Concentrating solar thermal power generation system optimization scheduling method considering solar irradiation uncertainty

Abstract

The invention relates to a concentrating solar thermal power generation system optimization scheduling method considering solar radiation uncertainty, which comprises the steps of centralized concentrating solar thermal power generation system modeling, solar radiation uncertainty modeling and double-layer optimization model establishment, and the objective of an upper-layer optimization model is to minimize the operation cost of the system and the optimal capacity configuration of a CSP system. The objective of the lower-layer optimization model is based on CSP capacity configuration determined by the upper layer, and the two-layer optimization model is determined and then solved. The method has the beneficial effects that capacity configuration and scheduling optimization are performed on the CSP system in combination with a double-layer optimization strategy, the solar radiation uncertainty model is constructed, and various possible solar radiation scenes can be comprehensively considered, so that the optimal capacity configuration and scheduling scheme is determined, and the robustness and stability of the system are improved.

Description

Technical Field

[0001] The invention belongs to a solar power generation system and relates to the operation scheduling of a concentrated solar thermal power generation system. Background Art

[0002] Concentrated solar thermal power generation technology has received extensive attention and research in recent years, especially in areas rich in solar energy resources such as Gansu, where the development momentum of CSP systems is rapid. CSP technology mainly uses collectors to focus solar radiation onto receivers, converts light energy into high-temperature heat energy, and transfers it to the thermal cycle system through heat exchange media (such as molten salt, thermal oil, etc.) to achieve power generation. Compared with traditional photovoltaic power generation (PV) technology, the significant advantage of CSP systems lies in their efficient heat energy conversion capability and strong thermal energy storage capacity, which can achieve smooth power output to a certain extent. The power generation efficiency of CSP systems depends not only on the performance of the collectors, but also on the real-time changes in solar irradiance. Since solar irradiance is affected by many factors such as season, climate, and geographical location, its changes have great uncertainty.

[0003] In actual operation, the power generation of CSP systems shows diurnal and seasonal variations. At the same time, weather conditions such as cloud cover, rainfall, and atmospheric transparency will also have a significant impact on solar irradiance, which in turn affects the power generation capacity of CSP systems. Traditional CSP system scheduling schemes are usually formulated based on the most unfavorable solar irradiation conditions throughout the year, without fully considering the differences in seasons, climate, and operating environment, which makes the scheduling strategy too conservative and unable to fully utilize the power generation potential of CSP systems. Ignoring the uncertainty of solar irradiation, it is difficult to cope with actual weather changes during operation, thus affecting the operating efficiency and stability of the system. To this end, recent studies have attempted to cope with changes in solar irradiance by real-time monitoring and scheduling optimization of CSP systems. In CSP systems, thermal energy storage technology is used to store excess thermal energy to supplement power generation when solar irradiation is insufficient. However, most of these traditional methods are based on deterministic models and do not fully consider the uncertainty of solar irradiation, making it difficult to obtain the best scheduling scheme in actual operation and operation.

[0004] In order to fully tap the power generation capacity of CSP systems, it is usually necessary to model the uncertainty of solar irradiance and formulate scheduling strategies based on actual environmental conditions. However, since the actual solar irradiance is affected by meteorological conditions, there is a large degree of randomness and uncertainty. Especially when the solar irradiance changes dramatically, it may cause the power output of the CSP system to deviate from the expected value, thus causing an imbalance in the supply and demand of the power system and even causing grid instability.

