Power generation side dispatching electric quantity optimization method and device considering multiple time scales

Abstract

The invention provides a power generation side dispatching electric quantity optimization method and device considering multiple time scales, and relates to the technical field of electric power grids. According to the method, the unified multi-scale collaborative framework is constructed, power generation curve optimization of the medium-and-long-term stage and power generation curve optimization of the spot stage are organically combined, the problem that the two stages in traditional dispatching are disjointed is solved, the day-degree power generation curve is generated in the medium-and-long-term stage, refined rolling optimization is carried out in the spot stage, and the dispatching efficiency is improved. Therefore, the final power generation scheme not only meets medium and long-term macroscopic targets, such as the total carbon quota, but also can adapt to real-time operation conditions of the spot market, the deviation between the plan and actual execution is reduced, and the accuracy of power generation scheduling is improved. The problem that the power generation side power generation plan is inaccurately formulated is solved, and the accuracy of formulating the multi-time-scale power generation side power generation plan is improved.

Description

Technical Field

[0001] This invention relates to the field of power grid technology, and in particular to a method and apparatus for optimizing power dispatch on the generation side that considers multiple time scales. Background Technology

[0002] With the continuous advancement of the green and low-carbon transformation of the energy system, the integration of a high proportion of renewable energy into the power system has become a development trend. The output of renewable energy sources such as wind and solar power exhibits significant intermittency, volatility, and uncertainty, posing a significant challenge to power system operation and dispatch. Simultaneously, carbon quota constraints are gradually becoming a rigid condition that power generation companies must consider in their production and operation, making it difficult for traditional power dispatching methods to simultaneously meet the dual requirements of renewable energy consumption and carbon emission reduction.

[0003] Currently, when formulating power generation plans, the power generation side typically separates medium- and long-term dispatch from spot market dispatch, lacking a coordinated optimization mechanism across time scales. Power generation plans formulated in the medium- and long-term phases struggle to adapt to real-time fluctuations in renewable energy in the spot market, leading to significant deviations between actual operation and plans. Furthermore, existing dispatch models are mostly limited to the single dimension of electricity, failing to fully consider the impact of carbon quota constraints on unit output, and thus unable to achieve coordinated optimization of electricity and carbon emissions.

[0004] In renewable energy output forecasting, traditional methods such as BP neural networks suffer from slow convergence speed and susceptibility to local optima, resulting in lower prediction accuracy. Regarding scheduling model solutions, the optimization models for electricity-carbon coupling often involve complex nonlinear constraints and integer variables, making direct solutions difficult. Existing methods often employ simplification, impacting the accuracy and practicality of the optimization results. Summary of the Invention

[0005] This invention provides a method and apparatus for optimizing power dispatch on the generation side that considers multiple time scales, which can solve the problem of inaccurate generation planning on the generation side and improve the accuracy of generation planning on the generation side at multiple time scales.

[0006] Firstly, this invention provides a method for optimizing power dispatch on the generation side considering multiple time scales. The method includes: acquiring medium- and long-term stage data and spot stage data; constructing a multi-scale collaborative framework based on the medium- and long-term stage data and the spot stage data, with the objectives of minimizing carbon trading costs in the medium- and long-term stage and minimizing power dispatch costs on the generation side in the spot stage; optimizing medium- and long-term carbon trading costs based on the multi-scale collaborative framework, solving for the daily power generation curve and carbon emission plan in the medium- and long-term stage; optimizing the power dispatch on the generation side in the spot stage using a two-layer dispatch model based on the daily power generation curve and carbon emission plan in the medium- and long-term stage, obtaining the power generation curve in the spot stage; and generating an optimized power dispatch scheme based on the daily power generation curve in the medium- and long-term stage and the power generation curve in the spot stage.

[0007] In one possible implementation, based on medium- and long-term data and spot market data, a multi-scale collaborative framework is constructed with the objectives of minimizing carbon trading costs in the medium- and long-term phase and minimizing generation-side dispatch costs in the spot market. This framework includes: constructing a medium- and long-term carbon cost optimization objective function based on medium- and long-term data, with the goal of minimizing carbon trading costs and using multi-source unit collaboration constraints and carbon quota trading safety boundaries as constraints; multi-source unit collaboration constraints include wind power output constraints, photovoltaic power output constraints, thermal power ramping constraints, and energy storage constraints; and carbon quota trading safety boundaries include carbon trading volume limits and annual compliance constraints. Based on spot market data, a two-tiered dispatch model is constructed, with the goal of minimizing generation-side dispatch costs in the spot market as the objective of the upper-level dispatch model and maximizing the social welfare of the joint clearing of the electricity carbon market as the objective of the lower-level clearing model. Finally, based on the medium- and long-term carbon cost optimization objective function and the two-tiered dispatch model in the spot market, a multi-scale collaborative framework is constructed.

[0008] In one possible implementation, wind power output constraints are used to limit the operating range of wind power output; photovoltaic power output constraints are used to limit the operating range of photovoltaic power output; thermal power ramping constraints are used to limit the operating range of thermal power output; thermal power ramping constraints are used to limit the range of output changes of thermal power units at adjacent times; energy storage constraints are used to limit the capacity and charging / discharging power of energy storage devices; carbon trading volume constraints are used to limit the total amount of carbon allowances traded each month; and annual compliance constraints are used to ensure that the actual annual carbon emissions do not exceed the allocated initial carbon allowance.

[0009] In one possible implementation, a multi-scale collaborative framework is used to optimize medium- and long-term carbon trading costs, and to solve for the daily power generation curve and carbon emission plan in the medium- and long-term stages. This includes: constructing an objective function based on the medium- and long-term carbon trading cost optimization objective function, guided by maximizing net revenue, and including a dual calculation mechanism for electricity revenue and carbon emission costs; constructing a nonlinear carbon emission cost function based on the objective function of the dual calculation mechanism for electricity revenue and carbon emission costs, as well as multi-source unit collaborative constraints and carbon quota trading safety boundaries; transforming the nonlinear carbon emission cost function into a mixed-integer linear function through piecewise linearization techniques; and solving for the medium- and long-term daily power generation curve and carbon emission plan based on the mixed-integer linear function.

