Photothermal power station robust optimization scheduling method based on polyhedral uncertain set

Abstract

The invention provides a photo-thermal power station robust optimization scheduling method based on a polyhedral uncertain set. The method comprises the following steps: step 1, describing a photo-thermal power station output error by using the polyhedral uncertain set; 2, optimizing the power grid dispatching model containing the photo-thermal power station; 3, establishing an adjustable robust optimization model by using an affine strategy; and 4, solving the constraint condition, wherein the result comprises the generation power of the photo-thermal power station in each time period, the heat energy storage or release condition and the affine coefficient needing to be adjusted. By introducing the polyhedral uncertain set and cooperatively establishing the affine strategy, the uncertain change of the photo-thermal power station can be flexibly coped with by scheduling, and the stability and flexibility of the power system are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of scheduling of solar thermal power plants and relates to a robust optimization scheduling method for solar thermal power plants. Background Art

[0002] The power generation of a CSP plant is directly affected by the intensity of sunshine, which has significant volatility and uncertainty in actual operation, making it impossible to accurately predict the power generation output of a CSP plant. Changes in sunshine intensity are determined by a variety of factors, including weather conditions, cloud distribution, seasonal changes, etc. These factors lead to inherent errors and uncertainties in predicting sunshine intensity. Due to the dependence of CSP plants on sunshine, their power generation often deviates greatly from the expected power during actual operation. This uncertainty poses great challenges to the flexibility management and scheduling of power systems, especially in ensuring power supply stability and coping with changes in power load. Traditional scheduling methods are difficult to effectively cope with this complex and uncertain environment, so it is necessary to quantify and analyze the power generation capacity of CSP plants more accurately to improve the system's adaptability to uncertain conditions.

[0003] The existing uncertainty optimization methods for power systems are mainly divided into two categories: stochastic optimization and robust optimization. Stochastic optimization methods usually assume that random variables conform to a known probability distribution and solve the optimal scheduling plan of the system by establishing multiple scenarios. This method can deal with the uncertainty of the system more accurately when the number of scenarios is sufficient. However, in practical applications, these probability distribution assumptions are often untenable due to insufficient historical data or unclear data correlation, which makes the results of stochastic optimization methods lack reliability, especially when dealing with highly uncertain renewable energy sources (such as solar thermal power stations and wind power).

[0004] In order to overcome the shortcomings of stochastic optimization methods, robust optimization has become another widely used method. Robust optimization does not rely on the specific probability distribution of uncertain quantities, but describes uncertainty by constructing uncertain sets. Compared with stochastic optimization, robust optimization has the advantages of not relying on historical data and probability distribution assumptions. Therefore, it is more robust when dealing with uncertainty, especially for complex systems that need to cope with multiple scenario changes. However, robust optimization is usually too conservative because it provides the system with robustness in the worst case, which may affect the economic efficiency of the scheduling results.

[0005] In robust optimization methods, commonly used sets to describe uncertainty include elliptical uncertainty sets, polyhedral uncertainty sets, etc. In existing studies, elliptical uncertainty sets use the geometric characteristics of an ellipsoid to describe the range of uncertainty to achieve a balance between system robustness and economy. For example, the range of uncertainty can be controlled by adjusting the size of the elliptical set. However, the use of elliptical sets has a limitation, that is, it may not be precise enough when describing highly irregular uncertainties, which can easily lead to an excessively large range of uncertainty sets, thereby increasing the conservatism of system operation results.

[0006] The present invention solves the uncertainty problem in power generation of solar thermal power stations by introducing a polyhedron uncertainty set. To this end, a scheduling method that can flexibly cope with changes in uncertainty is established, thereby significantly improving the stability and flexibility of the power system. Summary of the invention

[0007] The main purpose of the present invention is to overcome the problem of strong dependence of existing random optimization methods on probability distribution assumptions by introducing polyhedral uncertainty sets, while solving the problem of over-conservatism of traditional robust optimization, thereby providing a more efficient and flexible solar thermal power station power generation scheduling method and improving the operating stability and economy of the system.

