A double-layer optimization scheduling method for wind-solar-thermal-hydrogen storage system facing deep regulation risk

Abstract

The application provides a wind-solar-thermal-storage-hydrogen system double-layer optimization scheduling method for deep regulation risk, constructs a system net load sequence and a dynamic risk potential field based on power system source network load storage data, establishes an upper layer optimization model, takes risk mitigation and regulation resource optimization allocation as targets, solves power distribution boundaries of each period, constructs a lower layer optimization scheduling model with the minimum system comprehensive operation cost as the target, solves the power distribution boundaries as constraints to obtain a scheduling scheme, performs risk identification and constraint checking on the scheme, if there is risk or the constraint is not satisfied, updates the risk potential field and the upper layer model according to the risk information and iterates, if there is no risk, the constraint is satisfied and the convergence condition is reached, the final scheme is output. The method can perceive deep regulation risk evolution, adaptively adjust resource priority, and utilize lower layer risk information closed loop correction, and can consider new energy consumption, thermal power deep regulation inhibition, energy storage life, P2G consumption and power grid safety.

Description

Technical Field

[0001] This application relates to the field of power system optimization dispatch and integrated energy coordinated control, and in particular to a two-layer optimization dispatch method for wind, solar, thermal, storage and hydrogen systems facing deep dispatch risks. Background Technology

[0002] To reduce fossil fuel consumption and increase the proportion of renewable energy, building a new power system dominated by new energy sources has become an important development direction for the power industry. However, new energy sources such as wind power and photovoltaics are characterized by randomness, volatility, and temporal uncertainty. After large-scale grid connection, they can easily lead to rapid fluctuations in net load, a sharp increase in peak-shaving demand, and increased difficulty in balancing spinning reserves. For thermal power units that play a supporting role in the system, in scenarios with a high proportion of surplus new energy, they are prone to operating in the deep peak-shaving range for extended periods, which can lead to problems such as deteriorating coal consumption, increased rotor fatigue damage, and increased costs of fuel oil injection for combustion.

[0003] Existing research has explored numerous approaches to renewable energy integration and multi-source coordinated dispatch. For example, some studies coordinate wind power integration costs and thermal power deep-dispatch costs through unit combination and economic dispatch models, while others enhance system flexibility through energy storage and balance constraints. Some research also introduces P2G hydrogen production systems to absorb surplus renewable energy, achieving the conversion and utilization of electricity into hydrogen. While these solutions improve renewable energy integration capabilities in certain aspects, they primarily focus on single issues such as wind-solar-storage synergy, independent P2G integration, or thermal power deep-dispatch cost modeling.

[0004] Existing technologies generally still have the following shortcomings: First, most active wind curtailment, solar curtailment, or equivalent load correction strategies rely on fixed threshold triggers, which are prone to frequent switching and power correction abrupt changes near the threshold, making it difficult to adapt to the continuous evolution of deep-level adjustment risks. Second, the adjustment order between energy storage, P2G, and active curtailment often adopts a pre-set fixed priority, failing to dynamically adjust based on the health status of energy storage, hydrogen storage capacity, and the duration of future surplus renewable energy. Third, the lower-level safety verification results mostly only transmit the risk type or uniform correction amount, lacking information on risk location, risk severity, and the sensitivity of different adjustment resources to risk mitigation, resulting in insufficient targeting of upper-level corrections. Summary of the Invention

[0005] This application provides a two-layer optimal scheduling method for wind, solar, thermal, and hydrogen storage systems to address deep scheduling risks. To solve the aforementioned technical problems, this application adopts the following technical methods: This application provides a two-layer optimization scheduling method for wind, solar, thermal, and hydrogen storage systems to address deep scheduling risks, including: S101: Obtain basic operational data of power system sources, grids, loads, and storage during the dispatching cycle; S102: Based on the power system's source-grid-load-storage basic operation data, construct the system net load sequence; and construct a dynamic risk potential field according to the system net load sequence; S103: Construct an upper-level optimization model with the goal of risk mitigation and adjusting the optimal allocation of resources; S104: Using the output of the dynamic risk potential field as input, solve the upper-level optimization model to determine the power allocation boundary of each regulation resource in each time period; S105: Construct a lower-level optimization scheduling model with the goal of minimizing the overall system operating cost; S106: Substitute the power allocation boundary as a constraint into the lower-level optimization scheduling model to obtain the optimization scheduling scheme; S107: Identify operational risks and verify constraints for optimized scheduling schemes; If there are operational risks or constraints are not met, the dynamic risk potential field and the upper-level optimization model are updated based on the risk information, and the iteration is returned. If there is no operational risk, the constraints are satisfied, and the risk convergence and iteration termination conditions are met, then the iteration stops and the final optimized scheduling scheme is output.

[0006] Optionally, the power system source-grid-load-storage basic operation data includes load forecast data, wind power forecast data, photovoltaic forecast data, thermal power unit parameters, energy storage system parameters, P2G hydrogen production system parameters, and power grid network parameters.

[0007] Optionally, the dynamic risk potential field consists of at least four of the following state quantities: the boundary distance between the current net load and the deep adjustment risk boundary, the trend approximation of the net load to the deep adjustment risk boundary within a future preset time window, the risk duration of the net load continuously located within the deep adjustment risk influence range, the historical deep adjustment memory formed by the cumulative duration or cumulative damage of the thermal power unit within the deep adjustment range within a preset historical time window, the reserve margin evolution formed by the current margin of the system's positive or negative reserve and its changing direction, the net load fluctuation, and the multi-resource adjustable margin.