[0010] In one possible implementation, based on the daily power generation curves and carbon emission plans in the medium- and long-term stages, a two-layer dispatch model is adopted to optimize the dispatched power volume on the generation side in the spot stage, thereby obtaining the power generation curve in the spot stage. This includes: dynamically decomposing the daily power generation curves and carbon emission plans in the medium- and long-term stages to obtain the output prediction sequence, which includes the thermal power output prediction sequence, wind power output prediction sequence, and photovoltaic power output prediction sequence; performing energy storage correction based on the output prediction sequence, as well as the actual wind power output and actual photovoltaic output, to determine the energy storage charging and discharging sequence; using the output prediction sequence and the energy storage charging and discharging sequence as the input of the upper-layer dispatch model in the two-layer dispatch model, and the generation-side dispatch price as the output of the upper-layer dispatch model, and the generation-side dispatch price and user load demand as the input of the lower-layer clearing model, and the actual power output plan of the generation-side units as the output, the two-layer dispatch model is solved to obtain the power generation curve in the spot stage.

[0011] In one possible implementation, dynamic decomposition is performed based on the daily power generation curve and carbon emission plan in the medium to long term to obtain the power output prediction sequence. This includes: using a BP neural network with optimized weights and thresholds via a genetic algorithm to predict wind and solar power output for future periods in the medium to long term, obtaining the predicted power output results; based on the predicted power output results, the monthly renewable energy contract power volume is proportionally decomposed into daily amounts, obtaining wind power output prediction sequences and solar power output prediction sequences; and the remaining electricity demand, excluding the monthly renewable energy contract power volume, is decomposed into daily thermal power generation according to the balancing schedule, obtaining a thermal power output prediction sequence.

[0012] In one possible implementation, the power output prediction sequence and energy storage charging and discharging sequence are used as inputs to the upper-level scheduling model in the two-level scheduling model, the generation-side scheduling price is used as the output of the upper-level scheduling model, the generation-side scheduling price and user load demand are used as inputs to the lower-level clearing model, and the actual power output plan of the generation-side units is used as the output. The two-level scheduling model is solved to obtain the power generation curve in the spot market stage. This includes: transforming the lower-level market clearing problem into an equivalent constraint based on KKT conditions; introducing the equivalent constraint into the upper-level optimization problem, transforming the original two-level model into a single-level mathematical programming problem, and obtaining a single-level mathematical model; linearizing the nonlinear terms in the single-level mathematical model by introducing auxiliary integer variables and the Big M method, and obtaining a mixed-integer linear model; and solving the mixed-integer linear model using linear programming to obtain the power generation curve in the spot market stage.

[0013] In one possible implementation, an optimized power dispatch scheme for the generation side is generated based on the daily power generation curves in the medium-to-long-term phase and the power generation curves in the spot phase. This includes: merging the medium-to-long-term daily power generation curves and the spot power generation curves along a time axis to generate the final power generation curve on the generation side; calculating the deviation rate between the actual carbon trading cost and the medium-to-long-term predicted cost, the comprehensive benefit deviation rate, the actual carbon emissions and quota usage, and the proportion of clean energy based on the final power generation curve on the generation side; generating a benefit assessment result based on the deviation rate between the actual carbon trading cost and the medium-to-long-term predicted cost, the comprehensive benefit deviation rate, the actual carbon emissions and quota usage, and the proportion of clean energy, including the comprehensive benefit increase, carbon quota utilization rate, and renewable energy absorption rate; and generating an optimized power dispatch scheme for the generation side based on the final power generation curve on the generation side and the benefit assessment result.

[0014] Secondly, embodiments of the present invention provide a generator-side dispatch power optimization device considering multiple time scales. This optimization device includes a communication module and a processing module. The communication module is used to acquire medium- and long-term stage data and spot stage data. The processing module is used to construct a multi-scale collaborative framework based on the medium- and long-term stage data and the spot stage data, with the objectives of minimizing carbon trading costs in the medium- and long-term stage and minimizing generator-side dispatch costs in the spot stage. Based on the multi-scale collaborative framework, it performs medium- and long-term carbon trading cost optimization, solving for the daily power generation curve and carbon emission plan in the medium- and long-term stage. Based on the daily power generation curve and carbon emission plan in the medium- and long-term stage, it adopts a two-layer dispatch model to optimize the generator-side dispatch power in the spot stage, obtaining the power generation curve in the spot stage. Based on the daily power generation curve in the medium- and long-term stage and the power generation curve in the spot stage, it generates a generator-side dispatch power optimization scheme.

[0015] In one possible implementation, the processing module is specifically used to construct a medium-to-long-term carbon cost optimization objective function based on medium-to-long-term data, with the goal of minimizing carbon trading costs and constraints on multi-source unit coordination and carbon quota trading security boundaries. The multi-source unit coordination constraints include wind power output constraints, photovoltaic power output constraints, thermal power ramp-up constraints, and energy storage constraints. The carbon quota trading security boundaries include carbon trading volume limits and annual compliance constraints. Based on spot market data, a two-tiered dispatch model is constructed, with the goal of minimizing the generation-side dispatch cost in the spot market as the objective of the upper-level dispatch model and maximizing the social welfare of the joint clearing of the electricity-carbon market as the objective of the lower-level clearing model. Finally, a multi-scale collaborative framework is constructed based on the medium-to-long-term carbon cost optimization objective function and the two-tiered dispatch model in the spot market.

[0016] Thirdly, embodiments of the present invention provide an electronic device including a memory and a processor. The memory stores a computer program, and the processor is configured to call and run the computer program stored in the memory to perform the steps of the method as described in the second aspect and any possible implementation thereof.

[0017] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, characterized in that, when executed by a processor, the computer program implements the steps of the method as described in the second aspect and any possible implementation thereof.