[0008] The technical solution of the present invention is a method for robust optimization scheduling of a CSP plant based on a polyhedral uncertainty set. The polyhedral uncertainty set is introduced, and the whole process is divided into the following four steps:

[0009] Step 1: Use a polyhedral uncertainty set to describe the output error of the CSP plant. According to the scatter plot of historical data, different envelopes are formed, namely, interval sets. The interval sets are then converted into polyhedral sets. The polyhedral sets set multidimensional boundaries to express the output fluctuation of the CSP plant in mathematical form as follows:

[0010] Among them, P csp,t is the actual heat collection power of the CSP power station at time t; P csp0,t is the expected value of solar thermal collector power at time t; is the maximum fluctuation in the output range; j,t is the jth uncertainty factor; ξ phd is a set of polyhedra; ξ csp，t is the uncertainty factor of the CSP plant; for Any power station j and any time period t;

[0011] Step 2: Optimize the grid dispatch model including CSP power stations, consider the general dispatch model including CSP, PV and thermal power systems, and the objective function is to minimize the total dispatch cost of the system, that is: Where P i,t ,PG,t ,P pv,t is the power generation of thermal power, solar thermal power and photovoltaic power station at time t, E(P i,t ,P G,t ,P pv,t ) is the power generation cost function;

[0012] The constraints for thermal power, photovoltaic power and power balance are as follows:

[0013] Where P i，min is the minimum output power of the thermal power station; P i,max is the maximum output power of the thermal power station; P i，t-1 is the power generation capacity of the thermal power station at time t-1; u i，t is the maximum ramp rate of the thermal power station; P pv,min is the minimum output power of the photovoltaic power station; P pv,max is the maximum output power of the photovoltaic power station; P load is the load of the power grid;

[0014] The dispatch model of a CSP power station is simplified into three parts: heat collection, heat storage and release, and power generation. Consider the CSP power received by the heat collection system: Among them, η SF Represents the photothermal conversion efficiency; S SF is the mirror field area, D t is the DNI at time t;

[0015] Part of this power is used for power generation and the other part is stored. The power balance relationship is as follows: in, represents the thermal power used for power generation at the CSP station at time t, P t TS,c ,P t TS,d They represent the heat storage and heat release powers at time t respectively;

[0016] For the heat storage system, the heat storage state equation is: in, represents the heat storage capacity of the system at time t, represents the heat storage capacity of the system at time t+1, η loss represents the attenuation coefficient of the heat storage system;

[0017] For the power generation part, thermal power is converted into electrical power. Among them, P Gt represents the electric power generated by the CSP plant at time t, η e is the thermoelectric conversion efficiency;

[0018] Step 3: Use the affine strategy to establish an adjustable robust optimization model. The output power error of the CSP power station model is expressed as: ΔP csp is the uncertainty of the output of the CSP power station; the affine strategy is used to deal with the uncertainty in the system. The specific scheduling strategy is as follows: The actual value of the heat storage and heat release power is expressed as: in, is the initial expected heat storage and release power, λ charge,t ,λ dis,t are the corresponding heat storage and heat release affine coefficients respectively; the part of the thermal power of the CSP plant used for power generation, the affine strategy is expressed as: in, is the expected thermal power of the initial CSP power station, λ g,t is the affine coefficient of the CSP power station; the total affine constraint is:

[0019]

[0020] Where η SF Represents the photothermal conversion efficiency; S SF is the mirror field area, D t is the DNI at time t, P load，t is the load of the power grid after affine transformation during period t, P load0，t is the original load of the power grid during period t, λ l1 is the affine factor of the uncertainty of photovoltaic output to grid load, λ l2 is the affine factor of the uncertainty of solar thermal output to grid load, ΔP pv is the uncertainty of photovoltaic output, P i0,t is the original output of the thermal power station at time t, λ i1 is the affine factor of the uncertainty of photovoltaic output to thermal power, λ i2 is the affine factor of the uncertainty of solar thermal output to thermal power;

[0021] Step 4: Solve the above constraints, and the solution results include the power generation of the CSP plant in each time period, the thermal energy storage or release, and the affine coefficients that need to be adjusted.

