Method for preventing power reverse peak regulation through primary frequency modulation coupling virtual inertia of network construction type draught fan

Abstract

The invention discloses a method for preventing power reverse peak regulation by coupling primary frequency modulation and virtual inertia of a network-forming type fan, and the method effectively prevents power reverse peak regulation by dynamically decoupling the coupling effect of primary frequency modulation and virtual inertia and adjusting an amplitude limiting threshold value based on a real-time working condition. The adaptive amplitude limiting control strategy adjusts an amplitude limiting threshold value in real time according to parameters such as power grid frequency deviation, frequency change rate and fan operation state, optimizes fan power output, and avoids excessive power fluctuation. The virtual inertia compensation module simulates inertia response of a traditional generator set, provides extra inertia support and stabilizes frequency fluctuation of a power grid. The hierarchical control architecture realizes the balance of virtual inertia and frequency modulation control through the synergistic effect of an upper decision layer and a bottom execution layer, and avoids the phenomenon of reverse peak regulation of power. The control strategy not only improves the frequency modulation response speed and precision, but also enhances the adaptive capacity of the fan to the frequency fluctuation of the power grid, and has important application value and wide market prospect.

Description

Technical Field

[0001] The present invention relates to the technical field of power system dispatching, and particularly to a method for preventing power reverse peaking by coupling virtual inertia in primary frequency modulation of grid-forming wind turbines. Background Art

[0002] With the increasing global demand for clean energy, wind power generation, as an important renewable energy source, has been widely used. With the gradual increase in the proportion of wind power generation, the traditional power system faces higher dispatching and frequency modulation pressures. Especially in the case of large-scale grid connection, the volatility of wind power generation poses a severe challenge to the stable operation of the power grid. Traditional wind turbines usually rely on external frequency modulation signals, with a slow frequency modulation response speed and poor adaptability to grid fluctuations, making it difficult to effectively cope with load fluctuations and other sudden grid events. Traditional grid-connected wind turbines usually operate in the maximum power point tracking (MPPT) mode. When the grid frequency is disturbed, the change in its active power output is small, and its ability to support the grid frequency is limited.

[0003] In recent years, the technology of grid-forming wind turbines (GFWT) has gradually emerged. By synchronizing with the power grid, it realizes autonomous regulation of power output and enhances the wind turbine's ability to regulate grid frequency fluctuations. The virtual inertia technology further enhances the stability of GFWT. Grid-forming wind turbines simulate the inertial response of traditional generators to enhance the adaptability of the wind turbine to short-term frequency changes. By simulating the operating characteristics of synchronous generators, grid-forming wind turbines have the ability to actively support the grid voltage and frequency, becoming a key technology for improving the stability of the power system and accommodating large-scale renewable energy. GFWT usually participates in the primary frequency modulation of the power grid through droop control and introduces a virtual inertia link in its control system to improve the system's response speed to frequency changes and suppress frequency deviation.

[0004] However, when GFWT participates in primary frequency modulation and provides virtual inertia simultaneously, when the grid frequency undergoes large and rapid disturbances, especially during the process of rapid frequency decline followed by rapid recovery, the active power output of GFWT may exhibit a "reverse peaking" phenomenon. Specifically, when the frequency drops, the primary frequency modulation and virtual inertia control will prompt GFWT to increase its active power output to support the frequency; but when the frequency rapidly recovers, if the control parameters are not properly set or the rate of frequency change is too fast, GFWT may rapidly reduce or even reverse its active power output within a short period, leading to a further expansion of the power deficit in the grid, exacerbating frequency fluctuations, and even potentially triggering a chain reaction, posing a threat to the safe and stable operation of the power grid.

[0005] At present, the research on the problem of reverse power regulation of GFWT mainly focuses on optimizing control parameters, designing more complex control structures, etc. However, existing methods often have difficulty achieving good suppression effects under various operating conditions and lack a direct prevention and adaptive adjustment mechanism for the phenomenon of reverse power regulation. Therefore, there is an urgent need for a limiter control strategy that can adaptively adjust according to the characteristics of grid frequency changes to effectively prevent the phenomenon of reverse power regulation when GFWT participates in primary frequency modulation and provides virtual inertia.

[0006] After retrieval, the invention patent with the publication number CN116562022A discloses a method and system for intelligent frequency modulation of a fan based on characteristic curve operation, which solves, fits and analyzes the extreme points of the characteristic curves of the fan under different frequency conditions to determine the operating frequency of the fan in the optimal energy consumption state and realizes energy-saving frequency modulation.

[0007] The technical comparison between the above-mentioned comparative document and the present application is as follows:

[0008] The present application proposes a dynamic decoupling control scheme for a grid-connected fan's primary frequency modulation coupled with virtual inertia to prevent reverse power regulation. Through an adaptive limiter control strategy under real-time working conditions, the limiter threshold is dynamically adjusted according to the grid frequency deviation, frequency change rate and the operating state of the fan, effectively preventing the phenomenon of reverse power regulation; at the same time, the virtual inertia compensation module simulates the inertial response of traditional generator sets to improve the grid frequency stability. While the above-mentioned comparative document focuses on using the fan characteristic curve to determine the optimal energy consumption and is mainly applied to the energy-saving control of fans in fields such as tunnel construction ventilation, the present application emphasizes more on eliminating power fluctuations and reverse power regulation problems during the frequency modulation process rather than only obtaining the optimal energy consumption by fitting the characteristic curve.

