A frequency support control method for coordinating shaguo desert base and multi-terminal flexible direct current system considering randomness of photovoltaic power generation

Abstract

This invention discloses a collaborative frequency support control method for a photovoltaic (PV) power generation base in a desert area and a multi-terminal flexible DC system, belonging to the field of power system frequency stability and control technology. First, system operating parameters are obtained, and the PV cluster is divided into maximum power point tracking (MPPT) and constant active power operation modes. Second, an active reconfiguration mechanism is introduced at the receiving end to break control clamping, and a continuous-time stochastic differential equation considering the stochasticity of PV output is established. Then, combined with the controller's sampling characteristics, this is transformed into a discrete-time stochastic difference equation. Next, the steady-state covariance matrix is ​​solved based on the discrete algebraic Lyapunov equation, and the theoretical envelope lower limit of the receiving-end frequency under random disturbances is quantified using a criterion. Finally, with the theoretical envelope not exceeding the limit as a safety constraint, a multi-objective collaborative optimization model is constructed and solved to jointly optimize the proportion of heterogeneous PV modes and the control parameters of each converter station. This achieves safe and economically optimized operation of the system under uncertainty.

Description

Technical Field

[0001] This invention relates to the field of power system frequency stabilization and control technology, and in particular to a frequency support control method for a multi-terminal flexible DC system in a desert base that takes into account the randomness of photovoltaic power generation. Background Technology

[0002] With the acceleration of the global energy transition, large-scale wind, solar, and hydropower renewable energy bases transmitting power across regions via multi-terminal flexible direct current (VSC-HVDC) technology has become a core feature of modern power grids. Although introducing additional frequency control measures at the converter station level can provide active power frequency support to the receiving-end grid in terms of physical mechanism, the continuous increase in renewable energy penetration presents the system with more severe uncertainties and challenges.

[0003] First, photovoltaic power output exhibits strong randomness and volatility. This uncertainty is coupled to the entire grid through the flexible DC network, causing the receiving-end frequency to no longer follow a definite evolutionary trajectory, but rather to exhibit random fluctuations with statistical characteristics. Existing deterministic analysis methods often cannot accurately quantify the frequency shift risk caused by this random fluctuation, leading to a blurring of the system's frequency security boundary.

[0004] Secondly, at the source-grid coordination level, there is a significant trade-off between security and economy: in order to ensure frequency security to the greatest extent, photovoltaic clusters often need to reduce their load significantly to reserve sufficient active power reserves, but this will bring serious economic penalties and reduce the renewable energy consumption rate; conversely, if the ultimate consumption level is pursued, the system is very likely to trigger low-frequency load reduction when faced with sudden disturbances due to insufficient reserves or random fluctuations.

[0005] Finally, multi-terminal flexible DC transmission systems involve the coordinated control of hydropower and photovoltaic heterogeneous resources at the sending end, as well as multiple converter stations. Complex nonlinear coupling relationships exist between the control parameters at each end (such as droop coefficients and virtual inertia coefficients) and the operating points on the source side. Traditional offline parameter tuning methods struggle to account for the dynamic safety constraints under photovoltaic stochasticity, failing to achieve optimal allocation of frequency regulation resources across the entire network in both space and time. Therefore, proposing a method capable of analytically quantifying stochastic risk boundaries and achieving multi-objective closed-loop optimization in source-grid coordination is crucial for improving the safe and economical operation of multi-terminal flexible DC transmission systems. Summary of the Invention

[0006] To address the random fluctuation risks and grid-source coordination challenges faced by multi-terminal flexible DC transmission systems with high photovoltaic (PV) grid integration, this invention provides a coordinated frequency support control method for desert and Gobi areas that considers the randomness of PV power generation. It aims to solve the technical problem that existing deterministic analysis methods cannot accurately quantify the frequency offset risks caused by random PV output, and the difficulty in balancing ensuring receiving-end frequency security with improving renewable energy absorption rates. This invention achieves safe and economical system operation under extreme uncertainty by constructing an analytical safety boundary and implementing multi-objective collaborative optimization.