[0018] This invention provides a method and apparatus for optimizing power generation scheduling on the generation side, considering multiple time scales. By constructing a unified multi-scale collaborative framework, this invention organically combines power generation curve optimization in the medium-to-long-term and spot market phases, overcoming the problem of disconnect between the two phases in traditional scheduling. By generating daily power generation curves in the medium-to-long-term phase and performing refined rolling optimization in the spot market phase, the final power generation scheme not only conforms to medium-to-long-term macroeconomic objectives, such as total carbon quotas, but also adapts to real-time operating conditions in the spot market, reducing the deviation between planned and actual execution and improving the accuracy of power generation scheduling. This invention solves the problem of inaccurate power generation planning on the generation side and improves the accuracy of power generation planning across multiple time scales. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating a power generation scheduling optimization method considering multiple time scales provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the training process of a BP neural network prediction model based on an improved genetic algorithm provided in an embodiment of the present invention; Figure 3 This is a schematic flowchart of a power generation-side scheduling optimization device considering multiple time scales provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0021] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0022] In the description of this invention, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The term "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Furthermore, "at least one" and "more than one" refer to two or more. The terms "first," "second," etc., do not limit the quantity or order of execution, and "first," "second," etc., do not necessarily imply differences.

[0023] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner to facilitate understanding.

[0024] Furthermore, the terms "comprising" and "having," and any variations thereof, used in the description of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or modules is not limited to the steps or modules listed, but may optionally include other steps or modules not listed, or may optionally include other steps or modules inherent to such process, method, product, or device.

[0025] To make the objectives, technical solutions, and advantages of the present invention clearer, the following description will be provided in conjunction with the accompanying drawings and specific embodiments.

[0026] like Figure 1 As shown, this embodiment of the invention provides a method for optimizing power dispatch on the generation side considering multiple time scales. The method includes steps S101-S105.

[0027] S101. Obtain medium- and long-term stage data and spot stage data.

[0028] S102. Based on medium- and long-term stage data and spot stage data, a multi-scale collaborative framework is constructed with the goal of minimizing carbon trading costs in the medium- and long-term stage and minimizing power generation-side dispatch costs in the spot stage.

[0029] As one possible implementation, step S102 can be specifically implemented as steps S1021-S1024.

[0030] S1021. Based on medium- and long-term data, with the goal of minimizing carbon trading costs and with constraints such as multi-source unit coordination and carbon quota trading security boundary, construct a medium- and long-term carbon cost optimization objective function.

[0031] In some embodiments, multi-source unit coordination constraints include wind power output constraints, photovoltaic power output constraints, thermal power ramping constraints, and energy storage constraints. The carbon quota trading safety boundary includes carbon trading volume limits and annual compliance constraints.

[0032] For example, wind power output constraints are used to limit the operating range of wind power output; photovoltaic power output constraints are used to limit the operating range of photovoltaic power output; thermal power ramping constraints are used to limit the operating range of thermal power output; thermal power ramping constraints are used to limit the range of output changes of thermal power units at adjacent times; energy storage constraints are used to limit the capacity and charging / discharging power of energy storage devices; carbon trading volume constraints are used to limit the total amount of carbon allowances traded each month; and annual compliance constraints are used to ensure that the actual carbon emissions for the whole year do not exceed the allocated initial carbon allowance.

[0033] For example, the objective function for optimizing carbon costs in the medium to long term prioritizes the economics of the medium to long term trading phase, including power generation revenue and carbon emission costs. Carbon market revenue considers both carbon allowance prices and carbon emission volumes. The objective function is expressed in the following formula.

[0034] ; in, For medium- to long-term returns; Cost of carbon emissions; For the revenue generated from electricity generation; This refers to the monthly power generation under medium- to long-term contracts. The electricity price for month m; The price of carbon allowances for month m; The estimated carbon emissions of thermal power units in month m; The estimated carbon quota for thermal power units in month m.

[0035] Expected carbon emissions from thermal power units for: ; This represents the thermal power output during the period m in month t. is the carbon factor of the thermal power unit, and T represents the number of time periods in the medium- and long-term phases.

[0036] For example, the constraints of the objective function for carbon cost optimization in the medium to long term are as follows: In the medium to long term, complex coupling relationships among various power generators are involved, and establishing constraints is a prerequisite for ensuring the rationality and scientific nature of decision-making. At the same time, to ensure the coordinated and stable operation of the carbon market, the carbon emission behavior of power generators needs to be considered. Therefore, based on the output characteristics of each unit and factors such as carbon quota trading volume, constraints on unit output and carbon quota trading volume are set.

[0037] The unit output constraints include the upper and lower limits of the operating output of wind power, photovoltaic, and thermal power units. They must operate within the specified range, as detailed below.

[0038] ; ; .

[0039] in, To generate power for wind power at the current moment, This is the upper limit of wind power output. To contribute to photovoltaic power generation at this moment, The lower limit of photovoltaic power output, This is the upper limit of photovoltaic power output; To generate power for thermal power plants at the current moment, This represents the lower limit of thermal power output. This is the upper limit of thermal power output.

[0040] Thermal power plant ramping constraints: ;in, The current equipment output status; This refers to the equipment output status at the previous moment. This represents the upper limit of effort required to climb the slope; This represents the lower limit of the effort required to climb the slope.

[0041] Energy storage constraints: ; in, Let t be the capacity of the energy storage device at time t; The maximum charging power of the energy storage device; This represents the maximum energy release power of the energy storage device. Let t be the power of the energy storage device at time t.

[0042] Carbon quota trading volume constraints: To prevent speculative behavior by power generators when prices are low, regulators set total trading volume constraints.

[0043] ; in, This represents the carbon allowance trading volume for month m. The minimum monthly quota for transactions; This is the maximum limit for monthly quota transactions.

[0044] Annual quota constraints: ; in, This represents the unit's carbon emissions in month m. The initial carbon quota allocated based on the baseline method.

[0045] To achieve coordinated trading across multiple time scales in the electricity and carbon markets, this paper analyzes the trading decisions of power generators in the spot market phase. This decision-making process is based on the long-term trading phase discussed earlier. In this phase, power generators participate in spot market and secondary carbon market trading constrained by the monthly contract decomposition results. This model considers the coordinated trading of power generators in the electricity and carbon markets, and is divided into a spot market power generator electricity-carbon trading decision-making model and an electricity-carbon market clearing model.

[0046] Upper Layer: Power Generator's Spot Market Trading Decisions. Upper-level power generator decisions aim to maximize profits, specifically by comprehensively considering the maximization of net revenue from both the spot electricity market and the secondary carbon market. Considering the connection between markets at different time scales, monthly contracts are decomposed; for example, monthly electricity contracts are broken down into daily (30-day) contracts, with the decomposed trading volume as one of the constraints. This allows for the determination of the power generator's electricity market bidding strategy and carbon market trading quota during the spot market trading phase.