[0009] After retrieval, the invention patent with the publication number CN107859647B discloses a frequency modulation and speed regulation interface circuit, a fan, a fan speed regulation system and a speed regulation method, which transmits the user input pulse signal to the micro control unit and realizes high-precision fan speed regulation under the influence of temperature through circuit designs such as pull-up resistors, so as to ensure that the fan output speed meets the predetermined target.

[0010] The technical comparison between the above-mentioned comparative document and the present application is as follows:

[0011] This application focuses on the coupled control of primary frequency regulation and virtual inertia of grid-forming wind turbines. By dynamically decoupling the effects of primary frequency regulation and virtual inertia and adjusting the power limit threshold based on real-time operating conditions, it effectively prevents the phenomenon of power reverse peak shaving. While simulating the inertial response of traditional generators, the virtual inertia compensation module optimizes the power output of the wind turbine, improving the grid frequency response speed and stability. Different from the above-mentioned comparative document which mainly focuses on the precise regulation of the wind turbine speed, with the technical emphasis on the speed regulation interface of the wind turbine and the guarantee of control accuracy, this application focuses on avoiding power fluctuations and reverse peak shaving problems during the overall frequency regulation process, achieving the coordinated support of the wind turbine for grid frequency fluctuations.

[0012] Upon retrieval, the invention patent with the publication number CN111062617A discloses a method and system for analyzing the output characteristics of offshore wind power. It mainly conducts time-series statistics, probability, and characteristic analysis on the wind speed data and wind turbine power data in the offshore wind area, constructs the relationship between wind power output and time and the reverse peak shaving characteristics, and focuses on evaluating the fluctuations of offshore wind power output and its statistical characteristics, providing data support for power and electricity balance and peak shaving balance.

[0013] The technical comparison between the above-mentioned comparative document and this application is as follows:

[0014] This application adopts a hierarchical control architecture, realizes the real-time coupling of virtual inertia and frequency regulation control during primary frequency regulation, adjusts the limit threshold based on real-time conditions through an adaptive limit control strategy and dynamic decoupling technology, and simulates the inertial response in real time through the virtual inertia compensation module, thereby preventing the phenomenon of power reverse peak shaving. This method not only improves the frequency regulation response speed and accuracy but also enhances the immediate adaptability of the wind turbine to grid frequency fluctuations, belonging to an online control and regulation scheme. Different from the above-mentioned comparative document which focuses on the offline analysis of wind power output characteristics, with the analysis method mainly being offline data processing and characteristic prediction, this application provides a more direct and real-time solution to prevent power reverse peak shaving for the power regulation problem during the primary frequency regulation process of grid-forming wind turbines. Summary of the Invention

[0015] Object of the Invention: The present invention provides a method for preventing power reverse peak shaving by coupling virtual inertia in the primary frequency regulation of grid-forming wind turbines. This method effectively prevents power reverse peak shaving while ensuring the frequency regulation ability and the safe operation of equipment by dynamically decoupling the coupling effect of primary frequency regulation and virtual inertia and adjusting the limit threshold based on real-time conditions, enhancing the frequency regulation ability of the wind turbine and the frequency regulation effect of the power grid.

[0016] To achieve the above object, the solution of the present invention includes the following steps:

[0017] S1. Adaptive dynamic limit control

[0018] Based on the frequency deviation, rate of change of frequency, current operating state of the wind turbine (such as rotational speed, pitch angle), and wind speed prediction data, the dynamic limit threshold is calculated in real time through fuzzy logic or dynamic weight algorithm. The fuzzy logic controller is used to correct the limit threshold and compensation coefficient online to adapt to the changes in different grid strengths and wind turbine operating conditions.

[0019] S2. Virtual inertia compensation module

[0020] Simulate the inertial response of traditional generating units to provide additional inertia support during grid frequency fluctuations, ensuring a smooth change in grid frequency. Through the feedforward compensation algorithm, the power reverse regulation component generated by virtual inertia control is offset to decouple the coupling effect of primary frequency modulation and virtual inertia.

[0021] S3. Coupling effect modeling and decoupling

[0022] Establish a frequency modulation power superposition model and use frequency domain sensitivity analysis to quantify the coupling coefficient. By jointly controlling primary frequency modulation and virtual inertia, balance their effects and avoid reverse power peaks.

[0023] S4. Hierarchical coordinated control architecture

[0024] Upper decision-making layer: Dynamically allocate the frequency modulation weights of virtual inertia and droop control according to the grid frequency deviation and the operating state of the wind turbine.

[0025] Lower execution layer: Use an adaptive limit module to correct the power command to ensure a smooth transition of the frequency modulation power and not exceed the safe range.

[0026] S5. Multi-time scale coordination

[0027] Apply a short-time window smoothing limit to the virtual inertia link to avoid sudden changes in power commands.

[0028] S6. Reverse peak regulation suppression logic

[0029] In the frequency recovery stage, introduce a reverse power gradient limit to force the frequency modulation power to exit according to an exponential decay curve, avoiding sudden changes in power.

