Power-carbon emission-green certificate multi-time scale coordinated rolling optimization method, device, equipment and medium considering carbon emission uncertainty

Abstract

This application discloses a method, apparatus, equipment, and medium for multi-timescale collaborative rolling optimization of electricity, carbon emissions, and green certificates, considering carbon emission uncertainty. It relates to the field of computer technology and includes: simulating power output changes in a first target time period based on historical wind, solar, and hydropower data; adjusting the annual operation plan using the obtained first simulated operation scenario; performing time-series production simulation based on the obtained target operation plan and target simulation constraints; determining the capacity configuration scheme of the target reserve pool for carbon quotas and green certificate plans using the obtained renewable energy penetration data; generating a second simulated operation scenario for a second target time period based on the power system's operational boundary data and using the XGBOOST algorithm; and performing multi-timescale collaborative rolling optimization of electricity, carbon emissions, and green certificates using a pre-set collaborative and mutually supportive mechanism to obtain a target executable plan. Through collaborative optimization of electricity and carbon certificates at different time scales, the uncertainty of carbon emissions is reduced.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a method, apparatus, equipment and medium for multi-timescale collaborative rolling optimization of electricity, carbon emissions and green certificates that takes into account the uncertainty of carbon emissions. Background Technology

[0002] As a crucial policy tool for addressing climate change, the carbon market has gained widespread application globally due to its economic efficiency and sustainability. The power industry, as a major participant in the carbon market, plays a vital role in achieving the "dual carbon" goals through its green and low-carbon transformation. Simultaneously, the power generation sector also needs to participate in the daily operation of the electricity market. The consistency of market participants and the convergence of emission reduction targets provide a foundation for the coordinated development of the carbon and electricity markets, and the green certificate market is gradually maturing. However, the uncertainty of carbon emissions remains a significant obstacle to the low-carbon transformation of the power system, making the balance of power generation and consumption in the medium to long term and in the short term even more difficult.

[0003] Current technologies generally treat carbon quota constraints as a constraint in an optimization problem at a specific time scale, limiting it by ensuring that the electricity generation multiplied by the carbon emission factor is less than the carbon quota value. Some methods further consider the possibility that some green certificates can be converted into carbon quotas, changing the constraint to the electricity generation multiplied by the carbon emission factor plus the redeemable carbon quotas being less than the carbon quota value. However, while this approach can ensure the optimization result meets the carbon quota limit, it fails to consider the varying timeframes of different markets. Carbon quotas are settled on an annual basis, green certificate markets on a monthly basis, and electricity markets on an annual basis. Short-term spot markets require day-ahead or real-time balancing, leading to a disconnect between electricity, carbon, and certificates in the electricity balance optimization process and failing to fully leverage market advantages. Furthermore, existing methods considering the coupling of electricity and carbon constraints are often limited to a specific time scale, such as one year or one day. Even when considering multi-time scale interactions, they often only pass the results from longer time scales to lower time scale programs as boundary conditions, failing to account for the inherent randomness of carbon emissions due to the uncertainty of new energy sources. This can result in discrepancies between actual and expected boundaries, leading to insufficient carbon quotas.

[0004] As can be seen from the above, how to reduce the uncertainty of carbon emissions through the synergistic optimization of different time scales of carbon certificates is an urgent problem to be solved. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a method, apparatus, device, and medium for multi-timescale collaborative rolling optimization of electricity-carbon emissions-green certificates, considering carbon emission uncertainty. This method can reduce the uncertainty of carbon emissions through collaborative optimization of electricity carbon certificates at different timescales. The specific solution is as follows: Firstly, this application provides a multi-timescale collaborative rolling optimization method for electricity, carbon emissions, and green certificates that considers carbon emission uncertainty, including: Historical wind, solar, and hydropower data, along with operational boundary data of the power system, are collected. Based on this historical data, the power output changes within a first target time period are simulated to obtain a corresponding first simulated operational scenario. The historical wind, solar, and hydropower data includes wind turbine output data, photovoltaic output data, and hydropower station inflow data. The operational boundary data includes relevant data on electricity, carbon emissions, and green certificates. The annual operation plan is adjusted using the first simulated operation scenario to obtain the target operation plan; the annual operation plan is an operation plan determined based on the annual electricity, power generation and carbon quota. Based on the target operation plan and target simulation constraints, time-series production simulation is performed to obtain the penetration of new energy sources in each period, and the carbon emission uncertainty coefficient in each period is determined using the new energy penetration data. The capacity configuration scheme of the target reserve pool of the carbon quota and green certificate plan is then determined using the carbon emission uncertainty coefficient. Based on the historical wind, solar, and hydropower data and the operational boundary data, and using the XGBOOST algorithm, a second simulated operational scenario corresponding to the second target time period is generated. Based on the capacity configuration scheme, the first simulated operational scenario, and the second simulated operational scenario, and using a preset collaborative and mutually supportive mechanism, multi-timescale collaborative rolling optimization of electricity, carbon emissions, and green certificates is performed to obtain a corresponding target executable plan. The time length corresponding to the first target time period is greater than the time length corresponding to the second target time period. The target executable plan includes a power generation plan, a carbon quota usage plan, and a green certificate circulation plan.

[0006] Optionally, the collection of historical wind, solar, and hydropower data and power system operational boundary data, and the simulation of power output changes within a first target time period based on the historical wind, solar, and hydropower data, to obtain a corresponding first simulated operational scenario, includes: The wind turbine output data, photovoltaic output data, and electricity load data are normalized and filtered to obtain processed data. The K-Means algorithm is then used to classify and cluster the processed data to obtain the corresponding target curve. The Markov transition matrix corresponding to the target curve is statistically analyzed, and the Markov transition matrix is ​​sequentially sampled and superimposed to generate the corresponding landscape scene. Based on the inflow data of the hydropower station and using the ARIMA model to predict the time-series runoff curve in the first time period in the future, the time-series runoff curve is decomposed to obtain the corresponding decomposition results. The decomposition results and seasonal factors are used to generate corresponding hydropower station scenarios, and the corresponding first simulated operation scenario is constructed based on the wind and solar scenarios and the hydropower station scenarios.

[0007] Optionally, adjusting the annual operation plan using the first simulated operation scenario to obtain the target operation plan includes: The annual power surplus / deficit data are determined based on the adjustable power, power transmission and reception, reserve power, and power load demand within the target area; the adjustable power within the target area is determined based on the total capacity, maintenance capacity, and out-of-use capacity of all generator units within the target area. The annual electricity surplus / deficit data is determined using the total power generation, transmitted and received power, abandoned water power, and total user electricity demand within the target area. The total power generation within the target area is determined based on the adjustable power within the target area, the annual power generation hours, and the under-full-load coefficient. The under-full-load coefficient is the ratio of the actual output of the generator set to its rated output. Carbon allowance profit and loss data are determined based on carbon allowances, carbon emission factors, actual annual power generation of thermal power units, carbon emission correction factor, theoretical annual generation of green certificates, exchange factor between green certificates and carbon allowances, and green certificate correction factor within the target area. The corresponding profit and loss situation is determined based on the annual electricity profit and loss data, the annual electricity volume profit and loss data, and the carbon quota profit and loss data. If the profit and loss situation is not within the target profit and loss range, the power generation side and power consumption side in the annual operation plan are adjusted using a preset adjustment strategy until the profit and loss situation corresponding to the adjusted operation plan is within the target profit and loss range, so as to obtain the target operation plan.

[0008] Optionally, the target simulation constraints include backup power constraints, power balance constraints, transmission line constraints, unit power generation output constraints, unit ramp rate constraints, unit operating status constraints, energy storage operation constraints, hydropower operation constraints, and new energy output constraints. The constraints are as follows: the reserve power constraint is that the actual reserve capacity is not less than the required reserve capacity; the power balance constraint is that the sum of the total electricity consumption and the total transmission volume is equal to the total power generation volume; the total power generation volume includes the output of thermal power units, renewable energy output, hydropower unit output, pumped storage power station output, and nuclear power output; the transmission line constraint is that the actual transmission power of the transmission line is within the target transmission power range; the generating unit output constraint is that the actual output of the generating unit is within the target generating unit output range; the generating unit ramp rate constraint is that the output power of the generating unit per unit time is within the target ramp power range; the generating unit operating status constraint is that the number of generating units in operation is consistent with the dispatch instructions; the energy storage operation constraint is that the current energy storage capacity is within the target energy storage range, and the energy storage status is charging or discharging; the hydropower operation constraint is that the hydropower generation water consumption and reservoir capacity meet the preset matching conditions; and the renewable energy output constraint is that the renewable energy output does not exceed the target renewable energy output threshold.

[0009] Optionally, the step of determining the carbon emission uncertainty coefficient for each time period using the new energy penetration situation, and determining the capacity configuration scheme of the target reserve pool for the carbon quota and green certificate plan using the carbon emission uncertainty coefficient, includes: The new energy penetration rate is determined by the ratio of new energy power generation to actual electricity consumption in the aforementioned new energy penetration situation. The carbon emission uncertainty coefficient for each time period is determined based on the new energy penetration rate and the corresponding target upper and lower limit thresholds. The planned values ​​for the annual public green certificate pool and the annual public carbon quota pool are determined using the carbon emission uncertainty coefficient and the preset carbon emission uncertainty prevention coefficient. The capacity configuration scheme of the target reserve pool for the carbon quota and green certificate plan is determined based on the planned value of the annual public green certificate pool and the planned value of the annual public carbon quota pool.

[0010] Optionally, the step of performing multi-timescale collaborative rolling optimization of electricity, carbon emissions, and green certificates based on the capacity configuration scheme, the first simulated operation scenario, and the second simulated operation scenario, and utilizing a preset collaborative and mutually supportive mechanism, to obtain a corresponding target executable plan, includes: A target-level rolling optimization model is constructed with annual, monthly, weekly, and daily windows. Based on the target-level rolling optimization model and the annual public green certificate pool plan value and annual public carbon quota pool plan value in the capacity configuration scheme, the first simulated operation scenario is used as the boundary input for annual window optimization, and a first objective function is constructed to minimize the annual comprehensive cost. The annual comprehensive cost is the cost determined based on the circulation cost of electricity, carbon emissions, and green certificates, the unit operating cost, and the transmission cost throughout the year. The first objective function is solved based on the circulation constraints of electricity, carbon emissions, and green certificates, contract deviation penalties, and non-contract penalties to obtain the annual window optimization results including the annual carbon emission plan and the annual contracted electricity circulation. Based on the target-level rolling optimization model, the annual window optimization result is used as the boundary input for the monthly window optimization, and a second objective function is constructed by minimizing the monthly circulation cost. The second objective function is solved using the boundary violation penalty term to obtain the monthly window optimization results, which include the monthly carbon emission plan, the monthly green certificate plan, and the monthly subdivided electricity flow. Based on the target-level rolling optimization model, the monthly window optimization result is used as the boundary input of the weekly window optimization to construct a third objective function. The target additional reserve constraint corresponding to the target special date is used to solve the third objective function to obtain the weekly window optimization result including the weekly carbon emission plan, the weekly green certificate plan, and the weekly subdivided circulating electricity. Based on the target-level rolling optimization model, the weekly window optimization result is used as the boundary input of the daily window optimization to construct a fourth objective function. The fourth objective function is solved to obtain the daily window optimization results, including the daily power generation plan, the daily carbon quota usage plan, and the daily green certificate circulation plan. Based on a preset collaborative and mutually supportive mechanism, the annual window optimization results, the weekly window optimization results, the monthly window optimization results, and the daily window optimization results are collaboratively and continuously optimized to obtain the corresponding target executable plan.