[0022] Preferably, the solution method for step 4 is to convert the scheduling model containing the solar thermal power station and the affine strategy into a mixed integer linear programming model for solution.

[0023] Step 3: The objective function is:

[0024] Beneficial effects of the present invention: 1. Accurately describe the uncertainty of the output of a solar thermal power station by using a polyhedral uncertainty set modeling method: introduce a polyhedral uncertainty set to describe the volatility and uncertainty of the output of a solar thermal power station. This method can more accurately characterize the power fluctuation range of a solar thermal power station caused by external factors, thereby providing an accurate mathematical basis for subsequent scheduling optimization. Compared with traditional uncertainty description methods (such as random distribution assumptions or elliptical uncertainty sets), polyhedral sets have significant advantages in flexibility and accuracy.

[0025] 2. A robust optimization model based on the combination of CSP power station model optimization and affine strategy is established, which is used to flexibly schedule and optimize the output of CSP power stations under uncertain conditions. The present invention first deeply analyzes the operating mechanism of CSP power stations, appropriately optimizes and adjusts its operating model, and simplifies the relationship between heat storage, heat release and power generation. On this basis, the output of the CSP power station is linearly adjusted through the affine strategy, so as to flexibly respond to uncertainty when the sunshine intensity fluctuates and achieve power balance of the system. Through this affine adjustable robust optimization model that combines model optimization with affine strategy, the present invention can effectively deal with the uncertainty scheduling problem in CSP power generation, and ensure that the system can still obtain the optimal scheduling strategy when facing the worst uncertainty scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a schematic diagram of interval set transformation;

[0027] Figure 2 It is a two-dimensional polyhedron uncertain set coordinate diagram;

[0028] Figure 3 It is a diagram of energy flow in a CSP plant;

[0029] Figure 4 It is a simplified diagram of energy flow in a CSP plant. DETAILED DESCRIPTION

[0030] About uncertain sets of polyhedra and affine strategies:

[0031] (I) Polyhedron Uncertain Set

[0032] Polyhedral uncertainty sets have a wide range of applications in fields such as optimization theory and control engineering. In practical systems, many parameters are often uncertain, and polyhedral uncertainty sets provide an effective mathematical framework for describing these uncertain parameters. For example, in the load forecasting of power systems, due to the influence of various factors such as economic activities and weather changes, the power load demand is uncertain, and this uncertainty can be modeled by polyhedral uncertainty sets. From a mathematical point of view, a polyhedral uncertainty set is a geometric region defined by a set of linear inequalities, which can limit uncertain parameters to a specific polyhedron range. For example, for a two-dimensional polyhedral uncertainty set, it can be expressed as:

[0033]

[0034] Among them, x and y are uncertain parameters, and a and b with subscripts are given constants. This representation method allows us to analyze and solve optimization problems involving uncertainty under the constraints of polyhedrons, thus taking into account the impact of uncertainty on the system to a certain extent.

[0035] (II) Affine Strategy

[0036] Affine strategies play an important role in dealing with control and optimization problems of uncertain systems. When faced with uncertainties in the system, affine strategies deal with them by constructing decision variables or control laws that have an affine relationship with the uncertain parameters. The basic form of an affine strategy can be written as:

[0037] u(ξ)=u0+Lξ (2)

[0038] Among them, u(ξ) represents the scheduling decision variable, which is an affine function of the uncertainty parameter ξ; u0 is a constant term, which represents the decision value in the absence of uncertainty; L is a weight matrix, which is used to adjust the relationship between the decision variable and the uncertainty parameter. Through this affine relationship, the system can achieve the optimal response to uncertain factors by adjusting L when facing uncertainty.