[0030] Furthermore, in step S1 of the adaptive dynamic limit control, multi-dimensional information such as frequency deviation, rate of change of frequency, current operating state of the wind turbine (such as rotational speed, pitch angle), and wind speed prediction data is fully considered. An intelligent calculation model is constructed using fuzzy logic or dynamic weight algorithm to calculate the dynamic limit threshold in real time. The fuzzy logic controller continuously corrects the limit threshold and compensation coefficient online based on the real-time input information, enabling it to closely fit different grid strengths and the complex and changeable operating conditions of the wind turbine, ensuring the accuracy and adaptability of the limit.

[0031] In the frequency modulation control of a wind farm, the amplitude limiting threshold is a crucial parameter that determines the output range of the frequency modulation power, avoids excessive power regulation or large fluctuations, and ensures the stability of the grid frequency. Traditional methods for calculating the amplitude limiting threshold are usually based on fixed frequency changes or the basic operating conditions of the wind turbines, and may not fully consider the dynamic changes of the grid, load fluctuations, and the real-time response of the wind turbines. Therefore, the present invention proposes an adaptive amplitude limiting threshold calculation method, which combines the real-time operating conditions of the grid, load changes, and the response characteristics of the wind turbines, and maximizes the frequency modulation efficiency and reduces the power reverse peaking phenomenon by dynamically calculating the amplitude limiting threshold.

[0032] The operating state of the wind turbine directly affects its frequency modulation ability, and the present invention at least includes the following factors:

[0033] Rotational speed N: The rotational speed of the wind turbine affects its inertial characteristics.

[0034] Pitch angle θ: The pitch angle affects the power output ability of the wind turbine.

[0035] Wind speed prediction data v uind : The wind speed directly affects the maximum output power of the wind turbine.

[0036] The calculation formula of the traditional amplitude limiting threshold generally adopts the combined output based on the grid frequency deviation and the frequency modulation power and virtual inertia.

[0037] The adaptive amplitude limiting threshold calculation formula proposed by the present invention considers multiple real-time factors, dynamically adjusts the amplitude limiting threshold, and adapts to the changes in the grid and wind turbine states. The improved amplitude limiting threshold formula is:

[0038]

[0039] Where:

[0040] P limit (t) is the dynamic amplitude limiting threshold, representing the maximum frequency modulation power at time ttt;

[0041] P wind (t) is the wind turbine output power, reflecting the current power state of the wind turbine;

[0042] Rate of change of grid frequency, representing the speed of frequency change;

[0043] Δf(t) is the grid frequency deviation, reflecting the deviation between the frequency and the target frequency;

[0044] ΔP load (t) is the change in grid load, representing the fluctuation of grid load;

[0045] ΔP wind (t) is the fluctuation of wind turbine power, representing the change amplitude of wind turbine power;

[0046] γ1(t) is an adaptive adjustment coefficient, which is dynamically adjusted according to the magnitude of the grid frequency change rate to ensure that the amplitude limiting threshold increases when the frequency fluctuation is large.

[0047] The calculation of the adaptive adjustment coefficient can adopt the following dynamic formula:

[0048]

[0049] Where:

[0050] γ 1,0 is the initial adaptive adjustment coefficient;

[0051] α1 is a coefficient adjustment factor used to control the influence degree of the frequency change rate on γ1(t);

[0052] In the case of drastic changes in the grid frequency, the frequency modulation power needs to respond quickly to restore the grid stability faster. Therefore, the frequency change rate As a key factor, it will significantly affect the dynamic adjustment of the amplitude limiting threshold. Especially when the frequency change rate is large, the adaptive adjustment coefficient γ1(t) will increase the amplitude limiting threshold, enabling the frequency modulation power to respond more quickly to frequency fluctuations and avoiding the reverse peak shaving phenomenon.

[0053] The change of the grid load directly affects the frequency stability of the grid. When the grid load increases, the frequency may decrease, and at this time the frequency modulation system needs to provide more power to maintain the frequency stability. On the contrary, when the load decreases, the frequency modulation power should be appropriately reduced to avoid excessive power generation and reverse peak shaving. Therefore, the load change amount ΔP load (t) can reflect the influence of the grid load fluctuation on the amplitude limiting threshold in real time.

[0054] The fluctuation of the fan power is caused by the change of the wind speed and the response characteristics of the fan. Excessive fan power fluctuation may lead to reverse peak shaving of power and affect the frequency stability of the grid. To reduce this fluctuation, the improved amplitude limiting threshold formula incorporates the fan power change amount ΔP wind (t) into the calculation to adjust the amplitude limiting threshold in real time and avoid reverse peak shaving caused by excessive fan regulation.

[0055] To adjust the amplitude limiting threshold more precisely, the present invention introduces an adaptive adjustment coefficient γ1(t), which changes dynamically according to the grid frequency change rate. When the grid frequency fluctuates violently, the coefficient increases, thereby increasing the amplitude limiting threshold and enabling the frequency modulation power to respond more quickly to the frequency change; when the frequency fluctuation is relatively stable, the coefficient decreases, thus avoiding excessive power adjustment.

[0056] To improve the adaptability of the control system, a fuzzy logic controller (FLC) can be used to correct the clipping threshold. Fuzzy logic can dynamically adjust the clipping threshold and compensation coefficient according to real-time input such as the operating conditions of the wind turbine, frequency deviation, and frequency change rate.