[0005] Among the existing technical solutions, some studies have used linear programming, dynamic programming and heuristic algorithms to optimize the capacity configuration and scheduling of CSP systems. These methods are usually based on certain solar irradiation conditions, taking into account the thermal energy storage characteristics of the CSP system and the power balance constraints of the power system. For example, some studies use historical solar irradiation data for scheduling optimization, but these methods cannot cope with the uncertainty of solar irradiation in real time. In addition, some studies have introduced probabilistic scenario generation technology to model and optimize solar irradiation uncertainty. For example, by generating a series of possible solar irradiation scenarios and analyzing these scenarios using optimization algorithms, the corresponding scheduling solutions are obtained. However, these methods have high computational complexity in practical applications, and the robustness and accuracy of the optimization results still need to be improved. Summary of the invention

[0006] The purpose of this invention is to propose an optimization scheduling method for CSP system, aiming to solve the problem of unstable power output caused by solar irradiation uncertainty. By introducing a two-layer optimization model, the uncertainty of solar irradiation and the heat storage characteristics of the CSP system are comprehensively considered to find the optimal capacity configuration and scheduling scheme.

[0007] The technical solution of the present invention is a method for optimizing the scheduling of a concentrated solar thermal power generation system considering the uncertainty of solar radiation, which adopts the following steps: 1) Modeling a concentrated solar thermal power generation system, according to the key components of the CSP system, namely, the power balance relationship P of the concentrated solar thermal system, the heat storage system, the electric heating device, and the power generation system t S-H +P t E-H -P t H-P +P t T-H -P t H-T =0, establish a model, where P t S-H is the thermal power absorbed by the heat transfer medium, P t E-H is the power transferred from the electric heating device to the heat transfer medium, P t H-P is the thermal power transferred from the heat transfer medium to the power generation system, P t T-H and P t H-T is the heat exchange power between the heat transfer medium and the thermal storage system. The model simulates the operation of the system under different solar irradiation conditions.

[0008] 2) Modeling of solar irradiance uncertainty: Establishing the Ito process model to describe the fluctuation of solar irradiance and simulate different solar irradiation scenarios. Assuming that the predicted solar output is The actual solar output is P s (t), the error between them is ξ(t), which satisfies: According to the Ito process, the above equation can be written as a stochastic differential equation: dξ(t)=μ(ξ(t),t)dt+σ(ξ(t),t)dW t , where ξ(t) is the error between the actual output and the predicted output, W t is the standard Wiener process, μ(ξ(t), t) is the drift term of the Ito process, and σ(ξ(t), t) is the diffusion term of the Ito process;

[0009] 3) Establish a two-layer optimization model. The goal of the upper optimization model is to minimize the system operation cost and the optimal capacity configuration of the CSP system. The upper optimization goal is minf1 = fG, run + f G,startup +f csp，run +f load-cut , the lower layer optimization target is minf2 = P csp，max , the optimization variable is P csp,max , where f G,run is the operating cost of traditional thermal power units, f G,startup is the startup cost of the thermal power unit, f csp,run is the operating cost of the CSP power plant, f load-cut is the load shedding cost, P csp,max It is the upper capacity configuration of CSP;

[0010] The constraints of the CSP system that need to be considered are:

[0011] CSP constraint 1: output upper and lower bound constraints, In the formula, I t,csp is the binary variable of the switch state of the CSP system; P t,csp is the scheduling output of CSP in time period t; P csp,min With P csp,max are the lower and upper limits of CSP output respectively;

[0012] CSP constraint 2: minimum start and stop time constraints, In the formula, I t-1，csp The switching state of the CSP system in the t-1 period (binary variable), T csp,off is the duration of CSP shutdown, T csp,on is the duration of CSP operation, I j，csp is the switch state of the CSP system in the jth period (binary variable);

[0013] CSP Constraint 3: Instantaneous thermal power constraint, P t,csp / η PB,csp +P cha,t,csp≤P dis,t,csp +P t,fore,csp , where P t，csp is the scheduling output of CSP in time period t, η PB,csp is the power efficiency of CSP, P cha,t,csp is the charging output power of CSP in time period t, P dis,t,csp is the output power of the thermal storage system in CSP during time period t, P t,fore,csp is the available solar thermal power in time period t;

