Frequency control method of photovoltaic-photo-thermal combined power generation system

Abstract

The invention relates to the technical field of operation control of a power system, in particular to a frequency control method of a photovoltaic-photothermal combined power generation system, and the method comprises the steps: solving an optimal solution through iterative calculation, feedback correction and rolling optimization according to the characteristics that an MPC controller can give consideration to multiple targets and multiple constraints; a double-layer MPC control framework is adopted for establishing the photovoltaic-photo-thermal combined power generation system model, the upper layer mainly considers output distribution and frequency fluctuation, the lower layer mainly considers CSP internal coordinated operation, the advantages of an MPC control strategy are fully utilized, control targets can be formulated more clearly and controlled more rapidly and finely through use of the double-layer MPC, and the control efficiency is improved. And the frequency modulation effect is better than that of a single-layer MPC. Meanwhile, the double-layer MPC can consider more targets and constraints, and can adjust the adjustment margins of the PC and the CSP in time according to the AGC requirements, so that the method can cope with various frequency fluctuation conditions, and is more targeted when control is implemented.

Description

Technical Field

[0001] The present invention relates to the technical field of power system operation control, and in particular to a frequency control method for a photovoltaic-solar-thermal combined power generation system. Background Art

[0002] Concentrated solar power generation technology converts solar energy into thermal energy and then into electricity. Concentrated solar power systems are equipped with thermal storage tanks that store thermal energy, enabling stable output with greater regulation margin and flexibility. Photovoltaic power generation technology also uses solar energy as an energy source to achieve photoelectric conversion. Therefore, its construction environment and location requirements are similar to those of concentrated solar power plants. Large-scale joint development of the two technologies offers lower costs. In many new projects, combining these technologies for power generation has become a popular trend and holds enormous potential.

[0003] The combined power generation of photovoltaic and solar thermal power generation systems can fully utilize the advantages of both and form complementarity to a certain extent. It combines the high energy efficiency of photovoltaic power generation systems with the high flexibility of solar thermal power generation systems, improves dynamic response performance, reduces the dependence of the combined power generation system's frequency regulation on weather conditions, and can utilize the abandoned light of photovoltaic power generation systems, so that the power generation efficiency and new energy absorption capacity are improved overall, which is more conducive to the safe and stable operation of the power grid.

[0004] Currently, when performing load frequency control (LFC), the system frequency deviation is often calculated through automatic generation control (AGC). The resulting area control error (ACE) signal is input into the PI controller, thereby forming control commands for the units participating in LFC regulation within the system, achieving secondary frequency regulation. However, for photovoltaic-thermal combined power generation systems, the internal structure, dynamic characteristics, component composition, and related control quantities of the power generation system are significantly different from traditional thermal power units. The combined power generation system contains multiple nonlinear links, involves multiple constraints, and needs to consider multiple objectives. It is difficult for PI controllers to achieve such complex and precise control. Summary of the Invention

[0005] The present invention provides a frequency control method for a photovoltaic-photothermal combined power generation system, which can effectively solve the problems in the background technology.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is:

[0007] A frequency control method for a photovoltaic-solar thermal combined power generation system comprises the following steps:

[0008] Develop a two-dimensional dynamic load reduction strategy for the photovoltaic-solar thermal power generation system. Based on the positive and negative values of the automatic generation control (AGC) instructions, the direct solar irradiance (DNI), and the state of energy (SOE) of the thermal storage tank, the load reduction rate of the photovoltaic (PV) and solar thermal power generation (CSP) systems is dynamically adjusted to reserve active reserve capacity.

[0009] An electric heater (EH) is configured in the combined power generation system to convert the abandoned solar energy generated by PV load shedding into thermal energy through electric heating and store it in the CSP thermal storage tank;

[0010] Establish a frequency response model for the photovoltaic-thermal combined power generation system, derive the system's state space model, and generate a model prediction sequence within the future prediction time domain;

[0011] A two-layer model predictive control architecture is designed. The upper-layer controller dynamically allocates PV and CSP output commands based on frequency deviation and AGC requirements, while the lower-layer controller coordinates the control of the CSP's internal subsystems.

[0012] Based on the rolling optimization mechanism, the optimal control sequence is solved in each control cycle, and the first control variable is applied to the system to achieve frequency stability control through iterative feedback.

[0013] Furthermore, the formulation of a two-dimensional dynamic load reduction strategy for a photovoltaic-solar thermal combined power generation system includes:

[0014] When the AGC instruction is positive and higher than 5% of the total system capacity, DNI>1000W / m 2 , set PV and CSP to retain 10% active reserve capacity;

[0015] When the AGC instruction is positive and less than 5% of the total system capacity, DNI>900W / m 2 , set PV and CSP to retain 5% active reserve capacity each, and give priority to PV to increase output to the maximum power point tracking MPPT mode;

[0016] When the AGC command is negative, the CSP output is limited to no less than its lower limit, the PV switches to droop control mode to reduce the output, and the SOE of the heat storage tank is maintained in the range of [0.1, 0.9] through the EH.

