A low-carbon power system planning method, system, equipment and medium

By feeding back oxygen produced as a byproduct of electrolysis into thermal power units and allocating carbon dioxide to the path of methanol electrolysis, combined with a multi-stage stochastic programming model, the problems of high energy consumption of air separation and idle carbon dioxide in oxygen-enriched combustion retrofitting were solved, achieving economical and sustainable low-carbon retrofitting and improving the economy and power supply security of the retrofitted units.

CN123121845APending Publication Date: 2026-11-17XI AN JIAOTONG UNIV +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611350415.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-02
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In existing technologies, carbon capture schemes based on oxygen-enriched combustion lack robustness and face problems such as high energy consumption of air separation and idle carbon dioxide, making it difficult to achieve economically sustainable low-carbon transformation under uncertain conditions.

Method used

By feeding back the oxygen produced as a byproduct of electrolysis to the thermal power unit to replace the oxygen production load of the air separation unit, and distributing the captured carbon dioxide to geological storage and electrochemical methanol production pathways, a two-way carbon-oxygen cycle is established. Combined with a multi-stage stochastic programming model to optimize investment and operating costs, an adaptive planning scheme is formed.

Benefits of technology

It reduced the energy consumption penalty of oxygen-enriched combustion retrofit, increased the utilization value of carbon dioxide, achieved an economically sustainable low-carbon transformation, and improved the economy and power supply security of the retrofitted units.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123121845A_ABST
    Figure CN123121845A_ABST
Patent Text Reader

Abstract

This invention discloses a low-carbon power system planning method, system, equipment, and medium, relating to the field of low-carbon power system planning technology. The method includes the following steps: obtaining the net carbon dioxide emissions of thermal power units and the oxygen produced as a byproduct of electrolysis during methanol production; feeding the oxygen produced as a byproduct back to the thermal power units via an oxygen pipeline network to replace the oxygen production load of the air separation unit in the thermal power units; allocating the net carbon dioxide emissions captured by the thermal power units to two paths: geological storage and use as raw material for methanol production, thus obtaining the operating cost; with the objective of minimizing the net present value of investment and operating costs within the planning period, representing the long-term uncertainty of carbon price and new energy penetration rate as a scenario tree, and establishing a multi-stage stochastic programming model; solving the multi-stage stochastic programming model to obtain a phased adaptive planning scheme. This invention, while ensuring power supply security, simultaneously alleviates the two bottlenecks of air separation energy consumption penalty and carbon dioxide idleness in oxygen-enriched combustion retrofitting.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of low-carbon power system planning technology, and in particular to a low-carbon power system planning method, system, equipment and medium. Background Technology

[0002] Low-carbon retrofitting of coal-fired power units is an important path to achieve deep emission reduction in the power industry. Among existing carbon capture technologies, oxygen-enriched combustion is increasingly favored in the retrofitting of existing units because it can increase the concentration of carbon dioxide in flue gas, simplify the capture process, and reduce capture costs.

[0003] In existing technologies, in order to better address the carbon dioxide produced by oxy-fuel combustion, the electro-to-X technology provides a way to utilize carbon dioxide in oxy-fuel combustion. Among them, the coupling potential of electro-methanol production with coal assets is particularly prominent: methanol is liquid at normal temperature and pressure, has high energy density, can utilize existing storage and transportation facilities, and has a broad market as a basic chemical raw material. Its sales revenue can offset part of the capture cost. In an integrated system, the oxygen produced as a by-product of electrolysis can be fed back to the oxy-fuel boiler to replace part of the air separation oxygen production load. At the same time, the captured carbon dioxide is converted into methanol, thereby simultaneously alleviating the two bottlenecks mentioned above in the oxy-fuel combustion retrofit.

[0004] However, the realization of the aforementioned synergistic potential depends on systematic long-term planning. Due to the deep uncertainty of key external conditions such as carbon price trends and the penetration rate of new energy sources, the technical and economic assessment of a single link or short-term operational optimization is difficult to depict the dynamic process of changes in these key external conditions. The resulting solutions lack robustness or face the risk of asset stranding. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of the prior art by providing a low-carbon power system planning method, system, equipment, and medium to solve the problems in the prior art.

[0006] The present invention specifically provides the following technical solution: A low-carbon power system planning method includes the following steps: To obtain the net carbon dioxide emissions of thermal power units and the oxygen produced as a byproduct of electrolysis during the production of methanol by electricity; The oxygen produced by electrolysis is fed back to the thermal power unit via the oxygen pipeline network to replace the oxygen production load of the air separation unit in the thermal power unit. The net carbon dioxide emissions captured by the thermal power unit are allocated to two paths: geological storage and as raw material for electro-methanol production. The operating cost is obtained by combining the cost of geological storage with the revenue from the dual-level pricing of electro-methanol synthesis. The objective function of the power network model is to minimize the net present value of investment costs and operating costs within the planning period. The fluctuations of carbon price and new energy penetration rate are represented as scenario trees, and a multi-stage stochastic programming model is established. The multi-stage stochastic programming model is solved to obtain a staged adaptive programming scheme.

[0007] Preferably, obtaining the net carbon dioxide emissions of the thermal power unit specifically involves: For thermal power units, the net carbon dioxide emissions of thermal power units are obtained by using the logical coupling constraints between the modification decision and the operation mode, the capture rate related to the operation mode, the oxygen demand in the oxygen-enriched mode, and the energy consumption penalty generated by air separation and carbon dioxide compression.

[0008] Preferably, in the thermal power unit, oxygen injection is only required in the oxygen-enriched mode, and the oxygen demand is adjusted by the air-to-volume ratio and oxygen purity; the net output of the thermal power unit is obtained by subtracting the plant power consumption and energy consumption penalty from the gross power generation, and the energy consumption penalty of air separation and carbon dioxide compression is only generated when the oxygen-enriched mode is activated.

[0009] Preferably, when obtaining oxygen as a byproduct of electrolysis during the process of producing methanol by electrolysis, the location and capacity of the methanol production station are constrained by the upper and lower limits of the single station capacity and the maximum number of stations in the system. Hourly operation is constrained by the installed capacity and electrolysis ramp-up. The amount of hydrogen produced and the amount of byproduct oxygen are determined by the electrolysis efficiency and stoichiometric ratio. After adiabatic compression of hydrogen and carbon dioxide to the synthesis pressure, methanol is synthesized according to stoichiometric ratio and comprehensive conversion efficiency. The methanol output within the green methanol cycle demand is priced at the green methanol price, and the excess is priced at the gray methanol price. Green methanol and gray methanol are distinguished according to the source of raw materials and carbon emissions.

[0010] Preferably, the step of feeding back the oxygen produced as a byproduct of electrolysis to the thermal power unit via an oxygen pipeline network to replace the oxygen production load of the air separation unit in the thermal power unit specifically involves: A unified discrete-time model is adopted for storage tanks, taking into account self-loss, filling and discharging rates, capacity upper and lower limits and periodic boundary conditions; for dedicated transportation pipelines, the pipeline is constructed only when the facilities at both ends of the carbon dioxide or oxygen pipeline are completed, the gas transportation of the pipeline is constrained by the pipeline capacity, and the power of the booster is taken into account. Synergy is achieved through carbon-oxygen bidirectional cyclic coupling constraints: the carbon dioxide captured by the thermal power unit is distributed to two routes: geological storage and transmission to the methanol-to-electricity plant via pipeline. The oxygen demand of the oxygen-enriched unit is met by the oxygen fed back from the methanol-to-electricity plant via pipeline and the local air separation unit.