[0014] CSP constraint 4: State of charge constraint, E t,csp =E t-1,csp +η TES,csp P cha,t,csp -P dis,t,csp / η TES,csp , where E t,csp is the amount of electricity stored in the CSP thermal storage system during time period t, E t-1，csp is the amount of electricity stored in the CSP thermal storage system during time period t-1; η TES,csp is the efficiency coefficient of the heat storage system;

[0015] CSP constraint 5: Thermal storage system boundary constraint, E csp,min ≤E t,csp ≤E csp,max , where E csp,min and E csp,max They are the lower and upper limits of the state of charge of the thermal storage system in CSP, respectively;

[0016] CSP constraint 6: CSP ramp rate constraint, -R csp,d ≤P t,csp -P t-1,csp ≤R csp,u , where R csp,d and R csp,u are the minimum rising rate and maximum falling rate of the CSP system, P t-1，csp is the scheduling output of CSP in time period t-1;

[0017] CSP Constraint 7: Power Balance Constraint, Where P t,G,i is the output of the i-th thermal power unit, N is the number of thermal power units, P t,pv is the photovoltaic output, P t,wp is wind power output, P t,csp is the CSP system output, P t,load is the load, P t,load -cut is the load shedding amount;

[0018] 4) Solving the two-layer optimization model: the upper optimization model is solved by a non-dominated multi-objective genetic algorithm, and the lower optimization model is based on the capacity configuration determined by the upper layer and takes into account the actual operating constraints of the power system and the operating constraints of the CSP system. The CPLEX solver is used to output the corresponding scheduling plan. In each iteration of the upper optimization, the lower optimization algorithm (using the linear programming algorithm in the CPLEX solver) is called to calculate the optimal scheduling plan under the given capacity configuration.

[0019] The drift term and diffusion term are calculated by the following formula: Where a is the time constant of cloud shadow change, b is the empirical deviation between the mean value of the Ito process and the predicted curve, and t r and t s are the sunrise and sunset times, t represents the time, sin() represents the solar elevation angle, c

[0020] is the intensity of uncertainty, d and e are the intervals of solar irradiance variation.

[0021] Beneficial effects of the invention: This patent combines a two-layer optimization strategy to perform capacity configuration and scheduling optimization on the CSP system. By constructing a solar irradiation uncertainty model, this patent can comprehensively consider various possible solar irradiation scenarios in scheduling optimization, thereby determining the optimal capacity configuration and scheduling scheme, and improving the robustness and stability of the system. The two-layer optimization model realizes capacity configuration optimization of the CSP system at the upper layer and real-time scheduling optimization at the lower layer, so that the CSP system can maintain a high operating efficiency under different solar irradiation conditions. Compared with existing optimization methods, the method of this patent can better cope with changes in solar irradiation, reflect the complexity of the actual operating environment, more accurately reflect the operating status of the CSP system under different irradiation conditions, provide more stable power output, improve power generation efficiency and grid stability, effectively reduce the operating cost of the system, effectively reduce the scheduling risk of the power system, and improve the stability of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 is a flow chart of the method of the present invention;

[0023] Figure 2 It is a schematic diagram of CSP system modeling;

[0024] Figure 3 It is the IEEE 14-node simulation structure diagram;

[0025] Figure 4 It is the result of 20 generations of optimization using the NSGAⅡ algorithm. DETAILED DESCRIPTION

[0026] In concentrated solar thermal power generation (CSP) systems, due to the uncertainty of solar irradiation, traditional scheduling and capacity configuration methods may not achieve optimal performance in actual operation. To solve this problem, this patent proposes a two-layer optimization model for CSP systems to achieve optimal scheduling and capacity configuration of power systems. The implementation of this patent mainly includes the following 4 parts, namely:

[0027] 1. Modeling of centralized solar thermal power generation system

[0028] 2. Solar irradiance uncertainty modeling

[0029] 3. Establishment of the two-layer optimization model

[0030] 4. Solution method of two-layer optimization model

[0031] The logical relationship between the above four processes can be seen in Figure 1 The following are four parts in detail.