[0047] Lower Tier: Power Generator Market Clearing Model. The lower tier market clearing aims to maximize social welfare, with constraints such as power balance and carbon quota trading, enabling power generators to flexibly exit the spot market and secondary carbon market.

[0048] The medium- and long-term contract decomposition model is affected by various uncontrollable factors such as weather and terrain, due to the inherent characteristics of wind and solar power generation, including random fluctuations, uncertainties, and anti-peak shaving. Therefore, in the spot market, the output deviation of renewable energy is also an important factor that must be considered on the power generation side to achieve optimal economic performance. This invention adopts a dynamic proportional decomposition method, adjusting the daily decomposition amount based on wind and solar power output forecasts.

[0049] This paper proposes a prediction model based on a genetic algorithm-improved BP neural network. A BP neural network is a multi-layer feedforward neural network, commonly using a three-layer structure: input layer, single hidden layer, and output layer. It exhibits strong nonlinear modeling capabilities in pattern recognition and time series prediction. The optimization approach involves first mapping the input signal feature data to the hidden layer (implemented by activation functions), then to the output layer (using a linear transfer function by default) to obtain the desired output value. The desired output value is compared with the actual measured value to calculate the error function J. The error is then backpropagated, and algorithms such as gradient descent are used to adjust the weights and thresholds of the BP network. This process is repeated until a set target error or maximum iteration count is met, at which point training stops. However, because it heavily relies on the selection of initial weights during training, it suffers from local optima traps or unstable convergence speeds caused by excessively large initial weights, thus affecting prediction accuracy. Since genetic algorithms possess excellent global search capabilities, they can optimize the weights and thresholds of BP neural networks. Therefore, this paper proposes a BP neural network prediction model improved by a genetic algorithm. Figure 2 As shown.

[0050] The optimization process of the BP neural network prediction model based on genetic algorithm optimization consists of the following steps: Step 1: Initialize the population weights and thresholds; Step 2: Calculate fitness, the fitness function is: ;in, The actual value to be output; is the predicted output value, where n is the population size.

[0051] Step 3: Selection operation. The smaller the fitness, the smaller the error. When selecting superior individuals from the population, the probability of an individual being selected is: ; in, The probability of an individual being selected; This represents the fitness value corresponding to the i-th individual; Let k be the rank of the i-th individual, and k be a normal coefficient, for example, k=1 or k is the population size n.

[0052] Step 4: Crossover operation, perform crossover operation on the t-th and v-th chromosomes at position k: ;in, For the middle position after the crossover operation at position k chromosome, For the middle position after the crossover operation at position k chromosome; It is a random number located in the interval [0,1].

[0053] Step 5: Mutation operation, perform mutation operation on the i-th gene in the t-th chromosome: ;in, It is a mutated gene; This represents the upper limit of mutated genes; This represents the lower limit of the mutated gene; Here, g represents the mutation rate, and g represents the number of iterations. This represents the maximum number of evolutions. A random number located in the interval [0,1]. It is a random number located in the interval [0,1].

[0054] Step 6: Return to step 3 and repeat the above operation to output the optimal weight and threshold.

[0055] S1022. Based on spot market data, a two-layer scheduling model is constructed with the goal of minimizing the dispatch cost on the generation side in the spot market as the objective of the upper-layer scheduling model and maximizing the social welfare of the joint clearing of the electricity carbon market as the objective of the lower-layer clearing model.

[0056] For example, a two-layer scheduling model, namely the power generation curve optimization model in the spot market stage: Upper layer: Power generator scheduling optimization model, the decision of power generators in the spot market aims to minimize costs, and under the premise of medium and long-term contract decomposition, the optimization cycle is 30 days.

[0057] The objective function is: ; in, Costs at the spot stage; The electricity trading price for time period t; The amount of electricity generated by the power generator during time period t; For generator cost function; Let t be the carbon price during the time period. This represents the amount of carbon allowances sold during period t. This represents the amount of carbon allowances purchased during period t.

[0058] Upper-level transaction decisions must simultaneously consider constraints such as the daily contract breakdown volume and unit operating conditions.

[0059] Monthly contract decomposition constraints: In the spot market, medium- and long-term contracts are decomposed into daily contracts, which serve as monthly constraints.

[0060] ;in, This represents the lower limit of the power generation declared by power generators during time period t. This refers to the maximum power generation capacity declared by the power generator during time period t. This refers to the amount of electricity generated by the power generator in month m.

[0061] Carbon quota settlement constraints: According to the carbon quota trading system, the annual carbon quota trading volume of a power generator shall not exceed the initial quota and the net carbon emissions from carbon quota trading.

[0062] ; in, This represents the carbon emissions of thermal power units during the t-th period of month m. For the initial carbon allowance, This represents the amount of carbon allowances purchased by power generators during the t-th period of month m. This represents the amount of carbon allowances sold by generators during the t-th period of month m.

[0063] Price constraints for power generators: ; ; in, For power generators, the segmented bidding prices in the electricity spot market are represented by k, where k is the segment number. This is the lower limit of the price for the k-th segment. This represents the upper limit of the price for the k-th segment.

[0064] Unit output constraints: This formula represents the upper and lower limits of the operating output of thermal power units, which must be satisfied to operate within the specified range.

[0065] in, Let t be the output of the thermal power unit. This represents the lower limit of the output of thermal power units. This is the upper limit of the output of thermal power units.

[0066] Energy storage dynamic constraints: ; in, Let be the remaining capacity of the energy storage device at time d; This represents the remaining capacity of the energy storage device at time d-1. The charging efficiency of energy storage devices. The discharge efficiency of energy storage devices. Let be the amount of charge the energy storage device receives at time d. Let be the discharge amount of the energy storage device at time d. This represents the minimum remaining capacity of the energy storage device. This represents the maximum remaining capacity of the energy storage device.

[0067] For example, the lower layer of the two-layer scheduling model in the spot market stage is: joint clearing of the power generation and carbon market by power generators. In the spot market stage, power generators participate in the clearing of the day-ahead market and the secondary carbon market, with the optimization objective of maximizing social welfare.