[0017] Furthermore, the operation of the electric heating device (EH) satisfies the following conditions:

[0018] When PV is abandoned and the SOE of the thermal storage tank is less than 0.9, the EH is started for energy conversion. The EH converts light energy into heat energy as shown in the following formula:

[0019] Q HOT =Q LIGHT η

[0020] Among them, Q HOT is the heat energy converted into the heat storage tank, Q LIGHT is the light energy absorbed by the electric heating device, and η is the energy conversion efficiency;

[0021] EH was stopped when SOE ≥ 0.9 and restarted when SOE < 0.8.

[0022] Furthermore, a state space model including the dynamic characteristics of the photovoltaic-thermal combined power generation system is constructed, which is specifically expressed as follows:

[0023]

[0024] Where x is the state variable of the regional system, including system frequency deviation, CSP main steam pressure and heat storage tank SOE; is the time derivative of the regional system state variable; u is the control variable of the regional system, including the photovoltaic inverter power command and the CSP turbine throttle opening; y is the measurement value of the regional system; A is the state matrix of the combined power generation system, B is the control input matrix, C is the output matrix, and D is the direct transfer matrix.

[0025] Furthermore, the forward Euler method is used to discretize the matrix A and the matrix B, which can be expressed as follows:

[0026]

[0027] Among them, Δx(k)=x(k)-x(k-1) represents the increment of the system state variable at time k.

[0028] Furthermore, the system state with m control steps and p prediction steps is predicted, which is specifically expressed as:

[0029]

[0030] Among them, y p (k+p|k) represents the output at time k+p predicted by the current time k;

[0031] The calculation method of the system output sequence of the next p steps is specifically expressed as:

[0032] Y p (k) = S x (x)Δx(k)+τy(k)+S u u(k);

[0033]

[0034]

[0035] Among them, I nc×ncis the identity matrix with the same dimensions as the matrix C.

[0036] Furthermore, the objective function used in the quadratic programming optimization when calculating the optimal control sequence of the regional system at the current moment based on the model prediction sequence is specifically expressed as:

[0037] J(z k )=J y (z k )+J u (z k );

[0038]

[0039] Among them, Z k is the result obtained by solving the quadratic programming objective function, ε k is the slack variable at the control step k; n y is the number of system output variables; y j (k+i|k) is the predicted value of the j-th output variable at time k+i; r j (k+i|k) is the control target of the j-th output variable at time k; S j is the scale factor of the jth output variable, which is determined by the upper and lower limits of the change of the variable during the control process; n u is the number of system input control variables; u j (k+i|k) is the size of the j-th input control variable at time k; u j，target (k+i|k) is the target of the j-th input control variable at time k;

[0040] The constraints for optimizing the objective function are:

[0041]

[0042] Among them, y j，min (i) and y j，max (i) is y j The upper and lower bounds of (k+i|k), and u j，max (i) is u j The upper and lower bounds of (k+i-1|k).

[0043] Furthermore, the objective function of the upper controller is specifically expressed as:

[0044]

[0045] Where (k+j|k) represents the predicted value at time k+j based on the state quantity at time k within the prediction time domain; a1, a2, and a3 are all weighting coefficients, representing the optimized proportions of frequency fluctuation, photovoltaic power generation power, and solar thermal power generation power during the frequency modulation process, respectively; N is the prediction time domain;

[0046] The weight coefficients a1, a2, and a3 are adaptively optimized and adjusted. The specific optimization method is as follows:

[0047] a=a0+k a (e |Δf|+sign(Δf)·dΔfldt -1);

[0048] Among them, a0 is the initial value of the weighting coefficient; k a is the weighted adjustment coefficient; is the system frequency difference change rate;

[0049] The formula for associating CSP weight with SOE is as follows:

[0050]

[0051] Among them, b0 is the initial value of the weighting coefficient; k b is the weighted adjustment coefficient; SOE is the state quantity of the heat storage tank of the CSP system;

[0052] The PV weight compensation formula is expressed as:

[0053] c=c0+b0-b;

[0054] Among them, c0 is the initial value of the weighting coefficient.

[0055] Furthermore, the objective function of the lower-level controller is specifically expressed as:

[0056] min J=M1||P CSP -P CSP.ref || 2 +M2|||p T -1|| 2 +M3|SOE-0.5|| 2 ;

[0057] Among them, M1, M2, and M3 are all weighting coefficients.