[0032] 1) Modeling of centralized solar thermal power generation system

[0033] In order to achieve optimal scheduling of the CSP system, it is necessary to first establish a mathematical model of the CSP system. The model includes the key components of the CSP system, namely, the concentrating solar collector system, the heat storage system, the electric heating device and the power generation system, such as Figure 2 As shown. The power balance relationship of the CSP system is:

[0034] P t S-H +P t E-H -P t H-P +P t T-H -P t H-T =0(0)

[0035] Among them, P t S-H is the thermal power absorbed by the heat transfer medium, P t E-H is the power transferred from the electric heating device to the heat transfer medium, P t H-P is the thermal power transferred from the heat transfer medium to the power generation system, P t T-H and P t H-T It is the heat exchange power between the heat transfer medium and the heat storage system.

[0036] The modeling process needs to consider the solar energy collection and conversion efficiency of the concentrated solar collector system, the thermal capacity of the thermal storage system, and the power output characteristics of the entire thermal cycle system. This model will serve as the basis for optimization calculations and is used to simulate the operating status of the system under different solar irradiation conditions.

[0037] 2) Solar irradiance uncertainty modeling

[0038] Since solar radiation has great randomness and uncertainty in the actual engineering operation environment, this patent introduces solar radiation uncertainty modeling to ensure the robustness of CSP system scheduling optimization. By establishing the Ito process model ( The model is used to describe the fluctuation of solar irradiance and simulate different solar irradiation scenarios. This model is used to describe the variation range of solar irradiance in different time periods and provide necessary input data for subsequent optimization calculations.

[0039] Assume the predicted solar output is The actual solar output is P s (t), the error between them is ξ(t), which satisfies:

[0040]

[0041] according to Process, formula (2) can be written as the following stochastic differential equation:

[0042] dξ(t)=μ(ξ(t),t)dt+σ(ξ(t),t)dW t (3)

[0043] Where ξ(t) is the error between the actual output and the predicted output, W t is a standard Wiener process, μ(ξ(t), t) is The drift term of the random process, σ(ξ(t), t) is Diffusion term of random process. Drift term and diffusion term are calculated by the following formula:

[0044]

[0045] Where a is the time constant of cloud shadow change, b is the empirical deviation between the mean value of the Ito process and the predicted curve, and t r and t s are the sunrise and sunset times, t represents the time, sin() represents the solar elevation angle, c

[0046] is the intensity of uncertainty, d and e are the intervals of solar irradiance variation.

[0047] 3) Establishment of a two-layer optimization model

[0048] In order to achieve the optimal scheduling of the CSP system, this patent constructs a two-level optimization model. The goal of the upper optimization model is to minimize the operating cost of the system and the optimal capacity configuration of the CSP system. Analysis and research have revealed that these two optimal goals conflict with each other, making this problem a multi-objective optimization challenge. The first optimization goal is to minimize the system operating cost after the integration of the CSP power plant, mainly considering the operating cost of the unit, the startup cost of the unit and the load shedding cost; the second optimal goal is to minimize the CSP system capacity; the decision variable is CSP

[0049] System capacity. From this we get:

[0050]

[0051] Among them, f G,run is the operating cost of traditional thermal power units, f G,startup is the startup cost of the thermal power unit, f csp,run is the operating cost of the CSP power plant, f loa-cut is the load shedding cost, P csp,max It is the upper capacity configuration of CSP.

[0052] The goal of the lower optimization model is to optimize the dispatch of the power system based on the CSP capacity configuration determined by the upper layer. In addition to considering the conventional operation constraints of the power system, the relevant constraints of the CSP system must also be considered in the solution.