[0007] To achieve the above objectives, this invention provides a method for coordinated frequency support control of a desert-based photovoltaic power generation system and a multi-terminal flexible DC system, taking into account the randomness of photovoltaic power generation. The method includes the following steps: S1. Real-time acquisition of operating parameters of multi-terminal flexible DC transmission system, dividing the sending-end photovoltaic cluster into maximum power point tracking operation mode cluster and constant active power operation mode cluster; and acquisition of operating status parameters of hydropower base; S2. When a frequency drop occurs at the receiving end, frequency supplementary control is introduced at the receiving end active power station to provide initial support; simultaneously, an active reconfiguration mechanism is introduced at the receiving end DC voltage station to lower the DC voltage reference value to eliminate control conflicts between parallel converter stations; and the DC voltage drop caused by active reconfiguration is transmitted to the sending end network as a physical signal without communication delay. S3. After the sending-end converter station senses the voltage drop, it changes the active power. Based on the resulting frequency drop at the sending end, the sending-end hydropower base and the photovoltaic cluster in constant active power mode start additional control to coordinate and increase active power generation. At the same time, the randomness of the photovoltaic output in the maximum power point tracking mode is characterized as a continuous-time stochastic differential equation driven by Gaussian white noise. Combined with the controller sampling period, it is discretized and transformed into a discrete-time stochastic difference equation to adapt to digital calculation and analysis. S4. Based on the discrete-time stochastic difference equation, the steady-state covariance matrix of the system state variables is solved using the discrete Lyapunov equation. The theoretical envelope boundary of the receiving end frequency under random disturbance is quantitatively calculated, and the minimum value of the envelope is extracted as the lowest point of the system frequency operation. S5. Construct a collaborative optimization model aimed at improving the minimum operating frequency of the system and reducing curtailment; jointly optimize the proportion of photovoltaic heterogeneous mode operation and the control coefficients of each converter station, and issue optimal control commands.

[0008] Preferably, in S1, the multi-terminal flexible DC transmission system has a four-terminal network topology, with the sending end including two converter stations respectively connected to the new energy base, and the receiving end including a constant active power converter station and a constant DC voltage converter station.

[0009] Preferably, in S1, the four-terminal network topology includes four converter stations, specifically including: Receiving-end constant active power converter station: Monitoring receiving-end frequency deviation The increment of its active power command output to the receiving-end power grid is: ; In the formula, and These are the primary frequency regulation and virtual inertia control coefficients, respectively; their action causes an energy imbalance on the DC side, resulting in a passive drop in DC voltage. Receiving-end constant DC voltage (Vdc) converter station: To eliminate the control clamping effect with the active power station, an active reconfiguration mechanism is adopted. The dynamic voltage command value issued in real time is: ; In the formula, Rated voltage, It is the droop factor; by actively following the frequency drop through commands, it suppresses the reverse power transmission to the DC network, thus releasing support space; Sending-end hydroelectric converter station: Sensing DC voltage signals are transmitted along the DC line without communication delay. When the local DC voltage is detected... When the value deviates from the rated value and exceeds the dead zone, DC voltage supplementary control is triggered, increasing the active power command increment injected into the DC network by: ; In the formula, This is the DC voltage-frequency droop factor. This is the virtual inertia coefficient.

[0010] Sending-end V / F converter station: V / F grid control is adopted to establish a connection between DC voltage and AC frequency on the photovoltaic side. The angular velocity deviation value is obtained by the DC voltage deviation and frequency droop coefficient. The operating frequency of the AC bus is actively changed to convert the DC side signal into an AC side frequency signal and transmit it to the photovoltaic cluster.

[0011] Preferably, in step S3, the additional control-based coordinated generation of active power by the sending-end hydropower base and the photovoltaic cluster in constant active power mode specifically includes: Dynamic modeling of the sending-end hydropower cluster: considering the turbines and their water intake system, relays, and governors using proportional-integral control; utilizing the inertial time constant of the water flow. Relay response time constant Proportional gain Integral gain and permanent state difference coefficient Characterizing the power output variation of the hydroelectric speed controller Its frequency dynamic response after disturbance is represented by the swing equation: ; In the formula, The equivalent inertial time constant of the hydropower cluster, This is the damping coefficient of the water turbine. This represents the change in active power demand at the converter station. This refers to the frequency deviation of the hydroelectric busbar; Photovoltaic cluster in constant active power mode: After detecting a frequency drop on the AC bus of converter station No. 3, primary frequency regulation and virtual inertia control are triggered, and the frequency support active power command increment is: ; In the formula: The active power command increment is provided to support the response frequency of the photovoltaic cluster in the fixed active power mode. This represents the primary frequency regulation droop coefficient of the photovoltaic cluster. Frequency deviation of the AC bus of the photovoltaic cluster; This refers to the virtual inertia control coefficient of the photovoltaic cluster; This represents the frequency deviation rate of the AC bus of the photovoltaic cluster.

[0012] Preferably, in step S3, the randomness of the photovoltaic output in the maximum power point tracking mode is characterized as a continuous-time stochastic differential equation driven by Gaussian white noise, and then discretized into a discrete-time stochastic difference equation in conjunction with the controller sampling period. Specifically, this includes: Stochastic Modeling of Photovoltaic Clusters: The stochastic fluctuations in the output of a maximum power point tracking (MPPT) photovoltaic cluster that maintains maximum power operation are characterized as follows: in, The stochastic differential change in the active power output of the photovoltaic cluster. This represents the percentage of photovoltaic clusters operating in maximum power point tracking mode. The intensity of random fluctuations in photovoltaic power output. For the differential terms of the standard Wiener process; By coupling the above-mentioned hydropower speed regulation dynamics and flexible DC control dynamics with the photovoltaic fluctuation model, the continuous-time stochastic differential equations of the system are constructed: ; Combined with the fixed sampling period of the system physical controller The continuous-time stochastic differential equation is transformed into a discrete-time stochastic difference equation using the difference method: ; In the formula, This includes sampling period information and hydropower response time constant. The discretized system state transition matrix of the converter station additional control coefficients.