[0068] The objective function is: ; in, Let d be the electricity consumption utility function for user i during time period d; Let i be the power generation cost for generator i during time period d.

[0069] Lower-level clearing differs from upper-level transactions, primarily considering the constraints during the clearing process.

[0070] Power balance constraints: Power generators use segmented pricing in the spot market, where their bids must satisfy the following: ; in, The actual contribution of power generators during period d; This is for the user's day-ahead load demand.

[0071] Renewable energy output restrictions: ; in, For wind power generators, the actual cleared electricity during period d, This represents the actual cleared electricity volume of photovoltaic power generators during period d. This represents the predicted power generation of wind power generators during period d. This represents the predicted power generation of the photovoltaic power generator during time period d.

[0072] S1023. Based on the objective function for optimizing carbon costs in the medium and long term, and the two-layer scheduling model in the spot market, a multi-scale collaborative framework is constructed.

[0073] S103. Based on a multi-scale collaborative framework, optimize the medium- and long-term carbon trading costs and solve for the daily power generation curve and carbon emission plan in the medium- and long-term stages.

[0074] As one possible implementation, step S103 can be specifically implemented as steps S1031-S1034.

[0075] S1031. Based on the objective function for optimizing carbon trading costs in the medium and long term, and guided by maximizing net revenue, construct an objective function that includes both electricity revenue and carbon emission costs.

[0076] S1032. Based on the objective function of dual calculation mechanism of electricity revenue and carbon emission cost, as well as multi-source unit coordination constraints and carbon quota trading security boundary, a nonlinear carbon emission cost function is constructed.

[0077] S1033. The nonlinear carbon emission cost function is transformed into a mixed integer linear function through piecewise linearization technology.

[0078] S1034. Based on mixed integer linear functions, the daily power generation curve and carbon emission plan for the medium and long term are obtained by solving.

[0079] S104. Based on the daily power generation curve and carbon emission plan in the medium and long term, a two-level dispatch model is adopted to optimize the dispatch of power generation on the spot market side, so as to obtain the power generation curve in the spot market.

[0080] As one possible implementation, step S104 can be specifically implemented as steps S1041-S1044.

[0081] S1041. Based on the daily power generation curves and carbon emission plans in the medium and long term, dynamic decomposition is performed to obtain the power output prediction sequence.

[0082] In some embodiments, the power output prediction sequence includes a thermal power output prediction sequence, a wind power output prediction sequence, and a photovoltaic power output prediction sequence.

[0083] For example, step S1041 can be specifically implemented as steps A1-A3.

[0084] A1. A BP neural network with optimized weights and thresholds using a genetic algorithm is used to predict the wind and solar power output in the medium to long term, and the predicted output results are obtained.

[0085] A2. Based on the predicted power output results, the monthly contracted electricity volume of renewable energy is proportionally decomposed into daily volumes to obtain the wind power output prediction sequence and the photovoltaic power output prediction sequence.

[0086] A3. Decompose the remaining electricity demand, excluding the monthly contracted electricity volume of renewable energy, into the daily power generation of thermal power according to the equilibrium schedule to obtain the thermal power output forecast sequence.

[0087] For example, in the decomposition of medium- and long-term contracts, due to the significant uncertainty of wind and solar power output, which varies at different times, the medium- and long-term contract decomposition model constructed in this paper will prioritize decomposition to trading days with high wind and solar power output to ensure the consumption of renewable energy. Furthermore, it still guarantees the effective execution of conventional thermal power contract electricity volumes.

[0088] The breakdown of medium- and long-term renewable energy contract volumes is based on the proportion of predicted output per day: ;in, This refers to the daily power generation after the decomposition of renewable energy sources. Monthly electricity generation from renewable energy sources. This represents the predicted renewable energy generation for period d.

[0089] The breakdown of medium- and long-term contracted electricity volume for thermal power will be handled by thermal power to fill any remaining demand, while ensuring a balanced progress: ;in, This represents the daily power generation on day d after the thermal power plant is decomposed. This represents the monthly contracted electricity volume after the thermal power plant is decomposed.

[0090] S1042. Based on the power output prediction sequence, as well as the actual power output of wind power and photovoltaic power, perform energy storage correction and determine the energy storage charging and discharging sequence.

[0091] For example, energy storage collaborative correction uses energy storage to smooth out deviations in wind and solar power output and calculates charging and discharging demand.

[0092] ; in, This represents the deviation value of wind and solar power output. The amount of charge for energy storage. This refers to the amount of energy discharged from the stored energy source. This refers to the daily power generation of the wind turbine. For daily photovoltaic power generation, This is the predicted daily power generation value of the wind turbine. This represents the predicted daily power generation from photovoltaic systems.

[0093] S1043. Using the power output prediction sequence and energy storage charging and discharging sequence as the inputs to the upper-level dispatch model in the two-level dispatch model, the power generation side dispatch price as the output of the upper-level dispatch model, the power generation side dispatch price and user load demand as the inputs to the lower-level clearing model, and the actual power output plan of the power generation units as the output, the two-level dispatch model is solved to obtain the power generation curve in the spot market stage.

[0094] For example, step S1043 can be specifically implemented as steps B1-B4.

[0095] B1. Transform the lower market clearing problem into equivalent constraints based on KKT conditions.

[0096] B2. Introduce equivalent constraints into the upper-level optimization problem, transforming the original two-level model into a single-level mathematical programming problem, thus obtaining a single-level mathematical model.

[0097] B3. By introducing auxiliary integer variables and the Big M method, the nonlinear terms in the single-layer mathematical model are linearized to obtain a mixed-integer linear model.

[0098] B4. By using a mixed-integer linear model, linear programming is performed to obtain the power generation curve in the spot market stage.