[0053] CSP Constraint 1: Output Upper and Lower Constraints

[0054]

[0055] In the formula, I t,csp is the binary variable of the switch state of the CSP system; P t,csp is the scheduling output of CSP in time period t; P csp,min With P csp,max are the lower and upper limits of the CSP output, respectively. This constraint ensures that the output power of the CSP power plant remains within the specified upper and lower limits in each time period.

[0056] CSP Constraint 2: Minimum start and stop time constraints

[0057]

[0058] In the formula, I t-1，csp is the switching state of the CSP system in the t-1 period, T csp,off is the duration of CSP shutdown, T csp,on is the duration of CSP operation, I j，cspis the on / off state of the CSP system in the jth period (binary variable). This constraint enforces the minimum duration for which the CSP plant must remain on or off after startup or shutdown to ensure operational stability and avoid frequent switching.

[0059] CSP Constraint 3: Instantaneous Thermal Power Constraint

[0060] P t,csp / η PB,csp +P cha,t,csp ≤P dis,t,csp +P t,fore,csp (8)

[0061] Where P t，csp is the scheduling output of CSP in time period t, η PB,csp is the power efficiency of CSP, P cha,t,csp is the charging output power of CSP in time period t, P dis,t,csp is the output power of the thermal storage system in CSP during time period t, P t,fore,csp is the available solar thermal power in time period t. This constraint ensures the balance between thermal power and electricity generation in the CSP system.

[0062] CSP Constraint 4: State of Charge Constraint

[0063] E t,csp =E t-1,csp +η TES,csp P cha,t,csp -P dis,t,csp / η TES,csp (9)

[0064] In the formula, E t,csp is the amount of electricity stored in the CSP thermal storage system during time period t, E t-1，csp is the amount of electricity stored in the CSP thermal storage system during time period t-1; η TES,csp is the efficiency coefficient of the thermal storage system, reflecting the loss during the storage process. This constraint controls the state of charge of the thermal storage system to ensure that it remains within the allowable range of efficient operation.

[0065] CSP Constraint 5: Thermal Storage System Boundary Constraints

[0066] E csp,min ≤E t,csp ≤E csp,max (10)

[0067] In the formula, E csp,min and E csp,max are the lower and upper limits of the state of charge of the thermal storage system in CSP, respectively. This constraint ensures that the state of charge of the thermal storage system remains within its upper and lower limits, preventing overcharging or deep discharge that could damage the system.

[0068] CSP Constraint 6: CSP ramp rate constraint

[0069] -R csp,d ≤P t,csp -P t-1,csp ≤R csp,u (11)

[0070] In the formula, R csp,d and R csp,u are the minimum rising rate and maximum falling rate of the CSP system, P t，csp is the scheduling output of CSP in time period t, P t-1，csp is the scheduling output of CSP in time period t-1. This constraint limits the speed at which the CSP system increases or decreases production to ensure the stability of the CSP system.

[0071] CSP Constraint 7: Power Balance Constraint

[0072]

[0073] Where P t,G,i is the output of the i-th thermal power unit, N is the number of thermal power units, P t,pv Photovoltaic (PV)

[0074] Output, P t,wp is wind power (WP) output, P t,csp is the CSP system output, P t,load is the load, P t,load-cut is the load rejection. t,pv With P t,csp Both introduce uncertainty in solar irradiance. This constraint ensures that the total power meets the demand, taking into account any load shedding if necessary.

[0075] The above two-layer optimization model not only considers the power generation efficiency of the system, but also integrates the system operation stability under different irradiation conditions, ensuring the optimal performance of the CSP system under various weather conditions.

[0076] 4) Two-layer optimization model solution method

[0077] The larger the CSP system capacity configuration, the higher the initial investment and the lower the possibility of load shedding.

[0078] The smaller the system capacity configuration, the lower the initial investment, the higher the possibility of load shedding, and the higher the corresponding load shedding cost. It can be seen that there is a mutually exclusive relationship between the two objectives of this patent, so the upper-level optimization adopts the non-dominated multi-objective genetic algorithm (NSGAⅡ) to solve and optimize the capacity configuration of the CSP system.