[0011] Optionally, the step of performing collaborative rolling optimization on the annual window optimization results, the weekly window optimization results, the monthly window optimization results, and the daily window optimization results based on a preset collaborative mutual assistance mechanism to obtain the corresponding target executable plan includes: The solution process for each window optimization result is monitored to obtain the corresponding monitoring results; If the monitoring result indicates that the solution corresponding to the target window optimization result has a feasible solution, and there is a target window with a carbon quota gap or a green certificate gap, then the target window is gap-filled based on the capacity configuration scheme and the candidate windows in the target window optimization result to obtain the window optimization result after filling; the candidate window is the window corresponding to the target window optimization result that has a carbon quota surplus and a green certificate surplus. If the monitoring result indicates that the solution corresponding to the target window optimization result is unsolvable, or there is an unfillable carbon quota gap or green certificate gap in the target window, then the corresponding gap data is fed back to the upper-level window of the target window, so as to adjust the upper-level window optimization result corresponding to the upper-level window based on the gap data, so as to obtain the adjusted upper-level window optimization result. The adjusted optimization results of the previous window are sent to the target window, and the objective function corresponding to the target window is solved based on the adjusted optimization results of the previous window to obtain the target optimization results. If there is a feasible solution for the solution corresponding to the target optimization results, and there is no target window with a carbon quota gap or green certificate gap, then a corresponding target executable plan is constructed based on the target optimization results.

[0012] Secondly, this application provides a multi-timescale collaborative rolling optimization device for electricity, carbon emissions, and green certificates that considers carbon emission uncertainty, including: The operation scenario simulation module is used to collect historical wind, solar, and hydropower data and power system operation boundary data, and simulate the output changes within a first target time period in the future based on the historical wind, solar, and hydropower data to obtain the corresponding first simulated operation scenario; the historical wind, solar, and hydropower data includes wind turbine output data, photovoltaic output data, and hydropower station inflow data; the operation boundary data includes relevant data on electricity, carbon emissions, and green certificates; The operation plan adjustment module is used to adjust the annual operation plan using the first simulated operation scenario to obtain the target operation plan; the annual operation plan is an operation plan determined based on the annual electricity, power generation and carbon quota. The configuration scheme determination module is used to perform time-series production simulation based on the target operation plan and target simulation constraints to obtain the new energy penetration situation in each period, and use the new energy penetration situation to determine the carbon emission uncertainty coefficient in each period, and use the carbon emission uncertainty coefficient to determine the capacity configuration scheme of the target reserve pool of the carbon quota and green certificate plan. The rolling optimization module is used to generate a second simulated operation scenario corresponding to a future second target time period based on the historical wind, solar, and hydropower data and the operation boundary data, using the XGBOOST algorithm. Based on the capacity configuration scheme, the first simulated operation scenario, and the second simulated operation scenario, and using a preset collaborative and mutually supportive mechanism, it performs multi-time-scale collaborative rolling optimization of electricity, carbon emissions, and green certificates to obtain a corresponding target executable plan. The time length corresponding to the first target time period is greater than the time length corresponding to the second target time period. The target executable plan includes a power generation plan, a carbon quota usage plan, and a green certificate circulation plan.

[0013] Thirdly, this application provides an electronic device, comprising: Memory, used to store computer programs; A processor is used to execute the computer program to implement the aforementioned collaborative rolling optimization method for electricity-carbon emissions-green certificates across multiple time scales, taking into account carbon emission uncertainties.

[0014] Fourthly, this application provides a computer-readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the aforementioned collaborative rolling optimization method for electricity-carbon emissions-green certificates across multiple time scales, taking into account carbon emission uncertainties.

[0015] This application collects historical wind, solar, and hydropower data and power system operation boundary data. Based on the historical wind, solar, and hydropower data, it simulates the output changes within a first target time period in the future to obtain a corresponding first simulated operation scenario. The historical wind, solar, and hydropower data includes wind turbine output data, photovoltaic output data, and hydropower station inflow data. The operation boundary data includes relevant data on electricity, carbon emissions, and green certificates. The first simulated operation scenario is used to adjust the annual operation plan to obtain a target operation plan. The annual operation plan is an operation plan determined based on the annual electricity, power generation, and carbon quotas. Based on the target operation plan and target simulation constraints, time-series production simulation is performed to obtain the penetration of new energy sources in each time period, and the penetration of new energy sources is used to determine... The carbon emission uncertainty coefficient for each time period is used to determine the capacity configuration scheme of the target reserve pool for the carbon quota and green certificate plan. Based on the historical wind, solar and hydropower data and the operational boundary data, and using the XGBOOST algorithm, a second simulated operation scenario corresponding to the second target time period is generated. Based on the capacity configuration scheme, the first simulated operation scenario and the second simulated operation scenario, and using a preset collaborative and mutual assistance mechanism, multi-time-scale collaborative rolling optimization of electricity, carbon emission and green certificates is performed to obtain the corresponding target executable plan. The time length corresponding to the first target time period is greater than the time length corresponding to the second target time period. The target executable plan includes a power generation plan, a carbon quota usage plan and a green certificate circulation plan.

[0016] As can be seen from the above, this application collects historical wind, solar, hydro, and power system operation boundary data to generate a first simulated operation scenario in the medium to long term. This scenario can restore the natural fluctuation characteristics of wind, solar, and hydropower output and the core operational constraints of the power system, providing a realistic basis for planning on an annual scale. Based on the first simulated operation scenario, the application conducts a profit and loss analysis of annual electricity, power generation, and carbon quotas to identify supply and demand gaps and carbon quota surplus / deficit situations in advance, and obtains the target operation plan through adjustments. Based on the target operation plan and target simulation constraints, the application conducts time-series production simulations to calculate the penetration rate of new energy and convert it into a carbon emission uncertainty coefficient. This coefficient is then used to configure the carbon quota and green certificate reserve pool capacity, transforming the abstract carbon emission uncertainty into quantifiable and dispatchable reserve resources. The application uses the XGBOOST algorithm to generate a highly accurate short-term second simulated operation scenario, and combines long-term and short-term scenarios, reserve pool schemes, and collaborative mechanisms to perform multi-timescale rolling optimization on a year, month, day, and week basis. In this way, by combining short-term and long-term scenarios to compensate for the errors of single-scale forecasts, and by optimizing and matching the cycles of different markets through multiple time scales, the final power generation, carbon quotas, and green certificate plans are more in line with actual operating conditions, breaking down the disconnect between electricity, carbon emissions, and green certificates, and effectively addressing the long-term uncertainty of carbon emissions. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, 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 embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0018] Figure 1 This application discloses a flowchart of a multi-timescale collaborative rolling optimization method for electricity, carbon emissions, and green certificates that considers carbon emission uncertainty. Figure 2 This application provides a schematic diagram of the generation of a first simulated running scenario; Figure 3 A schematic diagram of a collaborative rolling optimization mechanism for electricity, carbon emissions, and green certificates provided in this application; Figure 4 A schematic diagram of a multi-timescale collaborative rolling optimization method for electricity, carbon emissions, and green certificates that considers carbon emission uncertainty, provided for this application; Figure 5 A flowchart of a multi-timescale collaborative rolling optimization method for electricity, carbon emissions, and green certificates that takes into account carbon emission uncertainties is provided in this application; Figure 6 This application provides a schematic diagram illustrating the annual electricity supply and demand balance; wherein, Figure 6 (a) is a schematic diagram of the annual electricity profit and loss. Figure 6(b) is a schematic diagram illustrating the annual electricity supply surplus or deficit; Figure 7 A schematic diagram of carbon emission uncertainty coefficient provided for this application; Figure 8 A schematic diagram of current flow results provided in this application; Figure 9 This application provides a schematic diagram illustrating the circulation results of green certificates. Figure 10 A schematic diagram of carbon flow results provided for this application; Figure 11 This application provides a schematic diagram of carbon emissions after using a coordinated mechanism of electricity, carbon emissions, and green certificates. Figure 12 A schematic diagram illustrating the comparison of carbon quotas provided in this application; Figure 13 A scatter plot illustrating the comparison between single-step optimization and collaborative rolling optimization provided for this application; Figure 14 A statistical diagram illustrating the optimization results provided in this application; Figure 15 A schematic diagram of a public carbon quota pool inventory curve provided for this application; Figure 16 This application provides a schematic diagram illustrating the variation of the monthly carbon quota boundary. Figure 17 This is a schematic diagram of a multi-timescale collaborative rolling optimization device for electricity, carbon emissions, and green certificates that takes into account the uncertainty of carbon emissions, as disclosed in this application. Figure 18 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Currently, carbon quota constraints are generally considered as a constraint in an optimization problem at a specific time scale. However, this approach fails to account for the varying cycle lengths of different markets. Carbon quotas are settled on an annual basis, green certificate markets on a monthly basis, and electricity markets on an annual basis. Short-term spot markets also require day-ahead or real-time balancing. This leads to a disconnect between electricity, carbon, and certificates in the electricity balance optimization process, failing to fully leverage market advantages. Furthermore, existing methods for considering the coupling of electricity and carbon constraints are often limited to a specific time scale, such as one year or one day. Even when considering the interaction across multiple time scales, they often only consider passing the results from longer time scales to lower time scale programs as boundary conditions. They fail to account for the inherent randomness of carbon emissions due to the uncertainty of new energy sources, resulting in discrepancies between actual and expected boundaries and ultimately, insufficient carbon quotas. To address this, this application provides a multi-timescale collaborative rolling optimization method for electricity, carbon emissions, and green certificates that considers carbon emission uncertainty. This method combines long-term and short-term scenarios to compensate for the errors in single-scale predictions. By rolling optimization across multiple timescales to match the cycles of different markets, the final power generation, carbon quota, and green certificate plans are made more in line with actual operating conditions. This breaks down the disconnect between electricity, carbon emissions, and green certificates and effectively addresses the long-term carbon emission uncertainty problem.

[0021] See Figure 1 As shown, this invention discloses a multi-timescale collaborative rolling optimization method for electricity, carbon emissions, and green certificates that considers carbon emission uncertainties, including: Step S11: Collect historical wind, solar and hydropower data and power system operation boundary data, and simulate the output changes in the first target time period in the future based on the historical wind, solar and hydropower data to obtain the corresponding first simulated operation scenario; the historical wind, solar and hydropower data includes wind turbine output data, photovoltaic output data and hydropower station inflow data; the operation boundary data includes relevant data on electricity, carbon emissions and green certificates.

[0022] In this embodiment, a scenario framework is required to perform power balance analysis at different time scales. Since the objectives of generating medium- and long-term and short-term operating scenarios differ, medium- and long-term scenarios emphasize maintaining consistency with the probabilistic characteristics of historical data, while short-term scenarios emphasize predicting potential future scenarios based on existing actual operating scenarios. Figure 2This embodiment provides a schematic diagram of the generation of a first simulated operation scenario. For the generation of wind turbine output in a medium-to-long-term scenario, the historical wind turbine output data is first normalized and filtered to extract the wind turbine fluctuation curves in the wind turbine output curves. Then, the K-Means (k-means clustering algorithm) algorithm is used to classify the wind turbine fluctuation curves into large fluctuations (such as strong wind days), medium fluctuations (such as ordinary wind days), and small fluctuation curves (such as light wind days). The Markov transition matrix between the wind turbine fluctuation curves is calculated, which can be done monthly. Then, a Gaussian function is used to fit each type of wind turbine fluctuation curve to obtain the probability distribution between each type of wind turbine fluctuation curve each month. Based on the probability distribution, wind turbine output data is generated sequentially, and this process is repeated 3760 times to obtain the wind turbine scenario for the whole year. Note that when different fluctuations are superimposed, the starting point of the next fluctuation should be the center point of the previous fluctuation, not the ending point of the previous fluctuation.

