the original objective function corresponding to the day-ahead electricity energy market clearing sub-model is converted into a first set of inequalities about the value function;

[0031] Among them, the original objective function corresponding to the day-ahead electricity energy market clearing sub-model is:

[0032]

[0033] The first set of inequalities about the value function is:

[0034]

[0035] The original objective function corresponding to the day-ahead electricity energy market clearing sub-model and the first set of inequalities about the value function both satisfy the following constraints:

[0036]

[0037] λ d,k,t represents the quotation of load d in time period t and capacity segment k; α i,b,t represents the slope of the bth segment of the bidding curve of generator set i in period t, is the output of generator i in time period t and capacity segment b in the day-ahead electricity market, γ i,b,t Indicates the ordinate value corresponding to the starting point of the bth segment in the quotation curve of generator set i in time period t; It represents the output of load d in time period t and capacity segment k;

[0038] Represents the set of generator sets at node n; represents the set of load components on node n; represents the set of all nodes connected to node n, B n,m represents the conductance of the line between node n and node m, represents the phase angle of the day-ahead market node n in period t, and represents the upper and lower limits of the generating unit i in the capacity segment b, and represents the upper and lower limits of the output of generator i in time period t in the day-ahead market, and represents the upper and lower limits of load d in time period t and capacity segment k, It represents the upper limit of the line transmission power between node n and node m.

[0039] Preferably, the lower model of the two-layer optimization model includes: a clearing sub-model of the real-time electric energy market;

[0040] When converting the two-layer optimization model into a single-layer optimization model based on the value function approximation method, the original objective function corresponding to the clearing sub-model of the real-time electric energy market is converted into a second set of inequalities about the value function;

[0041] Among them, the original objective function corresponding to the clearing sub-model of the real-time electric energy market is:

[0042]

[0043] The second set of inequalities about the value function is:

[0044]

[0045] The original objective function corresponding to the clearing sub-model of the real-time electric energy market and the second set of inequalities about the value function both satisfy the following constraints:

[0046]

[0047] represents the output of load d in scenario ω, time period t, and capacity segment k, represents the phase angle of real-time power market node n in scenario ω and time period t, and It represents the upper and lower limits of the output of generator i in the real-time electricity market in scenario ω and time period t.

[0048] Preferably, the two-layer optimization model also satisfies the following constraints:

[0049]

[0050] in, Indicates the price of the green certificate, Represents the quantity-price relationship coefficient of green certificates, It indicates the penalty price when the quota obligation subject fails to complete the quota task. It indicates the price that the quota obligor is willing to pay when the supply of green certificates is sufficient. represents the quota ratio required by the renewable energy quota system, Indicates the maximum demand for green certificates.

[0051] Based on the above method embodiments, the present invention provides corresponding device embodiments.

[0052] An embodiment of the present invention provides a unit output optimization device based on maximizing the interests of power producers, including: a double-layer optimization model building module, a model conversion module, a unit output generation module and a generator unit control module;

[0053] The two-layer optimization model building module is used to obtain data of power generators, market clearing data and power network data, and build a two-layer optimization model to maximize the interests of power generators; wherein the two-layer optimization model includes segment range constraints of generator set output, range constraints of generator set output, segment range constraints of load output, line transmission power range constraints and range constraints of node phase angles;

[0054] The model conversion module is used to convert the double-layer optimization model into a single-layer optimization model based on a value function approximation method;

[0055] The unit output generation module is used to solve the single-layer optimization model under the segment range constraints of the generator unit output, the range constraints of the generator unit output, the segment range constraints of the load output, the line transmission power range constraints and the range constraints of the node phase angle, and generate the unit output when the power generator benefits the most;

[0056] The generator set control module is used to control the generator set according to the output of the set.