[0079] In each upper-level optimization iteration, the lower-level optimization algorithm is called to calculate the optimal scheduling solution under a given capacity configuration. Through iterative optimization, the optimal capacity configuration and scheduling solution that simultaneously meets the requirements of power generation efficiency and stability are obtained. This solution method can effectively deal with the uncertainty of solar irradiation and improve the overall operating efficiency of the CSP system.

[0080] The lower-level optimization is based on the capacity configuration determined by the upper level and takes into account the actual operating constraints of the power system and the operating constraints of the CSP system, and uses the CPLEX solver to output the corresponding scheduling plan.

[0081] In the existing CSP system capacity configuration and scheduling optimization technical solutions, since the uncertainty of solar irradiation is not fully considered, when the actual solar irradiation conditions deviate greatly from the design value, the power generation output of the CSP system is often difficult to achieve the expected. This deviation may lead to a decrease in the operating efficiency of the system, while increasing the complexity and risk of power system scheduling. Compared with the prior art, this patent introduces uncertainty modeling of solar irradiation to make the scheduling optimization of the CSP system more robust and adaptable. First, in practical applications, since the uncertainty of solar irradiation directly affects the power generation performance of the CSP system, the model of this patent can more accurately reflect the operating status of the CSP system under different irradiation conditions and provide more stable power output. Secondly, the two-layer optimization model ensures the coordination of the capacity configuration of the CSP system and the scheduling of the power system, and realizes the efficient utilization of the CSP system by optimizing the configuration of the upper limit of the CSP system capacity. At the same time, the model of this patent shows a high optimization efficiency in the simulation experiment, effectively reducing the operating cost of the system and improving the stability of the power system.

[0082] Traditional dispatching methods usually rely on fixed estimates of solar irradiation and fail to effectively cope with the dynamic changes in solar irradiation. For example, under conditions of sufficient sunlight, traditional dispatching may be too conservative and fail to fully utilize the power generation potential of the CSP system; while under conditions of insufficient sunlight or sudden changes, the dispatching strategy may not be adjusted in time, resulting in unstable system operation.

[0083] The two-layer optimization model proposed in this patent can more accurately reflect the complexity of the actual operating environment by modeling the uncertainty of solar irradiation. The upper layer optimization determines the optimal capacity configuration of the CSP system to ensure that high power generation efficiency can be maintained under different solar irradiation conditions; the lower layer optimization comprehensively considers the actual changes in solar irradiation in real-time scheduling, dynamically adjusts the system operation strategy, and maximizes the utilization efficiency of the system.

[0084] Compared with the prior art, the method of the present invention shows higher efficiency and stability in practical applications through comprehensive analysis and optimization of solar irradiation uncertainty. It not only improves the power generation efficiency of the CSP system, but also effectively reduces the dispatch risk of the power system and ensures the reliable operation of the power grid. This effective treatment of uncertainty factors gives this patent a significant advantage in actual engineering and provides a more scientific and practical solution for the dispatch and capacity configuration of the CSP system.

[0085] Features of this application: 1) It overcomes the problem of insufficient consideration of solar irradiation uncertainty in the prior art; this patent introduces solar irradiation uncertainty modeling and establishes a double-layer optimization model that considers solar irradiation changes. Through the analysis and modeling of the randomness of solar irradiation, this patented technology can achieve the optimal scheduling of the CSP system under different solar irradiation conditions, improving the stability and power generation efficiency of the system in actual operation.

[0086] 2) Provides an effective CSP system capacity configuration and scheduling optimization solution; This patent constructs a two-layer optimization model, the upper layer optimization realizes the capacity configuration of the CSP system; the lower layer optimization comprehensively considers the solar radiation changes and power system operation constraints in real-time scheduling to ensure that the system can maintain optimal performance under different environmental conditions. This method improves the operational flexibility and power generation efficiency of the CSP system.