[0008] Optionally, the output of the dynamic risk potential field includes the current time period risk level, risk evolution direction indicator, and power correction requirement; the solution process for this output includes: Standardize the state variables of the dynamic risk potential field to obtain standardized state components. Based on the standardized state components, a comprehensive risk index for the system is constructed. Based on the comprehensive risk indicators of the system, the time-period power correction requirements, risk evolution direction indicators, and current time-period risk levels are determined. Based on the time-segmented power demand and the risk evolution direction indicator, the total power correction demand for the current time period is determined.

[0009] Optionally, S104 specifically includes: The current period risk level, risk evolution direction indicator, and current period total power correction requirement are taken as inputs from the dynamic risk potential field output, and the current period total power correction requirement is taken as the total power quota to be allocated. Calculate the available regulation capacity of four types of regulation resources in the current period: energy storage, P2G hydrogen production system, active wind curtailment, and active solar curtailment. Furthermore, based on the available adjustable capacity, according to the risk level prediction results for future periods, the lifespan of the energy storage equipment, and the start-up and shutdown frequency information of the electric-to-gas electrolyzer, the power operation limit, adjustment response weight, and task participation ratio corresponding to the adjustable capacity of the energy storage system and the P2G hydrogen production system are dynamically adjusted. By combining the adjustable capacity, adjustment cost, risk mitigation contribution, cross-period adjustable margin, and equipment life depreciation factors of each adjustment resource, a comprehensive score for each adjustment resource is constructed, and the score is non-negatively processed. The allocation ratio is determined based on the comprehensive score ratio of various regulatory resources, and the total power quota to be allocated is distributed according to the allocation ratio to obtain the initial allocation power of each regulatory resource. By combining the upper limit of the available power of each regulating resource, the initial allocated power is capped to obtain the preliminary actual allocated power of each regulating resource; Under the constraint framework of the upper-level optimization model, the principle of equilibrium matching is based on the sum of the initial actual allocated power of various adjustment resources to match the total power of the target to be allocated; After optimization, the power allocation boundaries of each regulation resource in each time period are determined.

[0010] Optionally, the dynamic adjustment further includes: When the predicted risk level for a future period is higher than the risk level for the current period, the current available power limit of the energy storage and P2G hydrogen production system is reduced. When the energy storage lifespan depreciation index exceeds the preset first lifespan threshold, the regulation response weight of the energy storage system is reduced and its operating power boundary is tightened. When the start-up and shutdown frequency of the electrolyzer exceeds the preset start-up and shutdown threshold, the power sharing ratio of the P2G hydrogen production system participating in high-frequency regulation tasks is reduced.

[0011] Optionally, the constraints of the lower-level optimized scheduling model include the following: Constraints include: overall power balance, upper and lower limits of thermal power unit output, thermal power unit ramp rate, thermal power unit start-up and shutdown time, thermal power unit start-up and shutdown cost, system reserve, and grid security.

[0012] Optionally, the step of identifying operational risks and verifying constraints in the optimized scheduling scheme includes: Based on the distribution of various regulating resources and the actual output of thermal power units, the power flow distribution of power grid branches, the system reserve capacity level, and the operating conditions of thermal power units are deduced. Operational risks are identified and constraints are checked from three dimensions: whether the branch power flow exceeds the limit, whether the system reserve capacity meets the operating requirements, and whether the thermal power unit operating conditions are close to the deep peak shaving range. Based on the verification results, determine whether there are operational risks. If there are power flow exceeding limits, insufficient reserve capacity, or thermal power units approaching deep peak shaving operation conditions, then proceed to the iterative update process; if there are no such situations and the iteration termination conditions are met, then end the iteration and output the final scheduling plan.

[0013] Optionally, when an operational risk is determined to exist, the various regulatory resources are sorted in descending order of risk mitigation efficiency, and their maximum adjustable correction power is allocated to each regulatory resource in turn. After each type of resource allocation is completed, the allocated power is deducted, and the remaining correction requirements of the system and the remaining adjustable space of each resource are updated synchronously. This process is repeated iteratively until the remaining correction amount is zero or all adjustable resources have reached their adjustment limits. During the allocation process, the resource power boundary, energy storage and hydrogen storage capacity, equipment lifespan depletion and start-stop frequency constraints are simultaneously verified to complete the update of the risk severity and adjustment resource allocation results for this round. Based on the updated risk severity and adjustment resource allocation status, the relevant risk information and resource constraint parameters are fed back to the upper-level optimization model to enter a new round of double-layer closed-loop iteration.

[0014] This application has the following beneficial effects: The method proposed in this application is an optimized scheduling method for a combined wind, solar, thermal, energy storage, and hydrogen system that can sense the evolution of deep-modulation risks, adaptively adjust the priority of regulation resources based on multiple state variables, and perform closed-loop correction by combining lower-level risk positioning and sensitivity information. This method can improve the level of new energy consumption while taking into account the suppression of deep-modulation of thermal power units, the maintenance of energy storage life, the flexible consumption of P2G, and the safe operation of the power grid. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating a two-layer optimization scheduling method for a wind, solar, thermal, and hydrogen storage system, which addresses the risks of deep scheduling, as provided in an embodiment of this application. Detailed Implementation

[0016] To facilitate understanding by those skilled in the art, the present application will be further described below in conjunction with embodiments and accompanying drawings. The content mentioned in the embodiments is not intended to limit the present application.