[0057] The following beneficial effects are achieved by implementing the present invention:

[0058] The embodiment of the present invention provides a method and device for optimizing the output of a unit based on maximizing the interests of a power generator, the method comprising: obtaining data of the power generator, data of market clearing, and data of the power network, and constructing a double-layer optimization model for maximizing the interests of the power generator; converting the double-layer optimization model into a single-layer optimization model based on the value function approximation method; solving the single-layer optimization model under the segmented range constraints of the output of the power generator, the range constraints of the output of the power generator, the segmented range constraints of the load output, the line transmission power range constraints, and the range constraints of the node phase angle, and generating the output of the unit when the power generator's interests are maximized; regulating the power generator according to the output of the unit. The present invention applies the value function approximation method to the solution of the double-layer optimization model. After constructing the double-layer optimization model for maximizing the interests of the power generator, the double-layer optimization model is converted into a single-layer optimization model according to the value function approximation method, and the double-layer model is converted into a single-layer model, so that the single-layer model can be solved and the output of the unit when the power generator's interests are maximized can be generated. The present invention converts a double-layer optimization model into a single-layer optimization model according to the value function approximation method, and can convert the double-layer model into a single-layer model without introducing new integer variables, and obtains the corresponding unit output when the power generator's profit is maximized through the single-layer model solution, so that the present invention can automatically obtain the unit output without manual calculation and screening, and the final unit output can meet the maximum profit of the power generator, and make the total operating cost of the power unit corresponding to the unit output reach the minimum, which is more in line with the needs of analyzing the actual power market. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a flow chart of a method for optimizing unit output based on maximizing the interests of power generators provided by one embodiment of the present invention.

[0060] Figure 2 It is a structural schematic diagram of a unit output optimization device based on maximizing the interests of power generators provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0061] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0062] like Figure 1 FIG. 1 is a flow chart of a method for optimizing unit output based on maximizing the interests of a power generator provided by an embodiment of the present invention. A method for optimizing unit output based on maximizing the interests of a power generator provided by an embodiment of the present invention includes:

[0063] Step S1: Acquire data of power generators, market clearing data and power network data, and construct a two-layer optimization model for maximizing the interests of power generators; wherein the two-layer optimization model includes segment range constraints of generator output, range constraints of generator output, segment range constraints of load output, line transmission power range constraints and range constraints of node phase angles;

[0064] Step S2: converting the double-layer optimization model into a single-layer optimization model based on a value function approximation method;

[0065] Step S3: Under the constraints of the segment range of the generator set output, the range of the generator set output, the segment range of the load output, the line transmission power range and the range of the node phase angle, the single-layer optimization model is solved to generate the unit output when the power generator benefits the most;

[0066] Step S4: regulating the generator set according to the output of the set.

[0067] For step S1, in a preferred embodiment, data of power generators, market clearing data and power network data are obtained to construct a two-layer optimization model for maximizing the interests of power generators; wherein the two-layer optimization model includes segment range constraints of generator set output, range constraints of generator set output, segment range constraints of load output, line transmission power range constraints and range constraints of node phase angles;

[0068] Power generators participate in the competition in the electricity market and need to submit a quotation in the market to maximize their own interests. If the quotation is too high, the number of winning bids will be small, and the interests will not be maximized. If the quotation is too low, it is not conducive to maximizing one's own interests. Therefore, how to determine one's own optimal quotation and unit output under various constraints (power network constraints, the role of the green certificate market, the role of the carbon market, the influence of multi-party games, etc.), the present invention can solve this problem by solving the double-layer optimization model, and provide a reference for the optimal decision of power generators.

[0069] Specifically, the upper model of the two-layer optimization model for maximizing the interests of power generators is that power generators pursue maximum interests, and the maximization of interests of power generators in the upper model is determined by the power generators' income in the day-ahead electricity energy market, the income in the real-time electricity energy market, the income in the green certificate market, the income in the carbon market, and the total unit operation cost of the power generators; the lower model of the two-layer optimization model includes the day-ahead electricity energy market clearing sub-model, the real-time electricity energy market sub-model and the green certificate market clearing sub-model;

[0070] In the two-layer optimization model, the upper layer is for power generator j to pursue maximum profit, that is,

[0071]

[0072] in, is the revenue of generator j in the day-ahead electricity market, is the revenue of generator j in the real-time electric energy market, represents the income of renewable energy generator j in the green certificate market, represents the income of traditional power producer j in the carbon market, C j is the total unit operating cost of generator j;

[0073] The specific calculation process of the profit and power generation cost of generator j in each market is as follows:

[0074]

[0075] in, represents the marginal electricity price of node n in the day-ahead electricity market in time period t, T represents a set of multiple time periods t (such as 1-24h), Ω j represents the set of units of generator j; n Represents a collection of nodes. represents the winning bid capacity of generator i in time period t and capacity segment b in the day-ahead electricity market, W represents the set composed of various scenarios ω, π ω represents the probability of scenario ω occurring, represents the marginal electricity price of node n in the real-time electricity market in scenario ω and time period t, represents the winning capacity of generator i in the real-time electric energy market in scenario ω, time period t, and capacity segment b, Indicates the price of the green certificate, represents the number of green certificates obtained by renewable energy generator j under scenario ω, λ CT represents the carbon trading price in the carbon market, σ j represents the carbon quota decomposition factor of power producer j, represents the carbon quota of power producer j, represents the carbon emission factor of generator set i, a i represents the quadratic coefficient of the operating cost of generator set i, b i represents the first-order coefficient of the operating cost of generator set i, c i The constant term representing the operating cost of generator set i, Ω RE represents the set of renewable energy generators, Ω T represents the set of traditional power generators, Used to mark whether power generator j participates in the green certificate market, with a value of 0 or 1;

[0076] Generator j adopts piecewise linear bidding method in the electric energy market, and the bidding curve must meet the following constraints:

[0077]

[0078] α i,b,t represents the slope of the bth segment of the bidding curve of generator set i in period t, Indicates the minimum slope of the bth segment of the bidding curve of generator set i, Indicates the maximum slope of the bth segment of the bidding curve of generator set i, β i,t represents the intercept of the quotation curve of generator set i in period t, It represents the maximum intercept value of the quotation curve of generator set i, Indicates the minimum intercept value of the quotation curve of generator set i, Ω j Represents the set of generator sets of generator j.

[0079] In the two-layer model of the electricity market, the lower model includes the day-ahead electricity energy market clearing sub-model and the real-time electricity energy market sub-model;

[0080] The objective function of the day-ahead electricity market clearing submodel is to minimize the day-ahead negative total social welfare, that is,

[0081]

[0082] The day-ahead electricity market clearing needs to satisfy constraints (13)-(20):

[0083]

[0084]

[0085] Among them, λ d,k,t represents the quotation of load d in time period t and capacity segment k; α i,b,t represents the slope of the bth segment of the bidding curve of generator set i in period t, is the output of generator i in time period t and capacity segment b in the day-ahead electricity market, γ i,b,t Indicates the ordinate value corresponding to the starting point of the bth segment in the quotation curve of generator set i in time period t; It represents the output of load d in time period t and capacity segment k;

[0086] Represents the set of generator sets at node n; represents the set of load components on node n; represents the set of all nodes connected to node n, B n,m represents the conductance of the line between node n and node m, represents the phase angle of the day-ahead market node n in period t, and represents the upper and lower limits of the generating unit i in the capacity segment b, and represents the upper and lower limits of the output of generator i in time period t in the day-ahead market, and represents the upper and lower limits of load d in time period t and capacity segment k, It represents the upper limit of the line transmission power between node n and node m.

[0087] It should be noted that formula (13) calculates the price corresponding to the starting point of each segment in the quotation curve, formula (14) ensures the power balance of each node, formula (15) limits the output of the generator set to not exceed the upper and lower limits of each segment power, formula (16) ensures that the winning capacity of the generator set does not exceed the predicted upper and lower limits of the output, formula (17) ensures that the load output does not exceed the upper and lower limits of each segment power, formula (18) ensures that the line transmission power does not exceed the limit, formula (19) limits the node phase angle to not exceed the upper and lower limits, and formula (20) selects node 1 as the reference node.

[0088] The two-layer optimization model includes the segmented range constraints of generator set output, the range constraints of generator set output, the segmented range constraints of load output, the line transmission power range constraints and the range constraints of node phase angles. In a preferred embodiment, the specific constraints are the segmented range constraints and range constraints of generator set output in the day-ahead electricity energy market clearing sub-model, as well as the segmented range constraints of load output of the day-ahead electricity energy market clearing sub-model, the line transmission power range constraints of the day-ahead electricity energy market clearing sub-model and the range constraints of node phase angles of the day-ahead electricity energy market clearing sub-model.