[0011] Preferably, the objective function of the power network model is to minimize the net present value of investment costs and operating costs within the planning period. The fluctuations in carbon prices and renewable energy penetration rates are represented as scenario trees, and a multi-stage stochastic programming model is established. Specifically: The power grid adopts a DC power flow model, which takes into account the line power flow and thermal stability limits, the upper and lower limits of the output of the unit combination, the ramp rate, the minimum start and stop time and start and stop logic, the upper limit of wind curtailment and load shedding, and the node power balance. The objective function is the sum of the net present value of investment costs and operating costs at each stage within the planning period, discounted at the discount rate. Investment costs include unit modification, construction of the electric methanol production station and pipeline network, and operating costs include fuel and start-up, carbon tax and storage, transportation, wind curtailment and load shedding penalties, and methanol revenue included in the negative. The carbon tax and the penetration rate of new energy sources are solved as random variables along the scenario tree to obtain a multi-stage stochastic programming model.

[0012] Preferably, the step of solving the multi-stage stochastic programming model to obtain a staged adaptive planning scheme specifically involves: The continuous state variables in the multi-stage stochastic programming model are discretized by binary approximation, and the upper and lower bounds and the cutting plane pool are initialized. The scene sequence from root to leaf is sampled from the scene tree. The forward process is executed, and the stage problem is solved stage by stage within the current solver tolerance. The future cost function is approximated by Lagrange cut and the statistical upper bound is updated. The backward process is executed, and local replication variables are introduced and unexpected constraints are relaxed to solve the Lagrange dual. The reinforcement cut is generated by the dual multiplier and the Lagrange function value and added to the previous stage problem. At the end of each iteration, the relative gap between the upper and lower bounds is compared. When the gap meets the target tolerance, the solver tolerance is tightened to the minimum and an additional forward or backward verification step is performed. The process terminates only when the verification gap still meets the tolerance; otherwise, the convergence is rejected and the scaling factor is penalized. When the gap does not meet the tolerance and the lower bound improvement stagnates, the development stage is entered and the problem tolerance is tightened, resulting in a phased adaptive planning scheme for the transformation of thermal power units, the construction of electric methanol production stations, and the investment in transportation pipelines.

[0013] This invention provides a low-carbon power system planning system, comprising: The data acquisition module is used to acquire the net carbon dioxide emissions of thermal power units and the oxygen produced as a byproduct of electrolysis during the methanol electrolysis process. The carbon-oxygen coupling module is used to feed back the oxygen produced by electrolysis to the thermal power unit via the oxygen pipeline network, replacing the oxygen production load of the air separation unit in the thermal power unit, and to allocate the net carbon dioxide emissions captured by the thermal power unit to two paths: geological storage and as raw material for electro-methanol production. The operating cost is obtained by combining the cost of geological storage with the revenue from the dual-level pricing of electro-methanol synthesis. The stochastic programming modeling module is used to take the minimum net present value of investment costs and operating costs within the planning period as the objective function of the power network model, and to represent the fluctuations of carbon price and new energy penetration rate as a scenario tree to establish a multi-stage stochastic programming model. The solution output module is used to solve the multi-stage stochastic programming model and obtain a staged adaptive planning scheme.

[0014] The present invention also provides a computer device, characterized in that it includes a memory and a processor, wherein the memory stores a program, and when the program is executed by the processor, the processor performs the steps of the aforementioned low-carbon power system planning method.

[0015] The present invention also provides a storage medium storing a computer program thereon, characterized in that the computer program, when executed by a processor, implements the steps of the above-described low-carbon power system planning method.

[0016] Compared with the prior art, the present invention has the following significant advantages: This invention feeds back the byproduct oxygen from electrolysis directly to the oxygen-enriched boiler as an optimizable decision variable, replacing part of the oxygen production load of the high-energy-consuming air separation unit. The high-purity carbon dioxide captured by the thermal power unit no longer relies solely on geological sequestration, but is instead allocated to the electrochemical methanol synthesis pathway, transforming it into a marketable methanol product. This transforms carbon capture from an environmental cost item into a revenue-generating item, reducing the plant power consumption penalty and carbon dioxide idleness issues caused by oxygen-enriched combustion retrofitting from the planning source. The two processes are transformed from independent steps into a coupled and synergistic whole, thus ensuring power supply security while providing an economically sustainable transformation path for the deep decarbonization of existing coal-fired power units, improving the economics of the retrofitted units. Attached Figure Description

[0017] Figure 1 This is a flowchart of a low-carbon power system planning method according to the present invention; Figure 2 This is a schematic diagram of the overall framework for the carbon-oxygen cycle utilization of the low-carbon power system of the present invention; Figure 3 This is a schematic diagram of the multi-stage planning process of the present invention; Figure 4 This is a schematic diagram of the long-term uncertainty scenario tree of the present invention; Figure 5 This is a convergence characteristic diagram of the adaptive precision random dual dynamic integer programming algorithm of the present invention. Detailed Implementation

[0018] 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, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0019] Addressing climate change and promoting the low-carbon transformation of the energy system has become a global consensus. As a major source of greenhouse gas emissions, the power industry is at the core of emission reduction efforts. For countries with large-scale coal-fired power generation capacity and relatively new units, large-scale shutdowns of coal-fired units in the short term would jeopardize energy security and economic stability. Therefore, implementing low-carbon retrofits on existing units is generally regarded as a more realistic transformation path.

[0020] The low-carbon retrofit routes for coal-fired power units are mainly divided into two categories: fuel substitution and carbon capture, utilization, and storage (CCUS). Among them, carbon capture technology, which achieves deep emission reduction while preserving the regulation capacity of the power unit, has become an important choice. Carbon capture mainly includes three routes: post-combustion capture, pre-combustion capture, and oxy-fuel combustion capture. Post-combustion capture technology is mature, but solvent regeneration has high energy consumption, significantly reducing unit efficiency; pre-combustion capture system retrofitting is complex and requires high investment. Oxy-fuel combustion, which burns in an oxygen-enriched atmosphere, results in a high concentration of carbon dioxide in the flue gas, simplifying the downstream capture process and reducing capture costs, and is increasingly favored in retrofitting applications. However, oxy-fuel combustion retrofitting still faces two bottlenecks: first, the cryogenic air separation device required for oxygen supply consumes a lot of electricity, resulting in a significant energy penalty; second, the high-purity carbon dioxide obtained from capture lacks high-value utilization, and relying solely on geological storage wastes the economic value of this carbon stream, and is also limited by the resources of storage sites.

[0021] See Figure 1 and Figure 2 This invention addresses existing problems by proposing a low-carbon power system planning method that integrates oxygen-enriched combustion carbon capture power plants with methanol electrolysis. In the proposed system, existing thermal power units are retrofitted for oxygen-enriched combustion, and a new methanol electrolysis plant is planned in conjunction with this system, thereby achieving the recycling of carbon and oxygen. In the methanol electrolysis plant, renewable electricity is used for electrolysis to produce green hydrogen, captured carbon dioxide is converted into methanol, and the oxygen produced as a byproduct of electrolysis is fed back to the oxygen-enriched boiler to partially replace the oxygen supply from the air separation unit. The method specifically includes the following steps.