[0013] Preferably, step S4 specifically includes: Based on the obtained discrete-time stochastic difference equation, the discrete algebraic Lyapunov equation satisfied by the steady-state covariance matrix of the system is constructed as follows: ; By analytically solving the Lyapunov equation, the steady-state variance corresponding to the AC frequency deviation at the receiving end is extracted. Using Gaussian distribution The criteria, combined with the deterministic mean trajectory of the dynamic characteristics evolution of hydropower reversal regulation, are used to determine the underlying principles. The theoretical lower bound of the receiving-end frequency deviation at the corresponding confidence level was quantitatively calculated: .

[0014] Preferably, in step S5, the collaborative optimization model and the issuance of optimal control commands specifically include: S51. Objective Function Calculation: Construct a comprehensive objective function that considers both the system's disturbance rejection security and the economic viability of photovoltaic backup. ; In the formula, This indicates the comprehensive curtailment penalty for photovoltaic power generation when it exits maximum power point tracking mode and switches to constant active power mode, with reserves reserved for backup. This is a penalty term for the lowest point deviation of the frequency envelope obtained by quantization; This is a penalty item for frequency deviation of the hydropower supply at the sending end; Weighting coefficients for multi-objective optimization; S52, Parameter optimization: While satisfying system power balance constraints and claim 4... Under dynamic safety boundary constraints, an optimization algorithm is used to solve the objective function. The global optimal solution was calculated, which included the best photovoltaic configuration ratio. Optimal frequency regulation coefficient set for converter stations The decision vector is included; and based on this vector, the expected frequency minimum point increase and curtailment rate of the system under extreme disturbances are calculated. S53. Control command issuance: The decision vector is converted into the underlying physical control command and issued in real time to the controllers of each converter station and the source-side power station coordination control unit of the four-terminal flexible DC system through the communication network. S54, Closed-loop collaborative execution: The photovoltaic base at the sending end... Adjust the actual active power output point to release support margin; each converter station controller adjusts the received data... Real-time correction of the primary frequency modulation droop coefficient at the receiving end Receiving end active reconstruction coefficient and the voltage-frequency co-droop coefficient at the sending end It achieves closed-loop support for the frequency deficit at the receiving end through AC / DC energy linkage.

[0015] Preferably, in S51, the analytical expression of the photovoltaic reserve penalty term in the comprehensive objective function is: ; in, The unit curtailment penalty cost coefficient, This is the predicted maximum power generation capacity of the photovoltaic cluster.

[0016] The advantages and beneficial effects of this invention compared to the prior art are: 1. This invention constructs and discretizes a continuous-time stochastic differential equation that takes into account the randomness of photovoltaic power output, analyzes the steady-state covariance matrix based on the discrete algebraic Lyapunov equation, and uses the 3σ criterion to accurately quantify the theoretical lower bound of the receiving-end frequency under random disturbances. This solves the technical problem that existing deterministic analysis methods cannot accurately assess the frequency offset risk under high-proportion photovoltaic access, and provides an analytical theoretical basis for the dynamic frequency safety boundary of the system.

[0017] 2. This invention introduces an active reconfiguration mechanism at the receiving-end constant DC voltage converter station to lower the DC voltage reference value to break the control clamp between the constant active power converter station and the receiving-end constant DC voltage converter station, suppress reverse power extraction and release frequency modulation support space; at the same time, it uses the DC voltage drop as a physical signal without communication delay to be transmitted to the sending end, triggering the hydropower base and the constant PQ mode photovoltaic cluster to coordinate and increase active power generation, realizing the precise source-grid linkage and frequency deficit complementary support of the multi-end flexible DC system in the spatiotemporal dimension.

[0018] 3. This invention uses the theoretical envelope band not exceeding the limit as a hard safety constraint, constructs a comprehensive objective function that takes into account both improving the minimum operating frequency and reducing curtailment penalties, and jointly optimizes the proportion of photovoltaic MPPT and PQ heterogeneous modes as well as the additional control coefficients of each converter station. While ensuring the dynamic safety of the receiving end frequency under extreme disturbances, it significantly reduces curtailment costs and improves the renewable energy absorption rate, achieving the optimal synergy between system safety operation and economic benefits.

[0019] 4. This invention transforms the optimized decision vector into underlying physical control commands in real time and issues them for execution, enabling the photovoltaic base to adaptively adjust its active power output points to release frequency regulation margin. Each converter station dynamically corrects its frequency regulation coefficient and reconfiguration parameters. Through AC / DC energy linkage, a closed-loop collaborative support is formed for the frequency deficit at the receiving end, effectively improving the adaptive capability and long-term operational resilience of high-proportion new energy flexible DC systems under uncertain environments.