[0023] Understandably, for generating medium- to long-term photovoltaic (PV) output scenarios, the process begins by determining the PV deviation curve between theoretical and actual output based on historical PV output data. Then, the K-Means algorithm is used to classify these deviation curves into four categories: sunny days, cloudy / rainy days, sunny-to-cloudy days, and cloudy-to-sunny days. The Markov transition matrix between each type of deviation curve is calculated, which can be done monthly. A three-component Gaussian probability distribution is fitted to each type of deviation curve. Finally, sequential sampling is used to generate PV output, noting the need to correct for sunrise and sunset times on sunny days. The net airspace output model for sunny days is fitted into three straight lines to obtain the annual PV output data, thus yielding the PV scenario. For generating load scenarios in medium- to long-term scenarios, historical load data is first normalized, and then load features including daily peak-to-valley difference, daily load rate, and the time of occurrence of daily maximum load are extracted. The load features are then clustered to obtain load curves. The Markov transition matrix between various types of load curves is calculated monthly, and probability statistics are performed on each type of load curve. Then, through sequential sampling, curves that conform to the probability distribution of their intra-class features are gradually generated based on the monthly Markov transition matrix to obtain the load scenario.

[0024] Furthermore, for the generation of medium- and long-term scenarios for hydropower stations, considering that inflow runoff has obvious seasonal characteristics compared to wind and solar loads, and that it has clearly defined scenarios such as low inflow (i.e., natural inflow) - high inflow - low inflow - and extremely low inflow, these different scenarios are manifested over a long time scale. Therefore, the ARIMA (Autoregressive Integrated Moving Average) model, suitable for long-term forecasting, is used for scenario prediction to obtain the time-series runoff curve. In one specific implementation, the time-series runoff curve is decomposed into three components: a trend-cyclical component reflecting the annual increase or decrease in inflow, a seasonal component reflecting the periodic regularity of the flood season and dry season, and an irregular component reflecting the sudden changes in inflow caused by extreme weather. The corresponding formulas are as follows: ; in, The time-series runoff curve; For trend-cycle components; It is a seasonal ingredient; The irregular component is obtained by using the center-moving average method to extract the trend-cyclic component from the time-series runoff curve. The trend-cyclic component is then removed from the time-series runoff curve to obtain the remaining component. The remaining component is grouped according to its seasonal position, and the average value of each group is determined. The seasonal position is, for example, January, February, etc. The average value is used to determine the initial seasonal index, which is then normalized to obtain the seasonal component. The seasonal component is then removed from the remaining component to obtain the irregular component.

[0025] Specifically, the collection of historical wind, solar, and hydropower data, along with power system operational boundary data, and the simulation of power output changes within a first target time period based on this historical data to obtain a corresponding first simulated operational scenario, includes: normalizing and filtering wind turbine output data, photovoltaic output data, and electricity load data to obtain processed data; classifying and clustering the processed data using the K-Means algorithm to obtain corresponding target curves; statistically analyzing the Markov transition matrix corresponding to the target curves and performing sequential sampling and overlay on the Markov transition matrices to generate corresponding wind and solar scenarios; predicting the time-series runoff curve within the first time period based on hydropower station inflow data and using the ARIMA model; decomposing the time-series runoff curve to obtain corresponding decomposition results; generating corresponding hydropower station scenarios using the decomposition results and seasonal factors; and constructing a corresponding first simulated operational scenario based on the wind and solar scenarios and the hydropower station scenarios.

[0026] Step S12: Adjust the annual operation plan using the first simulated operation scenario to obtain the target operation plan; the annual operation plan is an operation plan determined based on the annual electricity, power generation and carbon quota.

[0027] In this embodiment, before decomposing the annual carbon allowance into daily carbon allowance consumption, an analysis of the annual power balance and carbon allowance surplus / deficit should be conducted. This is because the installation, commissioning, and maintenance plans of thermal power units and new energy units may result in a certain power shortage. In this case, no matter how the carbon allowance is decomposed, the power balance cannot be met. Therefore, adjustments need to be made at the level of the overall annual operation plan. If the target area is defined as within the province, the annual power surplus / deficit data is determined based on the province's adjustable power, the power exchanged with the external power grid, the reserve power reserved to cope with sudden failures and load fluctuations, and the total power load that the power grid needs to meet. The province's adjustable power is determined based on the total capacity, maintenance capacity, and out-of-use capacity of all generating units within the province. The out-of-use capacity includes the capacity of units that cannot operate due to maintenance or failure. Then, the annual power surplus / deficit is determined using the province's total power generation, power transmission and reception, hydropower that is forced to be abandoned due to insufficient grid absorption capacity or peak-shaving demand, and the total power demand of users. Data; the total power generation in the province is determined based on the province's adjustable power supply, annual power generation hours, and underutilization coefficient; the underutilization coefficient is the proportion of generator units not operating at full capacity; carbon quota profit and loss data are determined based on the province's carbon quota, carbon emission factor, annual actual power generation of thermal power units, carbon emission correction factor, annual theoretical green certificate generation, green certificate-carbon quota exchange factor, and green certificate correction factor; wherein, the carbon emission correction factor takes into account the changes in carbon emission factor during unit start-up and shutdown and the changes in non-steady-state carbon emission factor, and the green certificate correction factor takes into account the green certificate-carbon quota exchange limit restricted by external influences.

[0028] Understandably, the profit and loss situation is determined based on the annual electricity profit and loss data, the annual power output profit and loss data, and the carbon quota profit and loss data. If the profit and loss situation indicates an imbalance between electricity and power output, then from the perspective of power generation, electricity can be purchased from the external grid to supplement the provincial gap, or the output of thermal power units can be increased within a preset safety range to improve supply capacity. Adjustments can also be made to the maintenance plan, such as postponing non-emergency maintenance to keep units running and increase output, or utilizing energy storage systems for peak shaving to improve the grid's absorption capacity. Then, from the perspective of electricity consumption, users are guided to stagger their electricity consumption and reduce peak demand to alleviate demand pressure. If the profit and loss situation indicates a carbon quota shortage, priority is given to absorbing renewable energy power, reducing thermal power output, or purchasing additional quotas to fill the gap.

[0029] Specifically, adjusting the annual operation plan using the first simulated operation scenario to obtain the target operation plan includes: determining annual power surplus / deficit data based on adjustable power, transmitted and received power, reserve power, and power load demand within the target area; determining the adjustable power within the target area based on the total capacity, maintenance capacity, and out-of-service capacity of all generator units within the target area; determining annual power surplus / deficit data using total power generation, transmitted and received power, abandoned water power, and total user power demand within the target area; determining the total power generation within the target area based on the adjustable power within the target area, annual power generation hours, and underutilization coefficient; and determining the underutilization coefficient... The number is the ratio of the actual output of the generator set to its rated output; carbon quota profit and loss data is determined based on carbon quotas, carbon emission factors, actual annual power generation of thermal power units, carbon emission correction coefficients, theoretical annual green certificates, exchange coefficients between green certificates and carbon quotas, and green certificate correction coefficients within the target area; the corresponding profit and loss situation is determined based on the annual power profit and loss data, the annual electricity profit and loss data, and the carbon quota profit and loss data; if the profit and loss situation is not within the target profit and loss range, the power generation and power consumption sides in the annual operation plan are adjusted using a preset adjustment strategy until the profit and loss situation corresponding to the adjusted operation plan is within the target profit and loss range, so as to obtain the target operation plan.

[0030] Step S13: Perform time-series production simulation based on the target operation plan and target simulation constraints to obtain the penetration of new energy sources in each time period, and use the penetration of new energy sources to determine the carbon emission uncertainty coefficient in each time period, and use the carbon emission uncertainty coefficient to determine the capacity configuration scheme of the target reserve pool of the carbon quota and green certificate plan.

[0031] In this embodiment, after adjusting the annual operation plan, a time-series production simulation is performed. By simulating the time-series operation of the power grid, information such as the renewable energy absorption rate and monthly carbon emissions is calculated to provide guidance for subsequent optimization. In the time-series production simulation, the simulation is performed time-by-time with the first simulated operation scenario as the boundary. The objective function of the time-series production simulation is as follows: ; in, The number of aggregated power grids; The number of types of thermal power units in each aggregated power grid; The representation covers all time periods throughout the year; Let $t$ be the operating cost of the $j$-th type of thermal power unit in the aggregated power grid $n$ during time period $t$. This represents the number of thermal power units that are started. The cost of starting a thermal power unit once; This represents the number of thermal power units that are shut down. Let $C$ be the cost of a single shutdown of a thermal power unit. The objective function described above aims to minimize the total annual cost of thermal power units by optimizing their start-up, shutdown, and output.

[0032] Understandably, to ensure grid security, it is necessary to set corresponding constraints. The primary constraint is the backup power constraint to ensure the grid's safety redundancy in response to sudden load changes or generator failures. The corresponding formula is as follows: ; in, and These are the designated positive rotation standby and negative rotation standby, respectively; and These are the upper and lower limits of the power generation output of the j-th type of generating unit in the aggregated power grid n during the t-th time period; Let be an integer variable representing the number of units of type j operating in the aggregated power grid during time period t; This aggregates the output of renewable energy sources in the power grid n during time period t, such as the output of fluctuating power sources like wind power and photovoltaics. To aggregate the output of other power sources in the power grid during time period t, such as the output of stable power sources like hydropower, nuclear power, and energy storage; Let n be the aggregated power of the grid during time period t.

[0033] Furthermore, to ensure that the sum of the power output and power exchanged between power lines in each time period of the aggregated power grid is equal to the load demand, a power balance constraint is set, and the corresponding formula is as follows: ; in, The power output of type j generating units in the aggregated power grid during time period t; This represents the number of units of type j operating in the aggregated power grid n during the t-th time period; This refers to the renewable energy output of the aggregated power grid n during time period t. The hydropower output of the aggregated power grid n in time period t; The power output of the pumped storage power station in time period t is the power generation of the aggregated power grid n. The nuclear power output of the aggregated power grid n during time period t; Let n be the number of transmission lines connecting aggregated grid n to other aggregated grids; Let be the transmission power of the aggregated power grid n on the i-th tie line in time period t, with positive values ​​for inflows into the region and negative values ​​for outflows; This represents the total load demand of the aggregated power grid n during time period t, i.e., the total electricity consumption of users.

[0034] In this embodiment, in order to limit the transmission power of inter-regional tie lines and prevent line overload or exceeding the safe operating range, corresponding transmission line constraints are set, and the corresponding formulas are as follows: ; in, and These are the lower limit and upper limit of transmission power, respectively; Let be the actual transmission power of the i-th transmission line during time period t. To constrain the output range of thermal power units and ensure they operate within a safe output range, a power generation output constraint is set, with the corresponding formula as follows: ; ; in, The optimal power output of the j-th type of thermal power unit in the aggregated power grid n during time period t; The power output of type j generating units in the aggregated power grid during time period t; This represents the number of units of type j operating in the aggregated power grid n during the t-th time period; and These represent the upper and lower limits of the output of the j-th type of thermal power unit in the aggregated power grid n during time period t.