[0022] Step 1: Obtain the net carbon dioxide emissions of the thermal power unit and the oxygen produced as a byproduct of the electrolysis process in the production of methanol.

[0023] For thermal power units, the net carbon dioxide emissions of thermal power units are obtained by using the logical coupling constraints between the modification decision and the operation mode, the capture rate related to the operation mode, the oxygen demand in the oxygen-enriched mode, and the energy consumption penalty generated by air separation and carbon dioxide compression.

[0024] The fuel consumption of thermal power units is described by a quadratic characteristic curve, and the carbon dioxide generated is proportional to the fuel consumption, as shown in equations (1) and (2): (1); (2); The modification decision and operation mode are coupled through logical constraints, that is, the unit can only operate in oxygen-enriched mode after the oxygen-enriched combustion modification is completed, as shown in Equation (3); the capture rate is determined by the operation mode and expressed in linear form. This linearization is accurate enough on a long-term planning scale, as shown in Equation (4); the carbon dioxide capture and emission of the unit are shown in Equations (5) and (6): (3); (4); (5); (6); Oxygen injection is only required in oxygen-enriched mode. The oxygen demand is adjusted by the air-to-volume ratio and oxygen purity, as shown in Equation (7). The net output of the unit is obtained by subtracting the plant power consumption and energy consumption penalty from the gross power generation. The energy consumption penalty of air separation and carbon dioxide compression is only generated when oxygen-enriched mode is activated, as shown in Equations (8) to (12). (7); (8); (9); (10); (11); (12); The above model shows that the local air separation load can be reduced through external oxygen feedback, thus laying the foundation for carbon-oxygen coupling in step S3.

[0025] In the electro-methanol plant, hydrogen production and by-product oxygen are obtained through electrolysis efficiency and stoichiometry. After compression, methanol is synthesized. The methanol output is priced using a dual-tier pricing system of green methanol price and gray methanol price, based on the demand of the green methanol cycle.

[0026] The site selection and capacity determination of the methanol electrolysis station are constrained by the upper and lower limits of the single station capacity and the maximum number of stations in the system, as shown in equations (13) and (14); the hourly operation is constrained by the installed capacity and the electrolysis ramp-up, and the hydrogen production and by-product oxygen production are determined by the electrolysis efficiency and stoichiometric ratio, as shown in equations (15) to (18): (13); (14); (15); (16); (17); (18); Hydrogen and carbon dioxide are adiabatically compressed to the reactor operating pressure before synthesis, with compression power as shown in equations (19) to (21). In the reactor, hydrogen and carbon dioxide are stoichiometrically synthesized into methanol using the comprehensive conversion efficiency (CO2 + 3H2 → CH3OH + H2O). Methanol is priced using a green / gray dual-level pricing system based on the green methanol cycle demand: the portion not exceeding the green methanol demand is priced at the green price, and the excess portion is priced at the lower gray price. Green methanol and gray methanol are distinguished by raw material source and carbon emissions, as shown in equations (22) to (25). (19); (20); (twenty one); (twenty two); (twenty three); (twenty four); (25); Step 2: The oxygen produced by electrolysis is fed back to the thermal power unit through the oxygen pipeline network to replace the oxygen production load of the air separation unit in the thermal power unit. The net carbon dioxide emissions captured by the thermal power unit are allocated to two paths: geological storage and as raw material for electro-methanol production. The operating cost is obtained by combining the cost of geological storage and the revenue from the dual pricing of electro-methanol synthesis.

[0027] Construct a model of storage tanks and dedicated transportation pipelines, and establish a two-way carbon-oxygen cycle coupling constraint (oxygen recycling to replace air separation, carbon dioxide sequestration or utilization and distribution).

[0028] For storage tanks containing materials such as hydrogen, oxygen, and carbon dioxide, a unified discrete-time model is adopted, taking into account self-loss, filling and discharging rates, upper and lower capacity limits, and cyclic boundary conditions of the operating cycle, as shown in equations (26) to (30): (26); (27); (28); (29); (30); For dedicated transport pipelines, to avoid ineffective investment, a pipeline segment can only be constructed when the facilities at both ends of the carbon dioxide or oxygen pipeline segment have been completed, as shown in equations (31) and (32); the gas transport of the pipeline segment is constrained by the pipeline capacity, as shown in equations (33) and (34); the booster is used to compensate for pressure drop and maintain pipeline pressure, and its power is shown in equations (35) to (38): (31); (32); (33); (34); (35); (36); (37); (38); Based on this, a carbon-oxygen bidirectional cycle coupling constraint is established: the carbon dioxide captured by the thermal power unit is distributed to two routes: geological storage and transmission to the power plant via pipeline, as shown in equations (39) and (40); the oxygen demand of the oxygen-enriched unit is met by the oxygen fed back from the power plant via pipeline and the local air separation unit, where equation (41) is the explicit planning constraint for oxygen feedback to replace part of the air separation oxygen production load, and the by-product oxygen of the power plant is distributed in the network, with the surplus being vented, as shown in equations (41) and (42). (39); (40); (41); (42); By writing oxygen substitution as an explicit planning constraint through equation (41), the amount of oxygen substitution is directly incorporated into the investment decision, thereby simultaneously alleviating the two bottlenecks of air separation energy consumption penalty and carbon dioxide idleness in oxygen-enriched combustion transformation. This is the key difference between the present invention and the prior art.

[0029] Step 3: Using the minimum net present value of investment costs and operating costs within the planning period as the objective function of the power network model, the long-term uncertainty (i.e., long-term fluctuation) of carbon price and new energy penetration rate is represented as a scenario tree, and a multi-stage stochastic programming model is established.

[0030] Among them, the multi-stage stochastic programming process is as follows: Figure 3As shown. The power network adopts a DC power flow model. The line power flow is determined by the phase angle difference between the two nodes and is constrained by the thermal stability limit, as shown in equations (43) to (45). The unit combination constraint describes the operating behavior of the generator, including the upper and lower limits of output, ramping, minimum start-stop time and start-stop logic, as shown in equations (46) to (50). The upper limits of wind curtailment and load shedding are shown in equations (51) and (52). The node power balance is established between power generation, wind power, line power flow, methanol power generation load and transmission boost load, as shown in equation (53).

[0031] (43); (44); (45); (46); (47); (48); (49); (50); (51); (52); (53); The objective function is the sum of the net present value of investment costs and operating costs at each stage within the planning period, discounted at the discount rate, as shown in Equation (54). Investment costs include unit modification, electric methanol production station and pipeline construction, as shown in Equation (55). Operating costs include fuel and start-up, carbon tax and storage, transportation, wind curtailment and load shedding penalties, and methanol revenue included in negative terms, as shown in Equation (56). Among them, the long-term uncertainty scenario tree is as follows: Figure 4 As shown, carbon tax and new energy penetration rate are solved as uncertainties along the scenario tree: (54) (55); (56); The low-carbon transition spans decades, and carbon prices and the penetration rate of new energy have evolved gradually rather than being known in advance. Therefore, the planning model must commit to investment in stages, and each decision must be adjusted based on the information already revealed. Compared with deterministic mixed-integer linear programming that commits to all options at once and overly conservative robust optimization, multi-stage stochastic programming portrays the transition as a sequence of follow-up decisions on a scenario tree, which is more in line with the phased nature of the transformation. For any stage τ and the uncertainty ξ_τ of realization, the stage problem is as shown in Equation (57), which minimizes the sum of the current stage cost and the expected future cost under the constraints of state transition, investment domain and operational constraints under each realization. The planning strategy is a mapping from the uncertainty history to the investment decision, as shown in Equation (58):

[0032] (57); (58); Step 4: The adaptive precision stochastic dual dynamic integer programming algorithm is used to solve the multi-stage stochastic programming model to obtain a phased adaptive planning scheme for the renovation of thermal power units, the construction of methanol power plants, and the investment in transportation pipelines.