[0039] This paper proposes a scheduling and optimization method for a CSP plant based on a polyhedral uncertainty set, which constructs a scheduling model step by step, establishes a robust optimization strategy, and finally solves the optimal scheduling solution to cope with the uncertainty of the CSP plant output. The whole process can be divided into the following four steps:

[0040] 1. Using polyhedral uncertainty sets to describe the output error of CSP plants

[0041] First, the power generation of a CSP plant is directly affected by the intensity of sunlight, which has a high degree of uncertainty. In order to quantify these uncertainties, the present invention introduces a polyhedron uncertainty set to describe the error range of the CSP plant output.

[0042] According to the scatter plot of historical data, different envelopes can be formed, that is, interval sets. Then the interval set is converted into a polyhedron set, such as Figure 1 shown.

[0043] The polyhedron set expresses the output fluctuation of the CSP power station in mathematical form by setting multi-dimensional boundaries:

[0044]

[0045] P csp,t is the actual heat collection power of the CSP power station at time t; P csp0,t is the expected value of solar thermal collector power at time t; is the maximum fluctuation in the output range; j,t is the jth uncertainty factor; ξ phd is a set of polyhedra; ξ csp，t is the uncertainty factor of the CSP plant; For any power station j and any time period t, its envelope range is shown as follows Figure 2 As shown; τ is the uncertainty.

[0046] The polyhedron set is not only used to define the range of output error, but also serves as a constraint in the subsequent optimization model to ensure that all scheduling decisions are still feasible under the worst conditions. Combined with the affine strategy, scheduling decisions can be adjusted within the constraints of the polyhedron set to cope with the various uncertainties faced by CSP plants in actual operation.

[0047] 2. Optimize the grid dispatch model including CSP plants

[0048] Considering the general dispatch model of solar thermal, photovoltaic and thermal power systems, the objective function is to minimize the total dispatch cost of the system, that is:

[0049]

[0050] Where P i,t ,P G,t ,P pv,t is the power generation of thermal power, solar thermal power and photovoltaic power station at time t, E(P i,t ,P G,t ,P pv,t ) is the power generation cost function. The constraints of thermal power, photovoltaic power and power balance are as follows:

[0051]

[0052] Where P i,min is the minimum output power of the thermal power station; P i,max is the maximum output power of the thermal power station; P i,t-1 is the power generation capacity of the thermal power station at time t-1; u i，t is the maximum ramp rate of the thermal power station; P pv,min is the minimum output power of the photovoltaic power station; P pv,max is the maximum output power of the photovoltaic power station; P load is the load of the power grid;

[0053] For a CSP plant, first, a dispatching model for the CSP plant is established. Figure 3 As shown, first, solar energy is captured and converted into thermal energy through the thermal collection system. Part of the converted thermal energy is stored in the thermal storage device, and the release of thermal energy is adjusted according to the demand of the power grid. The other part is transported to the power generation system together with the stored thermal energy and converted into electrical energy. The power generation efficiency will directly affect the energy output of the entire power station.

[0054] against Figure 3 The energy flow of the heat tank can be merged into the collection system, so the operation of the CSP power station can be simplified into three main parts: heat collection, heat storage and power generation, such as Figure 4 This simplification does not affect the optimal scheduling of the CSP plant and helps to simplify the description of the flow of energy in the system and its conversion process, which is convenient for subsequent calculations.

[0055] Based on the predicted sunshine intensity, the expected power generation of the CSP plant can be calculated. However, due to the certain errors in sunshine forecast, the actual power generation of the CSP plant will deviate from the predicted results, which will affect the accuracy of the scheduling results. For CSP plants, first consider the CSP power received by the collector system:

[0056]

[0057] Among them, η SF Represents the photothermal conversion efficiency; S SF is the mirror field area, D t is the DNI at time t.

[0058] Part of this power is used for power generation and the other part is stored. Figure 4 , can be expressed by mathematical formula, and its power balance relationship is as follows:

[0059]

[0060] in, represents the thermal power used for power generation at the CSP station at time t, P t TS,c ,Pt TS,d They represent the heat storage and heat release powers at time t respectively.