[0020] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0021] Figure 1This is an overall flowchart of a frequency support control method for a desert base and a multi-terminal flexible DC system that takes into account the randomness of photovoltaic power generation, according to an embodiment of the present invention. Figure 2 This is a network topology diagram of a multi-terminal flexible DC transmission system according to an embodiment of the present invention; Figure 3 This is a comparison diagram of the receiving-end AC power grid frequency response under different control strategies according to embodiments of the present invention; Figure 4 This is a comparison diagram of the frequency response of the power grid on the sending end of the hydropower station under different control strategies according to an embodiment of the present invention; Figure 5 This is a comparison chart of the cost of photovoltaic primary frequency regulation backup curtailment under different control strategies according to embodiments of the present invention. Detailed Implementation

[0022] In the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product is in use. They are used only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," and "connect" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0023] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0024] This embodiment uses, as follows: Figure 2 The diagram shows a four-terminal flexible DC transmission system topology. The system comprises two receiving ends and two sending ends: Receiving end converter station 1 uses constant active power (PQ) control, and receiving end converter station 2 uses constant DC voltage (Vdc) control; sending end converter station 3 (connected to the photovoltaic base) uses V / F grid control, and sending end converter station 4 (connected to the hydropower base) uses constant active power (PQ) control. Figure 1 As shown, the specific system response and underlying control timing of the source-grid collaborative optimization method of the present invention, which takes into account photovoltaic random fluctuations, include the following steps: Step S1: Global runtime status awareness and heterogeneous mode partitioning.

[0025] Real-time acquisition of AC bus frequencies at both the receiving and sending ends, DC-side operating voltages at each converter station, and AC / DC-side power flow. The sending-end photovoltaic base is divided into clusters operating in Maximum Power Point Tracking (MPPT) mode (accounting for a certain percentage). Clusters operating in constant active power (PQ) mode (accounting for a certain percentage) (and obtain the current maximum photovoltaic power generation capacity prediction value). .

[0026] Step S2: Receiving end control clamp breaking and AC / DC physical linkage.

[0027] When a sudden load surge occurs in the receiving-end AC power grid, causing a frequency drop that exceeds the frequency regulation dead zone, the receiving end initiates active coordinated support: (1) Additional frequency support for receiver-end converter station No. 1 (fixed PQ): An additional frequency control is used to establish a relationship between the receiving-end grid frequency and the active power reference value: ; In the formula, This is the steady-state reference value. This refers to primary frequency modulation and virtual inertia coefficient. This command alters the AC side characteristics; its d-axis current reference value and AC side output power are: ; Ignoring internal losses, the DC side of Station 1 dynamically satisfies: ; Due to AC side output power The addition broke the DC capacitor. The balance of charging and discharging will inevitably lead to a decrease in the DC voltage at the receiving end. A passive fall occurred.

[0028] (2) Active reconfiguration of receiving end converter station No. 2 (fixed Vdc) to break clamp: To prevent reverse power extraction from Substation No. 2 due to its fixed rated voltage, an active reconfiguration mechanism is introduced to update the DC voltage command value in real time. ; The AC-side d-axis characteristics and DC-side charging / discharging dynamics of Station 2 are as follows: ; ; Through active reconfiguration, the error term of the outer loop PI controller at station No. 2 was corrected, suppressing the back-feeding power to the DC network and freeing up system frequency modulation support space. The DC voltage drop signal was transmitted to the sending-end network without communication delay.

[0029] Step S3: Sending end heterogeneous collaborative response and multi-source dynamic modeling.

[0030] The sending-end converter station senses a DC voltage drop, triggering a heterogeneous collaborative response on the source side: (1) Coordinated response between the sending end converter station No. 4 (stationary PQ) and hydropower: Station No. 4 monitored the local DC voltage After the drop, the DC voltage additional control is triggered to extract active power: ; This active power extraction caused a decrease in the frequency of the sending-end hydroelectric bus. This leads to the hydropower units participating in frequency regulation. Considering the turbine's water intake system, servo motor, and PI speed governor, the hydropower speed regulation increment... The frequency response is characterized by the following continuous transfer function and the swing equation: ; ; In the formula, It is the inertial time constant of the water flow (characterizing the water hammer effect). is the time constant of the relay.

[0031] (2) Coordinated response between the sending-end converter station No. 3 (V / F) and the photovoltaic cluster: Station No. 3 does not directly extract power; instead, it converts the sensed DC voltage deviation into angular velocity deviation and actively lowers the AC bus frequency. The fixed-PQ mode photovoltaic cluster detected a frequency drop. Then, an active response is triggered: ; Meanwhile, for MPPT mode photovoltaics that maintain maximum power operation, the random fluctuations in output influenced by natural conditions are characterized by continuous-time stochastic differential equations driven by Gaussian white noise: ; In the formula, For fluctuation intensity, This represents the differential term of the standard Wiener process. Combining the above AC / DC dynamics, we construct a global continuous-time state-space model of the system: ; Step S4: Derivation of discretization of multi-source heterogeneous model based on finite difference method.