[0058] Furthermore, the constraints of rolling optimization include:

[0059] Frequency deviation limit: |Δf|≤0.5Hz;

[0060] Solar thermal output upper and lower limits: P CSP,min ≤P CSP ≤P CSP,max ;

[0061] Heat storage tank charging and discharging power limit: |Q 储热 ∣≤Q max .

[0062] The beneficial effects of the present invention are:

[0063] The present invention first builds a dynamic response model of the power system frequency of the photovoltaic-solar thermal power generation system, takes the system frequency and the output of the photovoltaic power generation system and the solar thermal power generation system as the upper optimization control target, and constructs the objective function. At the same time, the different dynamic response characteristics of the photovoltaic and solar thermal power generation systems are comprehensively considered. In the lower-level control, the main emphasis is on the coordination of the steam turbine subsystem and the steam generation system inside the solar thermal power generation system. On this basis, the steam turbine throttle opening and the molten salt flow rate of the solar thermal power generation system are used as the main controlled quantities, so that the solar thermal power generation system can accurately and flexibly respond to the load demand instruction. When using the model predictive control strategy, the discrete state space model of the system is first derived and iteratively calculated to obtain the model prediction sequence for a period of time in the future. The objective function is solved to obtain the optimal control sequence, and the first variable of the sequence is applied to the control object. The corresponding target weights and control parameters are adjusted according to the changes in the system frequency during the frequency modulation process and the changes in the state quantity of the heat storage tank of the solar thermal power generation system. At the next sampling moment, the optimization window is moved forward and solved again to achieve rolling optimization. When performing load frequency control, the frequency variation is divided into different stages. Combined with the state of the CSP system's heat storage tank, the controlled variable weight coefficients and MPC controller parameters are adjusted in a timely manner to minimize the difference between actual and predicted outputs. Compared to traditional load frequency control methods, this method fully considers the dynamic characteristics of the photovoltaic-thermal power generation system, offering greater flexibility and robustness in both coordinated control and frequency adjustment, resulting in superior control effectiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0065] Figure 1 This is a frequency control structure diagram of a photovoltaic-solar thermal combined power generation system based on a double-layer model predictive control in the present invention;

[0066] Figure 2 It is the regional frequency response model of the photovoltaic-thermal combined power generation system in the present invention. DETAILED DESCRIPTION

[0067] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0068] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly attached to the other element or there may be an intermediate element. When an element is referred to as being "connected to" another element, it may be directly connected to the other element or there may be an intermediate element. The terms "vertical," "horizontal," "left," "right," and similar expressions used herein are for illustrative purposes only and do not represent the only implementation methods.

[0069] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0070] The present invention discloses a frequency control method for a photovoltaic-thermal combined power generation system. Figure 1 The figure shows the frequency control structure of photovoltaic-solar thermal power generation system based on double-layer model predictive control, and Figure 2 The regional frequency response model of the photovoltaic-thermal combined power generation system includes the following steps:

[0071] A two-dimensional dynamic load shedding strategy for the photovoltaic-solar thermal power generation system is formulated. The load shedding rates of the photovoltaic (PV) and solar thermal (CSP) systems are dynamically adjusted according to the positive and negative values and magnitude of the automatic generation control (AGC) instructions, the direct solar radiation intensity (DNI), and the state of energy (SOE) of the thermal storage tank, reserving active reserve capacity. An electric heater (EH) is configured in the combined power generation system to convert the abandoned solar energy generated by PV load shedding into thermal energy through electric heating and store it in the CSP thermal storage tank. A frequency response model of the photovoltaic-solar thermal power generation system is established, and the state space model of the system is derived to generate a model prediction sequence in the future prediction time domain. A two-layer model predictive control architecture is designed. The upper-layer controller dynamically allocates the output instructions of PV and CSP according to the frequency deviation and AGC demand, and the lower-layer controller coordinates the control of the internal subsystems of the CSP. Based on the rolling optimization mechanism, the optimal control sequence is solved in each control cycle, and the first control variable is applied to the system to achieve frequency stability control through iterative feedback.

[0072] Specifically, a two-dimensional dynamic load reduction strategy for the photovoltaic-solar thermal power generation system was developed, taking into account the intensity of solar radiation and the state of the thermal storage tank of the solar thermal power generation system. Based on the different AGC requirements of the system, the DNI and the SOE of the thermal storage tank were discussed separately. Different load reduction rates were set for the photovoltaic and solar thermal power generation subsystems under different DNI and SOE conditions, that is, different active reserve capacities were reserved. An electric heating device was added to the combined power generation system, and the abandoned solar energy caused by the load reduction of the photovoltaic power generation subsystem was converted into thermal energy in the thermal storage tank of the solar thermal power generation subsystem through electric heating, thereby increasing the frequency regulation flexibility and frequency regulation capability of the combined power generation system and promoting the absorption of solar energy. A frequency response model of the photovoltaic-solar thermal power generation system was constructed, and a state-space model of the system was derived. The state-space model was iteratively calculated to obtain a model prediction sequence for a period of time in the future.