[0033] Because the binary investment variables make the future cost function non-convex and discontinuous, traditional stochastic dual dynamic programming relying on Benders' cut is not applicable. Therefore, a stochastic dual dynamic integer programming algorithm incorporating Lagrange cut is adopted; and an adaptive precision variant is proposed: maintaining a relaxed tolerance during the exploration phase, and tightening the tolerance to enter the development phase when the lower bound trajectory stagnates. The algorithm includes binary approximation, initialization, sampling, forward process, backward process, and convergence and verification checks. The convergence characteristics of the adaptive precision stochastic dual dynamic integer programming algorithm are as follows: Figure 5 As shown.

[0034] (1) Binary approximation: To ensure the effectiveness and compactness of the Lagrange cut, the state variables connecting adjacent stages are required to be binary variables. Therefore, the continuous state variables in the multi-stage stochastic programming model are discretized into binary variables, as shown in equations (59) to (61): (59); (60); (61); (2) Initialization: Set the lower bound to negative infinity and the upper bound to positive infinity. The cut pool of the future cost function is initialized to empty.

[0035] (3) Sampling: Sample several root-to-leaf scene sequences from the scene tree so that the original subproblems in the forward process can be solved under deterministic conditions.

[0036] (4) Forward process: From the first stage to the last stage, the stage problem is solved stage by stage under the current solver tolerance, and the future cost is approximated by the following approximation function, as shown in Equation (62) and Equation (63); after all sampling scenarios are simulated, the statistical upper bound is updated by the sample mean and standard deviation according to the confidence interval, as shown in Equation (64): (62); (63); (64); (5) Backward process: From the last stage to the second stage, a local replication variable is introduced for each uncertainty and the unexpected constraint is relaxed to obtain the Lagrange dual, as shown in Equation (65); the slope of the cut is obtained by the dual multiplier of the local replication constraint, and the intercept is obtained by evaluating the value of the Lagrange function, as shown in Equations (66) and (67); the average of the sampled scenarios is taken to generate a reinforcement cut and added to the problem of the previous stage, as shown in Equation (68): (65); (66); (67); (68); (6) Convergence and Verification: At the end of each iteration, compare the relative gap between the upper and lower bounds with the target tolerance, as shown in equation (gap). If the gap meets the tolerance, trigger the verification step: tighten the solver tolerance to the minimum and perform an additional forward / backward process, terminating only if the verification gap still meets the tolerance; otherwise, reject the convergence and penalize the scaling factor. If the gap does not meet the tolerance and the lower bound improvement stagnates, enter the development phase and tighten the main problem tolerance according to equation (tighten):

[0037] ; ; This mechanism ensures that the reported convergence reflects the true optimal gap, rather than an illusion caused by relaxation tolerance.

[0038] Example verification: To verify the effectiveness of this invention, analysis was conducted on an improved IEEE 24-node system. This system comprises 12 thermal power units, all considered candidates for oxy-fuel combustion carbon capture retrofitting; nodes {7, 16, and 22} were selected as candidate sites for methanol-to-electricity plants, with a baseline load of 3200 MW; and the IEEE 118-node system was used to verify scalability. The planning period is divided into three phases, each representing five years. The long-term uncertainties of carbon prices and renewable energy penetration rates are characterized by a scenario tree, as shown in Table 1.

[0039] Table 1. Scenario Tree of Carbon Price and New Energy Penetration Rate stage node Parent node Conditional probability Path probability Carbon price (¥ / t) Penetration 1 S1 (root) - 1.00 1.00 100 0.30 2 S2-L S1 0.40 0.40 120 0.45 2 S2-H S1 0.60 0.60 160 0.55 3 S3-LL S2-L 0.30 0.12 160 0.55 3 S3-LM S2-L 0.40 0.16 200 0.65 3 S3-LH S2-L 0.30 0.12 240 0.80 3 S3-HL S2-H 0.30 0.18 160 0.55 3 S3-HM S2-H 0.40 0.24 200 0.65 3 S3-HH S2-H 0.30 0.18 240 0.80 The results of the multi-stage investment plan are shown in Table 2. The plan presents a clear "renovation first, then conversion" sequence: In the first stage, a conservative strategy was adopted under high initial uncertainty, and units {3, 5, 6} were renovated, with a cumulative renovation capacity of 760MW. Due to insufficient penetration, the construction of the electro-methanol plant was temporarily suspended. In the second stage, the plan was differentiated along different branches, with the renovation capacity of the high-penetration branch exceeding that of the low-penetration branch. In the third stage, as the penetration rate reached its peak, electro-methanol plants were deployed in all scenarios, and the site selection was adaptive according to the implementation branch (node ​​16 was selected for the low-penetration branch, and node 22 was selected for the high-penetration branch).

[0040] Table 2. Improved Multi-Stage Investment Planning Scheme for the IEEE 24-Node System stage node New and upgraded units Cumulative retrofit capacity (MW) Site of electro-methanol production station Electrolysis capacity (MW) 1 S1 (root) {3,5,6} 760 - 0 2 S2-L {11} 1201 - 0 2 S2-H {7,10} 1360 - 0 3 S3-LL - 1201 16 125.8 3 S3-LM - 1201 16 135.8 3 S3-LH {1} 1501 16 163.0 3 S3-HL - 1360 22 133.6 3 S3-HM - 1360 22 136.0 3 S3-HH {1} 1660 22 139.1 The cost structure of each node is shown in Table 3. The total expected cost decreases along the planning cycle, from approximately RMB 3,409 million in Phase 1 to approximately RMB 2,369 million in Phase 3. The investment in the EPC methanol plant in Phase 3 is offset by methanol revenue, which is approximately -RMB 418.6 to -RMB 428.2 million. This transforms the EPC methanol plant from a pure load to a profitable asset, thereby achieving economic sustainability through deep decarbonization.

[0041] Table 3. Cost breakdown of each node in the improved IEEE 24-node system (million RMB) node invest Power generation operation Carbon related wind abandonment punishment Other operations Methanol revenue Total stage cost S1 (root) 66.88 3065.2 275.9 0.84 0.00 0 3408.8 S2-L 39.12 2616.9 265.4 0.00 0.51 0 2922.0 S2-H 53.03 2375.0 313.4 23.86 0.68 0 2766.1 S3-LL 50.46 2534.1 344.8 14.92 5.72 -420.36 2529.6 S3-LM 54.51 2326.3 385.6 49.71 5.82 -426.93 2395.0 S3-LH 92.07 2045.2 350.3 106.51 6.96 -428.24 2172.8 S3-HL 53.86 2535.7 344.5 15.59 0.62 -418.63 2531.7 S3-HM 54.56 2318.7 386.2 49.46 0.76 -420.65 2389.0 S3-HH 82.62 2063.8 338.9 112.17 1.99 -425.49 2174.0 To contribute to the integrated isolation design and carbon-oxygen cycle mechanism, this invention (C4) is compared with three benchmark schemes: no investment benchmark (C1), pure capture scheme with complete sequestration after capture (C2), and scheme with oxygen feedback disabled (C3), as shown in Table 4.