[0035] Understandably, in order to ensure the consistency of start-up and shutdown actions across all units, unit operating state constraints are set, with the corresponding formula as follows: ; in, A variable ranging from 0 to 1 indicates a stop command; A variable ranging from 0 to 1, representing the start command; This represents the maximum number of units of type j that can be operated in the aggregated power grid during time period t. This represents the number of units of type j operating in the aggregated power grid n during the t-th time period; Let be the number of units of type i in the aggregated power grid n that are in operation during time period t-1 (i.e., the previous time period); when the dispatch issues a start-up command, =0, When the value is 1, the above constraint inequality becomes: At this point, the units can only be started, meaning the number of operating units can only increase or remain unchanged; when the dispatcher issues a shutdown command, the above constraint inequality becomes: At this point, the unit can only be shut down, meaning the number of operating units can only decrease or remain unchanged.

[0036] Furthermore, to ensure the safe and reasonable operation of the energy storage system, constraints are set from four dimensions: charge / discharge mutual exclusion, dynamic capacity changes, upper and lower capacity limits, and power range. The corresponding formulas are as follows: ; ; ; ; in, It is a 0-1 variable, representing the charging state; A variable of 0-1, representing the discharge state; Let n be the remaining capacity of energy storage in the aggregated power grid during time period t. Let n be the remaining capacity of energy storage in the aggregated power grid n during time period t-1; This represents the net power of energy storage during time period t; charging is positive and discharging is negative. The duration of a single time period; Let n be the minimum allowable remaining energy storage capacity in the aggregated power grid n; This represents the maximum allowable remaining energy storage capacity in the aggregated power grid n; and These represent the maximum and minimum charging power of energy storage in the aggregated power grid n, respectively; and These represent the maximum and minimum discharge power of energy storage in the aggregated grid n, respectively.

[0037] In this embodiment, the hydropower operation constraints cover five dimensions: unit output, flow control, reservoir scheduling, head relationship, and water transfer from cascade power stations, ensuring that hydropower operates within a safe and efficient range. The corresponding formulas are as follows: ; ; ; ; ; ; ; ; ; in, and These are the target theoretical maximum output and target theoretical minimum output of the hydropower unit, respectively. This represents the actual power generation output of the hydropower station during time period t. For the NHQ function of the hydroelectric generator unit; Let be the power generation flow of the hydropower station during time period t; The head of the hydropower station during time period t; This refers to the maximum allowable power generation flow rate of the hydropower station; This represents the total outflow from the hydropower station during time period t. This refers to the maximum allowable outflow from the hydropower station. This represents the water-limited (discarded) flow rate of the hydropower station during time period t. This represents the current reservoir capacity of the hydropower station during time period t. The reservoir capacity of the hydropower station is the reservoir capacity in the period preceding time t-1; The inflow rate of the hydropower station during time period t; and These are the maximum and minimum allowable storage capacities, respectively. The head-reservoir capacity function of a hydropower station; This refers to the total runoff from the cascade hydropower projects. The inflow attenuation factor represents the proportion of the outflow from the upstream hydropower station that is converted into the inflow from the downstream hydropower station. It is mainly used to account for water resource losses caused by factors such as infiltration, irrigation, and evaporation of the outflow from the upstream hydropower station. For upstream hydropower stations during the period Outbound flow; The inflow time delay factor characterizes the time from when the upstream hydropower station opens its gates to when the downstream hydropower station releases water. It is related to the water flow velocity and the distance between the upstream and downstream stations. This represents the inflow and outflow of tributaries between the upstream and downstream sections of a cascade hydropower station. A positive value indicates inflow, and a negative value indicates outflow.

[0038] Understandably, the power output constraint for new energy sources is used to keep the power output within a reasonable range, and the corresponding formula is as follows: ; in, For the actual output of new energy in time period t; The target simulation constraints include: reserve power constraints, power balance constraints, transmission line constraints, unit power generation constraints, unit ramp rate constraints, unit operating status constraints, energy storage operation constraints, hydropower operation constraints, and new energy power generation constraints. Specifically, the reserve power constraint requires that the actual reserve capacity is not less than the required reserve capacity; the power balance constraint requires that the sum of total electricity consumption and total transmission capacity equals total power generation; the total power generation includes thermal power unit output, new energy output, hydropower unit output, pumped storage power station output, and nuclear power output; and the transmission line constraint requires that the actual power generation capacity of the transmission line be equal to the required capacity. The actual transmission power is within the target transmission power range; the unit power generation output constraint is that the actual output of the unit is within the target unit output range; the unit ramp rate constraint is that the output power of the unit per unit time is within the target ramp power range; the unit operation status constraint is that the number of units in operation is consistent with the dispatch instructions; the energy storage operation constraint is that the current energy storage capacity is within the target energy storage range, and the energy storage status is charging or discharging; the hydropower operation constraint is that the hydropower generation water consumption and the reservoir capacity meet the preset matching conditions; the new energy output constraint is that the new energy output does not exceed the target new energy output threshold.

[0039] In this embodiment, after completing the time-series production simulation for each period of the year, the renewable energy consumption and renewable energy power penetration are obtained. The renewable energy penetration rate is calculated using the ratio of renewable energy power generation to actual electricity consumption in the renewable energy penetration data. Since carbon emission uncertainty is essentially caused by renewable energy uncertainty, the carbon emission uncertainty coefficient for each period of the system can be obtained based on the renewable energy power penetration data and after normalization. The corresponding formula is as follows: ; in, The carbon emission uncertainty coefficient for time period t ranges from 0 to 1. A larger value indicates a higher risk of carbon emission fluctuations during that time period. Let t be the penetration rate of new energy sources during time period t; This represents the minimum penetration rate of new energy sources. This represents the maximum penetration rate of new energy sources; the formula corresponding to the new energy penetration rate is as follows: ; in, This represents the total load demand during time period t. To contribute to new energy sources during time period t; This refers to the number of new energy generating units. To address carbon emission uncertainties, a target reserve pool corresponding to carbon quotas and green certificates is established. The capacity of the reserve pool is dynamically adjusted based on the annual planned value of the public green certificate pool and the annual planned value of the public carbon quota pool within the target reserve pool to buffer carbon quota and green certificate gaps during periods of high uncertainty. The formulas corresponding to the annual planned values ​​of the public green certificate pool and the annual planned values ​​of the public carbon quota pool are as follows: ; ; in, The annual public green certificate pool plan value; The annual public carbon allowance pool plan value; and It is a fixed constant, i.e., the benchmark reserve level; Let be the carbon emission uncertainty coefficient for time period t; and To pre-determine the uncertainty prevention factor for carbon emissions; The entire year period; This is because it is necessary to ensure that the total carbon allowance and green certificate inventory pools for the entire year are zero. The planned values ​​of the public carbon allowance pool and public green certificate pool for other time scales are derived from the optimization results of the public carbon allowance pool and public green certificate pool for the previous time scale.

[0040] Specifically, the step of determining the carbon emission uncertainty coefficient for each time period using the new energy penetration situation, and determining the capacity configuration scheme of the target reserve pool for the carbon quota and green certificate plan using the carbon emission uncertainty coefficient, includes: determining the new energy penetration rate using the ratio of new energy power generation to actual electricity consumption in the new energy penetration situation; determining the carbon emission uncertainty coefficient for each time period based on the new energy penetration rate and the corresponding target upper and lower limit thresholds; determining the annual public green certificate pool plan value and the annual public carbon quota pool plan value using the carbon emission uncertainty coefficient and the preset carbon emission uncertainty prevention coefficient; and determining the capacity configuration scheme of the target reserve pool for the carbon quota and green certificate plan based on the annual public green certificate pool plan value and the annual public carbon quota pool plan value.

[0041] Step S14: Based on the historical wind, solar, and hydropower data and the operational boundary data, and using the XGBOOST algorithm, generate a second simulated operational scenario corresponding to the second target time period in the future. Based on the capacity configuration scheme, the first simulated operational scenario, and the second simulated operational scenario, and using a preset collaborative and mutually supportive mechanism, perform multi-time-scale collaborative rolling optimization of electricity, carbon emissions, and green certificates to obtain a corresponding target executable plan. The time length corresponding to the first target time period is greater than the time length corresponding to the second target time period. The target executable plan includes a power generation plan, a carbon quota usage plan, and a green certificate circulation plan.

[0042] In this embodiment, after obtaining the first simulated operation scenario for the medium to long term, a short-term operation scenario, namely the second simulated operation scenario, is generated using supervised learning. A short-term original time series is constructed based on historical wind, light, and water data. Based on this original time series, and using a sliding window technique, the original time series is sliced ​​into subsequences with fixed time windows. These subsequences are used as input features, and the corresponding data from the next time step of the same length are used as output labels to construct a training set. The construction method is as follows: ; in, The original time series; The size of the sliding window can be set according to the actual situation; The characteristic matrix; The target vector.

[0043] It is understandable that training based on the XGBOOST (eXtreme Gradient Boosting, i.e., an optimized distributed gradient boosting library) algorithm aims to optimize the regularization loss function. The formula for the model's target loss function is as follows: ; in, For all parameters of the model; Squared loss function Used to measure predicted values Compared with the true value The error; This is a tree complexity control term; The penalty coefficient for the number of leaf nodes; The number of leaf nodes in a single decision tree; The weight penalty coefficient for leaf nodes; The weight vector for the leaf nodes; For the first The prediction function of the decision tree. The XGBOOST algorithm is trained using the training set and the target loss function to obtain the target XGBOOST model, and the target XGBOOST model is used to predict the wind, solar and water data for the next day / week to obtain the second simulated operation scenario.

[0044] In this embodiment, Figure 3 This embodiment provides a schematic diagram of a collaborative rolling optimization mechanism for electricity, carbon emissions, and green certificates. Through the collaborative rolling optimization model for electricity and carbon certificates, the decomposition and optimization process is transformed from a single long-term scale into a four-level process: annual carbon quota optimization, monthly green certificate optimization, weekly plan correction, and daily electricity generation optimization. This reduces the impact of carbon emission uncertainty. The collaborative rolling optimization model for electricity and carbon certificates has four layers, each with a different time window length. The four stages of rolling optimization are: the annual optimization layer (one-year window) outputs the annual carbon emission plan and annual contracted electricity generation; the monthly optimization layer (one-month window) outputs the monthly carbon emission plan, monthly green certificate plan, and monthly subdivided electricity generation; the weekly optimization layer (one-week window) outputs the weekly carbon emission plan, weekly green certificate plan, and weekly subdivided electricity generation; and the daily optimization layer (24-hour window) outputs the daily carbon emission plan, daily green certificate plan, and daily power generation plan. The results of the optimization decomposition at the previous layer provide the boundary for the optimization at the next layer. Assuming the window lengths for the year, month, day, and week are respectively... , , , Their respective boundary parameters are as follows: , , , ; , , If these are carbon emissions, green certificates, and power generation plans, then their pseudocode flow is as follows: While t Tc do; If t=0 then; Cy = RunCERlayer(Ac); End if; If t mod Tr = 0 then; Cm,Rm=RunREClayer(Ac, Cy); End if; If t mod T w =0 then; Rw,Ew=RunWeeklayer(Aw, Cm,Rm); End if; If t mod Te = 0 then; Cd,Rd,Ed=RunElectricitylayer(Ae, Cw,Rw); End if; T = t + 1; End while.