[0057] Set the input variables of the fuzzy logic controller as: frequency deviation, frequency change rate, wind turbine speed deviation, and wind speed deviation. The fuzzy logic controller outputs a clipping correction factor ΔP limit (t), and this correction factor is combined with the clipping threshold:

[0058] P th,adjusted = P limit (t)+ΔP limit (t)

[0059] Through real-time calculation of fuzzy logic, ΔP limit (t) can be automatically corrected according to the change of operating conditions to ensure that the clipping threshold can effectively adapt to different grid conditions.

[0060] Through the above improvement, the clipping threshold is no longer a fixed constant, but is dynamically calculated according to real-time grid frequency changes, load fluctuations, and wind turbine power fluctuations, which can better adapt to complex grid environments, improve frequency modulation efficiency, and avoid the phenomenon of reverse power peak shaving. This improvement method not only improves the performance of the wind farm frequency modulation system, but also provides a more reliable guarantee for the stable operation of the power grid.

[0061] Virtual inertia refers to using wind turbines or other renewable energy units to simulate the inertial response ability of traditional synchronous generator sets, thereby enhancing the power grid's anti-disturbance ability to frequency fluctuations. The modeling method of virtual inertia is crucial in practical applications because it can provide a rapid power response during grid frequency fluctuations, playing a role in alleviating frequency drops or rises. However, traditional virtual inertia modeling methods are usually relatively simple and may not fully reflect the complexity and variability of grid frequency fluctuations. For this reason, an improved virtual inertia modeling method is proposed, which can better simulate the inertial response during the process of grid frequency change while taking into account the stability of the power grid and the operating safety of the wind turbine.

[0062] Traditional virtual inertia modeling methods are usually based on simple linear models, in which the power response of virtual inertia and grid frequency change are achieved through a certain fixed proportional relationship. Although this method is simple, it has several limitations: assuming that the relationship between virtual inertia and frequency change rate is fixed, lacking dynamic adaptability to the characteristics of grid frequency fluctuations. Grid frequency changes are not constant, and the intensity and duration of frequency fluctuations will affect the effectiveness of virtual inertia, which the traditional model fails to fully consider. Operating parameters such as the rotational speed and pitch angle of the wind turbine have an important impact on the ability to provide virtual inertia, and the traditional model fails to dynamically model this.

[0063] To address the deficiencies of traditional virtual inertia modeling methods, an improved virtual inertia modeling method is proposed. This method comprehensively considers multiple factors of grid frequency change and dynamically adjusts the virtual inertia providing ability according to the real-time operating state of the wind turbine. The improved virtual inertia modeling method is as follows:

[0064] Introduce a dynamic virtual inertia coefficient H virtual (t) to dynamically adjust the providing intensity of virtual inertia according to the actual frequency change of the grid and the operating state of the wind turbine. The change of the dynamic virtual inertia coefficient considers the frequency change rate and the operating condition of the wind turbine.

[0065] To further improve the accuracy of virtual inertia response, consider the coupling effect between the operating state of the wind turbine and the grid frequency fluctuation. When the grid frequency changes violently, the virtual inertia response of the wind turbine may be affected by factors such as wind speed change and wind turbine speed adjustment. Therefore, a condition-frequency coupling model is proposed, considering the two-way influence of the wind turbine operating state and frequency fluctuation.

[0066]

[0067] Wherein:

[0068]

[0069] Wherein:

[0070] H base is the basic virtual inertia coefficient;

[0071] k1 and k2 are weight coefficients for adjusting the frequency change rate and frequency deviation;

[0072] R wind (t) is the relationship coefficient between the rotational speed (or pitch angle) of the wind turbine and the maximum output power, used to adjust the output ability of the wind turbine and reflect the operating state of the wind turbine.

[0073] The contribution of the wind turbine operating state to virtual inertia is crucial. Especially under different operating conditions such as wind speed, rotational speed, and pitch angle, the inertial response ability of the wind turbine will be different. Traditional wind turbine operating state modeling methods often only consider the simple relationship of wind speed or rotational speed and fail to fully explore the influence of the wind turbine on virtual inertia under complex operating conditions. However, the operating state of the wind turbine not only depends on wind speed or rotational speed but is also closely related to parameters such as pitch angle and power generation. Therefore, an improved wind turbine operating state modeling method is proposed. This method can more comprehensively consider the influence of different operating states of the wind turbine on virtual inertia output and optimize the virtual inertia response ability of the wind turbine in real time according to these parameters. A comprehensive model based on multiple parameters is proposed, which can evaluate the virtual inertia response ability of the wind turbine in real time.

[0074]

[0075] Wherein:

[0076] v max is the maximum wind speed;

[0077] ω max is the maximum rotational speed;

[0078] θ max is the maximum value of the pitch angle;

[0079] P max is the maximum power of the wind turbine;

[0080] α1, α2, α3, α4 are adjustment coefficients, used to represent the contribution degrees of different parameters to the virtual inertia.

[0081] This weighting coefficient reflects the influence of the current operating state of the wind turbine on the virtual inertia response, and can flexibly adjust the virtual inertia response ability.