[0042] Table 4 Comparison between the improved IEEE 24-node system and the baseline scheme plan Investment (M¥) Carbon cost (M¥) Methanol revenue (M¥) Net emissions (kt / a) Capture rate (%) Oxygen substitution rate (%) Total cost (M¥) gap(%) C1 0 3354.0 0 6660 0 - 11070.3 +28.6 C2 170.1 873.4 0 1607 76.9 - 9279.8 +7.8 C3 114.0 1009.1 -363.7 2064 66.8 0 8630.9 +0.3 C4 (This invention) 178.0 931.2 -423.4 1849 74.7 39.95 8605.2 0 (benchmark) The results show that while the pure capture scheme (C2) achieves the greatest emission reduction, it completely seals the carbon stream without any benefit, and its total cost is 7.8% higher than that of this invention. When oxygen feedback is disabled (C3), the air separation unit operates at full capacity, driving up the cost of oxygen enrichment. The model is forced to reduce oxygen enrichment operations, resulting in a decrease in the capture rate from 74.7% to 66.8% and an increase in carbon-related costs from RMB 931.2 million to RMB 1009.1 million. With oxygen feedback enabled (C4), electrolytic oxygen replaces 39.95% of the air separation demand, achieving a higher capture rate and greater methanol production while reducing the marginal cost of capture, and reducing the total cost by approximately 28.6% compared to the baseline without investment (C1). This invention achieves the lowest expected total cost of RMB 8605.2 million.

[0043] Regarding performance, the adaptive precision algorithm of this invention is compared with traditional stochastic programming (CSP) and standard stochastic dual dynamic integer programming (standard SDDiP), as shown in Table 5. This invention converges in 23 iterations and approximately 21616 seconds in 3 stages and 9 scenarios, with the lower bound monotonically increasing and stabilizing at approximately 8468 million, and the relative gap shrinking to approximately 1.85%. CSP directly constructs deterministic equivalent problems, and within practical time limits (exceeding 10... 5 The solution cannot return a feasible solution in seconds; the two SDDiP variants converge to essentially the same bound. Compared to the standard SDDiP's 33 iterations, 30808 seconds, and 1.88% gap, this invention achieves a comparable 1.85% gap in 23 iterations and 21616 seconds, reducing both the number of iterations and solution time by about one-third, while generating fewer cuts (121 vs. 177) and maintaining the same solution quality.

[0044] Table 5. Solution performance of different methods for improving the IEEE 24-bus system method CSP Standard SDDiP This invention, AP-SDDiP Total time (s) <![CDATA[>10 5 ]]> 30808 21616 Number of iterations - 33 23 Final Lower Bound (M¥) - 8468 8468 Ultimate Upper Boundary (M¥) - 8630 8627 Generate cut number - 177 121 Final gap (%) - 1.88 1.85 Furthermore, the results on the IEEE 118-node system are consistent with those on the 24-node system: the model also prioritizes the modification of baseload units, postpones investment in electro-methanol production to the high-penetration stage, and adapts site locations according to branches; the total cost decreased from approximately RMB 17,136 million in the first phase to approximately RMB 13,189 million in the third phase, verifying the scalability of the invention on large-scale systems.

[0045] This invention establishes a carbon-oxygen two-way closed loop through an oxygen feedback mechanism and targeted carbon dioxide utilization. The oxygen feedback from electrolysis byproducts is written as an explicit planning constraint, allowing oxygen substitution to be directly incorporated into investment decisions. This simultaneously alleviates the two bottlenecks of air separation energy consumption penalties and carbon dioxide idleness in oxygen-enriched combustion retrofits. The long-term transformation is characterized as a multi-stage stochastic planning on a scenario tree, with carbon price and new energy penetration rate as stochastic processes. Retrofit and conversion investments adapt sequentially as uncertainties are gradually revealed, transforming static techno-economic assessments into dynamic deployment planning. The resulting solution is both economical and robust.

[0046] The following explains the symbols present in the above formula: Table 6 General Subscripts, Sets and Symbols Formula range Symbols or abbreviations Suggested interpretation The whole article g;Ω_G g represents the thermal power unit index, and Ω_G represents the set of thermal power units. The whole article k;Ω_P2M k represents the index of the candidate station for electro-methanol production, and Ω_P2M represents the set of candidate stations. The whole article n、n′;Ω_N n and n′ represent nodes in the power system, and Ω_N represents the set of nodes. The whole article l;Ω_Line l represents a transmission line, Ω_Line represents the set of transmission lines; l(n,n′) represents the line connecting nodes n and n′. The whole article t、t′;Ω_T t and t′ represent runtime segments, and Ω_T represents the set of runtime segments. Planning Model τ;T τ represents the planning stage, and T represents the total number of planning stages. Energy storage model i; Ω_gas i represents a gas storage tank or storage medium, and Ω_gas represents the set of tanks under consideration. Algorithm Model s, σ, m, j, k These represent the sampling scenario / sample, random implementation, backward process sample, binary bit or cut number, and algorithm iteration number, respectively; they should be confirmed in conjunction with the original algorithm. Node balancing Ω_n^G, Ω_n^L, Ω_n^P These represent the sets of generators, lines, and methanol-producing stations connected to node n, respectively. General notation underline, overline These represent the lower and upper limits of a variable or parameter, respectively. General notation Δ、^、*、⊤ Δ represents the increment or change; the cap represents the sample estimate; the asterisk represents the optimal value; and ⊤ represents the transpose of a vector or matrix.

[0047] Table 7 Symbols in Thermal Power Units and Oxygen-Enriched Combustion Models Formula range Symbols or abbreviations Suggested interpretation Equation (1) F_g,t The fuel consumption of unit g during time period t. Equation (1) P_g,t^gen The gross power generation of unit g during time period t; gen represents generation. Equation (1) a_g, b_g, c_g The coefficients of the quadratic, linear, and constant terms of the unit's fuel consumption quadratic characteristic curve. Equation (2) <![CDATA[m_g,t^(CO2,gen)]]> The amount of carbon dioxide produced by unit g due to fuel combustion during time period t. Equation (2) δ^coal The carbon dioxide generation coefficient or emission factor corresponding to coal consumption should be specified in the model caliber, with the specific dimensions indicated. Equation (3) u_g,t^oxy Whether unit g is in oxygen-enriched operation mode during time period t is a 0-1 variable; oxy represents oxy-fuel. Equation (3) x_g^retro The variable is 0-1 indicating whether unit g has completed the oxygen-enriched combustion carbon capture retrofit; retro means retrofit. Equation (4) η_g,t^cap;η^base;η^oxy These represent the actual capture rate, the capture rate under normal operating conditions, and the capture rate under oxygen-enriched conditions, respectively; cap stands for capture. Equations (5)-(6) <![CDATA[m_g,t^(CO2,cap)、m_g,t^(CO2,emit)]]> These represent the amount of carbon dioxide captured and the net emissions of unit g during time period t, respectively. Equations (5)-(6) e^coal The carbon dioxide emission factor for coal combustion calculated based on power generation should be consistent with the δ^coal in equation (2). Equation (7) <![CDATA[m_g,t^(O2,req)]]> The oxygen demand for oxygen-enriched combustion of unit g during time period t; req indicates required. Equation (7) <![CDATA[r^O2、r^exc、e^pur]]> It is recommended to define them as the theoretical oxygen demand coefficient per unit of power generation, the excess air correction coefficient, and the oxygen purity correction coefficient, respectively, and specify the dimensions. Equations (8)-(9) P_g,t^net、P_g,t^aux、P_g,t^pen、P_g,t^post These represent net output, conventional plant power consumption, energy consumption penalty power, and post-treatment energy consumption power in non-oxygen-enriched mode, respectively. Equations (10)-(12) P^CPU, P^ASU; λ^post, λ^CPU, λ^ASU CPU represents the carbon dioxide purification / compression unit, and ASU represents the air separation unit; each λ represents the power consumption coefficient per unit processing capacity of the corresponding unit.