[0073] Furthermore, a two-tier control architecture was designed based on the frequency response model. The upper layer uses an MPC control strategy to allocate the output of the PV and CSP subsystems based on the AGC requirements. This strategy also dynamically adjusts the load shedding ratios of the two subsystems to meet varying demands, taking into account frequency regulation costs. The upper layer optimization objective function considers the system frequency deviation and the difference in active output between the PV and CSP subsystems, assigning different weights to minimize system frequency fluctuations.

[0074] Based on the output command from the upper layer, the lower layer selects the PV subsystem mode, operating it in MPPT or droop control mode. Simultaneously, the MPC controller coordinates the turbine control and steam generation subsystems within the CSP subsystem, controlling the turbine throttle valve opening and the molten salt flow valve opening as controlled variables. The lower-layer MPC optimization objective—minimizing the difference between the CSP system's active power output, internal main steam pressure, and the heat storage tank's SOE and reference values—requires adjustment of the MPC controller's target weights and control parameters at different frequency modulation stages.

[0075] When applying the MPC controller for solving, the first variable of the obtained optimal control sequence is taken to act on the control target. Through iteration and feedback correction, the optimization window is moved forward at the next sampling time to solve again, realizing rolling optimization to obtain the final solution.

[0076] In this embodiment, the formulation of a two-dimensional dynamic load reduction strategy for a photovoltaic-solar thermal combined power generation system includes:

[0077] When the AGC instruction is positive and higher than 5% of the total system capacity, DNI>1000W / m 2 , set PV and CSP to retain 10% active reserve capacity; when the AGC instruction is positive and less than 5% of the total system capacity, DNI>900W / m 2, set PV and CSP to retain 5% active reserve capacity respectively, and give priority to PV to increase output to the maximum power point tracking MPPT mode; when the AGC instruction is negative, limit the CSP output to no less than its lower limit, PV switches to droop control mode to reduce output, and maintain the SOE of the heat storage tank in the range of [0.1, 0.9] through EH.

[0078] In practice, since PV primarily converts sunlight to electricity while CSP converts sunlight to heat, their solar energy utilization mechanisms differ. The global horizontal irradiance (GHI) has a greater impact on PV, while the direct normal irradiance (DNI) has a greater impact on CSP. To facilitate the development of control strategies, the GHI is converted into the corresponding DNI value for analysis.

[0079] According to the positive and negative and high and low values of the system AGC instructions, the DNI is divided into zones and discussed, and the output and active reserve capacity of PV and CSP are allocated in combination with the SOE of the heat storage tank.

[0080] When the system AGC command is positive and exceeds 5% of the total system capacity, it indicates a rapid increase in grid load. To ensure that the combined power generation system can cope with the frequency fluctuations caused by the sudden load increase, sufficient active reserve capacity is required, and high requirements are placed on solar intensity. In this case, both PV and CSP maintain 10% active reserve capacity, and a DNI > 1000 is required to ensure that the combined power generation system can provide sufficient support for grid frequency. At the same time, if the SOE of the thermal storage tank is within the range of [0.1, 0.9], the electric heating device is operating normally, converting the abandoned solar energy generated by PV load shedding into heat energy stored in the CSP thermal storage tank.

[0081] When the system AGC command is positive but less than 5% of the total system capacity, the grid load increase is minimal. To mitigate grid frequency fluctuations, both PV and CSP require a certain amount of active reserve capacity. However, considering the cost of curtailment, a 5% active reserve capacity is set for both PV and CSP. In this case, a DNI > 900 indicates that the combined power generation system has sufficient regulation capacity to cope with load fluctuations. Because CSP is equipped with a heat storage tank and has greater regulation capability than PV, PV power is first increased to operate in MPPT mode, followed by CSP power to make up the remaining power difference. Furthermore, as long as the SOE of the heat storage tank remains within the normal operating range, the electric heating system continues to operate normally, converting the curtailed solar power generated by PV load shedding into heat energy that is stored in the CSP heat storage tank.

[0082] When the system AGC instruction is negative, the combined power generation system needs to reduce its output. In this case, it is necessary to ensure that the CSP output is always higher than its output lower limit to ensure that it does not exit operation. Under this premise, the PV adopts droop control to reduce the output. At the same time, if the SOE is still within the normal range, the electric heating device keeps running until the EH stops running when the SOE reaches the upper limit.