[0045] Understandably, since the prediction error of source load gradually decreases with the shortening of the time scale, rolling optimization can improve the accuracy of scheduling predictions. Rolling optimization divides a year into multiple time-scale windows, constructs an optimization model for each window, and solves it to obtain the optimal scheduling scheme. When the next time window arrives, rolling optimization dynamically adjusts the relevant constraints and objective function in the optimization model based on the latest acquired measured data and the prediction data from a shorter time scale, and re-solves the model to output a new optimal scheduling strategy. Because the time scale is shorter and the new prediction data is more accurate, the scheduling scheme obtained after rolling optimization based on the new data will be closer to the actual system state, resulting in more accurate optimization. The first objective function corresponding to the annual window optimization layer is as follows: ; in, This refers to the circulating electrical quantity; , , These are the price-time coefficients for current circulation, green certificate circulation, and carbon circulation, respectively. The circulation volume of green certificates during time period t; Let t be the carbon quota flow during time period t; This represents the operating cost of the thermal power unit during time period t. and These represent the start-up and shutdown costs of thermal power unit i during time period t; This includes inter-regional power transmission fees (including network loss fees). For cross-regional power transmission; This is a penalty for deviations from the contracted electricity volume. Other non-contractual penalties.

[0046] Furthermore, the constraints corresponding to the first objective function are as follows: First, there is the power balance constraint, and the corresponding formula is as follows: ; in, To provide power to new energy sources during time period t; The thermal power output during time period t; The hydropower output during time period t; Let be the energy storage discharge power during time period t; The energy storage charging power for time period t; For cross-regional power transmission and reception during time period t; The total power consumption of the user; the formula corresponding to the green certificate circulation constraint is as follows: ; in, To provide power to new energy sources during time period t; The amount of carbon allowances redeemed with green certificates during time period t; The amount of green certificates available for circulation during time period t.

[0047] In this embodiment, the constraints of the first objective function also include: carbon quota and green certificate exchange constraints, and the corresponding formulas are as follows: ; in, Let t be the carbon quota flow during time period t; Carbon emission factor for thermal power units; The thermal power output during time period t; The proportion of green certificates converted into carbon allowances; The amount of carbon allowances redeemed with green certificates during time period t; Let t be the available carbon allowance for time period t; the total annual power transmission and reception capacity of inter-regional transmission lines is determined using the following formula: ; in, The annual window length; For cross-regional power transmission and reception during time period t; This represents the actual amount of electricity transmitted and received during the year. The annual available electricity gap is defined by the following formula: ; ; in, for; The actual amount of electricity delivered and received in a given year; This refers to the medium- to long-term contract electricity volume, specifically the contract electricity volume for the first target time period.

[0048] It is understandable that the contracted electricity deviation penalty item The corresponding formula is as follows: ; ; in, The actual amount of electricity delivered and received in a given year; For medium- and long-term contract electricity volume; This is the contract deviation coefficient. The other non-contract penalty items... The corresponding formula is as follows: ; in, This refers to the number of connecting lines, i.e., the number of lines connecting different areas or external power grids; The annual window length; The constraint relaxation penalty factor is applied to the priority planning of the connecting line channel; Let k be the power flow relaxation variable for the tie-line channel priority plan; To relax the penalty factor for network trend constraints; and For time period t, these are the positive and negative power flow relaxation variables; The number of new energy generating units; This is a penalty factor for the amount of electricity wasted from renewable energy sources. The amount of electricity wasted by renewable energy unit i during time period t; The number of hydroelectric power stations; This is a penalty factor for water wastage at hydropower stations. Let be the water discharge flow of hydropower station i in time period t. The constraints of the first objective function also include a total carbon quota consumption constraint, the corresponding formula of which is as follows: ; in, The annual window length; The available carbon allowance for time period t; This represents the total available carbon allowance.

[0049] In this embodiment, boundary parameters are passed between different layers. The total amount of the annual window layer is passed to the monthly window layer. The corresponding formula is as follows: ; in, The carbon quota usage for time period t is optimized for the annual window layer output. The monthly window layer optimizes the carbon allowance usage for period t. The total annual carbon allowance usage equals the sum of the carbon allowance usage for all 12 months, ensuring that the monthly plan does not exceed the annual limit. The parameters of the monthly window layer are then passed to the weekly window layer, with the following formula: ; ; in, Optimize the carbon quota usage for time period t in the monthly window layer output; To optimize the carbon quota usage for time period t in the weekly window layer output; The green certificate circulation volume for time period t is optimized for the monthly window layer output; The green certificate circulation volume for time period t is optimized for the weekly window layer output.

[0050] Understandably, the parameters of the weekly window layer are passed to the daily window layer, and the corresponding formula is as follows: ; ; in, To optimize the carbon quota usage for time period t in the weekly window layer output; The carbon quota usage for time period t is optimized for the output of the daily window layer; The green certificate circulation volume for time period t is optimized for the weekly window layer output; The green certificate circulation volume for time period t is the output of the daily window optimization layer.

[0051] In this embodiment, after annual window optimization, the annual window optimization result, including the annual carbon emission plan and the annual contracted electricity circulation, is obtained. This annual window optimization result is then used as the boundary input for monthly window optimization. The second objective function corresponding to the monthly window optimization is: ; in, The length of the moon window; , , These are the price-time coefficients for current circulation, green certificate circulation, and carbon circulation, respectively. This refers to the circulating electrical quantity; The circulation volume of green certificates during time period t; Let t be the carbon quota flow during time period t; This represents the operating cost of the thermal power unit during time period t. and These represent the start-up and shutdown costs of thermal power unit i during time period t; This includes inter-regional power transmission fees (including network loss fees). For cross-regional power transmission; This is a penalty for violating the previous optimization layer; Other non-contractual penalties; This refers to the penalty for contracted electricity volume deviation. The formula for the penalty for violating the previous layer's penalty is as follows: ; in, This refers to the violation of the carbon emission plan boundary, i.e., the deviation between the actual carbon emissions and the annual plan. This refers to the amount of violation of the reservoir boundary capacity. This refers to the amount of energy storage boundary capacity violation. , , These are the corresponding penalty coefficients.

[0052] It is understandable that after performing monthly window layer optimization, the corresponding monthly window optimization result is obtained, and this result is used as the boundary input for weekly window optimization. The third objective function corresponding to the weekly window optimization is: ; in, The length of the weekly window; , , These are the price-time coefficients for current circulation, green certificate circulation, and carbon circulation, respectively. This refers to the circulating electrical quantity; The circulation volume of green certificates during time period t; Let t be the carbon quota flow during time period t; This represents the operating cost of the thermal power unit during time period t. and These represent the start-up and shutdown costs of thermal power unit i during time period t; This includes inter-regional power transmission fees (including network loss fees). For cross-regional power transmission; This is a penalty for violating the previous optimization layer; Other non-contractual penalties; This is a penalty for deviations from the contracted electricity volume.

[0053] Furthermore, the weekly window optimization will include additional backup constraints for special dates (such as high summer loads and important holidays), with the corresponding formula as follows: ; in, This represents the maximum power output limit of a single thermal power unit of type j in the aggregated power grid n during time period t. Let be the number of type j thermal power units in the aggregated power grid n during time period t; To provide power to new energy sources during time period t; Total power consumption of the user; Reserve coefficient for special dates; This is for the rotating reserve capacity of conventional systems.

[0054] In this embodiment, after performing weekly window layer optimization, the corresponding weekly window optimization result is obtained, and the weekly window optimization result is used as the boundary input for daily window optimization. The fourth objective function corresponding to the daily window optimization is: ; in, The length of the day window; , , These are the price-time coefficients for current circulation, green certificate circulation, and carbon circulation, respectively. This refers to the circulating electrical quantity; The circulation volume of green certificates during time period t; Let t be the carbon quota flow during time period t; This represents the operating cost of the thermal power unit during time period t. and These represent the start-up and shutdown costs of thermal power unit i during time period t; This includes inter-regional power transmission fees (including network loss fees). For cross-regional power transmission; This is a penalty for violating the previous optimization layer; Other non-contractual penalties; This is a penalty for deviations from the contracted electricity volume.

[0055] Understandably, high carbon emission uncertainty can be categorized into two main types: short-term carbon emission uncertainty caused by insufficient carbon allowances due to fluctuations in wind turbines, photovoltaics, and loads; and long-term carbon emission uncertainty caused by sudden and severe shortages of carbon allowances in extreme scenarios such as extremely low water levels or continuous cloudy and windless days. For short-term carbon emission uncertainty, carbon allowances are treated as a special resource, borrowed or postponed across different days, i.e., a horizontal coordination mechanism. For short-term carbon emission uncertainty, it is necessary to seek assistance from the next higher-level time window and replan the entire time window, i.e., a vertical coordination mechanism. Based on the aforementioned horizontal and vertical coordination mechanisms, a pre-defined collaborative mutual assistance mechanism is constructed. For short-term carbon emission uncertainty, such as insufficient carbon allowances on certain days, adjustments are made using the horizontal coordination mechanism. Specifically, the other non-contractual penalty terms of the objective function corresponding to the target window optimization layer with short-term carbon emission uncertainty are expanded, and the corresponding formula is as follows: ; in, This refers to the number of connecting lines, i.e., the number of lines connecting different areas or external power grids; The window time corresponding to the objective function; The constraint relaxation penalty factor is applied to the priority planning of the connecting line channel; Let k be the power flow relaxation variable for the tie-line channel priority plan; To relax the penalty factor for network trend constraints; and For time period t, these are the positive and negative power flow relaxation variables; The number of new energy generating units; This is a penalty factor for the amount of electricity wasted from renewable energy sources. The amount of electricity wasted by renewable energy unit i during time period t; The number of hydroelectric power stations; This is a penalty factor for water wastage at hydropower stations. Let be the discharge flow rate of hydropower station i during time period t; , This is the penalty coefficient; , These are the deviation values ​​for the public carbon quota pool and the public green certificate pool, respectively.

[0056] Furthermore, after expanding the other non-contractual penalty items, the green certificate circulation constraint is expanded, and the corresponding formula is as follows: ; in, To provide power to new energy sources during time period t; The amount of carbon allowances redeemed with green certificates during time period t; The amount of green certificates extracted from the public green certificate pool during time period t is used to supplement the green certificate shortage in the current window. Let be the amount of tradable green certificates in time period t. Then, the carbon quota and green certificate exchange constraints are extended, and the corresponding formula is as follows: ; in, Let t be the carbon quota flow during time period t; Carbon emission factor for thermal power units; The thermal power output during time period t; The amount of carbon allowances redeemed with green certificates during time period t; The amount of carbon allowance deposited into the public carbon allowance pool during time period t, i.e., the carbon allowance surplus in the current window is deposited into the pool. The amount of carbon allowances extracted from the public carbon allowance pool during time period t is used to supplement the carbon allowance gap in the current window. Let t be the amount of available carbon allowance during time period t.

[0057] In this embodiment, constraints are added regarding the public carbon quota pool and the public green certificate pool, and the corresponding formulas are as follows: ; ; in, For time period t, the public carbon allowance pool; The amount of green certificates deposited into the public green certificate pool during time period t, i.e., the surplus green certificates in the current window; This refers to the number of green certificates extracted from the public green certificate pool during time period t, i.e., the number of green certificates retrieved from the public green certificate pool when there is a gap in the current window. For time period t, the public green certificate pool; The amount of carbon allowances deposited into the public carbon allowance pool during time period t is the carbon allowance surplus in the current window. This represents the amount of carbon allowances extracted from the public carbon allowance pool during time period t, i.e., the carbon allowances called from the public carbon allowance pool when there is a current window gap.