[0082] This method comprehensively considers multiple factors such as wind speed, rotational speed, pitch angle, and power output, so that the virtual inertia response ability can more accurately reflect the actual working conditions of the wind turbine. By dynamically adjusting the virtual inertia coefficient, the virtual inertia response can be optimized in real time according to the operating state of the wind turbine and the change of the grid frequency, thereby improving the frequency regulation ability of the grid. After introducing the operating condition - frequency coupling model, this method can more accurately simulate the inertial response of the wind turbine under different frequency fluctuations, improving the accuracy and flexibility of the frequency modulation control. In addition, the improved modeling method provides a more accurate virtual inertia response reference for the control system of the wind turbine, thereby optimizing the operating strategy of the wind turbine and enhancing the adaptability of the wind turbine to the grid frequency fluctuations.

[0083] In the frequency modulation control of a wind farm, the feed - forward compensation algorithm is mainly used to predict and compensate for the influence of the output power change of the wind turbine (such as wind speed fluctuation, etc.) on the grid frequency. The traditional feed - forward compensation method usually performs linear compensation based on the relationship between the power output of the wind turbine and the grid frequency deviation. However, this method may be affected by environmental changes, load fluctuations, and changes in the characteristics of the wind turbine, resulting in unstable or lagging compensation effects. To solve these problems, the present invention proposes an improved feed - forward compensation algorithm, which improves the accuracy and response speed of the feed - forward compensation by introducing non - linear modeling and a multi - variable dynamic adjustment mechanism, ensures the smooth output of the frequency modulation power, and avoids power reverse peaking.

[0084] Traditional feedforward compensation algorithms usually assume that the response of the wind turbine is linear, that is, the change in grid frequency is proportional to the adjustment of the wind turbine output power. However, this assumption ignores the non-linear characteristics of the wind turbine under different operating conditions. To solve the above problems, the present invention proposes a feedforward compensation algorithm based on non-linear modeling and adaptive adjustment of the wind turbine operating conditions. The improved feedforward compensation algorithm makes the compensation amount dynamically adjustable by introducing parameters such as real-time wind speed, rotational speed, and pitch angle, so as to more accurately predict and respond to grid frequency changes. By considering the real-time operating state of the wind turbine (such as rotational speed, pitch angle) and the characteristics of grid frequency fluctuations, the algorithm dynamically adjusts the feedforward compensation coefficient, improving the compensation accuracy and response speed.

[0085] The improved feedforward compensation algorithm can be expressed as:

[0086]

[0087] To make up for the time delay in the wind turbine response process, the improved feedforward compensation algorithm adds a delay compensation correction mechanism. Due to the certain lag in the wind turbine response (such as the mechanical delay in pitch angle adjustment), this mechanism ensures that the wind turbine can respond faster to grid frequency changes by adjusting the predicted compensation amount. This correction can be achieved by introducing a delay function T delay as follows:

[0088]

[0089] By adjusting the frequency data at the compensation moment, the feedforward compensation can effectively cancel the error caused by the wind turbine response lag.

[0090] To further optimize the effect of the compensation algorithm, the improved algorithm introduces an adaptive adjustment mechanism. Based on the historical data of the actual operation of the wind turbine, the algorithm can adjust the compensation coefficient in real time and automatically correct the compensation strategy according to the frequency characteristics of grid frequency fluctuations. The adaptive mechanism is based on the reinforcement learning algorithm and can continuously optimize the compensation strategy through real-time feedback to adapt to different grid frequency fluctuation patterns.

[0091] θ adapt (t) = θ adapt (t - 1) + γ·(Δf(t) - Δf(t - 1))

[0092] where: γ is the learning rate, which can adjust the current compensation strategy according to historical data.

[0093] When a wind farm with grid-forming wind turbines conducts primary frequency regulation and virtual inertia coupling control with the power grid, there is an important challenge - how to effectively decouple the interaction between primary frequency regulation and virtual inertia. These two are usually highly coupled and have an impact on each other's response to the grid frequency, easily leading to problems such as reverse power peaking or overshooting of the frequency. Therefore, designing an effective decoupling algorithm for the coupling effect is one of the key technologies to improve the frequency regulation accuracy of wind turbines and ensure the stable operation of the power grid.

[0094] Traditional frequency regulation and virtual inertia control strategies usually adopt simple linear superposition or weighted average methods to synthesize the effects of primary frequency regulation and virtual inertia. Generally speaking, the system will add the control signals of the two to form a unified control instruction. This method is simple and intuitive, but due to the different action mechanisms of primary frequency regulation and virtual inertia, direct superposition may lead to mutual interference between their effects. For example, when the virtual inertia control generates an inertial response, it may conflict with the fast power response generated by primary frequency regulation, resulting in reverse power peaking or frequency overshoot.

[0095] To solve the coupling problem in the traditional method, the present invention proposes a coupling effect decoupling algorithm based on frequency-domain sensitivity analysis and dynamic optimization decoupling mechanism. This algorithm accurately models the coupling relationship between primary frequency regulation and virtual inertia power, and dynamically adjusts the decoupling parameters in combination with real-time feedback information to ensure that the two can work together and avoid reverse power peaking and frequency overshoot.

[0096] First, establish a coupling model of the two powers. By performing frequency-domain analysis on the dynamic response of the wind turbine and the frequency response of the power grid, the coupling coefficient between primary frequency regulation power and virtual inertia power is quantified. This process can be completed through the following steps:

[0097] By analyzing the sensitivity of the wind turbine power response and the change of the grid frequency, the interaction relationship between frequency regulation power and virtual inertia power in the frequency domain is obtained. Assuming that the frequency-domain response models of the frequency change are G f (ω), G v (ω), then the coupling coefficient C(ω) of the system can be expressed as:

[0098]

[0099] By analyzing these functions, the coupling degree between the two can be quantified.