[0089] The objective function of clearing the real-time electric energy market is to minimize the real-time negative total social welfare, that is, the original objective function corresponding to the clearing sub-model of the real-time electric energy market is:

[0090]

[0091] Real-time electricity market clearing needs to satisfy constraints (22)-(28):

[0092]

[0093] in, represents the output of load d in scenario ω, time period t, and capacity segment k, represents the phase angle of real-time power market node n in scenario ω and time period t, and It represents the upper and lower limits of the output of generator i in the real-time electricity market in scenario ω and time period t.

[0094] In a preferred embodiment, the two-layer optimization model also includes the clearing of the green certificate market, which is simulated by the Cournot model and described by equations (29)-(31);

[0095] The two-layer optimization model also satisfies the following constraints:

[0096]

[0097]

[0098] in, Indicates the price of the green certificate, Represents the quantity-price relationship coefficient of green certificates, It indicates the penalty price when the quota obligation subject fails to complete the quota task. It indicates the price that the quota obligor is willing to pay when the supply of green certificates is sufficient. represents the quota ratio required by the renewable energy quota system, Indicates the maximum demand for green certificates.

[0099] It should be noted that formula (29) establishes the inverse demand linear function in the Cournot model; formula (30) calculates the slope of the inverse demand linear function, which is determined by the quota ratio and historical data; formula (31) calculates the maximum green certificate demand of electricity users.

[0100] For step S2, in a preferred embodiment, in solving the electricity market quotation-clearing two-layer optimization model (i.e., the two-layer optimization model of the present invention), the traditional method converts the two-layer model into a single-layer model by introducing integer variables, and the number of introduced integer variables is proportional to the number of constraints in the lower model; when the model contains a large number of constraints, the computational burden brought by the integer variables is huge. To solve this problem, the present invention uses a value function approximation method to achieve a rapid solution to the electricity market quotation-clearing two-layer optimization model.

[0101] The value function approximation method is to use the polynomial basis and the least squares method to fit the input variables and output variables of the lower model to describe the response characteristics of the power company's quotation (input variable) in the lower model to the total social welfare value (output variable), and then convert the two-layer model into a single-layer model to avoid the iterative solution of the upper and lower models.

[0102] The present invention fits the output variables of the lower model in the two-layer optimization model to the linear combination of the polynomial basis functions of the input variables of the lower model:

[0103]

[0104] Among them, V represents the social welfare value in the lower model, α represents the slope and intercept of the bid curve of each power producer, is the mapping function, Representation value function The approximate value of represents the set of polynomial basis functions, φ l (α) represents the lth polynomial basis function, d l represents the coefficients of the lth polynomial to be solved.

[0105] Specifically, the set of polynomial basis functions It can be expressed as:

[0106]

[0107] Where N s Represents the total number of polynomial basis functions; N c represents the highest degree of the basis function; M represents the variable α m The dimension of . represents the tensor product operator; Represents the set of polynomial basis functions consisting of the pth variable.

[0108] Solving for coefficients of polynomial basis functions First, solve the higher-order basis function The root of Z ={α 1 ,α 2 ,...,α N} T ;

[0109] Given sampling point information α Z ={α 1 ,α 2 ,...,α N} T Solve the lower-level model (such as the day-ahead electricity market model (12)-(20)) to obtain the response under the current decision

[0110] According to the least squares method, the coefficients of the polynomial basis function D = {d1, d2, ..., d l}:

[0111]

[0112] In the formula, Indicates the Nth s The value of the polynomial basis function at the Nth sampling point; Φ is the polynomial basis function The matrix formed.