[0061] For the heat storage system, the heat storage state equation can be used to describe the change of heat storage capacity as follows:

[0062]

[0063] in, represents the heat storage capacity of the system at time t+1, represents the heat storage capacity of the system at time t, η loss

[0064] Represents the attenuation coefficient of the heat storage system.

[0065] Because heat storage and heat release cannot continue at the same time, there are the following constraints:

[0066] P t TS,c P t TS,d =0(9)

[0067] For the power generation part, the thermal power used for power generation will be converted into electrical power according to the following formula:

[0068]

[0069] Among them, P Gt represents the electric power generated by the CSP plant at time t, η e is the thermoelectric conversion efficiency.

[0070] 3. Use affine strategy to establish an adjustable robust optimization model

[0071] The basic idea of ​​the affine strategy is to use linear functions to describe the relationship between control variables and uncertainty, so that the dispatch system can respond more flexibly under uncertain conditions. Specifically, the affine strategy is used to adjust the dispatch of different power variables in the system so that it can cope with the volatility and uncertainty of the output of renewable energy power stations.

[0072] In the optimized CSP power station model, the collector power is affected by the fluctuation of sunlight intensity, and the output power error can be expressed as:

[0073]

[0074] ΔP csp The uncertainty of the output of the CSP plant is represented by an affine strategy to deal with the uncertainty in the system. The specific scheduling strategy is as follows:

[0075] According to the power balance equation of the CSP plant (7), the actual value of the thermal storage and release power can be expressed as:

[0076]

[0077] Among them, P t TS,c ,P t TS,d are the heat storage and heat release powers at time t, is the initial heat storage and heat release expected power, λ charge,t ,λ dis,t are the corresponding heat storage and heat release affine coefficients respectively.

[0078] Similarly, according to formula (7), for the part of the thermal power of the CSP plant used for power generation, the affine strategy is expressed as:

[0079]

[0080] in, is the expected thermal power of the initial CSP power station, λ g,t is the power generation affine coefficient of the CSP plant. This strategy allows the system to flexibly adjust the power generation according to the changes in the output of the CSP plant.

[0081] In order to ensure the power balance of the system, the affine coefficients need to meet the following constraints:

[0082] λ g,t +λ charge,t +λ dis,t =1 (15)

[0083] Accordingly, the constraints (3) and (5) of the original CSP power station should be transformed into:

[0084]

[0085] At this point, the uncertainty factors of CSP power stations are introduced into scheduling optimization. Similarly, for photovoltaic power stations:

[0086]

[0087] Among them, P pv0，t Indicates the expected output value of the photovoltaic power station, Indicates the output fluctuation of photovoltaic power station, ξ pv，t Represents the uncertainty factor of the photovoltaic power station, ΔP pv represents the uncertainty of the photovoltaic power station, which can be calculated according to (5)

[0088] The affine expressions of thermal power and load are derived from the formula:

[0089] P load,t =P load0,t +λ l2 ΔPcsp +λ l1 ΔP pv (20)

[0090] P i,t =P i0,t -λ i2 ΔP csp -λ i1 ΔP pv (twenty one)

[0091] Among them, P load，t is the load of the power grid after affine transformation during period t, P load0，t is the original load of the power grid during period t, λ l1 is the affine factor of the uncertainty of photovoltaic output to grid load, λ l2 is the affine factor of the uncertainty of solar thermal output to grid load, ΔP pv is the uncertainty of photovoltaic output, P i0，t is the original output of the thermal power station at time t, λ i1 is the affine factor of the uncertainty of photovoltaic output to thermal power, λ i2 is the affine factor of the uncertainty of solar thermal output to thermal power; at the same time, in order to meet the constraints of the original power balance, each parameter needs to meet the following conditions:

[0092]

[0093] After sorting, the objective function is:

[0094]

[0095] Among them, P i0，t represents the original output of the thermal power plant at time t, P G0,t represents the original output of the CSP plant at time t, P pv0，t represents the original output of the photovoltaic power station at time t, ξ j represents the uncertainty factor;