[0100] Through frequency-domain analysis, establish a coupling matrix M between primary frequency regulation power and virtual inertia power, such that

[0101]

[0102] The matrix M describes the mutual coupling relationship between the two, and the influence degree and transfer effect of the two can be accurately extracted by matrix operations.

[0103] Based on the above coupling modeling and sensitivity analysis results, the present invention proposes a decoupling control algorithm based on dynamic optimization. This algorithm dynamically adjusts the decoupling parameters between primary frequency regulation and virtual inertia power according to the real-time frequency change and the operating conditions data of the wind turbine, ensuring that the adjustment capabilities of the two do not interfere with each other and enabling coordinated regulation during power grid frequency changes.

[0104] The improved decoupling algorithm can be implemented through the following steps:

[0105] Adaptive adjustment of frequency change rate: Dynamically adjust the decoupling coefficient in the coupling matrix M according to the real-time data of the power grid frequency change rate. When the frequency changes violently, the decoupling coefficient M will be automatically adjusted so that the responses of primary frequency regulation and virtual inertia do not interfere with each other, avoiding the reverse power peaking caused by too fast a response.

[0106] Power distribution optimization: Dynamically optimize the power distribution between primary frequency regulation and virtual inertia according to the real-time power grid frequency deviation and the real-time operating state of the wind turbine. By introducing a machine learning algorithm, the decoupling coefficient of the system is adjusted online to optimize the distribution ratio of primary frequency regulation and virtual inertia. Assuming the optimization goal is to minimize the risk of reverse power peaking and frequency fluctuation, the following objective function can be used for optimization:

[0107]

[0108] Where: w1 and w2 are weight coefficients, representing the weights of frequency deviation and reverse power peaking respectively.

[0109] Feedback regulation mechanism: Dynamically correct the power distribution between primary frequency regulation and virtual inertia through a real-time feedback regulation mechanism. For example, when the power grid frequency recovers, a reverse power gradient limiting strategy is adopted to control the withdrawal of primary frequency regulation and virtual inertia, avoiding frequency overshoot caused by overcompensation.

[0110] Based on the above design, the implementation of the decoupling algorithm can be divided into the following steps:

[0111] (1) Real-time collection of power grid frequency data: Obtain power grid frequency deviation and frequency change rate data.

[0112] (2) Wind turbine state monitoring: Real-time monitor the rotational speed, pitch angle and power output of the wind turbine.

[0113] (3) Calculate the coupling matrix: Calculate the coupling matrix M based on frequency domain sensitivity analysis and dynamically adjust it according to the power grid frequency change and the wind turbine state.

[0114] (4) Optimize power distribution: Calculate the optimal power distribution strategy P f1 (t) and P v1 (t) through machine learning.

[0115] (5) Output the decoupled frequency regulation command: Generate the decoupled and optimized frequency regulation power command to ensure the stability and frequency regulation accuracy of the system.

[0116] Through frequency-domain sensitivity analysis and dynamic optimization decoupling mechanism, the improved coupling decoupling algorithm of the present invention can effectively solve the mutual interference problem between primary frequency regulation and virtual inertia, improve the frequency regulation accuracy, avoid the power reverse peaking phenomenon, and thus improve the grid connection stability and frequency regulation ability of wind turbines. This provides technical support for large-scale wind power grid connection scheduling and the stability of power systems.

[0117] At the critical stage of grid frequency recovery, that is, when the rate of frequency change changes from negative to positive, an innovative reverse power gradient limiting mechanism is introduced. By forcing the frequency regulation power to exit orderly according to the exponential decay curve, a smooth transition is achieved, avoiding secondary impacts on the grid caused by power mutations and ensuring the stability of the entire frequency regulation process.

[0118] The anti-reverse peaking suppression logic is used to effectively avoid the reverse fluctuation and mutation of frequency regulation power, especially in the grid frequency recovery stage. The frequency regulation system usually balances the grid frequency fluctuation through primary frequency regulation response. However, when the frequency recovery stage occurs, if the frequency regulation power fails to smoothly exit in time, it may lead to the reverse peaking phenomenon, causing a large fluctuation in the grid frequency. Therefore, the present invention proposes a reverse power gradient limiting mechanism, combined with the exponential decay exit curve, to effectively avoid power mutations and smoothly exit the frequency regulation mode.

[0119] In the grid frequency recovery stage, the rate of frequency change changes from negative to positive, that is, the grid frequency begins to recover to the rated value. At this time, the original frequency regulation power of the system may generate reverse fluctuations. To prevent such sudden fluctuations, the present invention introduces an adaptive reverse power gradient limiting method. The role of this limiting mechanism is to impose a limit on the rate of change of frequency regulation power to ensure that the frequency regulation power does not cause excessive fluctuations in the grid frequency due to over-response.

[0120] The grid load state and the fan response characteristics are the key factors determining the reverse power gradient limit. The change speed of the grid load and the frequency regulation response characteristics of the fan (such as factors like the fan speed, pitch angle, wind speed, etc.) will directly affect the power adjustment process. Therefore, the present invention proposes to dynamically adjust the coefficient of the reverse power gradient limit by combining the current state of the fan and the change of the grid load.