[0113] It should be noted that V represents the social welfare value in the lower model, that is, it can represent the output value of the day-ahead electricity energy market clearing sub-model and the output value of the real-time electricity energy market clearing sub-model. By fitting the output variables of the lower model in the two-layer optimization model to the linear combination of the polynomial basis functions of the input variables of the lower model, the original functions of the day-ahead electricity energy market clearing sub-model and the real-time electricity energy market clearing sub-model can be converted into linear combinations with polynomial basis functions, so that only the linear combinations with polynomial basis functions need to be solved to solve the day-ahead electricity energy market clearing sub-model and the real-time electricity energy market clearing sub-model; that is, the original objective functions of the day-ahead and real-time electricity energy market clearing sub-models are converted into multiple groups of inequalities about value functions. By determining the expression of the value function in each group of inequalities (including the combination form of the basis functions and the corresponding coefficients of the basis functions), the equivalent replacement of the day-ahead and real-time electricity market clearing sub-models can be achieved.

[0114] For step S3, in a preferred embodiment, under the segment range constraint of the generator set output, the range constraint of the generator set output, the segment range constraint of the load output, the line transmission power range constraint and the range constraint of the node phase angle, the single-layer optimization model is solved to generate the unit output when the power generator benefits the most:

[0115] Specifically, the two-layer model is converted into a single-layer model, and the approximate response of the lower-layer model under different input variables α is obtained according to formula (32): Finally, the two-layer optimization model of the electricity market can be converted into a single-layer optimization model.

[0116] Among them, the objective function of the single-layer model is still the pursuit of profit maximization by generator j, that is,

[0117]

[0118] The revenue and bid of generator j still need to satisfy equations (2)-(11), and the day-ahead electricity market clearing model can be converted into the form of equations (37)-(38).

[0119]

[0120] In the formula, represents the approximate value of negative social welfare in time period t in the day-ahead electricity energy market; represents the approximate value of the node marginal electricity price in time period t in the day-ahead electricity market;

[0121] Similarly, the day-ahead electricity market clearing model still needs to satisfy constraints (13)-(20), and the real-time electricity market clearing model can be converted into the form of (39)-(40).

[0122]

[0123] In the formula, It represents the negative social welfare approximation of scenario ω and time period t in the real-time electric energy market; Represents the approximate marginal electricity price of the node in the scenario ω and time period t in the real-time electricity energy market.

[0124] The real-time electricity energy market clearing model still needs to satisfy equations (22)-(28), and the green certificate market clearing model is described by equations (29)-(31).

[0125] The solver CONOPT is called to solve the single-layer model, and the optimal bidding strategy of generator j and the winning output of the generator set under this strategy are obtained, and the unit output when the generator benefit is maximized is obtained.

[0126] In a preferred embodiment, after fitting the output variables of the lower model in the two-layer optimization model to the linear combination of the polynomial basis functions of the input variables of the lower model, the method further includes:

[0127] Value Function The approximate accuracy of is crucial to the accurate solution of the two-layer model, and the lower the dimension of the input variable α, the better the value function Therefore, the present invention adopts the variable decomposition method to transform the input variable into a high-dimensional value function Convert multiple input variables into low-dimensional value functions This improves the solution accuracy of the model.

[0128] The input variables are transformed into high-dimensional value functions by Convert multiple input variables into low-dimensional value functions

[0129] The present invention decomposes the value function based on Taylor expansion theory, which mainly includes the following three steps:

[0130] (1) At a given point α C The pair value function Perform Taylor expansion:

[0131]

[0132] In the formula, the high-dimensional variable α={α1,α2,...,α n}; Representation value function For the k-th dimension α of variable α k Find the j-order partial derivative; R represents the expansion term with higher dimension;

[0133] (2) Constructing value functions A one-dimensional variable function of , and performs Taylor expansion at a given point.

[0134] make but It can be expressed as

[0135]

[0136] In the formula, Represents a univariate function with only one variable α1.

[0137] Similarly, we can get the value function , and Taylor expansion at a given point, as shown in formula (43).

[0138]

[0139] (3) The input variable is a high-dimensional value function Convert multiple input variables into low-dimensional value functions

[0140] Adding equation (42) and equation (43) yields

[0141]

[0142] Substitute equation (44) into equation (41) and discard the high-dimensional term R in Taylor expansion

[0143]

[0144] Will Recorded as Will Recorded as Then formula (45) can be converted into

[0145]

[0146] It can be further deduced from formula (46) that the value function Can be decomposed into multiple input variables with low-dimensional value functions As shown in equations (47)-(48), the value function Decompose into multiple input variables into low-dimensional value functions As shown in the following formula:

[0147]

[0148] n=n1r (48)

[0149] in, represents a low-dimensional variable-valued function, express The specific value at a certain point; n1 is α mThe dimension of .