[0099] In some embodiments, the constructed multi-timescale power generation curve optimization model for power generators covers both medium-to-long-term and spot timescales. In the medium-to-long-term stage, with cost minimization as the objective, a linear objective function is constructed by integrating power generation revenue and carbon emission costs, considering multi-dimensional physical and market rules such as unit output constraints, energy storage dynamic characteristics, and carbon quota trading volume constraints. The energy storage energy equation is linearized, and the nonlinear constraints are transformed into a mixed-integer linear programming (MILP) problem. Then, the global optimization problem is decomposed into monthly subproblems and solved in parallel. The constraint matrix is ​​used to construct and map variables to achieve an efficient expression of spatiotemporal coupling constraints. Finally, the linprog solver in MATLAB is used to obtain the global optimal solution and output the monthly power generation curve that satisfies the power supply and demand balance, unit operation limitations, and carbon quota constraints.

[0100] In the spot market phase, the constructed optimization model for the power generation curve of power generators in the spot market phase is a two-level mixed-integer nonlinear model, which is difficult to solve directly for its global optimum. Therefore, we first use KKT theorem to reformulate the two-level problem; then, we transform the lower-level market clearing problem into upper-level constraints, and use the KKT conditions of the lower-level problem to transform the two-level optimization into a single-level mathematical programming problem; then, we use a linearization approach to transform the single-level mixed-integer nonlinear model into a mixed-integer linear programming problem, and use the CPLEX solver in MATLAB to solve the problem.

[0101] Model transformation of the lower-level problem: The lower-level clearing model is transformed into a single-level nonlinear programming problem based on KKT conditions. A Lagrangian function is constructed for the lower-level problem, and the KKT conditions are obtained by taking the derivative.

[0102] ; in, For Lagrange multipliers constrained by power balance; For the Lagrange multipliers constraining the upper limit of wind power output, Lagrange multipliers constraining the upper limit of photovoltaic power output; Lagrange multipliers constraining the upper limit of energy storage charging power. Lagrange multipliers constraining the lower limit of energy storage charging power. For the Lagrange multipliers constrained by the upper limit of energy storage discharge power, The Lagrange multiplier is used to constrain the lower limit of the energy storage discharge power. For Lagrange functions, Let i be the electricity consumption utility function. For the power generation cost of power generator i, This represents the cumulative actual output of the power generator during period d. This represents the cumulative value of the user's met-up demands over the past few days. This is the predicted daily power generation value of the wind turbine. This represents the actual daily power generation of the wind turbine. This is the predicted daily photovoltaic power generation value. This represents the actual daily power generation from photovoltaic power plants. For energy storage discharge capacity, This represents the maximum energy storage discharge capacity. For energy storage charging capacity, This represents the maximum energy storage charging capacity.

[0103] By taking the partial derivatives with respect to the decision variables, we obtain the necessary conditions and complementary relaxation conditions of KKT.

[0104] ; ; ; The linearization solution of nonlinear problems uses the KKT conditions of the lower level as constraints for the upper level optimization problem, forming a single-level nonlinear programming model, in which the nonlinear terms are solved linearly.

[0105] Carbon emissions and carbon quotas can be represented by introducing 0-1 variables to indicate the cost direction, and the Big M method can be used to handle inequality constraints. Let... ,when The time indicated that power generators sell carbon allowances. When it is indicated that carbon allowances are being purchased, then: ; ; Where M is a sufficiently large constant. This represents the amount of carbon quota sold during period t. This represents the amount of carbon quota purchased during period t. Let t be the change in carbon quota during time period t.

[0106] The upper-level decision variables are updated by iteratively optimizing the generator's power generation curve. Initially, an initial power generation curve is given to the generator. Then, based on the lower-level market clearing results, the bid is continuously adjusted to gradually reduce the generator's costs. In each iteration, the parameters of the lower-level market are fixed, and the upper-level generator cost minimization problem is solved. Based on the current bid, the lower-level market clearing model is solved to obtain the electricity price, carbon price, and output of each unit. Given the generator's bid, the lower-level market clearing model with the goal of maximizing social welfare is solved to obtain the market equilibrium price and the optimal output of each unit.

[0107] S105. Based on the daily power generation curves in the medium- and long-term phases and the power generation curves in the spot phase, generate an optimized power dispatch scheme for the power generation side.

[0108] As one possible implementation, step S105 can be specifically implemented as steps S1051-S1054.

[0109] S1051. Merge the medium- and long-term daily power generation curves with the spot power generation curves along the time axis to generate the final power generation curve on the power generation side.

[0110] S1052. Based on the final power generation curve on the power generation side, calculate the deviation rate between the actual carbon trading cost and the medium- and long-term predicted cost, the comprehensive revenue deviation rate, the actual carbon emissions and quota usage, and the proportion of clean energy. S1053. Based on the deviation rate between actual carbon trading costs and medium- and long-term predicted costs, the comprehensive benefit deviation rate, actual carbon emissions and quota usage, and the proportion of clean energy, generate benefit assessment results.

[0111] In some embodiments, the benefit assessment results include the overall increase in revenue, carbon allowance utilization rate, and renewable energy integration rate.

[0112] S1054. Based on the final power generation curve on the power generation side and the benefit evaluation results, generate an optimization scheme for power dispatch on the power generation side.

[0113] This invention provides a method and apparatus for optimizing power generation scheduling on the generation side, considering multiple time scales. By constructing a unified multi-scale collaborative framework, this invention organically combines power generation curve optimization in the medium-to-long-term and spot market phases, overcoming the problem of disconnect between the two phases in traditional scheduling. By generating daily power generation curves in the medium-to-long-term phase and performing refined rolling optimization in the spot market phase, the final power generation scheme not only conforms to medium-to-long-term macroeconomic objectives, such as total carbon quotas, but also adapts to real-time operating conditions in the spot market, reducing the deviation between planned and actual execution and improving the accuracy of power generation scheduling. This invention solves the problem of inaccurate power generation planning on the generation side and improves the accuracy of power generation planning across multiple time scales.

[0114] Furthermore, this invention divides the transaction into a medium-to-long-term trading phase and a spot trading phase based on the time scale.

[0115] In the medium to long term, power generators need to simultaneously consider electricity sales revenue and annual carbon quota compliance, with the goal of maximizing their overall revenue. The model assumes that power generators have already established monthly medium to long-term contracts through bilateral negotiations and centralized bidding. Therefore, the key decision-making focus in the medium to long term is minimizing carbon trading costs. Specifically, high-carbon-emission, low-power-cost thermal power plants will generate electricity when carbon prices are low, thus meeting compliance requirements while reducing carbon costs.