[0017] To solve the above technical problems, such as Figure 1 As shown, this application proposes a two-layer optimal scheduling method for wind-solar-thermal-storage-hydrogen systems to address deep scheduling risks, including: S101: Obtain basic operational data of power system sources, grids, loads, and storage during the dispatching cycle; The power system used in this application is a vertically interconnected topology of source-grid-load-storage-hydrogen. This structure specifically includes thermal power units, wind turbines, photovoltaic arrays, integrated loads, energy storage systems, P2G (Power-to-Gas) hydrogen production systems, and the power grid network. The basic operational data for the power system's source-grid-load-storage system includes load forecast data, wind power forecast data, photovoltaic forecast data, thermal power unit parameters, energy storage system parameters, P2G hydrogen production system parameters, and power grid network parameters. On the source side, thermal power units serve as the steady-state core supporting system frequency and voltage, while wind and photovoltaic power generation contribute to low-carbon energy injection. On the storage side, the charging and discharging behavior of battery energy storage systems compensates for fluctuations in renewable energy output and smooths the net load curve. On the hydrogen side, the P2G hydrogen production system absorbs surplus renewable energy, achieving cross-time conversion and utilization of electrical energy to hydrogen energy. The P2G hydrogen production system parameters include the electrolyzer's rated power, minimum operating power, power ramp-up rate, electro-hydrogen conversion efficiency, and hydrogen storage capacity boundary. Energy storage system parameters include rated capacity, state of charge boundary, upper limit of charge and discharge power, cycle efficiency, health status parameters, and remaining equivalent cycle capacity.

[0018] S102: Based on the power system's source-grid-load-storage basic operation data, construct the system net load sequence; and construct a dynamic risk potential field according to the system net load sequence; A net load sequence is constructed based on the predicted load power, predicted wind power, predicted photovoltaic power, energy storage power, and P2G hydrogen production power. The calculation formula is as follows: (1) In the formula, The predicted load power for time period t. For wind power forecasting, For photovoltaic power prediction, For wind curtailment power, For abandoned light power ( , The initial iteration value is 0). For energy storage power, This refers to the power consumption of P2G. Specifically, this includes the power consumption during energy storage charging. During energy storage and discharge The target net load level corresponding to the conventional peak-shaving high-efficiency range of thermal power units is set as the economic operation benchmark value. The boundary for peak shaving at depth without oil injection or at depth with oil injection will be set as the risk boundary for deep peak shaving. .

[0019] To characterize the risk of deep-tuning operation of the system, a dynamic risk potential field is constructed based on the above system net load sequence. The dynamic risk potential field consists of at least four of the following state variables: the boundary distance between the current net load and the deep-tuning risk boundary. The trend of net load within the future preset time window approaching the risk boundary of deep adjustment. The duration of risk that net load remains continuously within the deep adjustment risk impact range The historical deep-tuning memory amount formed by the cumulative duration or cumulative damage of thermal power units within the preset historical time window in the deep-tuning interval. The evolution of reserve margin, which is formed by the current reserve or negative reserve of the system and its direction of change. Net load fluctuation and multi-resource adjustable margin measurement The state variables are defined as follows: (2) (3) (4) (5) (6) (7) (8) In the formula, H is the future preset time window length; The duration during which the net load remains within the range affected by the risk of a deep adjustment; To deepen the memory of history; This is the evolution amount for the reserve margin; This refers to net load fluctuations. For multi-resource adjustable margin measurement, The relative step size in the scheduling time domain.

[0020] The solution process for this dynamic risk potential field is as follows: Standardize the state variables of the dynamic risk potential field to obtain standardized state components. : (9) in, , representing the number of state variables involved in the construction of the risk potential field; and These are the maximum and minimum values ​​of the i-th type of state variable under the preset sample set or the operational experience boundary, respectively.

[0021] Weights are set based on the sensitive relationship between standardized state components and deep-tuning risk. Construct a comprehensive risk index Rt: (10) (11) Map the system's overall risk indicators to time-of-use power correction requirements. Let the threshold for the safe zone be R1, and the threshold for the warning zone be R2, where R2 > R1. Then we have: (12) in, The slope of the warning zone mapping. Let the slope of the intervention area be the mapping slope, and satisfy the following conditions: .

[0022] Based on the system's comprehensive risk indicators, in order to characterize the risk evolution trend within a future time window, the direction of risk evolution is determined. for: (13) In the formula, To set a comprehensive risk index for the end of a future time window or the average of future time windows. Threshold for risk evolution discrimination; =1 indicates increased risk. =0 indicates stable risk. =-1 indicates that the risk has decreased.

[0023] When the risk evolution direction within the future time window is enhancement, a feedforward amplification term is added to the time-sharing power correction requirement. To release adjustment resources in advance, otherwise the value is 0. (Add a feedforward amplification term after...) It can be represented as: (14) In the formula, This is the feedforward amplification factor.

[0024] Current period total power correction demand for: (15) Based on the system comprehensive risk indicators The size will determine the risk level for the current period. It is divided into three levels: safety, early warning, and intervention. (16) Therefore, the output of the dynamic risk potential field is: the risk level for the current time period. Indicators of risk evolution direction and the current period's total power correction demand .