[0032] In order for a digital microprocessor to execute the cooperative optimization algorithm proposed in this invention, a fixed sampling period of the controller must be used. The first-order Euler finite difference method is used to discretize the global continuous system, which includes multi-terminal flexible direct current systems and heterogeneous dynamics on the source side. The specific multi-stage discrete difference dynamic model is as follows: (1) Receiving end grid frequency and unit discrete dynamics.

[0033] Receiving end AC grid frequency under load disturbance Support power of converter station No. 1 The discrete swing equation is as follows: ; The discrete active power dynamics of the receiving-end equivalent unit, considering the governor limiting stage and the reheat prime mover stage, are as follows: ; ; In the formula, The fixed sampling period for the system controller; subscript and These represent the k-th and (k+1)-th discrete sampling times, respectively. The equivalent inertial time constant of the receiving-end AC power grid; The damping load factor of the receiving-end power grid; Let k be the load surge disturbance at time k; The governor's response time constant; The time constant of the reheat prime mover; The unit adjustment power coefficient; This represents the upper and lower limit functions of the speed controller output.

[0034] (2) The AC / DC side and additional control discrete dynamics of receiving end station 1 (fixed PQ).

[0035] The differential model between the AC side d-axis current and the DC side capacitor charging and discharging of converter station No. 1 is as follows: ; ; The additional frequency control PI regulation and inertia element discretization of Station No. 1 are as follows: ; In the formula, and These are the equivalent resistance and equivalent inductance on the AC side of converter station No. 1, respectively. The synchronous angular velocity of the receiving-end AC power grid; and These represent the changes in current along the d-axis and q-axis of the AC side, respectively. and These represent the d-axis voltage changes at the AC-side converter bus and the grid-side, respectively. This is the equivalent capacitance on the DC side of Station 1; This is the rated voltage of the DC network; The active power flowing into the DC network; and These are the primary frequency regulation and virtual inertia control coefficients added to Station No. 1.

[0036] (3) The AC / DC side and active reconfiguration discrete dynamics of receiving end station No. 2 (fixed Vdc).

[0037] Similarly, the intrinsic dynamic discretization of the AC / DC side of Station 2 is as follows: ; ; Its active reconfiguration of additional control commands is discretized and updated as follows: ; In the formula, , , These are the equivalent resistance, inductance, and DC capacitance of converter station No. 2, respectively. The DC voltage reference command value after active reconstruction at time k+1; The active reconfiguration voltage-frequency droop control coefficient for station No. 2.

[0038] (4) Dynamic coordination between the sending end No. 4 station (fixed PQ) and the hydropower cluster.

[0039] Converter station No. 4 directly senses DC voltage dips, and its additional active power command discrete model is as follows: ; The discrete response equation for the frequency deviation of the equivalent hydropower system at the sending end is: ; Taking into account the dynamics of the relay (opening degree) The time constant of water hammer effect The active power dynamic discretization of the hydropower unit is as follows: ; ; In the formula, and These are the DC voltage droop and virtual inertia control coefficients for Station 4, respectively. The equivalent inertial time constant of the sending-end hydropower system; The damping coefficient of the hydroelectric system; This represents the change in the opening degree of the turbine servo connector. The relay response time constant; and These are the proportional and integral gains of the PI controller for the hydroelectric speed regulator; The inertial time constant of the water flow characterizes the water hammer effect; This represents the change in the mechanical power of the hydroelectric generator unit.

[0040] (5) The heterogeneous discrete dynamics of the sending end station 3 (V / F) and the photovoltaic cluster.

[0041] Station No. 3 uses V / F network control to convert DC voltage drops into AC frequency deviations, and its signal mapping discretization is as follows: ; The active power response of a fixed-PQ mode photovoltaic cluster is discretized based on the above frequency signals: ; The continuous random fluctuations of the MPPT mode photovoltaic cluster are discretized into difference equations that follow a normal distribution: ; In the formula, It is a standard Gaussian white noise sequence. The voltage-frequency conversion coefficient for the V / F grid control of converter station No. 3; and The active power random fluctuations output by the photovoltaic cluster in MPPT mode at time k+1 and time k are respectively. (6) Construction of global state equations and Lyapunov analytical quantization.

[0042] Extract the AC / DC current, voltage, node frequencies, and controller integrals from the above equations into a global state variable sequence. Solving all the above difference equations simultaneously, we obtain the global discrete-time stochastic difference model: ; In the formula, The discretized state transition Jacobian matrix has internal elements that precisely map all the underlying physical and control parameters mentioned above.