[0116] Spot Market Stage: The spot market stage involves the coupled trading optimization of the electricity spot market and the secondary trading of the carbon market. The upper layer is the generator dispatch model, and the lower layer is the market joint clearing optimization. The upper-layer trading decision takes into account the uncertainty of wind and solar power, aims to minimize dispatch costs, and is constrained by the decomposition of monthly medium- and long-term electricity trading contracts; the lower-layer market clearing aims to maximize social welfare. The optimization of the upper and lower layers constitutes the daily power generation curve of generators in the electricity-carbon market environment.

[0117] In the medium- to long-term stage, the optimization model for power generation curves aims to maximize profits over a long timescale for power generators. In this paper, power generators have already drafted medium- to long-term contracts, thus their monthly electricity generation is fixed. However, considering that participation in secondary carbon market trading is a flexible activity for power generators, a fixed monthly carbon quota constraint is not imposed. Power generators calculate their projected carbon quotas based on electricity generation and make flexible decisions based on actual market price fluctuations. In other words, by locking in electricity revenue through medium- to long-term contracts, profits are maximized while minimizing carbon emission costs.

[0118] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0119] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0120] Figure 3 This diagram illustrates a structural schematic of a power generation-side scheduling optimization device considering multiple time scales, according to an embodiment of the present invention. The optimization device 200 includes a communication module 201 and a processing module 202.

[0121] Communication module 201 is used to acquire medium- and long-term stage data and spot stage data.

[0122] Processing module 202 is used to construct a multi-scale collaborative framework based on medium- and long-term stage data and spot stage data, with the goal of minimizing the carbon trading cost in the medium- and long-term stage and the generation-side dispatch cost in the spot stage. Based on the multi-scale collaborative framework, it optimizes the medium- and long-term carbon trading cost and solves the daily power generation curve and carbon emission plan in the medium- and long-term stage. Based on the daily power generation curve and carbon emission plan in the medium- and long-term stage, it adopts a two-level dispatch model to optimize the power dispatch on the generation side in the spot stage and obtain the power generation curve in the spot stage. Based on the daily power generation curve in the medium- and long-term stage and the power generation curve in the spot stage, it generates an optimized power dispatch scheme for the generation side.

[0123] Figure 4This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The electronic device 300 includes: a processor 301, a memory 302, and a computer program 303 stored in the memory 302 and executable on the processor 301. When the processor 301 executes the computer program 303, it implements the steps in the above-described method embodiments. Alternatively, when the processor 301 executes the computer program 303, it implements the functions of each module / unit in the above-described device embodiments.

[0124] For example, the computer program 303 may be divided into one or more modules / units, which are stored in the memory 302 and executed by the processor 301 to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program 303 in the electronic device 300.

[0125] The processor 301 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0126] The memory 302 can be an internal storage unit of the electronic device 300, such as a hard disk or memory of the electronic device 300. The memory 302 can also be an external storage device of the electronic device 300, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device 300. Furthermore, the memory 302 can include both internal and external storage units of the electronic device 300. The memory 302 is used to store the computer program and other programs and data required by the terminal. The memory 302 can also be used to temporarily store data that has been output or will be output.

[0127] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for optimizing power dispatch on the generation side considering multiple time scales, characterized in that, include: Acquire medium- and long-term data and spot market data, and based on these data, construct a multi-scale collaborative framework with the goal of minimizing carbon trading costs in the medium- and long-term phase and minimizing power generation-side dispatch costs in the spot market phase. Based on the aforementioned multi-scale collaborative framework, medium- and long-term carbon trading costs are optimized, and the daily power generation curve and carbon emission plan for the medium- and long-term stages are solved. Based on the daily power generation curves and carbon emission plans in the medium and long term, a two-level dispatch model is adopted to optimize the dispatch of power generation on the power generation side in the spot market, so as to obtain the power generation curve in the spot market. Based on the daily power generation curves in the medium- and long-term phases and the power generation curves in the spot phases, an optimized power dispatch scheme for the power generation side is generated.

2. The power generation scheduling optimization method considering multiple time scales as described in claim 1, characterized in that, Based on medium- and long-term data and spot market data, and with the goal of minimizing carbon trading costs in the medium- and long-term phases and minimizing power generation-side dispatch costs in the spot market, a multi-scale collaborative framework is constructed, including: Based on medium- and long-term data, and with the goal of minimizing carbon trading costs, a medium- and long-term carbon cost optimization objective function is constructed, with constraints on multi-source unit collaboration and carbon quota trading safety boundaries. The multi-source unit collaboration constraints include wind power output constraints, photovoltaic power output constraints, thermal power ramp-up constraints, and energy storage constraints. The carbon quota trading safety boundaries include carbon trading volume limits and annual compliance constraints. Based on spot market data, a two-layer scheduling model is constructed with the goal of minimizing the dispatch cost on the generation side in the spot market as the objective of the upper-layer scheduling model and maximizing the social welfare of the joint clearing of the electricity carbon market as the objective of the lower-layer clearing model. Based on the aforementioned medium- and long-term carbon cost optimization objective function and the two-layer scheduling model in the spot market, a multi-scale collaborative framework is constructed.

3. The power generation scheduling optimization method considering multiple time scales according to claim 2, characterized in that, The wind power output constraint is used to limit the operating range of wind power output; the photovoltaic output constraint is used to limit the operating range of photovoltaic output; the thermal power ramping constraint is used to limit the operating range of thermal power output; the thermal power ramping constraint is used to limit the range of output variation of thermal power units at adjacent times; the energy storage constraint is used to limit the capacity and charging / discharging power of energy storage devices; the carbon trading volume limit is used to limit the total monthly carbon allowance trading volume; the annual compliance constraint is used to ensure that the actual annual carbon emissions do not exceed the allocated initial carbon allowance.