[0058] Understandably, in addition to the parameter transfer between yearly, monthly, weekly, and daily window optimization layers, it is also necessary to transfer the capacity of the public carbon quota pool and the public green certificate pool. The corresponding formulas are as follows: ; ; ; in, Public green certificates for the annual window optimization layer plan; Optimize the actual public green certificate for the lunar window layer; It is the end of the nth month; The carbon quota pool capacity for the annual window optimization layer plan; Optimize the actual carbon quota pool capacity of the monthly window layer; The actual public green certificate for the weekly window optimization layer; To optimize the actual carbon quota pool capacity of the weekly window layer; For the nth weekend; For the daily window optimization layer, the actual public green certificate; The actual carbon quota pool capacity of the daily window optimization layer; This is the end of the nth day.

[0059] Furthermore, taking annual and monthly examples, it is required that the public carbon quota pool and public green certificate pool be consistent with the annual plan at the end of each month. Similarly, the public carbon quota pool and public green certificate pool should be consistent with the monthly plan at the end of each weekly window, and the public carbon quota pool and public green certificate pool should be consistent with the weekly plan at the end of each daily window. Through the horizontal collaboration mechanism, when there is a shortage of short-term carbon quotas, the public carbon quota pool or public green certificate pool can be consumed first to obtain the urgently needed carbon quotas. For long-term carbon emission uncertainties, a vertical collaboration mechanism is adopted. The purpose is to connect different time scales so that the results of lower time scale operations can be fed back to higher-level optimization decomposition levels, breaking the top-down single boundary information transmission mode. When encountering extreme scenarios such as continuous no light and no wind, or extremely low water levels, there may be situations where even the horizontal collaboration mechanism cannot solve the problem. In such cases, the actual operating data can be fed back to the higher-level time scale to resolve the optimization decomposition model at the higher level, thereby reducing the operating risk of extreme scenarios from a longer time scale. In addition, after the optimization of the lower-level time window is completed, if there is no solution due to a large difference between the actual short-term prediction boundary and the long-term generated scenario, the feedback will be sent directly to the next higher level. The feedback will include the actual operating output data, cumulative constraint boundary, actual electricity and carbon certificate plan execution results, and actual wind, solar and hydropower operating scenarios. The next higher level will generate a more accurate scenario based on the feedback information and recalculate. If there is still no solution, the feedback will continue to be sent to the next higher level.

[0060] Specifically, the step of performing multi-timescale collaborative rolling optimization of electricity, carbon emissions, and green certificates based on the capacity configuration scheme, the first simulated operation scenario, and the second simulated operation scenario, and utilizing a preset collaborative and mutually supportive mechanism to obtain a corresponding target executable plan, includes: constructing a target-level rolling optimization model with annual, monthly, weekly, and daily windows; based on the target-level rolling optimization model and the annual public green certificate pool plan value and the annual public carbon quota pool plan value in the capacity configuration scheme, using the first simulated operation scenario as the boundary input for annual window optimization, and constructing a first objective function with minimizing the annual comprehensive cost; the annual comprehensive cost is the cost determined based on the circulation cost, unit operating cost, and transmission cost of electricity, carbon emissions, and green certificates throughout the year; solving the first objective function based on the circulation constraints, contract deviation penalty term, and non-contract penalty term of electricity, carbon emissions, and green certificates to obtain the annual window optimization result including the annual carbon emission plan and the annual contracted circulation electricity; and using the annual window optimization result as the boundary input for monthly window optimization based on the target-level rolling optimization model. The system constructs a second objective function by minimizing monthly circulation costs. It then solves this second objective function using boundary violation penalties to obtain monthly window optimization results, including monthly carbon emission plans, monthly green certificate plans, and monthly subdivided circulation electricity. Based on the target-level rolling optimization model, the monthly window optimization results are used as boundary inputs for weekly window optimization to construct a third objective function. This third objective function is then solved using target-specific date-specific additional reserve constraints to obtain weekly window optimization results, including weekly carbon emission plans, weekly green certificate plans, and weekly subdivided circulation electricity. Based on the target-level rolling optimization model, the weekly window optimization results are used as boundary inputs for daily window optimization to construct a fourth objective function. Solving this fourth objective function yields daily window optimization results, including daily power generation plans, daily carbon quota usage plans, and daily green certificate circulation plans. Finally, based on a pre-defined collaborative and mutually supportive mechanism, the annual window optimization results, weekly window optimization results, monthly window optimization results, and daily window optimization results are collaboratively rolled and optimized to obtain corresponding target executable plans.

[0061] As can be seen from the above, this application collects historical wind, solar, hydro, and power system operation boundary data to generate a first simulated operation scenario in the medium to long term. This scenario can restore the natural fluctuation characteristics of wind, solar, and hydropower output and the core operational constraints of the power system, providing a realistic basis for planning on an annual scale. Based on the first simulated operation scenario, the application conducts a profit and loss analysis of annual electricity, power generation, and carbon quotas to identify supply and demand gaps and carbon quota surplus / deficit situations in advance, and obtains the target operation plan through adjustments. Based on the target operation plan and target simulation constraints, the application conducts time-series production simulations to calculate the penetration rate of new energy and convert it into a carbon emission uncertainty coefficient. This coefficient is then used to configure the carbon quota and green certificate reserve pool capacity, transforming the abstract carbon emission uncertainty into quantifiable and dispatchable reserve resources. The application uses the XGBOOST algorithm to generate a highly accurate short-term second simulated operation scenario, and combines long-term and short-term scenarios, reserve pool schemes, and collaborative mechanisms to perform multi-timescale rolling optimization on a year, month, day, and week basis. In this way, by combining short-term and long-term scenarios to compensate for the errors of single-scale forecasts, and by optimizing and matching the cycles of different markets through multiple time scales, the final power generation, carbon quotas, and green certificate plans are more in line with actual operating conditions, breaking down the disconnect between electricity, carbon emissions, and green certificates, and effectively addressing the long-term uncertainty of carbon emissions.

[0062] As can be seen from the above embodiments, this application achieves a balanced optimization of electricity, carbon emissions, and green certificates based on simulated operation scenarios and using multi-timescale rolling optimization. Therefore, the process of achieving a balanced optimization of electricity, carbon emissions, and green certificates based on simulated operation scenarios and using multi-timescale rolling optimization is described.

[0063] Combination Figure 4 and Figure 5 This invention discloses a specific multi-timescale collaborative rolling optimization method for electricity, carbon emissions, and green certificates that considers carbon emission uncertainties, including: In this embodiment, a long-term first simulated operation scenario is generated based on collected historical wind, solar, and hydropower data. Then, a profit and loss analysis is performed on the annual-scale electricity, carbon emission, and energy quotas in the annual operation plan. If a loss exists, the annual operation plan is adjusted until no loss exists. Next, time-series production simulation is performed based on the obtained target operation plan and target simulation constraints to obtain the penetration of new energy sources in each time period. The carbon emission uncertainty coefficient for each time period is determined using the new energy penetration data. Then, the carbon emission uncertainty coefficient is used to configure the target reserve pool for carbon quotas and green certificate plans to obtain a capacity configuration scheme. Following this, a short-term second simulated operation scenario is generated using the XGBOOST algorithm based on historical wind, solar, and hydropower data and power system operation boundary data. Multi-scale collaborative rolling optimization of electricity, carbon emission, and green certificate data is performed based on the above two simulated operation scenarios. During the optimization process, a preset collaborative mutual assistance mechanism is used to adjust for anomalies until optimization of all time windows is completed.

[0064] Understandably, simulations and result analysis were conducted on the IEEE-118 system. To highlight the uncertainties in the power system, photovoltaic, wind turbine, and hydropower units were added to the IEEE-118 system. Verification and analysis were performed from four aspects: feasibility verification, electricity-carbon-green certificate synergy mechanism, electricity-carbon certificate synergy rolling optimization mechanism, and feedback mechanism. All optimization problems were solved using Gurobi (a mathematical optimization solver). First, the annual electricity consumption and carbon allowance profit and loss were calculated. Figure 6 This embodiment provides a schematic diagram illustrating the annual electricity supply and demand balance; wherein, Figure 6 (a) is a schematic diagram of the annual electricity profit and loss. Figure 6 Figure (b) shows a diagram illustrating the annual electricity supply and demand imbalance. It can be seen that there is a seasonal electricity / power shortage throughout the year, with the largest electricity shortage occurring in the summer (June-August).

[0065] Furthermore, after performing time-series generation simulations, the carbon emission uncertainty coefficient is obtained, such as... Figure 7 As shown, the horizontal axis uses 15-minute granularity for time, meaning that planning and management are based on 15-minute units. It can be seen that the coefficient fluctuates drastically between 0 and 1, indicating that fluctuations in new energy output lead to high uncertainty in carbon emissions. Then, the feasibility of the collaborative rolling optimization of electricity carbon certificates was verified using 8760 hours of scenario data, yielding a total computation time of 10567.42 seconds (2.93 hours). Considering the long optimization time and difficulty in visualization under the 8760-hour scenario, subsequent experiments set a one-year window to 30 days and a one-month window to 10 days, totaling 720 hours for verification. To verify the algorithm's feasibility, it was assumed that accurate predictions could be made, i.e., that boundary data such as wind, solar, and hydropower would not change. The multi-timescale collaborative rolling optimization program of this scheme was run to obtain the corresponding results; among them, the current flow results in the running results are as follows... Figure 8 As shown, blue indicates a small circulation scale, and red indicates a large circulation scale; by Figure 8 It can be seen that the power flow was relatively high in the early stage (0-200 hours) (the color was darker), and the flow gradually decreased as time went on (the color became lighter). Figure 9 This is a schematic diagram of the results of green certificate circulation. The trend of green certificate circulation is basically synchronized with that of electricity circulation, with a higher circulation scale in the early stage and a gradual decline in the later stage. Figure 10 This is a schematic diagram of carbon flow results, showing complete synchronization with the current flow and green certificate flow diagrams. Therefore, when calculating the cumulative data at the window boundaries (such as carbon quota usage and the cumulative amount of the public carbon quota pool / green certificate pool), it was found that the cumulative data of the upper and lower windows are completely consistent, indicating that the multi-timescale collaborative rolling optimization method of this scheme has high reliability.

[0066] In this embodiment, the uncertainty of carbon emissions is reduced by converting some of the generated green certificates into usable carbon allowances. Figure 11 This is a schematic diagram of carbon emissions after using a coordinated mechanism of electricity-carbon emissions-green certificates. The red curve represents actual carbon emissions, and the blue curve represents carbon allowances redeemed from green certificates. Figure 12 This is a schematic diagram comparing carbon allowances. Using the electricity-carbon emission-green certificate market coordination mechanism, a total of 10,692,240 carbon allowances were consumed, and 1,519,850 carbon allowances were exchanged for green certificates. Without using the electricity-carbon emission-green certificate market coordination mechanism, a total of 10,872,530 carbon allowances were consumed. Using the electricity-carbon emission-green certificate market coordination mechanism of this scheme reduced carbon allowance consumption by 13.98%.

[0067] Understandably, in order to verify the effect of the combined rolling optimization of electric carbon certificates on suppressing the high uncertainty of carbon emissions, this embodiment generates 100 scenarios to address the randomness of the operating scenarios, assuming that the scenario parameter errors increase over time, i.e., the boundary parameters... , The law governing the variation of error over time is defined; among which, These are the system's boundary parameters (such as wind and solar power output, load demand, and inflow rate). express Follows a normal distribution. The predicted mean of the boundary parameters, The predicted variance of the boundary parameters; This is the error scaling factor; The current time; This is the initial time for the experiment. Then, a random scenario is generated. Figure 13 This is a scatter plot comparing single-window optimization and collaborative scrolling optimization. It can be seen that the horizontal axis represents the total cost of each scenario in single-window optimization, and the vertical axis represents the total cost of each scenario in collaborative scrolling optimization. The red line is a reference line. From the distribution of the scatter points, most of the blue points fall below the red line, that is, the vertical axis value is less than the horizontal axis value. This indicates that in most scenarios, the cost of multi-window scrolling optimization is lower than that of single-window optimization. Figure 14 This is a statistical diagram of optimization results. The average cost of rolling optimization is 3.85% lower than that of single optimization, and the standard deviation of rolling optimization is 81.65% lower than that of single optimization. The standard deviation represents the degree of fluctuation of the results. The smaller the value, the lower the operational risk. This proves that rolling optimization can significantly reduce the fluctuation of operational results caused by the uncertainty of carbon emissions.