[0043] Based on this model, the steady-state covariance matrix of the system is constructed. The discrete algebraic Lyapunov equations that are satisfied are: ; The equation is solved analytically, and the main diagonal elements corresponding to the receiving-end frequency are extracted. Combined with deterministic evolutionary trajectories Quantify the heterogeneous fluctuations of water and light Lower limit of the receiving end frequency security envelope: ; Step S5: Multi-objective constraint optimization and collaborative execution.

[0044] Under the constraints of overall network power balance and source-side upper and lower amplitude limits, including the large-base-four-terminal flexible DC system, a collaborative optimization objective function is constructed: ; In the formula, the analytical expression for the photovoltaic reserve penalty term is: .

[0045] by As a hard constraint, the optimal photovoltaic configuration is obtained by using the Gurobi solver. and the set of optimal control coefficients for each converter station Optimization decisions are issued and executed via the communication network; photovoltaic power plants are phased out of MPPT and released for reserve in proportion to their capacity; and converter stations operate in an optimal manner. The underlying controller is updated to achieve frequency support closed loop.

[0046] The method of the present invention will be illustrated below through a specific example.

[0047] like Figure 2 The four-terminal flexible DC transmission system shown consists of two sending ends and two receiving ends. The sending ends are a large-scale photovoltaic (PV) renewable energy base and a hydropower base. The sending-end hydropower converter station uses constant active power control, while the sending-end PV converter station uses V / F control. Power is transmitted to the receiving-end grid via DC lines. The two receiving-end converter stations use constant active power control and constant DC voltage control, respectively, and are connected in parallel to the same receiving-end grid. The total capacity of the sending-end PV renewable energy base is 6000MW, with an operating point of 4500MW, and a portion is reserved for frequency regulation. The total capacity of the sending-end hydropower base is 5000MW, with an operating point of 3770MW. The total capacity of the receiving-end grid is 20000MW. The equivalent receiving-end grid inertia is considered to be 8s, and a load surge disturbance of 1000MW (approximately 0.05 pu) is considered at this time. To highlight the beneficial effects of the present invention, the control strategy proposed in this embodiment introduces cooperative additional control at all four converter stations at both the sending and receiving ends; while the traditional control strategy, which serves as a comparison benchmark, only introduces additional control at the receiving end constant active power converter station.

[0048] To highlight the beneficial effects of the algorithm proposed in this invention, this embodiment includes the following two sets of comparative verifications: Control group: The system adopted a multi-terminal flexible DC active coordinated control mechanism, but parameter optimization was not performed. Frequency regulation and voltage droop parameters for each converter station. We take a fixed empirical value and do not consider optimizing the proportion of heterogeneous photovoltaic operation. This means that all of them operate in constant active power mode, and according to national standards, 10% of the rated power is generally reserved as a backup to support frequency regulation.

[0049] Experimental Group: Based on the collaborative control physical architecture, the discretization analytical quantization and multi-objective optimization model of this invention were strictly implemented. Simultaneously, the frequency regulation parameters of the converter station and the proportion of heterogeneous photovoltaic operation were also considered. Conduct joint optimization and strictly consider... Frequency random envelope band security constraint.

[0050] Example simulation verification analysis: (1) Frequency response and dynamic safety envelope band comparison Figure 3 The image shows a comparison of the frequency response of the receiving-end AC power grid. Figure 4 This is a comparison of the frequency response of the power grid on the sending-end hydropower side. Under a sudden load surge disturbance, the control group, which did not consider optimizing the control parameters of the converter station and the photovoltaic heterogeneity ratio, used fixed parameters, and the photovoltaic system operated in constant active power mode. It can be observed that after the disturbance, the maximum frequency deviation on the receiving-end power grid side was 0.131Hz, and the maximum frequency deviation on the sending-end hydropower side was 0.129Hz. For the experimental group, after considering the optimization of the converter station parameters and the photovoltaic heterogeneity ratio, the photovoltaic heterogeneity ratio was calculated. Since the optimized photovoltaic power station will operate in MPPT mode, the frequency will be affected by random fluctuations in photovoltaic power generation. Considering a 99.7% probability scenario, the maximum frequency deviation of the receiving-end AC grid frequency envelope is 0.129 Hz, and the maximum deviation of the sending-end hydropower grid frequency envelope is 0.120 Hz. In contrast, the experimental group using this invention, through joint optimization of the control coefficients at each end, significantly improved the sending-end frequency, reducing its maximum frequency deviation by 6.98%, and enabling precise coordination of system resources in both time and space dimensions.