[0096] The total affine constraint is

[0097]

[0098] 4. Solve the optimal scheduling strategy

[0099] The present invention solves the above-mentioned scheduling model and affine strategy of the solar thermal power station by converting it into a mixed integer linear programming (MILP) model. The goal of this process is to find the optimal scheduling strategy so that the operating cost of the system is minimized and the flexibility requirements can be met under the worst uncertainty scenario. Through the affine adjustable robust optimization model, the robust optimization problem is simplified to a linear programming problem that can be effectively solved, thereby obtaining the optimal scheduling solution.

[0100] The solution results will include the power generation of the CSP plant in each time period, the storage or release of thermal energy, and the affine coefficients that need to be adjusted. These results will provide specific guidance for the actual operation of the CSP plant, enabling it to operate flexibly and efficiently under uncertain conditions.

[0101] The present invention proposes an optimization method based on a polyhedral uncertainty set to better cope with the uncertainty of the output of renewable energy sources such as solar thermal power plants. The polyhedral uncertainty set can more accurately quantify the output volatility and uncertainty characteristics of solar thermal power plants by constructing a geometric polyhedron with clear boundaries. Compared with an elliptical set, the structure of a polyhedral set is more flexible, which can better adapt to the actual characteristics of uncertainty data and effectively reduce unnecessary conservatism. In addition, the present invention establishes an affine adjustable robust optimization model based on a polyhedral uncertainty set, which makes scheduling under uncertain conditions more flexible, and can control the balance between the robustness and economy of the system by adjusting the size of the uncertainty set.

[0102] The advantages of the present invention are: 1. Accuracy: The uncertainty of a CSP plant is described using a polyhedral uncertainty set. In the power system, the output of a CSP plant is affected by external factors (such as sunshine intensity), and there is significant uncertainty. Traditional uncertainty description methods, such as stochastic optimization, assume that these uncertainties follow a known probability distribution. In practice, the statistical characteristics of uncertainties often lack sufficient data support, and errors in this assumption may lead to inaccurate scheduling. Therefore, the inaccuracy of the uncertainty description of traditional methods will directly affect the performance of the system.

[0103] To solve this problem, the present invention introduces a polyhedral uncertainty set to describe the output uncertainty of a solar thermal power station. This set is defined by geometric boundaries, can flexibly adapt to different uncertainty factors and provide clear upper and lower limits, making the description of uncertainty closer to the actual situation and avoiding the error accumulation in traditional models. This method improves the system's quantification accuracy of uncertainty, makes the basic scheduling data more reliable, and directly improves the accuracy of the scheduling process.

[0104] 2. Flexibility: Adopt an affine strategy to build an adjustable robust optimization model. When faced with the uncertainty of the output of a solar thermal power station, the dispatch of the power system needs to be able to flexibly respond to various possible changes. Traditional robust optimization methods use an unchanging worst-case setting. Although this static method can ensure the safety of the system, it lacks flexibility and adaptability, resulting in unnecessary waste of resources when the system responds to dynamic changes.

[0105] The present invention adopts an affine strategy to construct an adjustable robust optimization model. The core idea of ​​the affine strategy is to express the relationship between uncertainty and control variables through linear functions, so that when the uncertainty changes, the scheduling can be adjusted accordingly. Compared with static scheduling strategies, this method can adjust various variables in real time to cope with different uncertain scenarios, enhance the dynamic adaptability of the system, and thus improve the flexibility of scheduling. This flexibility enables the system to operate efficiently under different environmental conditions.