[0150] The present invention first converts the power market quotation-clearing two-layer optimization model into a single-layer optimization model based on the value function approximation theory without introducing integer variables; secondly, the variable decomposition technology is used to decompose the value function with high-dimensional input variables into multiple value functions with low-dimensional input variables; finally, considering that the converted single-layer optimization model containing low-dimensional value functions is relatively simple, the solver CONOPT is directly called to solve it, so as to obtain the power generator's quotation curve and the unit's winning bid output. That is, the present invention uses the variable decomposition method to reduce the dimension of the value function input variables, improve the approximate accuracy of the value function, and realize accurate and efficient solution of the two-layer optimization model of the present invention; without introducing new integer variables, the two-layer model can be converted into a single-layer model, and the unit output corresponding to the maximum benefit of the power generator can be obtained by solving the single-layer model, so that the present invention can automatically obtain the unit output. The present invention efficiently solves the double-layer model based on the value function approximation theory and variable decomposition technology, and can automatically obtain the unit output without manual calculation and screening. The final unit output can meet the maximization of power generator interests and take into account the maximization of social welfare, which is more conducive to healthy competition in the power market.

[0151] For step S4, in a preferred embodiment, the generator set is regulated according to the output of the set;

[0152] Indicatively, the predicted unit output can be used to automatically control the generator set, and the power generator can maximize its profit in various power markets under the predicted unit output.

[0153] like Figure 2 As shown, based on the above-mentioned various embodiments of the unit output prediction method based on maximizing the interests of the power generator, the present invention provides a corresponding device item embodiment;

[0154] An embodiment of the present invention provides a unit output optimization device based on maximizing the interests of power producers, including: a double-layer optimization model building module, a model conversion module, a unit output generation module and a generator unit control module;

[0155] The two-layer optimization model building module is used to obtain data of power generators, market clearing data and power network data, and build a two-layer optimization model to maximize the interests of power generators; wherein the two-layer optimization model includes segment range constraints of generator set output, range constraints of generator set output, segment range constraints of load output, line transmission power range constraints and range constraints of node phase angles;

[0156] The model conversion module is used to convert the double-layer optimization model into a single-layer optimization model based on a value function approximation method;

[0157] The unit output generation module is used to solve the single-layer optimization model under the segment range constraints of the generator unit output, the range constraints of the generator unit output, the segment range constraints of the load output, the line transmission power range constraints and the range constraints of the node phase angle, and generate the unit output when the power generator benefits the most;

[0158] The generator set control module is used to control the generator set according to the output of the set.

[0159] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without creative work.

[0160] Those skilled in the art can clearly understand that, for the sake of convenience and simplicity, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0161] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for optimizing unit output based on maximizing the interests of power generators, characterized in that: include: Obtain data of power generators, market clearing data and power network data, and construct a two-layer optimization model to maximize the interests of power generators; wherein the two-layer optimization model includes segment range constraints of generator set output, range constraints of generator set output, segment range constraints of load output, line transmission power range constraints and range constraints of node phase angles; Converting the two-layer optimization model into a single-layer optimization model based on a value function approximation method; Under the constraints of the segmented range of generator set output, the range of generator set output, the segmented range of load output, the range of line transmission power and the range of node phase angle, the single-layer optimization model is solved to generate the unit output when the power generator benefits the most; regulating the generator set according to the output of the unit; The method of converting the double-layer optimization model into a single-layer optimization model based on the value function approximation method includes: Fitting the output variables of the lower model in the double-layer optimization model to the linear combination of the polynomial basis functions of the input variables of the lower model to generate a single-layer optimization model; When the lower model of the two-layer optimization model includes a day-ahead electricity energy market clearing sub-model, the output variables of the lower model in the two-layer optimization model are fitted as linear combinations of polynomial basis functions of the input variables of the lower model to generate a single-layer optimization model, including: The original objective function corresponding to the day-ahead electricity energy market clearing sub-model is converted into a first set of inequalities about the value function; wherein the first set of inequalities about the value function includes: Where f(x) is the original objective function corresponding to the day-ahead electricity market clearing sub-model, which is used to minimize the negative social welfare in the day-ahead electricity market; Represents the approximate value of negative social welfare in time period t in the day-ahead electricity energy market.