[0051] (2) Considering the comparison of backup costs for primary frequency regulation of the photovoltaic system, as shown in the figure below. Figure 5 As shown. The control group did not optimize the photovoltaic ratio. Often, an overly conservative strategy is adopted, reserving 10% of the rated power as a backup, resulting in extremely high curtailment penalties during normal operation. The experimental group, however, based its strategy on a comprehensive objective function... Through optimization, the algorithm adaptively calculates the optimal photovoltaic heterogeneous ratio. .contrast Figure 5 It can be seen that the cost of curtailment penalty dropped sharply from 1864.18 yuan in the control group to 1348.63 yuan, significantly improving economic efficiency by 27.65% while ensuring that the frequency safety envelope does not exceed the limit. This optimization result allows photovoltaic clusters to exit the MPPT mode and switch to the fixed PQ mode with only a small proportion, releasing enough supporting power to compensate for the frequency deficiency at the receiving end while minimizing the overall curtailment penalty item. The experimental group achieved extremely high economic benefits and renewable energy consumption rates while ensuring the dynamic security of the entire network.

[0052] In this application, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. In case of any inconsistency, the meaning set forth in this specification or derived from the content described herein shall prevail. Furthermore, the terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit the scope of this application.

[0053] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for coordinated frequency support control of a desert base and a multi-terminal flexible DC system considering the randomness of photovoltaic power generation, characterized in that, Includes the following steps: S1. Real-time acquisition of operating parameters of multi-terminal flexible DC transmission system, and division of sending-end photovoltaic cluster into maximum power point tracking operation mode cluster and constant active power operation mode cluster. And obtain the operating status parameters of the hydropower base; S2. When a frequency drop occurs at the receiving end, additional frequency control is introduced at the fixed active power station at the receiving end to provide initial support. An active reconfiguration mechanism is introduced at the receiving-end DC voltage station to lower the DC voltage reference value to eliminate control conflicts between parallel converter stations; and the DC voltage drop caused by active reconfiguration is transmitted to the sending-end network as a physical signal without communication delay. S3. After the sending-end converter station senses the voltage drop, it changes the active power. Based on the resulting frequency drop at the sending end, the sending-end hydropower base and the photovoltaic cluster in constant active power mode start additional control to coordinate and increase active power generation. At the same time, the randomness of the photovoltaic output in the maximum power point tracking mode is characterized as a continuous-time stochastic differential equation driven by Gaussian white noise. Combined with the controller sampling period, it is discretized and transformed into a discrete-time stochastic difference equation to adapt to digital calculation and analysis. S4. Based on the discrete-time stochastic difference equation, the steady-state covariance matrix of the system state variables is solved using the discrete Lyapunov equation. The theoretical envelope boundary of the receiving end frequency under random disturbance is quantitatively calculated, and the minimum value of the envelope is extracted as the lowest point of the system frequency operation. S5. Construct a collaborative optimization model aimed at improving the minimum operating frequency of the system and reducing curtailment; jointly optimize the proportion of photovoltaic heterogeneous mode operation and the control coefficients of each converter station, and issue optimal control commands. 2.The method of claim 1, wherein, In S1, the multi-terminal flexible DC transmission system has a four-terminal network topology. Its sending end includes two converter stations that are respectively connected to the new energy base, and the receiving end includes a constant active power converter station and a constant DC voltage converter station. 3.The method of claim 2, wherein, In S1, the four-terminal network topology includes four converter stations, specifically including: Receiving end active power converter station: monitor receiving end frequency deviation The active power instruction increment output to the receiving end power grid is: ； In the formula, With The first frequency modulation and virtual inertia control coefficient respectively; its action causes the imbalance of direct current side energy, resulting in passive drop of direct current voltage; Receiving-end constant DC voltage (Vdc) converter station: To eliminate the control clamping effect with the active power station, an active reconfiguration mechanism is adopted. The dynamic voltage command value issued in real time is: ； In the formula, Vn is the rated voltage, K is the droop coefficient; by following the frequency drop actively according to the instruction, the power sent back to the DC network is inhibited, and the support space is released. Sending end hydroelectric converter station: the sensing DC voltage signal is transmitted along the DC line without communication delay, and when the local DC voltage When the deviation from the rated value and the dead zone are exceeded, the DC voltage additional control is triggered, and the active power injection instruction increment to the DC network is increased: ； wherein is a DC voltage-frequency droop coefficient, is a virtual inertia coefficient; Sending-end V / F converter station: V / F grid control is adopted to establish a connection between DC voltage and AC frequency on the photovoltaic side. The angular velocity deviation value is obtained by the DC voltage deviation and frequency droop coefficient. The operating frequency of the AC bus is actively changed to convert the DC side signal into an AC side frequency signal and transmit it to the photovoltaic cluster.