[0121]

[0122] Wherein:

[0123] κ1 and κ2 are adjustment coefficients used to weight the contributions of grid load and wind turbine response.

[0124] To achieve more flexible regulation, the present invention further introduces an adaptive feedback mechanism, enabling the reverse power gradient limit coefficient to be adjusted adaptively according to the amplitude and rate of grid frequency fluctuations, changes in wind turbine response, and grid load status. The specific adjustment method is as follows:

[0125]

[0126] Where:

[0127] λ0 is the initial reverse power gradient limit coefficient;

[0128] α1, α2, and α3 are adaptive coefficients used to adjust the gradient limit intensity according to different grid frequency fluctuations, load changes, and wind turbine states;

[0129] In this way, the reverse power gradient limit coefficient will be adjusted dynamically according to the severity of grid frequency fluctuations, changes in load fluctuations, and fluctuations in wind turbine output power. For example, when the grid frequency fluctuates greatly, the reverse power gradient limit coefficient will increase to enhance the smooth transition of frequency modulation power; when the load changes drastically, the system will adjust the wind turbine frequency modulation response speed to avoid frequent power fluctuations.

[0130] To ensure the smooth exit of frequency modulation power and reduce the risk of secondary frequency dips, an improved smooth exit algorithm is introduced, which provides more refined regulation in combination with the actual states of grid load and wind turbine output. Specifically, the exit curve adopts the following form:

[0131] P exit (t) = P max ·exp(-α(t)·t)·(1 + β1·ΔP load + β2·ΔP wind )

[0132] Where: The exponential decay coefficient α(t) controls the decay rate during the power exit process and directly affects the response speed of the wind turbine to frequency recovery.

[0133] To more precisely adapt to different grid frequency fluctuations, wind turbine operating states, and load changes, the present invention proposes the design of a dynamic decay coefficient. This coefficient not only considers the dynamic response during the frequency recovery stage but also combines parameters such as the actual load of the grid, the rotational speed and pitch angle of the wind turbine to dynamically adjust the decay rate. Its mathematical expression is:

[0134]

[0135] Where: α0 is the initial decay coefficient;

[0136] The adaptive coefficients γ1, γ2, and γ3 in the above formula can be adjusted through an online feedback mechanism. The values of these coefficients can be dynamically optimized according to changes in the grid state, the operating state of the wind turbine, and load fluctuations to achieve the best power withdrawal effect.

[0137] Specifically, after each frequency modulation cycle, the system will feedback and adjust the attenuation coefficient according to the actual grid frequency recovery effect, load fluctuation conditions, and wind turbine response. By correcting γ1, γ2, and γ3 in real time, each frequency modulation operation can be made smoother and more efficient.

[0138] In the frequency modulation and virtual inertia coupling control system of grid-forming wind turbines, the hierarchical coordinated control architecture, as the core strategy, plays a key role. Its main purpose is to effectively coordinate the control of frequency modulation and virtual inertia through a hierarchical structure to ensure grid frequency stability and prevent the occurrence of power reverse peak shaving. This architecture realizes efficient frequency modulation of the wind turbine and virtual inertia control through the collaborative work of the upper decision-making layer and the lower execution layer.

[0139] The hierarchical coordinated control architecture realizes efficient frequency modulation control and wind turbine operation optimization by dividing the control strategy of the wind turbine into an upper decision-making layer and a lower execution layer. In the upper decision-making layer, the frequency modulation weights of virtual inertia and droop control are dynamically adjusted mainly according to the grid frequency deviation, the rate of change of frequency, and the real-time operating state of the wind turbine (such as rotational speed, pitch angle, etc.). By monitoring the frequency fluctuations of the grid and the state of the wind turbine in real time, this layer can intelligently allocate frequency modulation tasks, so as to ensure that the grid frequency fluctuations are adjusted in a timely and effective manner and avoid unnecessary waste of frequency modulation capacity.

[0140] The lower execution layer is responsible for specifically executing the frequency modulation instructions from the upper decision-making layer and making adjustments according to the real-time working conditions of the wind turbine. This layer dynamically corrects the power output of the wind turbine through an adaptive amplitude limiting control module to ensure that the power command does not exceed the safe range, and at the same time smooths the frequency modulation process to avoid excessive power fluctuations affecting the stable operation of the wind turbine. The role of the lower execution layer is to ensure that the wind turbine can respond quickly under grid frequency fluctuations and maintain a stable and reliable frequency modulation function.

[0141] Through this hierarchical coordinated control architecture, the upper decision-making layer and the lower execution layer achieve high coordination, optimizing the frequency modulation effect between virtual inertia and droop control and avoiding the occurrence of power reverse peak shaving. This architecture improves the frequency modulation response speed and accuracy, enhances the system's adaptability to various complex grid conditions, and ensures the long-term stable operation of the wind turbine. Overall, this architecture can not only improve the frequency modulation performance of the wind turbine but also effectively extend the service life of the equipment, with important practical application value.

[0142] To further improve the stability of the system and the frequency regulation effect, the present invention proposes a multi-time scale coordinated control mechanism. Under this mechanism, the response times of virtual inertia control and primary frequency regulation control are adjusted using different time scales respectively: Virtual inertia control: It responds quickly within a short time window to adapt to the rapid changes in frequency. Preferably, the time window is (10 - 100 ms). Primary frequency regulation control: It performs smooth control within a longer time window to avoid frequent power fluctuations. Preferably, the time window is (1 - 10 s). Through the allocation of different time scales, the system can respond quickly while avoiding the unstable impact on the power grid caused by sudden frequency changes.