2. A method for optimizing unit output based on maximizing the interests of power generators as claimed in claim 1, characterized in that: The two-layer optimization model is specifically: in, is the revenue of generator j in the day-ahead electricity market, is the revenue of generator j in the real-time electric energy market, represents the income of renewable energy generator j in the green certificate market, represents the income of traditional power producer j in the carbon market, C j is the total unit operating cost of generator j; represents the marginal electricity price of node n in the day-ahead electricity market in time period t, T represents a set of multiple time periods t (such as 1-24h), Ω j represents the set of units of generator j; n Represents a collection of nodes. represents the winning bid capacity of generator i in time period t and capacity segment b in the day-ahead electricity market, W represents the set composed of various scenarios ω, π ω represents the probability of scenario ω occurring, represents the marginal electricity price of node n in the real-time electricity market in scenario ω and time period t, represents the winning capacity of generator i in the real-time electric energy market in scenario ω, time period t, and capacity segment b, Indicates the price of the green certificate, represents the number of green certificates obtained by renewable energy generator j under scenario ω, λ CT represents the carbon trading price in the carbon market, σ j represents the carbon quota decomposition factor of power producer j, represents the carbon quota of power producer j, represents the carbon emission factor of generator set i, a i represents the quadratic coefficient of the operating cost of generator set i, b i represents the first-order coefficient of the operating cost of generator set i, c i The constant term representing the operating cost of generator set i, Ω RE represents the set of renewable energy generators, Ω T represents the set of traditional power generators, Used to mark whether power generator j participates in the green certificate market, with a value of 0 or 1; The two-level optimization model satisfies the following constraints: α i,b,t represents the slope of the bth segment of the bidding curve of generator set i in period t, Indicates the minimum slope of the bth segment of the bidding curve of generator set i, Indicates the maximum slope of the bth segment of the bidding curve of generator set i, β i,t represents the intercept of the quotation curve of generator set i in period t, It represents the maximum intercept value of the quotation curve of generator set i, Indicates the minimum intercept value of the quotation curve of generator set i, Ω j Represents the set of generator sets of generator j.

3. A method for optimizing unit output based on maximizing the interests of power generators as claimed in claim 2, characterized in that: The step of fitting the output variables of the lower model in the two-layer optimization model to a linear combination of polynomial basis functions of the input variables of the lower model comprises: The linear combination of the polynomial basis functions of the input variables of the lower model is fitted according to the following formula: Among them, V represents the social welfare value in the lower model, α represents the slope and intercept of the bid curve of each power producer, is the mapping function, Representation value function The approximate value of represents the set of polynomial basis functions, φ l (α) represents the lth polynomial basis function, d l represents the coefficients of the lth polynomial to be solved.

4. A method for optimizing unit output based on maximizing the interests of power generators as claimed in claim 3, characterized in that: After fitting the output variables of the lower model in the double-layer optimization model to the linear combination of the polynomial basis functions of the input variables of the lower model to generate a single-layer optimization model, the method further includes: The input variables are transformed into high-dimensional value functions by Convert multiple input variables into low-dimensional value functions At the preset point α C The pair value function Expanding it, we get the following formula: Among them, the high-dimensional variable α={α1,α2,...,α n }, Representation value function For the k-th dimension α of variable α k Find the j-order partial derivative, R represents the expansion term of the next dimension; The value function Decompose into multiple input variables into low-dimensional value functions As shown in the following formula: n=n1r in, represents a low-dimensional variable-valued function, express At the expansion point α C The specific value of n1 is α m The dimension of .