[0025] S103: Construct an upper-level optimization model with the goal of risk mitigation and adjusting the optimal allocation of resources; The core task of this upper-level optimization model is to achieve two main objectives from a global perspective: effectively mitigating various operational risks faced by the system and scientifically and rationally allocating limited regulatory resources. The model's input is the output of the aforementioned dynamic risk potential field. This model does not pursue the optimization of a single indicator but rather considers two interrelated objectives holistically. The first objective is to minimize the overall system risk, that is, to reduce the total expected loss from all possible risk events as much as possible by pre-configuring resources. The second objective is to minimize the imbalance in the allocation of regulatory resources, that is, to avoid situations where some areas have excessive resource redundancy while others suffer from severe resource shortages, striving to allocate limited regulatory resources according to the actual urgency of each area's needs, achieving a rational spatial distribution. The two objectives can be balanced through weighting coefficients, the specific values ​​of which can be adjusted according to the actual emphasis of the project. Ultimately, the upper-level optimization model determines a global regulatory resource allocation scheme.

[0026] In addition, the model must meet several constraints. A feasible allocation scheme must simultaneously comply with the following restrictions: First, a total quantity constraint, that is, the total number of resources allocated cannot exceed the total amount of resources actually available in the system; second, a security constraint, that is, the risk level of any node or line after allocation must be lower than the preset security threshold, which is also the bottom line requirement for achieving the risk mitigation goal; and third, a logical consistency constraint, that is, the allocation scheme given by the upper layer cannot exceed the actual execution capability of the lower layer model, ensuring that the scheme is feasible.

[0027] In this model, risk mitigation and resource allocation are not separate steps, but deeply coupled. Specifically, areas with higher risk levels automatically receive higher resource allocation weights in the upper-level optimization, meaning that the more concentrated the risk, the more adjustment resources are allocated. Conversely, the optimized resource allocation can proactively build up defensive capabilities in risk hotspots, thus completing proactive defenses before risk events actually occur and achieving a shift from passive response to proactive mitigation.

[0028] The upper-level optimization model constructed using the above methods can simultaneously achieve two core functions at the global optimal level: first, proactively controlling system operational risks within an acceptable safety range; and second, scientifically allocating limited adjustment resources according to risk distribution patterns and the urgency of needs. The allocation scheme output by this model will lay a stable, low-risk, and resource-balanced decision-making foundation for the precise regulation of the lower-level model.

[0029] It should be noted that the optimization objective described in this embodiment can be achieved through conventional multi-objective optimization methods, such as weighted summation or Pareto optimization. The specific mathematical expressions corresponding to the above objectives and constraints can be directly derived by those skilled in the art based on the textual description in this specification, and will not be elaborated in formula form here.

[0030] S104: Using the output of the dynamic risk potential field as input, solve the upper-level optimization model to determine the power allocation boundary of each regulation resource in each time period; Taking the current period's risk level, risk evolution direction indicator, and current period's total power correction demand from the dynamic risk potential field output as input, and using the current period's total power correction demand as the total power quota to be allocated, the upper-level optimization model allocates this correction demand among energy storage, P2G, wind curtailment, and solar curtailment, while satisfying: (17) in, To allocate corrected power to the energy storage system, To allocate corrected power to the P2G hydrogen production system, To compensate for the power allocated to active wind curtailment, This is the correction power allocated to active light rejection.

[0031] The specific allocation process is as follows: Calculate the available regulation capacity of four types of regulation resources in the current period: energy storage, P2G hydrogen production system, active wind curtailment, and active solar curtailment. Upper limit of available charging power of energy storage system Upper limit of available absorption capacity of P2G hydrogen production system Available correction limits for active wind curtailment and active solar curtailment , Its expression is (18)~(21): (18) (19) (20) (twenty one) In the formula, and These are the rated power limits for energy storage and P2G, respectively; It is in a state of energy storage charge; This represents the upper limit of the energy storage state of charge. Rated energy for energy storage; To improve energy storage charging efficiency; The scheduling time interval; and These represent the ramp-up margins for energy storage and P2G respectively during the current period; This represents the current available power of the P2G. This is the upper limit of the remaining hydrogen storage capacity after conversion to success rate; and These represent the wind power and solar power that can be actively reduced at present.

[0032] Based on the above, in order to avoid the problems of subsequent scheduling capacity loss and equipment wear caused by over-utilization of regulation resources, the power upper limit, response weight and regulation participation ratio of energy storage and P2G are reconstructed according to the prediction of future risk level, energy storage life status and P2G electrolyzer start-up and shutdown frequency. The reconstruction process is as follows: (1) When a higher risk level is predicted in the future, the current available cap for energy storage and P2G will be tightened: (twenty two) (twenty three) In the formula, and These are the margin retention factors for energy storage and P2G, respectively. .

[0033] (2) When the energy storage life loss index exceeds the threshold At this time, reducing the energy storage weight and tightening the energy storage boundary can be expressed as: (twenty four) In the formula, This is the adjusted energy storage weighting coefficient. This is the energy storage weight attenuation coefficient. Let be the boundary attenuation coefficient of energy storage power, and satisfy . , .

[0034] (3) When the start-up and shutdown frequency of the electrolytic cell exceeds the threshold At this time, reducing the proportion of P2G participating in high-frequency regulation can be expressed as: (25) In the formula, The adjusted P2G weighting coefficients. This is the P2G weight decay coefficient. Let be the power boundary attenuation coefficient of P2G, and satisfy . , .