[0087] 3) High efficiency in solving the optimization model: This patent adopts a nested optimization algorithm to solve the double-layer optimization model. Compared with the traditional method, the solution process of this patent is more efficient. By rationally designing the algorithm structure, a large number of repeated calculations are avoided, the computational complexity in the optimization process is reduced, and a faster optimization solution is achieved.

[0088] 4) Improved the stability of power grid operation: By optimizing the dispatch of the CSP system considering the uncertainty of solar radiation, this patent effectively reduces the impact of the output power fluctuation of the CSP system on the stability of the power system. It ensures that the power system maintains power balance when accepting CSP power generation, and improves the overall operational reliability of the power grid.

[0089] The optimization scheduling method and its solution algorithm proposed in this patent will provide strong support for the efficient operation of CSP systems in actual power systems, and are expected to become a key technology for improving the utilization rate of concentrated solar thermal power generation and grid stability.

[0090] To verify the effectiveness of this method, this patent conducted a simulation experiment on the IEEE 14-node system. The system includes traditional generators, loads and CSP systems, simulating the operation status of the power system under different solar irradiation conditions. By comparing and analyzing the performance of traditional scheduling methods and the method of this patent, the advantages of the two-layer optimization model in improving the power generation efficiency and operation stability of the CSP system are verified. The simulation results show that the optimization method of this patent can effectively reduce the system operation cost and improve the stability of the power system.

[0091] IEEE 14-bus system Figure 3 As shown, the new energy station is integrated in nodes 3 and 6, the photovoltaic power station is connected to bus 3, the wind farm is connected to bus 6, and the CSP power station with a thermal storage system is connected to node 8.

[0092] In the two-layer optimization, The process parameters were set as follows: a = 1600, b = -1.5, c = 0.84, d = -17.43, e = 37.84, t r =7, t s =18. The population size in the NSGAⅡ algorithm parameters is set to 50, and the maximum number of evolution generations is 20. The optimization results are as follows Figure 4 As shown, the horizontal axis f1 is the optimization target f1 in formula (5), and the vertical axis is the optimization target f2 in formula (5). Figure 4 The blue solid circles represent all individuals in the evolution process, and the black circles represent the non-dominated solutions obtained after 20 generations of optimization. Figure 4 It can be seen that in order to achieve more economical operation and scheduling, the maximum capacity of CSP should be configured at least above 200MW. Figure 4 The most economical cost can be achieved under the solution of f1=4100 USD and f2=335MW.