5. A method for optimizing unit output based on maximizing the interests of power generators as claimed in claim 4, characterized in that: The original objective function corresponding to the day-ahead electricity energy market clearing sub-model is: The first set of inequalities about the value function is: The original objective function corresponding to the day-ahead electricity energy market clearing sub-model and the first set of inequalities about the value function both satisfy the following constraints: λ d,k,t represents the quotation of load d in time period t and capacity segment k; α i,b,t represents the slope of the bth segment of the bidding curve of generator set i in period t, is the output of generator i in time period t and capacity segment b in the day-ahead electricity market, γ i,b,t Indicates the ordinate value corresponding to the starting point of the bth segment in the quotation curve of generator set i in time period t; It represents the output of load d in time period t and capacity segment k; Represents the set of generator sets at node n; represents the set of load components on node n; represents the set of all nodes connected to node n, B n,m represents the conductance of the line between node n and node m, represents the phase angle of the day-ahead market node n in period t, and represents the upper and lower limits of the generating unit i in the capacity segment b, and represents the upper and lower limits of the output of generator i in time period t in the day-ahead market, and represents the upper and lower limits of load d in time period t and capacity segment k, It represents the upper limit of the line transmission power between node n and node m.

6. A method for optimizing unit output based on maximizing the interests of power generators as claimed in claim 5, characterized in that: The lower model of the two-layer optimization model also includes: a clearing sub-model of the real-time electric energy market; When converting the two-layer optimization model into a single-layer optimization model based on the value function approximation method, the original objective function corresponding to the clearing sub-model of the real-time electric energy market is converted into a second set of inequalities about the value function; Among them, the original objective function corresponding to the clearing sub-model of the real-time electric energy market is: The second set of inequalities about the value function is: The original objective function corresponding to the clearing sub-model of the real-time electric energy market and the second set of inequalities about the value function both satisfy the following constraints: represents the output of load d in scenario ω, time period t, and capacity segment k, represents the phase angle of real-time power market node n in scenario ω and time period t, and It represents the upper and lower limits of the output of generator i in the real-time electricity market in scenario ω and time period t.

7. A method for optimizing unit output based on maximizing the interests of power generators as claimed in claim 6, characterized in that: The two-layer optimization model also satisfies the following constraints: in, Indicates the price of the green certificate, Represents the quantity-price relationship coefficient of green certificates, It indicates the penalty price when the quota obligation subject fails to complete the quota task. It indicates the price that the quota obligor is willing to pay when the supply of green certificates is sufficient. represents the quota ratio required by the renewable energy quota system, Indicates the maximum demand for green certificates.

8. A unit output optimization device based on maximizing the interests of power generators, characterized in that: include: Two-layer optimization model building module, model conversion module, unit output generation module and generator unit control module; The two-layer optimization model building module is used to obtain data of power generators, market clearing data and power network data, and build a two-layer optimization model to maximize the interests of power generators; wherein the two-layer optimization model includes segment range constraints of generator set output, range constraints of generator set output, segment range constraints of load output, line transmission power range constraints and range constraints of node phase angles; The model conversion module is used to convert the double-layer optimization model into a single-layer optimization model based on a value function approximation method; The unit output generation module is used to solve the single-layer optimization model under the segment range constraints of the generator unit output, the range constraints of the generator unit output, the segment range constraints of the load output, the line transmission power range constraints and the range constraints of the node phase angle, and generate the unit output when the power generator benefits the most; The generator set control module is used to control the generator set according to the output of the generator set; The method of converting the double-layer optimization model into a single-layer optimization model based on the value function approximation method includes: Fitting the output variables of the lower model in the double-layer optimization model to the linear combination of the polynomial basis functions of the input variables of the lower model to generate a single-layer optimization model; Wherein, when the lower model of the two-layer optimization model includes a day-ahead electricity energy market clearing sub-model, the output variables of the lower model in the two-layer optimization model are fitted as linear combinations of polynomial basis functions of the input variables of the lower model to generate a single-layer optimization model, including: The original objective function corresponding to the day-ahead electricity energy market clearing sub-model is converted into a first set of inequalities about the value function; wherein the first set of inequalities about the value function includes: Where f(x) is the original objective function corresponding to the day-ahead electricity market clearing sub-model, which is used to minimize the negative social welfare in the day-ahead electricity market; Represents the approximate value of negative social welfare in time period t in the day-ahead electricity energy market.

Images