[0048] Table 8 Symbols in the Electro-Methanol Production Station and Synthesis Model Formula range Symbols or abbreviations Suggested interpretation Equations (13)-(14) x_k^P2M、Z_k^P2M represents the 0-1 variables indicating whether candidate station k should be built as an electric methanol plant, and the installed capacity of the station, respectively; P2M represents power-to-methanol. Equations (13)-(14) Z̲、Z̄、N^P2M These represent the lower limit and upper limit of the single-site installed capacity, as well as the maximum number of sites that the system is allowed to build, respectively. Equations (15)-(16) P_k,t^EL; r^(EL,dn), r^(EL,up) These represent the operating power of the electrolytic cell and the downward and upward ramp rates of the electrolytic cell, respectively; EL stands for electrolyzer. Equation (17) <![CDATA[m_k,t^(H2,syn)、η^EL、LHV^H2]]> These represent the amount of hydrogen used in synthesis, the efficiency of hydrogen production by electrolysis, and the lower heating value of hydrogen, respectively. Equation (18) <![CDATA[m_k,t^(O2,syn)、r^(O2 / H2)]]> These represent the amount of oxygen produced as a byproduct of electrolysis, and the stoichiometric mass ratio of oxygen to hydrogen, respectively. Equations (19)-(20) <![CDATA[P^(H2,cmp)、P^(CO2,cmp)]]> These represent the compression power of hydrogen and carbon dioxide, respectively; cmp stands for compression. Equations (19)-(20) C_p, T^in, η^cmp These represent the specific heat capacity of the gas at constant pressure, the compressor inlet temperature, and the compression efficiency, respectively. Equations (19)-(20) <![CDATA[p^syn、p^EL、p^CO2]]> These represent the pressure of the synthesis reactor, the hydrogen inlet pressure, and the carbon dioxide inlet pressure, respectively. Equations (19)-(20) <![CDATA[κ_H2、κ_CO2]]> These represent the specific heat ratios of hydrogen and carbon dioxide, respectively. Equation (21) P_k,t^P2M The total power consumption of the electric methanol production station k during time period t. Equation (22) <![CDATA[m_k,t^(CO2,req)、M_ω^CO2、M_ω^H2]]> These represent the amount of carbon dioxide required for synthesis, and the molar mass of carbon dioxide and hydrogen, respectively. Equation (23) m_k,t^MeOH, η^MS, M_ω^MeOH These represent methanol yield, overall conversion efficiency of methanol synthesis, and molar mass of methanol, respectively; MS stands for methanol synthesis. Equations (24)-(25) m_k,t^green、D_k,t^green These represent the quantity of methanol priced at the green methanol price and the upper limit of green methanol demand for the corresponding period, respectively.

[0049] Table 9 Symbols in the tank, pipeline, and carbon-oxygen coupling model Formula range Symbols or abbreviations Suggested interpretation Equation (26) S_i,t, r_i^loss, m_i,t^in, m_i,t^out, Δt These represent the inventory, self-loss rate, filling flow rate, discharge flow rate, and single running time length of storage tank i, respectively. Equations (27)-(28) m̄_i^in、m̄_i^out These represent the maximum filling rate and maximum discharge rate of storage tank i, respectively. Equation (29) Z_i^store, e̲^store, ē^store These represent the construction status or capacity decision of the storage tank, as well as the lower and upper limits of the available storage capacity after construction; "store" represents storage. Equation (30) S_i,0、S_i,T These represent the initial and final inventory levels of the operating cycle, respectively. Their equality indicates the cycle boundary condition. Equations (31)-(32) <![CDATA[x_g,k^CO2、x_k,g^O2]]> These indicate whether the carbon dioxide pipeline section from unit g to station k and the oxygen pipeline section from station k to unit g have been constructed, respectively. Equations (33)-(34) <![CDATA[f_g,k,t^CO2、f_k,g,t^O2、f̄^pipe]]> These represent the flow rates of the carbon dioxide and oxygen pipeline sections, respectively, as well as the upper limit of the pipeline's transport capacity. Equations (35)-(36) <![CDATA[P^(CO2,tr)、P^(O2,tr)]]> These represent the boosting power for transporting carbon dioxide and oxygen, respectively; tr stands for transport. Equations (35)-(36) Δp, L_g,k, η^pump These represent the pressure increase per unit length, pipe section length, and booster efficiency, respectively. Equations (35)-(36) <![CDATA[m^CO2、m^O2]]> The physical meaning of the symbol without a subscript is unclear; if it represents density, specific volume, or mass-volume conversion factor, it should be renamed and its dimensions specified. Equations (37)-(38) P_k,t^tr、P_g,t^tr These represent the total power of transmission boosting undertaken by station k and unit g, respectively. Equations (39)-(40) <![CDATA[m_g,t^(CO2,store)、m_k,t^(CO2,in)]]> These represent the amount of carbon dioxide sent to geological storage and the amount of carbon dioxide entering station k for synthesis, respectively. Equation (41) <![CDATA[m_g,t^(O2,ASU)]]> The amount of oxygen supplied by the local air separation unit to unit g during time period t. Equation (42) <![CDATA[m_k,t^(O2,gen)]]> The total amount of oxygen produced by electrolysis at station k during time period t.

[0050] Table 10 Symbols in the Power Network and Unit Combination Model Formula range Symbols or abbreviations Suggested interpretation Equation (43) f_l,t、φ_n,t、φ_n′,t、X_l These represent the active power flow of line l, the phase angles of the nodes at both ends of the line, and the line reactance, respectively. Equation (44) φ_ref,t The phase angle of the reference node in time period t is fixed at 0; π is the allowable boundary of the node phase angle. Equation (45) f̄_l^max The upper limit of the thermally stable transmission capacity of line l. Equation (46) u_g,t、P̲_g、P̄_g These represent the unit's start-up status, minimum stable output, and maximum output, respectively. Equation (47) R_g^dn、R_g^up These represent the downward and upward ramp limits for unit g, respectively. Equations (48)-(49) T_g^on、T_g^off、v_g,t、w_g,t These represent the minimum startup time, minimum shutdown time, startup status variable, and shutdown status variable, respectively. Equation (50) u_g,t、v_g,t、w_g,t The logical relationship between the unit's operating status, startup status, and shutdown status. Equation (51) P_n,t^curt、P_n,t^Wind Let n represent the curtailed wind power and the available wind power at node n, respectively. Equation (52) P_n,t^shed、P_n,t^Load These represent the load shedding power and load demand of node n, respectively. Equation (53) P_n,t^tr The gas transport pressurization electrical loads are aggregated to node n.