[0068] Furthermore, the effectiveness of the horizontal-vertical coordination mechanism in the preset collaborative support mechanism was tested by setting up a continuous windless and cloudy day scenario. Starting from day 15 and day 20, a four-day continuous windless and cloudy day scenario was set. Without using the horizontal-vertical coordination mechanism, the system showed no solution. However, by using the horizontal-vertical coordination mechanism, a natural transition to the extreme scenario was achieved by consuming the accumulated public carbon quota pool. Figure 15 This diagram illustrates a public carbon quota pool inventory curve. It shows that in extreme scenarios, to compensate for insufficient renewable energy output and increased carbon emissions from thermal power due to windless and cloudy days, the system consumes the public carbon quota pool to borrow carbon quotas from future time windows. This reduces the impact of extreme scenarios and prevents the system from failing to maintain power balance due to carbon emission uncertainties. If even more extreme scenarios are set, such as a short-term load surge coupled with difficulties in renewable hydropower output, leading to a single-day inability to resolve the issue, a feedback mechanism, including a vertical coordination mechanism, is automatically triggered. In this example, after day 15, consecutive windless and cloudy days combined with a short-term load surge and insufficient water runoff trigger the vertical coordination mechanism. Figure 16 This diagram illustrates the change in the monthly carbon quota boundary. It shows that the carbon quota allocated increases significantly starting from day 15. While keeping the total carbon quota usage within the month constant, the daily carbon quota usage is redistributed at the monthly window scale, thus solving the unsolvable problem of continuous extreme scenarios.

[0069] As can be seen from the above, this application reduces the increased uncertainty risk caused by the expansion of prediction errors over long time scales through a rolling window mechanism. Rolling optimization divides a year into multiple time-scale windows, constructing and solving an optimization model for each window to obtain the optimal target executable plan. When the next time window arrives, rolling optimization dynamically adjusts the relevant constraints and objective function in the optimization model based on the latest acquired measured data and prediction data from a shorter time scale, and re-solves the model to output a new target executable plan. In this way, the time scale is shortened, the new prediction data is more accurate, and after rolling adjustments to the optimization model based on the new data, the final target executable plan will be closer to the actual system state, resulting in a more accurate optimization effect.

[0070] Accordingly, see Figure 17 As shown, this application also provides a multi-timescale collaborative rolling optimization device for electricity, carbon emissions, and green certificates that considers carbon emission uncertainties, including: The operation scenario simulation module 11 is used to collect historical wind, solar and hydropower data and power system operation boundary data, and simulate the output changes in the first target time period in the future based on the historical wind, solar and hydropower data to obtain the corresponding first simulated operation scenario; the historical wind, solar and hydropower data includes wind turbine output data, photovoltaic output data and hydropower station inflow data; the operation boundary data includes relevant data on electricity, carbon emissions and green certificates; The operation plan adjustment module 12 is used to adjust the annual operation plan using the first simulated operation scenario to obtain the target operation plan; the annual operation plan is an operation plan determined based on the annual electricity, power generation and carbon quota. The configuration scheme determination module 13 is used to perform time-series production simulation based on the target operation plan and target simulation constraints to obtain the new energy penetration situation in each period, and use the new energy penetration situation to determine the carbon emission uncertainty coefficient in each period, and use the carbon emission uncertainty coefficient to determine the capacity configuration scheme of the target reserve pool of the carbon quota and green certificate plan. The rolling optimization module 14 is used to generate a second simulated operation scenario corresponding to a future second target time period based on the historical wind, solar and hydropower data and the operation boundary data and using the XGBOOST algorithm. Based on the capacity configuration scheme, the first simulated operation scenario and the second simulated operation scenario, and using a preset collaborative and mutual assistance mechanism, it performs multi-time-scale collaborative rolling optimization of electricity, carbon emissions and green certificates to obtain a corresponding target executable plan. The time length corresponding to the first target time period is greater than the time length corresponding to the second target time period. The target executable plan includes a power generation plan, a carbon quota usage plan and a green certificate circulation plan.

[0071] In some specific embodiments, the running scenario simulation module 11 may specifically include: The data clustering unit is used to normalize and filter wind turbine output data, photovoltaic output data and electricity load data to obtain processed data. The K-Means algorithm is then used to classify and cluster the processed data to obtain the corresponding target curve. The matrix overlay unit is used to calculate the Markov transition matrix corresponding to the target curve, and to perform sequential sampling and overlay on the Markov transition matrix to generate the corresponding landscape scene. The curve decomposition unit is used to decompose the time-series runoff curve based on the inflow data of the hydropower station and the ARIMA model to predict the time-series runoff curve in the first time period in the future, so as to obtain the corresponding decomposition results. The first scenario construction unit is used to generate a corresponding hydropower station scenario using the decomposition results and seasonal factors, and to construct a corresponding first simulation operation scenario based on the wind and solar scenario and the hydropower station scenario.

[0072] In some specific embodiments, the operation plan adjustment module 12 may specifically include: The power surplus / deficit data determination unit is used to determine the annual power surplus / deficit data based on the adjustable power, transmitted and received power, reserve power, and power load demand within the target area; the adjustable power within the target area is determined based on the total capacity, maintenance capacity, and out-of-use capacity of all generator units within the target area. The power surplus / deficit data determination unit is used to determine the annual power surplus / deficit data using the total power generation, power transmission and reception, abandoned water power, and total user electricity demand within the target area. The total power generation within the target area is determined based on the adjustable power within the target area, the annual power generation hours, and the under-full-load coefficient. The under-full-load coefficient is the ratio of the actual output of the generator set to its rated output. The carbon allowance profit and loss data determination unit is used to determine carbon allowance profit and loss data based on carbon allowances within the target area, carbon emission factors, actual annual power generation of thermal power units, carbon emission correction coefficients, theoretical annual generation of green certificates, exchange coefficients between green certificates and carbon allowances, and green certificate correction coefficients. The profit and loss determination unit is used to determine the corresponding profit and loss situation based on the annual electricity profit and loss data, the annual electricity volume profit and loss data, and the carbon quota profit and loss data. The target calculation and determination unit is used to adjust the power generation side and power consumption side of the annual operation plan using a preset adjustment strategy if the profit and loss situation is not within the target profit and loss range, until the profit and loss situation corresponding to the adjusted operation plan is within the target profit and loss range, so as to obtain the target operation plan.

[0073] In some specific implementations, the target simulation constraints include backup power constraints, power balance constraints, transmission line constraints, unit power generation output constraints, unit ramp rate constraints, unit operating status constraints, energy storage operation constraints, hydropower operation constraints, and new energy output constraints.

[0074] In some specific embodiments, the configuration scheme determination module 13 may specifically include: The penetration rate determination unit is used to determine the new energy penetration rate by using the ratio of new energy power generation to actual electricity consumption in the new energy penetration situation. The coefficient determination unit is used to determine the carbon emission uncertainty coefficient for each time period based on the new energy penetration rate and the corresponding target upper and lower limit thresholds. The planned value determination unit is used to determine the annual planned value of the public green certificate pool and the annual planned value of the public carbon quota pool using the carbon emission uncertainty coefficient and the preset carbon emission uncertainty prevention coefficient. The configuration scheme determination unit is used to determine the capacity configuration scheme of the target reserve pool of the carbon quota and green certificate plan based on the annual public green certificate pool plan value and the annual public carbon quota pool plan value.

[0075] In some specific embodiments, the rolling optimization module 14 may specifically include: The first function construction unit is used to construct a target-level rolling optimization model for annual, monthly, weekly, and daily windows. Based on the target-level rolling optimization model and the annual public green certificate pool plan value and annual public carbon quota pool plan value in the capacity configuration scheme, the first simulated operation scenario is used as the boundary input for annual window optimization, and the first objective function is constructed by minimizing the annual comprehensive cost. The annual comprehensive cost is the cost determined based on the circulation cost of electricity, carbon emissions, and green certificates, the unit operating cost, and the transmission cost throughout the year. The first function solving unit is used to solve the first objective function based on the circulation constraints of electricity, carbon emissions and green certificates, contract deviation penalty terms and non-contract penalty terms, so as to obtain the annual window optimization results including the annual carbon emission plan and the annual contract circulation electricity. The second function construction unit is used to take the annual window optimization result as the boundary input of the monthly window optimization based on the target-level rolling optimization model, and construct the second objective function by minimizing the monthly circulation cost; The second function solving unit is used to solve the second objective function using the boundary violation penalty term to obtain the monthly window optimization results including the monthly carbon emission plan, the monthly green certificate plan, and the monthly subdivided electricity flow. The third function construction unit is used to construct a third objective function by taking the monthly window optimization result as the boundary input of the weekly window optimization based on the target-level rolling optimization model. The third objective function is solved by using the target additional reserve constraints corresponding to the target special date to obtain the weekly window optimization results including the weekly carbon emission plan, the weekly green certificate plan, and the weekly subdivided circulating electricity. The fourth function construction unit is used to construct a fourth objective function by taking the weekly window optimization result as the boundary input of the daily window optimization based on the target-level rolling optimization model, and to solve the fourth objective function to obtain the daily window optimization results including the daily power generation plan, the daily carbon quota usage plan, and the daily green certificate circulation plan. The optimization result optimization unit is used to perform collaborative rolling optimization on the annual window optimization result, the weekly window optimization result, the monthly window optimization result, and the daily window optimization result based on a preset collaborative mutual assistance mechanism, so as to obtain the corresponding target executable plan.

[0076] Furthermore, embodiments of this application also disclose an electronic device, Figure 18This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the electricity-carbon emission-green certificate multi-timescale collaborative rolling optimization method considering carbon emission uncertainty disclosed in any of the foregoing embodiments. Furthermore, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0077] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0078] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0079] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the electricity-carbon emission-green certificate multi-timescale collaborative rolling optimization method considering carbon emission uncertainty, which is executed by the electronic device 20 according to any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0080] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned disclosed multi-timescale collaborative rolling optimization method for electricity-carbon emission-green certificate considering carbon emission uncertainties. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0081] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0082] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0083] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0084] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0085] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A multi-timescale collaborative rolling optimization method considering carbon emission uncertainty for electricity, carbon emissions, and green certificates, characterized in that, include: Historical wind, solar, and hydropower data, along with operational boundary data of the power system, are collected. Based on this historical data, the power output changes within a first target time period are simulated to obtain a corresponding first simulated operational scenario. The historical wind, solar, and hydropower data includes wind turbine output data, photovoltaic output data, and hydropower station inflow data. The operational boundary data includes relevant data on electricity, carbon emissions, and green certificates. The annual operation plan is adjusted using the first simulated operation scenario to obtain the target operation plan; The annual operation plan is an operation plan determined based on the annual electricity, power generation and carbon quotas; Based on the target operation plan and target simulation constraints, time-series production simulation is performed to obtain the penetration of new energy sources in each period, and the carbon emission uncertainty coefficient in each period is determined using the new energy penetration data. The capacity configuration scheme of the target reserve pool of the carbon quota and green certificate plan is then determined using the carbon emission uncertainty coefficient. Based on the historical wind, solar, and hydropower data and the operational boundary data, and using the XGBOOST algorithm, a second simulated operational scenario corresponding to the second target time period is generated. Based on the capacity configuration scheme, the first simulated operational scenario, and the second simulated operational scenario, and using a preset collaborative and mutually supportive mechanism, multi-timescale collaborative rolling optimization of electricity, carbon emissions, and green certificates is performed to obtain a corresponding target executable plan. The time length corresponding to the first target time period is greater than the time length corresponding to the second target time period. The target executable plan includes a power generation plan, a carbon quota usage plan, and a green certificate circulation plan.