[0035] By combining the adjustable capacity, adjustment cost, risk mitigation contribution, cross-period adjustable margin, and equipment life depreciation factors of each adjustment resource, a comprehensive score for each adjustment resource is constructed, and the score is non-negatively processed. (26) (27) In the formula, Let m be a standardized indicator of the current adjustable capacity of resource m. Standardized indicators for unit adjustment costs To contribute standardized indicators to the marginal mitigation of target risks. To retain standardized demand metrics for adjustable margins across time periods, Standardized indicators for the impact of equipment lifespan reduction; As the scoring weight, and satisfying .

[0036] The allocation ratio is determined based on the comprehensive score ratio of various regulatory resources. The total power to be allocated is then distributed according to this ratio to obtain the initial allocation power of each regulatory resource, as shown in the following formula: (28) (29) In the formula, Let m be the initial allocation ratio for resource m. The initial power allocation for resource m in time period t.

[0037] By combining the available power limit of each regulating resource, the initial allocated power is capped to obtain the preliminary actual allocated power of each regulating resource: (30) In the formula, To account for the upper limit constraint of available resources, the preliminary actual power allocation of each adjustment resource is obtained without the aforementioned reconfiguration. Corresponding to , , and , For resource collection If a reconstruction is performed, the upper limit of the allocation or the allocation ratio shall be determined according to the upper limit or weight after reconstruction according to equations (22)-(25).

[0038] At this point, there is still an unallocated remaining amount. Under the constraint framework of the upper-level optimization model, the principle of balanced matching is based on the sum of the initial actual allocated power of various adjustment resources to match the total power of the target to be allocated. That is, the remaining adjustable capacity of each subject is reordered according to the comprehensive score, and iterative allocation is carried out. The cycle continues until any of the following termination conditions are met: the total correction requirement is completed, the maximum number of iterations is reached, all resources touch the constraint boundary and there is no further adjustment space. After optimization, the power allocation boundary of each adjustment resource in each time period is determined.

[0039] S105: Construct a lower-level optimization scheduling model with the goal of minimizing the overall system operating cost; The peak-shaving process of thermal power units can be analyzed into three stages: basic peak-shaving, deep peak-shaving without oil injection, and deep peak-shaving with oil injection. The peak-shaving process (from high load to low load) can be divided into three economic zones, and a different cost function is assigned to each zone: Phase I (Conventional Peak Shaving): When the unit is within its normal adjustable load range, the cost is mainly non-linear fuel consumption.

[0040] Phase II (Deep Peak Shaving without Oil Injection): The load is further reduced, and the unit begins to face life loss such as rotor wear.

[0041] Phase III (Oil Injection Depth Peak Shaving): When the load is too low, oil must be injected to assist combustion, which causes a sharp increase in costs.

[0042] Considering that the unit's coal consumption characteristics differ significantly from those of the conventional load period during low-load operation (deep peak shaving phase) due to reduced boiler thermal efficiency, increased plant power consumption, and turbine inlet pressure deviating from the rated value, this method employs the least squares method to piecewise fit historical operating data for three phases: conventional peak shaving, deep peak shaving without oil injection, and deep peak shaving with oil injection. This yields three sets of differentiated coal consumption coefficients. The formula for calculating segmented costs is as follows: (31) In the formula, P is the current output of the unit, and Pmin1 and Pmin2 are the load thresholds of the unit in different stages. , , (i=1,2) correspond to the coal consumption characteristic coefficients of different stages; Dror represents rotor fatigue loss; Coil represents the cost of fuel oil injection for combustion (only appears in the deep peak shaving stage).

[0043] The lower-level optimization scheduling model aims to minimize the overall system operating cost. It takes into account the segmented deep scheduling cost of thermal power units, the operation and maintenance cost of wind, solar and energy storage, the operating cost of P2G, the spinning reserve cost, and network security constraints, and performs comprehensive economic scheduling of the start-up and shutdown status and unit output of thermal power units.

[0044] With the goal of minimizing the total system operating cost Ctotal, including the segmented deep adjustment cost of thermal power units C1, the operation and maintenance cost of wind, solar and energy storage C2, the spinning reserve cost C3, the environmental protection tax cost C4, the environmental benefits of wind, solar and energy storage C5, and the P2G operating cost C6, the specific mathematical expressions are as follows: (32) (33) (34) (35) (36) (37) (38) (39) In the formula, Kw, Kpv, and Kesoc are the unit power operation and maintenance cost coefficients of wind power, photovoltaic, and energy storage systems, respectively, and PCDesoc1,t is the energy storage charging and discharging power. This is the system's spinning reserve cost factor. These are the prediction error rates for load, wind power, and photovoltaic power, respectively. This is the system's spinning reserve cost factor. The prediction error rates are for load, wind power, and solar power, respectively. Let t be the initial load of the receiving-end power grid at time t; Pcoal is the unit price of coal for fuel. , coal for unit , Production volume , The efficiency of the environmental protection device; Ds and Dn are... , Pollution equivalent value, B is the tax amount per pollution equivalent. Pcw and Pcpv are the environmental benefit coefficients for wind power and photovoltaic grid connection and consumption, and Pcesoc is the environmental value conversion factor for energy storage discharge. This represents the actual grid-connected discharge power of the energy storage system. For average electricity price, For the first Time period consumption, for unit price of raw materials For the first Methane production during the period This refers to the price of natural gas.