[0051] Table 11 Investment, Operating Costs, and Symbols in Multistage Stochastic Programming Formula range Symbols or abbreviations Suggested interpretation Equation (54) r^d、C_τ^inv、C_τ^op These represent the discount rate, the investment cost of stage τ, and the operating cost, respectively. Equation (55) <![CDATA[Δx_g,τ^retro、Δx_g,k,τ^CO2、Δx_k,g,τ^O2]]> These represent the investment decisions for the newly added unit upgrades, carbon dioxide pipeline sections, and oxygen pipeline sections in phase τ, respectively. Equation (55) ΔZ_k,τ^P2M The newly added installed capacity of electric methanol production in phase τ. Equation (55) c^(fix,oxy) and c^(var,oxy) Fixed investment costs for oxygen-enriched combustion retrofitting and variable investment cost coefficients calculated based on unit capacity. Equation (55) c^(fix,P2M) and c^(var,P2M) Fixed investment costs and variable investment cost coefficients based on installed capacity for electric methanol production stations. Equation (55) <![CDATA[c^retro、c^CO2、c^O2]]> It is suggested that these be interpreted as additional costs for the renovation and the unit length construction cost of the carbon dioxide and oxygen pipelines, respectively; the specific function of c^retro needs to be verified. Equation (56) c^coal, c^start These represent the unit fuel cost and the unit start-up cost, respectively. Equation (56) c^store, c_τ^tax, c^tr These represent the unit cost of carbon dioxide geological storage, the carbon tax price at stage τ, and the unit cost of gas transportation, respectively. Equation (56) c^curt、c^shed These represent the unit cost of wind curtailment penalty and the unit cost of load shedding penalty, respectively. Equation (56) c_τ^(meth,green), c_τ^(meth,grey) The sales prices of green methanol and gray methanol in stage τ. Equation (57) x_τ−1, x_τ, Δx_τ These represent the cumulative investment status of the previous stage, the cumulative investment status of the current stage, and the new investment decision vector of the current stage, respectively. Equation (57) y_τ、ξ_τ Let represent the operational decision vector of stage τ and the uncertain parameter vector realized in that stage, respectively. Equation (57) θ_τ、θ_τ+1 Let represent the value function of stage τ and its approximation of the expected future cost of the next stage, respectively. Equation (57) c_τ、d_τ These represent the cost coefficient vectors for investment and operational decisions at stage τ, respectively. Equation (57) G_τ、W_τ、h_τ、T_τ The coefficient matrix, right-hand vector, and uncertainty mapping matrix in the stage τ running constraints; their respective dimensions need to be specified in the model. Equation (57) Ω_τ、n_bin、n_con、ℝ_+ Let represent the feasible investment domain, the number of binary operating variables, the number of continuous non-negative variables, and the non-negative real number domain for stage τ, respectively. Equation (58) ℋ_τ、X_τ(ℋ_τ) These represent the historical uncertainty revealed up to stage τ, and the investment decisions made based on that history for stage τ, respectively.

[0052] Table 12 Symbols in the Adaptive Precision Stochastic Dual Dynamic Integer Programming Algorithm Formula range Symbols or abbreviations Suggested interpretation Equations (59)-(61) U, B, b_j, z These represent the upper bound of the continuous state variable to be discrete, the required number of binary bits, the j-th binary variable, and its binary reconstructed value, respectively. Equation (62) θ_τ,s^k、x_τ−1,s^k Let represent the subproblem value of stage τ in the k-th iteration, scenario s, and the previous stage state obtained from the forward process, respectively. Equation (62) 𝒳_τ、𝒲_τ、𝓜_τ The constraint matrix or mapping matrix in the sub-problem of the scenario stage should correspond to G_τ, W_τ, and T_τ in equation (57). Equation (62) 𝓛_τ+1^j、𝒥_τ Let represent the approximate cut function for the j-th future cost and the set of indices for the existing cuts in stage τ, respectively. Equation (63) Q_τ^k(x_τ), H_τ Let represent the approximate function for the future cost in the k-th iteration and the preset initial lower bound, respectively. Equation (63) N_τ+1, v_τ+1,n^j, π_τ+1,n^j Let represent the number of random implementations in the next stage, and the intercept and slope vector of the j-th cut under implementation n, respectively. Equation (64) N_s、𝒯_s^k Let represent the number of forward simulation samples and the total cost corresponding to sample s in the k-th iteration, respectively. Equation (64) μ̂_k, σ̂_k, UB_k, z_α / 2 These represent the standard normal quantiles corresponding to the sample mean, sample standard deviation, statistical upper bound, and confidence level, respectively. Equation (65) D_τ,σ^k,z_τ−1,π_τ,s Let represent the Lagrange dual problem value, the local copy variable of the previous stage state, and the dual multiplier of the unexpected constraint, respectively. Equation (66) L_τ,m^k(ρ_τ,m*) The Lagrangian function value evaluated at the k-th iteration, sample m, and optimal multiplier ρ_τ,m*. Equation (67) β̄_τ,s* The optimal intercept of the Lagrange cut in scenario s. Equation (68) ρ_τ,s*、β̄_τ,s* Let represent the optimal slope vector and intercept of the Lagrange cut in scenario s, respectively; the enhanced cut is formed by averaging the samples. (gap) Δ^k, UB^k, LB^k Let represent the relative optimal gap, upper bound, and lower bound of the k-th iteration, respectively. Tighten δ_solver^k, ε_min, ψ These represent the solver tolerance, the minimum allowable tolerance, and the tolerance scaling factor for the k-th iteration, respectively; when tightening, 0 < ψ < 1 is usually required.

[0053] Based on the above method, this invention proposes a low-carbon power system planning system, comprising: The system comprises several modules: a data acquisition module for obtaining net carbon dioxide emissions from thermal power units and oxygen produced as a byproduct of electrolysis during methanol production; a carbon-oxygen coupling module for feeding back the oxygen produced as a byproduct of electrolysis to thermal power units via an oxygen pipeline network to replace the oxygen production load of the air separator in the thermal power units, and for allocating the net carbon dioxide emissions captured by the thermal power units to two paths: geological storage and as a raw material for methanol production, with the operating cost calculated using the cost of geological storage and the revenue from the dual-level pricing of methanol production; a stochastic programming modeling module for using the minimum net present value of investment and operating costs within the planning period as the objective function of the power network model, representing the fluctuations in carbon price and new energy penetration rate as a scenario tree, and establishing a multi-stage stochastic programming model; and a solution output module for solving the multi-stage stochastic programming model to obtain a phased adaptive planning scheme.

[0054] The present invention provides a computer device, including a memory and a processor. The memory stores a program, and when the program is executed by the processor, the processor performs the steps of the aforementioned low-carbon power system planning method.

[0055] According to the disclosed embodiments, the computer device can communicate with one or more external devices (e.g., keyboard, pointing device, Bluetooth communication, etc.) or with any device that enables the computing device to communicate with one or more other computing devices (e.g., router, demodulator, etc.).

[0056] The present invention provides a storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the aforementioned low-carbon power system planning method.