2. The multi-timescale collaborative rolling optimization method for electricity, carbon emissions, and green certificates considering carbon emission uncertainties according to claim 1, characterized in that, The process involves collecting historical wind, solar, and hydropower data, as well as power system operational boundary data. Based on this historical data, the system simulates power output changes within a first target time period to obtain a corresponding first simulated operational scenario, including: The wind turbine output data, photovoltaic output data, and electricity load data are normalized and filtered to obtain processed data. The K-Means algorithm is then used to classify and cluster the processed data to obtain the corresponding target curve. The Markov transition matrix corresponding to the target curve is statistically analyzed, and the Markov transition matrix is ​​sequentially sampled and superimposed to generate the corresponding landscape scene. Based on the inflow data of the hydropower station and using the ARIMA model to predict the time-series runoff curve in the first time period in the future, the time-series runoff curve is decomposed to obtain the corresponding decomposition results. The decomposition results and seasonal factors are used to generate corresponding hydropower station scenarios, and the corresponding first simulated operation scenario is constructed based on the wind and solar scenarios and the hydropower station scenarios.

3. The multi-timescale collaborative rolling optimization method for electricity, carbon emissions, and green certificates considering carbon emission uncertainties according to claim 1, characterized in that, The step of adjusting the annual operation plan using the first simulated operation scenario to obtain the target operation plan includes: The annual power surplus / deficit data are determined based on the adjustable power, power transmission and reception, reserve power, and power load demand within the target area; the adjustable power within the target area is determined based on the total capacity, maintenance capacity, and out-of-use capacity of all generator units within the target area. The annual electricity surplus / deficit data is determined using the total power generation, transmitted and received power, abandoned water power, and total user electricity demand within the target area. The total power generation within the target area is determined based on the adjustable power within the target area, the annual power generation hours, and the under-full-load coefficient. The under-full-load coefficient is the ratio of the actual output of the generator set to its rated output. Carbon allowance profit and loss data are determined based on carbon allowances, carbon emission factors, actual annual power generation of thermal power units, carbon emission correction factor, theoretical annual generation of green certificates, exchange factor between green certificates and carbon allowances, and green certificate correction factor within the target area. The corresponding profit and loss situation is determined based on the annual electricity profit and loss data, the annual electricity volume profit and loss data, and the carbon quota profit and loss data. If the profit and loss situation is not within the target profit and loss range, the power generation side and power consumption side in the annual operation plan are adjusted using a preset adjustment strategy until the profit and loss situation corresponding to the adjusted operation plan is within the target profit and loss range, so as to obtain the target operation plan.

4. The multi-timescale collaborative rolling optimization method for electricity, carbon emissions, and green certificates considering carbon emission uncertainties according to claim 1, characterized in that, The target simulation constraints include backup power constraints, power balance constraints, transmission line constraints, unit power generation output constraints, unit ramp rate constraints, unit operating status constraints, energy storage operation constraints, hydropower operation constraints, and new energy output constraints. The constraints are as follows: the reserve power constraint is that the actual reserve capacity is not less than the required reserve capacity; the power balance constraint is that the sum of the total electricity consumption and the total transmission volume is equal to the total power generation volume; the total power generation volume includes the output of thermal power units, renewable energy output, hydropower unit output, pumped storage power station output, and nuclear power output; the transmission line constraint is that the actual transmission power of the transmission line is within the target transmission power range; the generating unit output constraint is that the actual output of the generating unit is within the target generating unit output range; the generating unit ramp rate constraint is that the output power of the generating unit per unit time is within the target ramp power range; the generating unit operating status constraint is that the number of generating units in operation is consistent with the dispatch instructions; the energy storage operation constraint is that the current energy storage capacity is within the target energy storage range, and the energy storage status is charging or discharging; the hydropower operation constraint is that the hydropower generation water consumption and reservoir capacity meet the preset matching conditions; and the renewable energy output constraint is that the renewable energy output does not exceed the target renewable energy output threshold.

5. The multi-timescale collaborative rolling optimization method for electricity-carbon emission-green certificate considering carbon emission uncertainty according to any one of claims 1 to 4, characterized in that, The process of determining the carbon emission uncertainty coefficient for each time period using the penetration of new energy sources, and then determining the capacity configuration scheme of the target reserve pool for the carbon quota and green certificate plan using the carbon emission uncertainty coefficient, includes: The new energy penetration rate is determined by the ratio of new energy power generation to actual electricity consumption in the aforementioned new energy penetration situation. The carbon emission uncertainty coefficient for each time period is determined based on the new energy penetration rate and the corresponding target upper and lower limit thresholds. The planned values ​​for the annual public green certificate pool and the annual public carbon quota pool are determined using the carbon emission uncertainty coefficient and the preset carbon emission uncertainty prevention coefficient. The capacity configuration scheme of the target reserve pool for the carbon quota and green certificate plan is determined based on the planned value of the annual public green certificate pool and the planned value of the annual public carbon quota pool.

6. The multi-timescale collaborative rolling optimization method for electricity, carbon emissions, and green certificates considering carbon emission uncertainties according to claim 5, characterized in that, The step of performing multi-timescale collaborative rolling optimization of electricity, carbon emissions, and green certificates based on the capacity configuration scheme, the first simulated operation scenario, and the second simulated operation scenario, and utilizing a preset collaborative and mutually supportive mechanism, to obtain a corresponding target executable plan, includes: A target-level rolling optimization model is constructed with annual, monthly, weekly, and daily windows. Based on the target-level rolling optimization model and the annual public green certificate pool plan value and annual public carbon quota pool plan value in the capacity configuration scheme, the first simulated operation scenario is used as the boundary input for annual window optimization, and a first objective function is constructed to minimize the annual comprehensive cost. The annual comprehensive cost is the cost determined based on the circulation cost of electricity, carbon emissions, and green certificates, the unit operating cost, and the transmission cost throughout the year. The first objective function is solved based on the circulation constraints of electricity, carbon emissions, and green certificates, contract deviation penalties, and non-contract penalties to obtain the annual window optimization results including the annual carbon emission plan and the annual contracted electricity circulation. Based on the target-level rolling optimization model, the annual window optimization result is used as the boundary input for the monthly window optimization, and a second objective function is constructed by minimizing the monthly circulation cost. The second objective function is solved using the boundary violation penalty term to obtain the monthly window optimization results, which include the monthly carbon emission plan, the monthly green certificate plan, and the monthly subdivided electricity flow. Based on the target-level rolling optimization model, the monthly window optimization result is used as the boundary input of the weekly window optimization to construct a third objective function. The target additional reserve constraint corresponding to the target special date is used to solve the third objective function to obtain the weekly window optimization result including the weekly carbon emission plan, the weekly green certificate plan, and the weekly subdivided circulating electricity. Based on the target-level rolling optimization model, the weekly window optimization result is used as the boundary input of the daily window optimization to construct a fourth objective function. The fourth objective function is solved to obtain the daily window optimization results, including the daily power generation plan, the daily carbon quota usage plan, and the daily green certificate circulation plan. Based on a preset collaborative and mutually supportive mechanism, the annual window optimization results, the weekly window optimization results, the monthly window optimization results, and the daily window optimization results are collaboratively and continuously optimized to obtain the corresponding target executable plan.

7. The multi-timescale collaborative rolling optimization method for electricity, carbon emissions, and green certificates considering carbon emission uncertainties according to claim 6, characterized in that, The method of performing collaborative rolling optimization on the annual window optimization results, weekly window optimization results, monthly window optimization results, and daily window optimization results based on a preset collaborative mutual assistance mechanism to obtain the corresponding target executable plan includes: The solution process for each window optimization result is monitored to obtain the corresponding monitoring results; If the monitoring result indicates that the solution corresponding to the target window optimization result has a feasible solution, and there is a target window with a carbon quota gap or a green certificate gap, then the target window is gap-filled based on the capacity configuration scheme and the candidate windows in the target window optimization result to obtain the window optimization result after filling; the candidate window is the window corresponding to the target window optimization result that has a carbon quota surplus and a green certificate surplus. If the monitoring result indicates that the solution corresponding to the target window optimization result is unsolvable, or there is an unfillable carbon quota gap or green certificate gap in the target window, then the corresponding gap data is fed back to the upper-level window of the target window, so as to adjust the upper-level window optimization result corresponding to the upper-level window based on the gap data, so as to obtain the adjusted upper-level window optimization result. The adjusted optimization results of the previous window are sent to the target window, and the objective function corresponding to the target window is solved based on the adjusted optimization results of the previous window to obtain the target optimization results. If there is a feasible solution for the solution corresponding to the target optimization results, and there is no target window with a carbon quota gap or green certificate gap, then a corresponding target executable plan is constructed based on the target optimization results.

8. A multi-timescale collaborative rolling optimization device considering carbon emission uncertainty, characterized in that, include: The operation scenario simulation module is used to collect historical wind, solar, and hydropower data and power system operation boundary data, and simulate the output changes within a first target time period in the future based on the historical wind, solar, and hydropower data to obtain the corresponding first simulated operation scenario; the historical wind, solar, and hydropower data includes wind turbine output data, photovoltaic output data, and hydropower station inflow data; the operation boundary data includes relevant data on electricity, carbon emissions, and green certificates; The operation plan adjustment module is used to adjust the annual operation plan using the first simulated operation scenario to obtain the target operation plan; The annual operation plan is an operation plan determined based on the annual electricity, power generation and carbon quotas; The configuration scheme determination module is used to perform time-series production simulation based on the target operation plan and target simulation constraints to obtain the new energy penetration situation in each period, and use the new energy penetration situation to determine the carbon emission uncertainty coefficient in each period, and use the carbon emission uncertainty coefficient to determine the capacity configuration scheme of the target reserve pool of the carbon quota and green certificate plan. The rolling optimization module is used to generate a second simulated operation scenario corresponding to a future second target time period based on the historical wind, solar, and hydropower data and the operation boundary data, using the XGBOOST algorithm. Based on the capacity configuration scheme, the first simulated operation scenario, and the second simulated operation scenario, and using a preset collaborative and mutually supportive mechanism, it performs multi-time-scale collaborative rolling optimization of electricity, carbon emissions, and green certificates to obtain a corresponding target executable plan. The time length corresponding to the first target time period is greater than the time length corresponding to the second target time period. The target executable plan includes a power generation plan, a carbon quota usage plan, and a green certificate circulation plan.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor is configured to execute the computer program to implement the multi-timescale collaborative rolling optimization method for electricity-carbon emissions-green certificates that considers carbon emission uncertainties as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store a computer program, wherein the computer program, when executed by a processor, implements the multi-timescale collaborative rolling optimization method for electricity-carbon emissions-green certificates that considers carbon emission uncertainties as described in any one of claims 1 to 7.

Images