[0083] Furthermore, the operation of the electric heating device (EH) satisfies the following conditions:

[0084] When PV is abandoned and the SOE of the thermal storage tank is less than 0.9, the EH is started for energy conversion. The EH converts light energy into heat energy as shown in the following formula:

[0085] Q HOT =Q LIGHT η

[0086] Among them, Q HOT is the heat energy converted into the heat storage tank, Q LIGHT is the light energy absorbed by the electric heating device, and η is the energy conversion efficiency;

[0087] EH was stopped when SOE ≥ 0.9 and restarted when SOE < 0.8.

[0088] That is, when PV power is curtailed and the CSP thermal storage tank SOE is less than 0.9, the EH is activated to convert the curtailed solar energy into heat, making the CSP output more sustainable. When SOE exceeds 0.9, the EH is stopped to ensure normal CSP operation. To prevent frequent EH starts and stops at critical points, the EH is set to restart only when SOE is less than 0.8.

[0089] In this embodiment, generating a model prediction sequence in the future prediction time domain includes the following steps:

[0090] Assume that the number of photovoltaic power generation systems in the region is n, and the number of solar thermal power generation systems is m. The overall capacity configuration ratio of photovoltaic power generation systems and solar thermal power generation systems in the combined system is 5:1.

[0091] A state space model including the dynamic characteristics of the photovoltaic-thermal combined power generation system is constructed, which is specifically expressed as follows:

[0092]

[0093] Where x is the state variable of the regional system, including system frequency deviation, CSP main steam pressure and heat storage tank SOE; is the time derivative of the regional system state variable; u is the control variable of the regional system, including the photovoltaic inverter power command and the CSP turbine throttle opening; y is the measurement value of the regional system; A is the state matrix of the combined power generation system, B is the control input matrix, C is the output matrix, and D is the direct transfer matrix.

[0094] Specifically, x=[ΔP PV,1 ,…ΔP PV,n ,ΔP CSP,1 ,…,ΔP CSP,m ,ΔP L ,Δf] T , ΔP PV，1 ,…ΔP PV，n is the difference in active power output from the prime movers of the first to n photovoltaic generator sets, ΔP CSP，1 ,…,ΔP CSP，m is the difference in active power output from the prime movers of the 1st to mth CSP generator sets, ΔP L is the load variation in the control area, Δf is the frequency deviation in the control area, and the superscript T indicates transposition; u=[u PV1 ,…,u PVn ,u CSP1 ,…,u CSPm ,…,u SOE1 ,…,u SOEm ,u L ] T ,u PV1 ,…,u PVn represents the control variable of the MPC controller for photovoltaic generator sets 1 to n, u CSP1 ,…,u CSPm represents the control variable of the MPC controller for the CSP generator sets 1 to m; u SOE1 ,…,u SOEm Indicates the initial state of the heat storage tank of the 1st to mth CSP generator sets, u L represents the load disturbance of the regional system; y=[Δf,ΔP PV，1 ,…,ΔP PV，n , ΔP CSP，1 ,…,ΔP CSP，m ] T .

[0095] Furthermore, the four matrices A, B, C, and D can be derived from the following differential equations that express the dynamic characteristics of the system:

[0096]

[0097] Where: K1 and K2 are the correlation coefficients of the internal links of the solar thermal power generation system; u CV 、u salt and uPV is the MPC controller instruction; H is the system inertia coefficient, D is the system load damping coefficient, which represents the degree of change of load power when the frequency changes; Δx1, Δx2, Δp D ,Δp T , Δx3 are intermediate state variables of the CSP system; T1 is the time constant of the CSP system speed regulator; T evap is the time constant of the evaporator of the CSP system; T CH is the steam capacity time constant; T RH is the reheating time constant; K RH is the reheat gain coefficient; P T,rated K is the rated value of the main steam pressure inside the CSP system; lass is the pressure loss coefficient of the superheater; C D is the heat storage time constant of the drum; C SH is the volume time constant of the superheater; ΔP CSP and ΔP PV are the output power changes of the solar thermal and photovoltaic power generation systems respectively; T PV Indicates the time constant of the photovoltaic power generation system grid-connected inverter action; R PV is the regulation coefficient of the photovoltaic power generation system; SOE ref is the initial state of the heat storage tank; E TES is the capacity of the heat storage tank; K lass is the pressure loss coefficient of the superheater.

[0098] During operation, the following constraints must also be observed:

[0099]

[0100] Where: Δf min , Δf max are the upper and lower limits of the frequency variation for the system to operate safely and stably; v CSP.min 、v CSP.max They are the upper and lower speed limits of the speed regulator of the solar thermal power generation system; are the upper and lower limits of the heat storage tank capacity of the CSP system; P t TES，c 、P t TES，f are the charging and discharging power of the heat storage tank respectively; is the maximum charging and discharging power of the heat storage link; P CSP.min 、P CSP.max are the upper and lower limits of the active output of the CSP system respectively; P PV.min 、P PV.max They are the upper and lower limits of the active output of the photovoltaic power generation system respectively.