4. The method according to claim 3, wherein, In step S3, the process of coordinating additional control to increase active power generation between the sending-end hydropower base and the photovoltaic cluster in constant active power mode specifically includes: Dynamic modeling of the sending-end hydropower cluster: considering the turbines and their water intake system, relays, and governors using proportional-integral control; utilizing the inertial time constant of the water flow. Relay response time constant Proportional gain Integral gain and permanent state difference coefficient Characterizing the power output variation of the hydroelectric speed controller Its frequency dynamic response after disturbance is represented by the swing equation: ； In the formula, The equivalent inertial time constant of the hydropower cluster, This is the damping coefficient of the water turbine. This represents the change in active power demand at the converter station. This refers to the frequency deviation of the hydroelectric busbar; Photovoltaic cluster in constant active power mode: After detecting a frequency drop on the AC bus of converter station No. 3, primary frequency regulation and virtual inertia control are triggered, and the frequency support active power command increment is: ； In the formula: The active power command increment is provided to support the response frequency of the photovoltaic cluster in the fixed active power mode. This represents the primary frequency regulation droop coefficient of the photovoltaic cluster. Frequency deviation of the AC bus of the photovoltaic cluster; The virtual inertia control coefficient for the photovoltaic cluster; This represents the frequency deviation rate of the AC bus of the photovoltaic cluster.

5. The frequency support control method for a desert base considering the randomness of photovoltaic power generation and a multi-terminal flexible DC system according to claim 4, characterized in that, In step S3, the randomness of the photovoltaic output in the maximum power point tracking mode is characterized as a continuous-time stochastic differential equation driven by Gaussian white noise, and then, in conjunction with the controller sampling period, it is discretized into a discrete-time stochastic difference equation, specifically including: Stochastic Modeling of Photovoltaic Clusters: The stochastic fluctuations in the output of a maximum power point tracking (MPPT) photovoltaic cluster that maintains maximum power operation are characterized as follows: in, The stochastic differential change in the active power output of the photovoltaic cluster. This represents the percentage of photovoltaic clusters operating in maximum power point tracking mode. The intensity of random fluctuations in photovoltaic power output. For the differential terms of the standard Wiener process; By coupling the above-mentioned hydropower speed regulation dynamics and flexible DC control dynamics with the photovoltaic fluctuation model, the continuous-time stochastic differential equations of the system are constructed: ； Combined with the fixed sampling period of the system physical controller The continuous-time stochastic differential equation is transformed into a discrete-time stochastic difference equation using the difference method: ； In the formula, This includes sampling period information and hydropower response time constant. The discretized system state transition matrix of the converter station additional control coefficients.

6. The frequency support control method for a desert base considering the randomness of photovoltaic power generation and a multi-terminal flexible DC system according to claim 5, characterized in that, Step S4 specifically includes: Based on the obtained discrete-time stochastic difference equation, the discrete algebraic Lyapunov equation satisfied by the steady-state covariance matrix of the system is constructed as follows: ； By analytically solving the Lyapunov equation, the steady-state variance corresponding to the AC frequency deviation at the receiving end is extracted. Using Gaussian distribution The criteria, combined with the deterministic mean trajectory of the dynamic characteristics evolution of hydropower reversal regulation, are used to determine the underlying principles. The theoretical lower bound of the receiving-end frequency deviation at the corresponding confidence level was quantitatively calculated: 。 7. The frequency support control method for a desert base considering the randomness of photovoltaic power generation and a multi-terminal flexible DC system according to claim 6, characterized in that, In step S5, the collaborative optimization model and the issuance of optimal control commands specifically include: S51. Objective Function Calculation: Construct a comprehensive objective function that considers both the system's disturbance rejection security and the economic viability of photovoltaic backup. ； In the formula, This indicates the comprehensive curtailment penalty for photovoltaic power generation when it exits maximum power point tracking mode and switches to constant active power mode, with reserves reserved for backup. This is a penalty term for the lowest point deviation of the frequency envelope obtained by quantization; This is a penalty item for frequency deviation of the hydropower supply at the sending end; Weighting coefficients for multi-objective optimization; S52, Parameter Optimization: While satisfying the system power balance constraints and the requirements of claim 4... Under dynamic safety boundary constraints, an optimization algorithm is used to solve the objective function. The global optimal solution was calculated, which included the best photovoltaic configuration ratio. Optimal frequency regulation coefficient set for converter stations The decision vector is included; and based on this vector, the expected frequency minimum point increase and curtailment rate of the system under extreme disturbances are calculated. S53. Issuance of control commands: The decision vector is converted into underlying physical control commands and sent in real time to the controllers of each converter station and the source-side power station coordination control unit of the four-terminal flexible DC system through the communication network. S54, Closed-loop collaborative execution: The photovoltaic base at the sending end... Adjust the actual active power output point to release support margin; each converter station controller adjusts the received data... Real-time correction of the primary frequency modulation droop coefficient at the receiving end Receiving end active reconstruction coefficient and the voltage-frequency co-droop coefficient at the sending end It achieves closed-loop support for the frequency deficit at the receiving end through AC / DC energy linkage.

8. The frequency support control method for a desert base considering the randomness of photovoltaic power generation and a multi-terminal flexible DC system according to claim 6, characterized in that, In S51, the analytical expression of the photovoltaic reserve penalty term in the comprehensive objective function is: ； in, The unit curtailment penalty cost coefficient, This is the predicted maximum power generation capacity of the photovoltaic cluster.

Images