4. The power generation scheduling optimization method considering multiple time scales as described in claim 1, characterized in that, The optimization of medium- and long-term carbon trading costs based on the multi-scale collaborative framework, and the solution of daily power generation curves and carbon emission plans in the medium- and long-term stages, include: Based on the aforementioned medium- and long-term carbon trading cost optimization objective function, and guided by maximizing net revenue, an objective function is constructed that includes a dual calculation mechanism for electricity revenue and carbon emission costs. Based on the objective function of the dual calculation mechanism of electricity revenue and carbon emission cost, as well as the multi-source unit coordination constraints and carbon quota trading security boundary, a nonlinear carbon emission cost function is constructed. The nonlinear carbon emission cost function is transformed into a mixed integer linear function using piecewise linearization techniques. Based on the aforementioned mixed-integer linear function, the daily power generation curve and carbon emission plan for the medium and long term are obtained by solving the problem.

5. The power generation scheduling optimization method considering multiple time scales according to claim 1, characterized in that, The aforementioned daily power generation curve and carbon emission plan based on the medium- and long-term phases employ a two-layer dispatch model to optimize the power dispatch on the power generation side during the spot phase, resulting in the power generation curve for the spot phase, including: Based on the daily power generation curves and carbon emission plans in the medium and long term, dynamic decomposition is performed to obtain the power output prediction sequence, which includes the thermal power output prediction sequence, the wind power output prediction sequence, and the photovoltaic power output prediction sequence. Based on the predicted output sequence, as well as the actual wind power output and actual photovoltaic output, energy storage correction is performed to determine the energy storage charging and discharging sequence. Using the power output prediction sequence and the energy storage charging and discharging sequence as inputs to the upper-level scheduling model in the two-level scheduling model, the generation-side scheduling price as the output of the upper-level scheduling model, the generation-side scheduling price and user load demand as inputs to the lower-level clearing model, and the actual power output plan of the generation-side units as the output, the two-level scheduling model is solved to obtain the power generation curve in the spot market stage.

6. The power generation scheduling optimization method considering multiple time scales according to claim 5, characterized in that, The daily power generation curves and carbon emission plans based on medium- to long-term phases are dynamically decomposed to obtain a power output prediction sequence, including: A BP neural network with optimized weights and thresholds using a genetic algorithm is used to predict wind and solar power output in the medium to long term, yielding the predicted output results. Based on the predicted output results, the monthly contracted electricity volume of renewable energy is proportionally allocated to each day to obtain the wind power output prediction sequence and the photovoltaic power output prediction sequence. The remaining electricity demand, excluding the monthly contracted electricity volume of renewable energy, is decomposed into the daily power generation of thermal power according to the equilibrium schedule, thus obtaining the thermal power output forecast sequence.

7. The power generation scheduling optimization method considering multiple time scales according to claim 5, characterized in that, The two-layer scheduling model is solved by using the power output prediction sequence and the energy storage charging and discharging sequence as inputs to the upper-layer scheduling model, the generation-side scheduling price as the output of the upper-layer scheduling model, the generation-side scheduling price and user load demand as inputs to the lower-layer clearing model, and the actual power output plan of the generation-side units as the output, to obtain the power generation curve in the spot market stage, including: The problem of clearing out the lower market is transformed into an equivalent constraint based on KKT conditions; By introducing the equivalent constraints into the upper-level optimization problem, the original two-layer model is transformed into a single-layer mathematical programming problem, resulting in a single-layer mathematical model. By introducing auxiliary integer variables and the Big M method, the nonlinear terms in the single-layer mathematical model are linearized to obtain a mixed-integer linear model. By using a mixed-integer linear model and solving linear programming problems, the power generation curve for the spot market stage can be obtained.

8. The power generation scheduling optimization method considering multiple time scales according to claim 1, characterized in that, The generation optimization scheme for power dispatch on the generation side, based on the daily power generation curves of the medium- and long-term phases and the power generation curves of the spot phase, includes: By merging the medium- and long-term daily power generation curves with the spot power generation curves along the time axis, the final power generation curve on the power generation side is generated. Based on the final power generation curve of the power generation side, calculate the deviation rate between the actual carbon trading cost and the medium- and long-term predicted cost, the comprehensive revenue deviation rate, the actual carbon emissions and quota usage, and the proportion of clean energy. Based on the deviation rate between the actual carbon trading cost and the medium- and long-term predicted cost, the comprehensive benefit deviation rate, the actual carbon emissions and quota usage, and the proportion of clean energy, a benefit assessment result is generated. The benefit assessment result includes the comprehensive benefit increase, carbon quota utilization rate, and renewable energy consumption rate. Based on the final power generation curve of the power generation side and the benefit evaluation results, an optimization scheme for power dispatching on the power generation side is generated.

9. A generator-side power dispatch optimization device considering multiple time scales, characterized in that, include: The communication module is used to acquire medium- and long-term data and spot market data. The processing module is used to construct a multi-scale collaborative framework based on medium- and long-term stage data and spot stage data, with the goal of minimizing carbon trading costs in the medium- and long-term stage and minimizing power generation-side dispatch costs in the spot stage. Based on the aforementioned multi-scale collaborative framework, medium- and long-term carbon trading costs are optimized by solving for the daily power generation curve and carbon emission plan in the medium- and long-term stages. Based on the daily power generation curve and carbon emission plan in the medium- and long-term stages, a two-layer scheduling model is adopted to optimize the power dispatch on the power generation side in the spot stage, thereby obtaining the power generation curve in the spot stage. Based on the daily power generation curve in the medium- and long-term stages and the power generation curve in the spot stage, an optimized power dispatch scheme for the power generation side is generated.

10. The power generation-side dispatch optimization device considering multiple time scales according to claim 9, characterized in that, The processing module is specifically used to construct a medium-to-long-term carbon cost optimization objective function based on medium-to-long-term data, with the goal of minimizing carbon trading costs and constraints on multi-source unit coordination and carbon quota trading security boundaries. The multi-source unit coordination constraints include wind power output constraints, photovoltaic power output constraints, thermal power ramp-up constraints, and energy storage constraints. The carbon quota trading security boundaries include carbon trading volume limits and annual compliance constraints. Based on spot market data, a two-tiered scheduling model is constructed, with the goal of minimizing the generation-side dispatch cost in the spot market as the objective of the upper-level scheduling model and maximizing the social welfare of the joint clearing of the electricity-carbon market as the objective of the lower-level clearing model. Finally, a multi-scale collaborative framework is constructed based on the medium-to-long-term carbon cost optimization objective function and the two-tiered scheduling model in the spot market.

Images