[0101] Finally, the state matrix A and control input matrix B can be derived:

[0102]

[0103] Where:

[0104]

[0105] B=[B PV.1 ,...,B PV.n , B CSP.1 ,...,B CSP.m ,B SOE , 1];

[0106] Where:

[0107]

[0108] The forward Euler method is used to discretize the matrix A and the matrix B, which can be expressed as follows:

[0109]

[0110] Among them, Δx(k)=x(k)-x(k-1), which represents the increment of the system state variable at time k

[0111] Predict the system state with m control steps and p prediction steps, which can be expressed as:

[0112]

[0113] Among them, y p (k+p|k) represents the output at time k+p predicted by the current time k;

[0114] The calculation method of the system output sequence of the next p steps is specifically expressed as:

[0115] Y p (k) = S x (x)Δx(k)+τy(k)+S u u(k);

[0116]

[0117] Among them, I nc×nc is the identity matrix with the same dimensions as the matrix C.

[0118] The objective function used in quadratic programming optimization when calculating the optimal control sequence of the regional system at the current moment based on the model prediction sequence is specifically expressed as:

[0119] J(z k )=Jy (z k )+J u (z k );

[0120]

[0121] Among them, Z k is the result obtained by solving the quadratic programming objective function, ε k is the slack variable at the control step k; n y is the number of system output variables; y j (k+i|k) is the predicted value of the j-th output variable at time k+i; r j (k+i|k) is the control target of the j-th output variable at time k; S j is the scale factor of the jth output variable, which is determined by the upper and lower limits of the change of the variable during the control process; n u is the number of system input control variables; u j (k+i|k) is the size of the j-th input control variable at time k; u j，target (k+i|k) is the target of the j-th input control variable at time k;

[0122] The constraints for optimizing the objective function are:

[0123]

[0124] Among them, y j，min (i) and y j，max (i) is y j The upper and lower bounds of (k+i|k), and u j，max (i) is u j The upper and lower bounds of (k+i-1|k).

[0125] In this embodiment, a two-layer control is implemented based on the above model. The upper control layer uses the MPC control strategy to allocate the output of the photovoltaic and solar thermal power generation systems according to the size of the AGC demand, and takes into account the frequency regulation cost so that the load reduction rate of the two changes dynamically under different demands.

[0126] When constructing the objective function, the system frequency, the output of the photovoltaic power generation system and the solar thermal power generation system are taken as the upper-level optimization control targets, and the MPC target weights and control parameters are set according to the specific situation at different stages of frequency modulation. The objective function of the upper-level controller is specifically expressed as follows:

[0127]

[0128] Where (k+j|k) represents the predicted value at time k+j based on the state quantity at time k within the prediction time domain; a1, a2, and a3 are weighting coefficients, representing the optimization proportions of frequency fluctuation, photovoltaic power generation power, and solar thermal power generation power during the frequency modulation process, respectively. The larger the weighting coefficient, the higher the priority of the item; N is the prediction time domain;

[0129] In the early stages of frequency regulation, system frequency fluctuations are expected to be large, so suppressing frequency fluctuations should be the primary control objective. In the later stages of frequency regulation, system frequency fluctuations gradually decrease, which can increase the priority of power increment control, ensuring frequency regulation while reducing the output increase. Furthermore, the output of CSP is directly affected by the SOE of its heat storage tank. If the SOE is low, the CSP output needs to be adjusted promptly to prevent it from being shut down. If the SOE is high, a higher frequency regulation ratio should be allocated to CSP.

[0130] Therefore, adaptive optimization adjustments are performed when setting the weight coefficients a1, a2, and a3 of the MPC controller to achieve a more accurate frequency modulation effect. The specific optimization method is as follows:

[0131] a=a0+k a (e |Δf|+sign(Δf)·dΔf / dt -1);

[0132] Among them, a0 is the initial value of the weighting coefficient; k a is the weighted adjustment coefficient; is the system frequency difference change rate;

[0133] When designing the adaptive weighting coefficient a, the main consideration is that the main optimization objectives in the early and late stages of frequency regulation should be different. In the early stage of frequency regulation, the frequency deviation should be minimized to the greatest extent, while in the late stage of frequency regulation, the active output of the combined power generation system needs to make up for the system power difference as quickly as possible.