[0045] In addition, the lower-level model also needs to be subject to corresponding constraints, including the following: power balance constraints of the entire network, upper and lower limits of thermal power unit output constraints, ramp rate constraints of thermal power units, start-up and shutdown time constraints of thermal power units, start-up and shutdown cost constraints of thermal power units, system reserve constraints, and grid security constraints (DC power flow), as shown in equations (40) to (48) below: (40) (41) (42) (43) (44) (45) (46) (47) (48) In the formula, For the first Taiwanese crew The start / stop status of the time period (1 for running, 0 for stopping); Pi,min and Pi,max are the minimum and maximum technical output of the unit, respectively; pi,t is the real-time output of the unit; Ri is the unit's ramp rate; ui,t / (t-1) is the start / stop status of the unit in the current / previous time period; Pi,max and Pi,min are the upper and lower limits of the unit's output; Ton,i and Toff,i are the minimum continuous running and stopping times of unit i, respectively; ui,t:t+Ton,i-1 represents the start / stop status of the unit for a continuous time period after startup; c ostHi,t and costJi,t represent start-up and shutdown costs, respectively; Hi and Ji represent unit start-up and shutdown cost coefficients, respectively; Cd1, Cw1, and Cpv1 represent positive reserve coefficients for load, wind power, and photovoltaic, respectively; and Cd2, Cw2, and Cpv2 represent negative reserve coefficients. The coefficient values ​​are set based on the prediction error rate to ensure that the reserve capacity can cover various power output fluctuations. Gk,i represents the power transfer distribution factor (PTDF) from node k to branch i; and PL,min / max,k represents the upper and lower limits of the power flow for the kth branch.

[0046] S106: Substitute the power allocation boundary as a constraint into the lower-level optimization scheduling model to obtain the optimization scheduling scheme; The four types of power allocation boundaries output by the upper-level optimization model—energy storage, P2G hydrogen production, active wind curtailment, and active solar curtailment—are used as constraints. The power allocation boundaries for energy storage and P2G hydrogen production limit the output upper limit of the corresponding adjustable resources in the lower-level model; the power allocation boundaries for active wind curtailment and active solar curtailment serve as constraints on the maximum active reduction quota for the corresponding renewable energy sources. These four types of allocation boundaries are then substituted into the lower-level optimization scheduling model as output limits for each resource. Under the constraints of overall grid power balance, unit ramping, start-up and shutdown, reserve capacity, and grid power flow, the optimal scheduling scheme is obtained.

[0047] S107: Identify operational risks and verify constraints for optimized scheduling schemes; If there are operational risks or constraints are not met, the dynamic risk potential field and the upper-level optimization model are updated based on the risk information, and the iteration is returned. If there is no operational risk, the constraints are satisfied, and the risk convergence and iteration termination conditions are met, then the iteration stops and the final scheduling scheme is output.

[0048] The optimized scheduling scheme also needs to undergo risk assessment and constraint verification. The specific risk assessment process is as follows: Based on the distribution of various regulating resources and the actual output of thermal power units, the power flow distribution of power grid branches, the system reserve capacity level, and the operating conditions of thermal power units are deduced. Then, operational risks are identified and constraints are checked from three dimensions: whether the power flow of branches exceeds limits, whether the system reserve capacity meets operational requirements, and whether the operating conditions of thermal power units are close to the deep peak-shaving range. Specifically: The criteria for determining whether a branch power flow exceeds the limit are as follows: When the power of a branch approaches or exceeds the upper limit, first find the branch with the highest severity of exceeding the limit, then calculate the sensitivity of each node and unit to the power of that branch. The higher the sensitivity, the greater the impact of the node or unit on the power of that branch, thus identifying the risk-associated branches. With nodes or units This is identified as the risk positioning target. The specific expression is as follows: (49) (50) (51) in, The active power injected into node or unit i. This represents the active power flow of a certain transmission line. Let k be the actual power flow of line k at time t. This represents the upper limit of the transmission capacity of line k. . The sensitivity of a node or unit to branch power. For the branch road with the most severe violation of the limit, The node or unit with the highest power flow sensitivity.

[0049] The determination of whether the system's backup capacity meets operational requirements is as follows: Positive reserve indicates the system's upward adjustment capability. When the system's positive reserve is insufficient, nodes or units with limited upward adjustment capability are targeted; negative reserve indicates the system's downward adjustment capability. When the system's negative reserve is insufficient, units with limited downward adjustment capability are targeted. The specific steps are as follows: First, calculate the positive reserve requirement for time period t. With positive reserve available quantity : (52) (53) The reserve gap is: (54) When the following formula is satisfied, it is determined that there is a risk of insufficient positive reserves in the current period: (55) If further unit-level positioning is required, calculate the upsizing margin for each unit: (56) The risk-associated units for positive standby can be defined as: (57) in, The margin warning threshold for the unit was raised.

[0050] Alternatively, the most tightly constrained unit can be defined as: (58) The above statement indicates that the units with the smallest upward margin are the main targets associated with the risk of insufficient positive reserve.

[0051] The assessment of insufficient negative reserves and its main related entities can be obtained by analogy with the steps described above.

[0052] The criteria for determining whether the operating conditions of a thermal power unit are close to the deep peak-shaving range are as follows: Units where deep adjustment is not feasible, but which meet the following conditions, are included in the risk assessment: (59) In the formula, This is the minimum load threshold for the deep adjustment range of the unit.

[0053] If any of the above three conditions occur, the iterative update process begins; if none of the above conditions occur and the iteration termination condition is met, the iteration ends and the final scheduling scheme is output. The iterative update process is as follows: After completing the risk assessment, a correction strategy is formulated based on the severity of the risk to the regulatory resources such as energy storage, P2G, wind curtailment, and solar curtailment, the sensitivity of each regulatory resource to mitigating the risk, and the risk mitigation efficiency per unit of regulatory cost.