Claims

1. A robust optimization scheduling method for a CSP plant based on a polyhedral uncertain set, characterized in that Introducing the polyhedron uncertain set, the whole process is divided into the following four steps: Step 1: Use a polyhedral uncertainty set to describe the output error of the CSP plant. According to the scatter plot of historical data, different envelopes are formed, namely, interval sets. The interval sets are then converted into polyhedral sets. The polyhedral sets set multidimensional boundaries to express the output fluctuation of the CSP plant in mathematical form as follows: Among them, P csp,t is the actual heat collection power of the CSP power station at time t; P csp0,t is the expected value of solar thermal collector power at time t; is the maximum fluctuation in the output range; j,t is the jth uncertainty factor; ξ phd is a set of polyhedra; ξ csp,t is the uncertainty factor of the CSP plant; for Any power station j and any time period t; Step 2: Optimize the grid dispatch model including CSP power stations, consider the general dispatch model including CSP, PV and thermal power systems, and the objective function is to minimize the total dispatch cost of the system, that is: Where P i,t ,P G,t ,P pv,t is the power generation of thermal power, solar thermal power and photovoltaic power station at time t, E(P i,t ,P G,t ,P pv,t ) is the power generation cost function; The constraints for thermal power, photovoltaic power and power balance are as follows: Where P i,min is the minimum output power of the thermal power station; P i,max is the maximum output power of the thermal power station; P i，t-1 is the power generation capacity of the thermal power station at time t-1; u i，t is the maximum ramp rate of the thermal power station; P pv,min is the minimum output power of the photovoltaic power station; P pv,max is the maximum output power of the photovoltaic power station; P load is the load of the power grid; The dispatch model of a CSP power station is simplified into three parts: heat collection, heat storage and release, and power generation. Consider the CSP power received by the heat collection system: Among them, η SF Represents the photothermal conversion efficiency; S SF is the mirror field area, D t is the DNI at time t; Part of this power is used for power generation and the other part is stored. The power balance relationship is as follows: in, represents the thermal power used for power generation at the CSP station at time t, P t TS ,c ,P t TS,d They represent the heat storage and heat release powers at time t respectively; For the heat storage system, the heat storage state equation is: in, represents the heat storage capacity of the system at time t, represents the heat storage capacity of the system at time t+1, η loss represents the attenuation coefficient of the heat storage system; For the power generation part, thermal power is converted into electrical power. Among them, P Gt represents the electric power generated by the CSP plant at time t, η e is the thermoelectric conversion efficiency; Step 3: Use the affine strategy to establish an adjustable robust optimization model. The output power error of the CSP power station model is expressed as: ΔP csp is the uncertainty of the output of the CSP power station; the affine strategy is used to deal with the uncertainty in the system. The specific scheduling strategy is as follows: The actual value of the heat storage and heat release power is expressed as: in, is the initial expected heat storage and release power, λ charge，t , dis，t are the corresponding heat storage and heat release affine coefficients respectively; the part of the thermal power of the CSP plant used for power generation, the affine strategy is expressed as: in, is the expected thermal power of the initial CSP power station, λ g,t is the affine coefficient of the CSP power station; the total affine constraint is: Where η SF Represents the photothermal conversion efficiency; S SF is the mirror field area, D t is the DNI at time t, P load，t is the load of the power grid after affine transformation during period t, P load0，t is the original load of the power grid during period t, λ l1 is the affine factor of the uncertainty of photovoltaic output to grid load, λ l2 is the affine factor of the uncertainty of solar thermal output to grid load, ΔP pv is the uncertainty of photovoltaic output, P i0，t is the original output of the thermal power station at time t, λ i1 is the affine factor of the uncertainty of photovoltaic output to thermal power, λ i2 is the affine factor of the uncertainty of solar thermal output to thermal power; Step 4: Solve the above constraints, and the solution results include the power generation of the CSP plant in each time period, the thermal energy storage or release, and the affine coefficients that need to be adjusted.

2. According to claim 1, a method for robust optimization scheduling of a CSP plant based on a polyhedral uncertain set is characterized by: The solution method for step 4 is to transform the scheduling model of the solar thermal power station and the affine strategy into a mixed integer linear programming model for solution.

3. According to claim 1, a method for robust optimization scheduling of a CSP plant based on a polyhedral uncertain set is characterized by: Step 3: The objective function is:

Images