[0134] The weight coefficient of the active output increment of the CSP system needs to be adaptively adjusted according to the system frequency fluctuation direction and the SOE of the heat storage tank. The formula for the correlation between CSP weight and SOE is specifically expressed as follows:

[0135]

[0136] Among them, b0 is the initial value of the weighting coefficient; k b is the weighted adjustment coefficient; SOE is the state quantity of the heat storage tank of the CSP system;

[0137] When the frequency drops, if the thermal storage tank SOE is greater than 0.5, it indicates that the CSP has sufficient frequency regulation capability and can generate a large amount of additional power, so a smaller weight coefficient can be set. If the thermal storage tank SOE is less than 0.5, it indicates that the CSP has limited frequency regulation capability, so a larger weight coefficient can be set to prevent the SOE from dropping too quickly and causing the CSP to stop operating. The same principle applies when the frequency rises. It can be seen that the adaptive weight coefficient of this embodiment can be flexibly adjusted based on frequency fluctuations and the thermal storage tank SOE, while also taking into account the CSP's active reserve capacity and its frequency regulation capability.

[0138] The PV weight compensation formula is expressed as:

[0139] c=c0+b0-b;

[0140] Among them, c0 is the initial value of the weighting coefficient

[0141] In summary, the weights of the CSP and PV active output increments satisfy b + c = b0 + c0, ensuring that the primary optimization objective can flexibly and stably switch between frequency fluctuations and the combined power generation system output increment. Setting different adaptive weight coefficients b and c also ensures adaptive distribution of CSP and PV output.

[0142] Furthermore, in this embodiment, the photovoltaic power generation system is selected to operate in the MPPT mode or the droop control mode. At the same time, the MPC controller is applied to coordinate the control of the turbine control subsystem and the steam generation subsystem within the CSP system. On this basis, the MPC control parameters and target weights are determined using the opening of the turbine throttle valve and the opening of the molten salt flow valve of the CSP system as the main controlled variables. The objective function of the lower-level controller is specifically expressed as follows:

[0143]

[0144] Among them, M1, M2, and M3 are all weighting coefficients.

[0145] The primary objective of the lower-level MPC controller is to coordinate the actions of the CSP system's internal subsystems based on the CSP output commands issued by the upper-level MPC controller, ensuring that the system's output is as close to the reference value as possible while minimizing fluctuations in the main steam pressure and maintaining the SOE around 0.5. Setting weighted coefficients for different objectives at different frequency modulation stages allows for more flexible and rapid control, while ensuring safe and stable system operation and balancing control cost and effectiveness.

[0146] The specific process of rolling optimization is: solving the objective function J(Z k ), and the optimal control sequence Z that takes into account both the objectives and constraints at the current moment is obtained k, and add the first variable into the system as the control input variable at the next moment. Through iteration and feedback correction, the newly obtained system state x(k) and system output y(k) are used to perform rolling optimization solution.

[0147] The constraints of rolling optimization include: frequency deviation limit: |Δf|≤0.5Hz; upper and lower limits of solar thermal output: P CSP,min ≤P CSP ≤P CSP,max ; Heat storage tank charging and discharging power limit: |Q 储热 ∣≤Q max .

[0148] Those skilled in the art will appreciate that the present invention is not limited to the foregoing embodiments. The foregoing embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A frequency control method for a photovoltaic-thermal combined power generation system, characterized in that: The following steps are involved: Develop a two-dimensional dynamic load reduction strategy for the photovoltaic-solar thermal power generation system. Based on the positive and negative values of the automatic generation control (AGC) instructions, the direct solar irradiance (DNI), and the state of energy (SOE) of the thermal storage tank, the load reduction rate of the photovoltaic (PV) and solar thermal power generation (CSP) systems is dynamically adjusted to reserve active reserve capacity. An electric heater (EH) is configured in the combined power generation system to convert the abandoned solar energy generated by PV load shedding into thermal energy through electric heating and store it in the CSP thermal storage tank; Establish a frequency response model for the photovoltaic-thermal combined power generation system, derive the system's state space model, and generate a model prediction sequence within the future prediction time domain; A two-layer model predictive control architecture is designed. The upper-layer controller dynamically allocates PV and CSP output commands based on frequency deviation and AGC requirements, while the lower-layer controller coordinates the control of the CSP's internal subsystems. Based on the rolling optimization mechanism, the optimal control sequence is solved in each control cycle, and the first control variable is applied to the system to achieve frequency stability control through iterative feedback.

2. The frequency control method of the photovoltaic-thermal combined power generation system according to claim 1, characterized in that: The formulation of a two-dimensional dynamic load reduction strategy for a photovoltaic-solar thermal combined power generation system includes: When the AGC instruction is positive and higher than 5% of the total system capacity, DNI>1000W / m 2 , set PV and CSP to retain 10% active reserve capacity; When the AGC instruction is positive and less than 5% of the total system capacity, DNI>900W / m 2 , set PV and CSP to retain 5% active reserve capacity each, and give priority to PV to increase output to the maximum power point tracking MPPT mode; When the AGC command is negative, the CSP output is limited to no less than its lower limit, the PV switches to droop control mode to reduce the output, and the SOE of the heat storage tank is maintained in the range of [0.1, 0.9] through the EH.