[0054] Risk mitigation efficiency The solution is shown in the following equation: (60) (61) (62) In the formula, Assess the severity of the risk. This represents the actual transmission power of the current line. This is the safe transmission power limit for this line; This represents the current reserve power deviation of the system. The "deep risk indicator" typically represents the current operating point of the system. , , The weight coefficients of the three sub-items satisfy the following conditions: , This indicates the sensitivity of resource m to mitigation of risk r. This represents the unit adjustment cost of resource m. A very small positive number is set to avoid the denominator being zero. This represents the efficiency of mitigating risk r with a unit cost of resource m.

[0055] When an operational risk is identified, various regulatory resources are sorted in descending order of risk mitigation efficiency, and their maximum adjustable correction power is allocated accordingly. After each type of resource is allocated, the allocated power is deducted, and the remaining correction requirements of the system and the remaining adjustable space of each resource are updated synchronously. This process is repeated iteratively until the remaining correction amount is zero or all regulatory resources reach their adjustment limits. During the allocation process, resource power boundaries, energy storage and hydrogen storage reserves, equipment lifespan depreciation, and start-stop frequency constraints are simultaneously verified to update the risk severity and regulatory resource allocation results for this round. Based on the updated risk severity and regulatory resource allocation status, relevant risk information and resource constraint parameters are fed back to the upper-level optimization model, entering a new round of two-layer closed-loop iteration.

[0056] The entire process is based on the rigid satisfaction of hard constraints related to power grid safety and equipment operation. On this basis, the iteration can be terminated if any of the following conditions are met: (1) The lower-level scheduling model fully meets all feasibility constraints such as power balance of the whole network, branch power flow, system reserve, upper and lower limits of unit output, ramp rate, minimum start and stop time and energy storage, and operating limits of P2G hydrogen production equipment. (2) The change in the comprehensive risk index between two consecutive iterations is less than the first preset threshold; (3) The change in power correction requirement between two consecutive iterations is less than the second preset threshold; (4) The risk severity index for a single time period is lower than the preset target risk limit; (5) The cumulative risk index within the preset rolling window is lower than the third preset threshold; (6) The number of iterations reaches the system's preset maximum iteration limit.

[0057] After the iteration is completed, the start-up and shutdown status of thermal power units, output of thermal power units, energy storage charging and discharging power, P2G hydrogen production power, wind and solar curtailment power and new energy consumption rate are output for each time period to form the final optimized scheduling scheme.

[0058] The aforementioned upper-level optimization model and lower-level optimization scheduling model can be established using the YALMIP toolbox in the MATLAB environment, and the CPLEX solver can be called for iterative solution. MATLAB, YALMIP, and CPLEX are merely one implementation of this invention and do not constitute a limitation on the scope of protection of this invention.

[0059] In summary, the method proposed in this application is an optimized scheduling method for a combined wind, solar, thermal, energy storage, and hydrogen system that can sense the evolution of deep-modulation risks, adaptively adjust the priority of regulation resources based on multiple state variables, and perform closed-loop correction by combining lower-level risk location and sensitivity information. This method can improve the level of new energy consumption while taking into account the suppression of deep-modulation of thermal power units, the maintenance of energy storage life, the flexible consumption of P2G, and the safe operation of the power grid.

[0060] In some embodiments, this application also provides a computer system including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0061] This application also provides a computer-readable storage medium for storing a computer program. This computer-readable storage medium can be applied to a computer device, and the computer program causes the computer device to execute the corresponding processes in the methods described above in the embodiments of this application; for brevity, further details are omitted here.

[0062] The above embodiments are preferred implementations of this application. In addition, this application can be implemented in other ways. Any obvious substitutions without departing from the concept of this technical solution are within the protection scope of this application.

[0063] To facilitate understanding by those skilled in the art of the improvements made by this application compared to the prior art, some of the accompanying drawings and descriptions have been simplified, and for clarity, some other elements have been omitted from this application. Those skilled in the art should realize that these omitted elements may also constitute the content of this application.

Claims

1. A two-layer optimal scheduling method for wind, solar, thermal, and hydrogen storage systems oriented towards deep scheduling risks, characterized in that, include: S101: Obtain basic operational data of power system sources, grids, loads, and storage during the dispatching cycle; S102: Based on the power system's source-grid-load-storage basic operation data, construct the system net load sequence; and construct a dynamic risk potential field according to the system net load sequence; the dynamic risk potential field is used to characterize the system's deep adjustment operation risk; the dynamic risk potential field consists of at least four of the following state quantities: the boundary distance between the current net load and the deep adjustment risk boundary, the trend approximation of the net load to the deep adjustment risk boundary within a future preset time window, the risk duration of the net load continuously located in the deep adjustment risk influence range, the historical deep adjustment memory formed by the cumulative duration or cumulative damage of thermal power units in the deep adjustment range within a preset historical time window, the reserve margin evolution formed by the current margin of the system's positive or negative reserves and its changing direction, the net load fluctuation, and the multi-resource adjustable margin. S103: Construct an upper-level optimization model with the goal of risk mitigation and adjusting the optimal allocation of resources; S104: Using the output of the dynamic risk potential field as input, solve the upper-level optimization model to determine the power allocation boundary of each regulation resource in each time period; S105: Construct a lower-level optimization scheduling model with the goal of minimizing the overall system operating cost; S106: Substitute the power allocation boundary as a constraint into the lower-level optimization scheduling model to obtain the optimization scheduling scheme; S107: Identify operational risks and verify constraints for optimized scheduling schemes; If there are operational risks or constraints are not met, the dynamic risk potential field and the upper-level optimization model are updated based on the risk information, and the iteration is returned. If there is no operational risk, the constraints are satisfied, and the risk convergence and iteration termination conditions are met, then the iteration stops and the final optimized scheduling scheme is output.