Claims

1. A method for optimizing the scheduling of a concentrated solar thermal power generation system considering the uncertainty of solar radiation, characterized in that The following steps are adopted: 1) Modeling of the centralized solar thermal power generation system, according to the key components of the CSP system, namely the power balance relationship P of the concentrating solar thermal system, the heat storage system, the electric heating device, and the power generation system t S-H +P t E-H -P t H-P +P t T-H -P t H-T =0, establish a model, where P t S-H is the thermal power absorbed by the heat transfer medium, P t E-H is the power transferred from the electric heating device to the heat transfer medium, P t H-P is the thermal power transferred from the heat transfer medium to the power generation system, P t T-H and P t H-T is the heat exchange power between the heat transfer medium and the thermal storage system. The model simulates the operation of the system under different solar irradiation conditions. 2) Modeling of solar irradiance uncertainty: Establishing the Ito process model to describe the fluctuation of solar irradiance and simulate different solar irradiation scenarios. Assuming that the predicted solar output is The actual solar output is Ps(t), and the error between them is ξ(t), which satisfies: According to the Ito process, the above equation can be written as a stochastic differential equation: dξ(t)=μ(ξ(t),t)dt+σ(ξ(t),t)dW t , where ξ(t) is the error between the actual output and the predicted output, Wt is the standard Wiener process, μ(ξ(t), t) is the drift term of the Ito process, and σ(ξ(t), t) is the diffusion term of the Ito process; 3) Establish a two-layer optimization model. The goal of the upper optimization model is to minimize the system operation cost and the optimal capacity configuration of the CSP system. The upper optimization goal is min f1 = f G，run +f G,startup +f csp,run +f load-cut , the lower layer optimization goal is min f2=P csp,max The optimization variable is P csp,max , where f G,run is the operating cost of traditional thermal power units, f G,startup is the startup cost of the thermal power unit, f csp,run is the operating cost of the CSP power plant, f load-cut is the load shedding cost, P csp,max It is the upper capacity configuration of CSP; The constraints of the CSP system that need to be considered are: CSP constraint 1: output upper and lower bound constraints, In the formula, I t,csp is the binary variable of the switch state of the CSP system; P t,csp is the scheduling output of CSP in time period t; P csp,min With P csp,max are the lower and upper limits of CSP output respectively; CSP constraint 2: minimum start and stop time constraints, In the formula, I t-1，csp is the switching state of the CSP system in the t-1 period, T csp,off is the duration of CSP shutdown, T csp,on is the duration of CSP operation, I j，csp is the switching state of the CSP system in the jth period; CSP Constraint 3: Instantaneous thermal power constraint, P t,csp / η PB,csp +P cha,t,csp ≤P dis,t,csp +P t,fore,csp , where P t，csp is the scheduling output of CSP in time period t, η PB,csp is the power efficiency of CSP, P cha,t,csp is the charging output power of CSP in time period t, P dis,t,csp is the output power of the thermal storage system in CSP during time period t, P t,fore,csp is the available solar thermal power in time period t; CSP constraint 4: State of charge constraint, E t,csp =E t-1,csp +η TES,csp P cha,t,csp -P dis,t,csp / η TES,csp , where E t,csp is the amount of electricity stored in the CSP thermal storage system during time period t, E t-1，csp is the amount of electricity stored in the CSP thermal storage system during time period t-1; η TES,csp is the efficiency coefficient of the heat storage system; CSP constraint 5: Thermal storage system boundary constraint, E csp,min ≤E t,csp ≤E csp,max , where E csp,min and E csp,max They are the lower and upper limits of the state of charge of the thermal storage system in CSP, respectively; CSP constraint 6: CSP ramp rate constraint, -R csp,d ≤P t,csp -P t-1,csp ≤R csp,u , where R csp,d and R csp,u are the minimum rising rate and maximum falling rate of the CSP system, P t-l，csp is the scheduling output of CSP in time period t-1; CSP Constraint 7: Power Balance Constraint, Where P t,G,i is the output of the i-th thermal power unit, N is the number of thermal power units, P t,pv is the photovoltaic output, P t,wp is wind power output, P t,csp is the CSP system output, P t,load is the load, P t,load-cut is the load shedding amount; 4) Solving the two-layer optimization model: the upper optimization model is solved using a non-dominated multi-objective genetic algorithm, and the lower optimization model is based on the capacity configuration determined by the upper layer and takes into account the actual operating constraints of the power system and the operating constraints of the CSP system. The CPLEX solver is used to output the corresponding scheduling plan. In each upper optimization iteration, the lower optimization algorithm is called to calculate the optimal scheduling plan under the given capacity configuration.

2. The method for optimizing the scheduling of a concentrated solar thermal power generation system considering the uncertainty of solar radiation according to claim 1 is characterized by: The drift term and diffusion term are calculated by the following formula: Where a is the time constant of cloud shadow change, b is the empirical deviation between the mean value of the Ito process and the predicted curve, and t r and t s are the sunrise and sunset times respectively, t represents time, sin() represents the solar elevation angle, c represents the intensity of uncertainty, and d and e are the intervals of solar irradiance change.

Images