3. The frequency control method of the photovoltaic-thermal combined power generation system according to claim 1, characterized in that: The operation of the electric heating device (EH) meets the following conditions: When PV is abandoned and the SOE of the thermal storage tank is less than 0.9, the EH is started for energy conversion. The EH converts light energy into heat energy as shown in the following formula: Q HOT =Q LIGHT the Among them, Q HOT is the heat energy converted into the heat storage tank, Q LIGHT is the light energy absorbed by the electric heating device, and η is the energy conversion efficiency; EH was stopped when SOE ≥ 0.9 and restarted when SOE < 0.

8.

4. The frequency control method of the photovoltaic-thermal combined power generation system according to claim 1, characterized in that: A state space model including the dynamic characteristics of the photovoltaic-thermal combined power generation system is constructed, which is specifically expressed as follows: Where x is the state variable of the regional system, including system frequency deviation, CSP main steam pressure and heat storage tank SOE; is the time derivative of the regional system state variable; u is the control variable of the regional system, including the photovoltaic inverter power command and the CSP turbine throttle opening; y is the measurement value of the regional system; A is the state matrix of the combined power generation system, B is the control input matrix, C is the output matrix, and D is the direct transfer matrix.

5. The frequency control method of the photovoltaic-thermal combined power generation system according to claim 4, characterized in that: The forward Euler method is used to discretize the matrix A and the matrix B, which can be expressed as follows: Among them, Δx(k)=x(k)-x(k-1) represents the increment of the system state variable at time k.

6. The frequency control method of the photovoltaic-thermal combined power generation system according to claim 5, characterized in that: Predict the system state with m control steps and p prediction steps, which can be expressed as: Among them, y p (k+p|k) represents the output at time k+p predicted by the current time k; The calculation method of the system output sequence of the next p steps is specifically expressed as: Y p (k)=S x (x)Δx(k)+τy(k)+S u u(k); Among them, I nc×nc is the identity matrix with the same dimensions as the matrix C.

7. The frequency control method of the photovoltaic-thermal combined power generation system according to claim 6, characterized in that: The objective function used in quadratic programming optimization when calculating the optimal control sequence of the regional system at the current moment based on the model prediction sequence is specifically expressed as: J(z k )=J y (from k )+J u (from k ); Among them, Z k is the result obtained by solving the quadratic programming objective function, ε k is the slack variable at the control step k; n y is the number of system output variables; y j (k+i|k) is the predicted value of the j-th output variable at time k+i; r j (k+i|k) is the control target of the j-th output variable at time k; S j is the scale factor of the jth output variable, which is determined by the upper and lower limits of the variable during the control process; n u is the number of system input control variables; u j (k+i|k) is the size of the j-th input control variable at time k; u j，target (k+i|k) is the target of the j-th input control variable at time k; The constraints for optimizing the objective function are: Among them, y j,min (i) and y j，max (i) is y j The upper and lower bounds of (k+i|k), and u j，max (i) is u j The upper and lower bounds of (k+i-1|k).

8. The frequency control method of the photovoltaic-thermal combined power generation system according to claim 1, characterized in that: The objective function of the upper controller is specifically expressed as: Where (k+j|k) represents the predicted value at time k+j based on the state quantity at time k within the prediction time domain; a1, a2, and a3 are all weighting coefficients, representing the optimized proportions of frequency fluctuation, photovoltaic power generation power, and solar thermal power generation power during the frequency modulation process, respectively; N is the prediction time domain; The weight coefficients a1, a2, and a3 are adaptively optimized and adjusted. The specific optimization method is as follows: a=a0+k a (and |Δf|+sign(Δf)·dΔf / dt -1); Among them, a0 is the initial value of the weighting coefficient; k a is the weighted adjustment coefficient; is the system frequency difference change rate; The formula for associating CSP weight with SOE is as follows: Among them, b0 is the initial value of the weighting coefficient; k b is the weighted adjustment coefficient; SOE is the state quantity of the heat storage tank of the CSP system; The PV weight compensation formula is expressed as: c=c0+b0-b; Among them, c0 is the initial value of the weighting coefficient.

9. The frequency control method of the photovoltaic-thermal combined power generation system according to claim 1, characterized in that: The objective function of the lower-level controller is specifically expressed as: min J=M1||P CSP -P CSP.ref || 2 +M2‖p T -1|| 2 +M3||SOE-0.5|| 2 ; Among them, M1, M2, and M3 are all weighting coefficients.

10. The frequency control method of the photovoltaic-thermal combined power generation system according to claim 1, characterized in that: The constraints for rolling optimization include: Frequency deviation limit: |Δf|≤0.5Hz; Solar thermal output upper and lower limits: P CSP,min ≤P CSP ≤P CSP,max ; Heat storage tank charging and discharging power limit: |Q 储热 |≤Q max .

Images