2. The method according to claim 1, characterized in that, The basic operational data of the power system source-grid-load-storage system includes load forecast data, wind power forecast data, photovoltaic forecast data, thermal power unit parameters, energy storage system parameters, P2G hydrogen production system parameters, and power grid network parameters.

3. The method according to claim 1, characterized in that, The output of the dynamic risk potential field includes the current risk level, risk evolution direction indicator, and power correction requirement; the solution process for this output includes: Standardize the state variables of the dynamic risk potential field to obtain standardized state components. Based on the standardized state components, a comprehensive risk index for the system is constructed. Based on the comprehensive risk indicators of the system, the time-period power correction requirements, risk evolution direction indicators, and current time-period risk levels are determined. Based on the time-segmented power demand and the risk evolution direction indicator, the total power correction demand for the current time period is determined.

4. The method according to claim 3, characterized in that, S104 specifically includes: The current period risk level, risk evolution direction indicator, and current period total power correction requirement are taken as inputs from the dynamic risk potential field output, and the current period total power correction requirement is taken as the total power quota to be allocated. Calculate the available regulation capacity of four types of regulation resources in the current period: energy storage, P2G hydrogen production system, active wind curtailment, and active solar curtailment. Furthermore, based on the available adjustable capacity, according to the risk level prediction results for future periods, the lifespan of the energy storage equipment, and the start-up and shutdown frequency information of the electric-to-gas electrolyzer, the power operation limit, adjustment response weight, and task participation ratio corresponding to the adjustable capacity of the energy storage system and the P2G hydrogen production system are dynamically adjusted. By combining the adjustable capacity, adjustment cost, risk mitigation contribution, cross-period adjustable margin, and equipment life depreciation factors of each adjustment resource, a comprehensive score for each adjustment resource is constructed, and the score is non-negatively processed. The allocation ratio is determined based on the comprehensive score ratio of various regulatory resources. The total power quota to be allocated is then distributed according to this allocation ratio to obtain the initial allocation power of each regulatory resource. By combining the upper limit of the available power of each regulating resource, the initial allocated power is capped to obtain the preliminary actual allocated power of each regulating resource; Under the constraint framework of the upper-level optimization model, the principle of equilibrium matching is based on the sum of the initial actual allocated power of various adjustment resources to match the total power of the target to be allocated; After optimization, the power allocation boundaries of each regulation resource in each time period are determined.

5. The method according to claim 4, characterized in that, The dynamic adjustment further includes: When the risk level for a future period is predicted to be higher than the risk level for the current period, the current available power limit of the energy storage and P2G hydrogen production system is reduced. When the energy storage lifespan depreciation index exceeds the preset first lifespan threshold, the regulation response weight of the energy storage system is reduced and its operating power boundary is tightened. When the start-up and shutdown frequency of the electrolyzer exceeds the preset start-up and shutdown threshold, the power sharing ratio of the P2G hydrogen production system participating in high-frequency regulation tasks is reduced.

6. The method according to claim 1, characterized in that, The constraints of the lower-level optimization scheduling model include the following: Constraints include: overall power balance, upper and lower limits of thermal power unit output, thermal power unit ramp rate, thermal power unit start-up and shutdown time, thermal power unit start-up and shutdown cost, system reserve, and grid security.

7. The method according to claim 5, characterized in that, The process of identifying operational risks and verifying constraints in the optimized scheduling scheme includes: Based on the distribution of various regulating resources and the actual output of thermal power units, the power flow distribution of power grid branches, the system reserve capacity level, and the operating conditions of thermal power units are deduced. Operational risks are identified and constraints are checked from three dimensions: whether the branch power flow exceeds the limit, whether the system reserve capacity meets the operating requirements, and whether the thermal power unit operating conditions are close to the deep peak shaving range. Based on the verification results, determine whether there are operational risks. If there are power flow exceeding limits, insufficient reserve capacity, or thermal power units approaching deep peak shaving operation conditions, then proceed to the iterative update process; if there are no such situations and the iteration termination conditions are met, then end the iteration and output the final scheduling plan.

8. The method according to claim 7, characterized in that, When an operational risk is identified, all types of regulatory resources are sorted in descending order of risk mitigation efficiency, and their maximum adjustable correction power is allocated to each regulatory resource in turn. After each type of resource allocation is completed, the allocated power is deducted, and the remaining correction requirements of the system and the remaining adjustable space of each resource are updated synchronously. This process is repeated iteratively until the remaining correction amount is zero or all adjustable resources have reached their adjustment limits. During the allocation process, the resource power boundary, energy storage and hydrogen storage capacity, equipment lifespan depletion and start-stop frequency constraints are simultaneously verified to complete the update of the risk severity and adjustment resource allocation results for this round. Based on the updated risk severity and adjustment resource allocation status, the relevant risk information and resource constraint parameters are fed back to the upper-level optimization model to enter a new round of double-layer closed-loop iteration.

Images