[0057] According to the disclosed embodiments, the storage medium can be a non-volatile computer-readable storage medium, such as, but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this invention, the storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0058] The above description, in conjunction with specific preferred embodiments, provides a more detailed explanation of the present invention. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such deductions or substitutions should be considered to fall within the scope of protection of the present invention.

Claims

1. A low-carbon power system planning method, characterized in that, Includes the following steps: To obtain the net carbon dioxide emissions of thermal power units and the oxygen produced as a byproduct of electrolysis during the production of methanol by electricity; The oxygen produced by electrolysis is fed back to the thermal power unit via the oxygen pipeline network to replace the oxygen production load of the air separation unit in the thermal power unit. The net carbon dioxide emissions captured by the thermal power unit are allocated to two paths: geological storage and as raw material for electro-methanol production. The operating cost is obtained by combining the cost of geological storage with the revenue from the dual-level pricing of electro-methanol synthesis. The objective function of the power network model is to minimize the net present value of investment costs and operating costs within the planning period. The fluctuations of carbon price and new energy penetration rate are represented as scenario trees, and a multi-stage stochastic programming model is established. The multi-stage stochastic programming model is solved to obtain a staged adaptive programming scheme.

2. The low-carbon power system planning method as described in claim 1, characterized in that, The acquisition of the net carbon dioxide emissions of the thermal power unit specifically involves: For thermal power units, the net carbon dioxide emissions of thermal power units are obtained by using the logical coupling constraints between the modification decision and the operation mode, the capture rate related to the operation mode, the oxygen demand in the oxygen-enriched mode, and the energy consumption penalty generated by air separation and carbon dioxide compression.

3. The low-carbon power system planning method as described in claim 1, characterized in that, In the aforementioned thermal power units, oxygen injection is only required in the oxygen-enriched mode, and the oxygen demand is adjusted by the air-to-volume ratio and oxygen purity. The net output of the thermal power units is obtained by subtracting the plant power consumption and energy consumption penalty from the gross power generation, and the energy consumption penalty of air separation and carbon dioxide compression is only generated when the oxygen-enriched mode is activated.

4. The low-carbon power system planning method as described in claim 1, characterized in that, When obtaining oxygen as a byproduct of electrolysis during the process of producing methanol by electrolysis, the site selection and capacity of the methanol production station are constrained by the upper and lower limits of the single station capacity and the maximum number of stations in the system. Hourly operation is constrained by the installed capacity and electrolysis ramp-up. The amount of hydrogen produced and the amount of byproduct oxygen are determined by the electrolysis efficiency and stoichiometric ratio. After adiabatic compression of hydrogen and carbon dioxide to the synthesis pressure, methanol is synthesized according to stoichiometric ratio and comprehensive conversion efficiency. The methanol output within the green methanol cycle demand is priced at the green methanol price, and the excess is priced at the gray methanol price. Green methanol and gray methanol are distinguished by the source of raw materials and carbon emissions.

5. The low-carbon power system planning method as described in claim 1, characterized in that, The process of feeding back the oxygen produced as a byproduct of electrolysis to the thermal power unit via an oxygen pipeline network to replace the oxygen production load of the air separation unit in the thermal power unit specifically involves: A unified discrete-time model is adopted for storage tanks, taking into account self-loss, filling and discharging rates, capacity upper and lower limits and periodic boundary conditions; for dedicated transportation pipelines, the pipeline is constructed only when the facilities at both ends of the carbon dioxide or oxygen pipeline are completed, the gas transportation of the pipeline is constrained by the pipeline capacity, and the power of the booster is taken into account. Synergy is achieved through carbon-oxygen bidirectional cyclic coupling constraints: the carbon dioxide captured by the thermal power unit is distributed to two routes: geological storage and transmission to the methanol-to-electricity plant via pipeline. The oxygen demand of the oxygen-enriched unit is met by the oxygen fed back from the methanol-to-electricity plant via pipeline and the local air separation unit.

6. The low-carbon power system planning method as described in claim 1, characterized in that, The objective function of the power network model is to minimize the net present value of investment and operating costs within the planning period. The fluctuations in carbon prices and renewable energy penetration are represented as scenario trees, and a multi-stage stochastic programming model is established, specifically as follows: The power grid adopts a DC power flow model, which takes into account the line power flow and thermal stability limits, the upper and lower limits of the output of the unit combination, the ramp rate, the minimum start and stop time and start and stop logic, the upper limit of wind curtailment and load shedding, and the node power balance. The objective function is the sum of the net present value of investment costs and operating costs at each stage within the planning period, discounted at the discount rate. Investment costs include unit modification, construction of the electric methanol plant and pipeline network, and operating costs include fuel and start-up, carbon tax and storage, transportation, wind curtailment and load shedding penalties, and methanol revenue included in the negative. The carbon tax and the penetration rate of new energy are solved as random variables along the scenario tree to obtain a multi-stage stochastic programming model.

7. A low-carbon power system planning method as described in claim 1, characterized in that, The solution to the multi-stage stochastic programming model yields a phased adaptive planning scheme, specifically as follows: The continuous state variables in the multi-stage stochastic programming model are discretized by binary approximation, and the upper and lower bounds and the cutting plane pool are initialized. The scene sequence from root to leaf is sampled from the scene tree. The forward process is executed, and the stage problem is solved stage by stage within the current solver tolerance. The future cost function is approximated by Lagrange cut and the statistical upper bound is updated. The backward process is executed, and local replication variables are introduced and unexpected constraints are relaxed to solve the Lagrange dual. The reinforcement cut is generated by the dual multiplier and the Lagrange function value and added to the previous stage problem. At the end of each iteration, the relative gap between the upper and lower bounds is compared. When the gap meets the target tolerance, the solver tolerance is tightened to the minimum and an additional forward or backward verification step is performed. The process terminates only when the verification gap still meets the tolerance; otherwise, the convergence is rejected and the scaling factor is penalized. When the gap does not meet the tolerance and the lower bound improvement stagnates, the development stage is entered and the problem tolerance is tightened, resulting in a phased adaptive planning scheme for the transformation of thermal power units, the construction of electric methanol production stations, and the investment in transportation pipelines.

8. A low-carbon power system planning system, characterized in that, include: The data acquisition module is used to acquire the net carbon dioxide emissions of thermal power units and the oxygen produced as a byproduct of electrolysis during the methanol electrolysis process. The carbon-oxygen coupling module is used to feed back the oxygen produced by electrolysis to the thermal power unit via the oxygen pipeline network, replacing the oxygen production load of the air separation unit in the thermal power unit, and to allocate the net carbon dioxide emissions captured by the thermal power unit to two paths: geological storage and as raw material for electro-methanol production. The operating cost is obtained by combining the cost of geological storage with the revenue from the dual-level pricing of electro-methanol synthesis. The stochastic programming modeling module is used to take the minimum net present value of investment costs and operating costs within the planning period as the objective function of the power network model, and to represent the fluctuations of carbon price and new energy penetration rate as a scenario tree to establish a multi-stage stochastic programming model. The solution output module is used to solve the multi-stage stochastic programming model and obtain a staged adaptive planning scheme.

9. A computer device, characterized in that, The system includes a memory and a processor, wherein the memory stores a program that, when executed by the processor, causes the processor to perform the steps of a low-carbon power system planning method as described in any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of a low-carbon power system planning method according to any one of claims 1 to 7.