[0143] Under the multi-time scale mechanism, the frequency regulation response of the wind turbine can be divided into short-term and long-term responses: The short-term response is that virtual inertia control and droop control give priority to handling the rapid changes in the grid frequency. The long-term response is that primary frequency regulation control handles the long-term fluctuations in the grid frequency. By optimizing the allocation of time scales, the frequency regulation accuracy of the system can be improved, and the negative impact caused by frequent adjustment of the frequency regulation power within a short time can be avoided.

[0144] The hierarchical coordinated control architecture realizes the dynamic coordination of virtual inertia and droop control through the close cooperation between the upper decision-making layer and the lower execution layer. The upper decision-making layer dynamically adjusts the frequency regulation weight according to the grid frequency deviation, the rate of change of frequency, and the operating state of the wind turbine, while the lower execution layer corrects the power command through the adaptive amplitude limiting control module to ensure a smooth transition of power and not exceed the safe range. Through multi-time scale coordinated control, the system can maintain a smooth change in the long-term frequency while responding quickly, thus effectively improving the frequency regulation ability and stability of the power grid.

[0145] Beneficial effects: The present invention has the following advantages:

[0146] The adaptive amplitude limiting control strategy for preventing power reverse peaking under the coupling of primary frequency regulation and virtual inertia of the grid-forming wind turbine proposed by the present invention effectively solves the problem of power reverse peaking caused by grid frequency fluctuations through a variety of innovative methods. First, through the adaptive amplitude limiting control strategy, the power output of the wind turbine is adjusted in real time to ensure that the wind turbine can flexibly respond to grid frequency changes and avoid the waste of frequency regulation capacity caused by a fixed amplitude limiting threshold. The introduction of the virtual inertia compensation module simulates the inertial response of traditional generator sets, provides additional inertia support for grid frequency fluctuations, and further smooths the grid frequency changes. Through the comprehensive modeling of the operating state of the wind turbine, the present invention can accurately adjust the virtual inertia coefficient, optimize the frequency regulation ability of the wind turbine, and improve the accuracy and flexibility of grid frequency regulation.

[0147] In addition, the improved feed-forward compensation algorithm and decoupling algorithm for coupling effects further enhance the stability and response ability of the wind turbine under frequency fluctuations, ensuring that the wind turbine can be effectively adjusted according to the real-time changes of the power grid and reducing the occurrence of reverse power peak regulation. The design of the hierarchical coordinated control architecture combines the dynamic regulation of the upper decision-making layer and the precise control of the lower execution layer, ensuring the smooth transition of the frequency modulation process and the safe operation of the wind turbine. Generally speaking, by optimizing the virtual inertia response and frequency modulation control strategy of the wind turbine, the present invention not only improves the power grid frequency regulation ability, but also enhances the adaptability of the wind turbine to power grid frequency fluctuations, having significant technical advantages and application prospects. Description of the Drawings

[0148] Figure 1 is the framework structure diagram of the adaptive amplitude limiting control strategy of the grid-forming wind turbine;

[0149] Figure 2 is the schematic diagram of the feed-forward compensation process of the virtual inertia compensation module;

[0150] Figure 3 is the frequency domain sensitivity analysis diagram of the coupling effect of primary frequency modulation and virtual inertia;

[0151] Figure 4 is the schematic diagram of the anti-reverse peak suppression logic. Detailed Embodiments

[0152] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings.

[0153] Embodiment

[0154] This embodiment is a method for primary frequency modulation coupling virtual inertia of a grid-forming wind turbine to prevent reverse power peak, including the following steps:

[0155] S1. Adaptive dynamic amplitude limiting control

[0156] Based on the frequency deviation, frequency change rate, current operating state of the wind turbine and wind speed prediction data, the dynamic amplitude limiting threshold is calculated in real time through the fuzzy logic or dynamic weight algorithm, and the amplitude limiting threshold and compensation coefficient are corrected online by the fuzzy logic controller;

[0157] The dynamic amplitude limiting threshold is calculated as follows:

[0158]

[0159] Where:

[0160] P limit (t) is the dynamic amplitude limiting threshold;

[0161] P wind (t) is the output power of the wind turbine;

[0162] is the power grid frequency change rate;

[0163] Δf(t) is the power grid frequency deviation;

[0164] ΔP load (t) is the change amount of the power grid load;

[0165] ΔP wind (t) is the fluctuation amount of the fan power;

[0166] γ1(t) is the adaptive adjustment coefficient;

[0167]

[0168] Where:

[0169] γ 1,0 is the initial adaptive adjustment coefficient;

[0170] α1 is the coefficient adjustment factor;

[0171] The input variables of the fuzzy logic controller are: frequency deviation, frequency change rate, fan speed deviation, wind speed deviation, and the fuzzy logic controller outputs a limit correction factor ΔP limit (t), and this correction factor is combined with the limit threshold:

[0172] P th,adjusted = P limit (t)+ΔP limit (t)

[0173] Through real-time calculation of fuzzy logic, ΔP limit (t) is automatically corrected according to the change of working conditions;

[0174] The compensation coefficient is corrected as follows:

[0175] θ adapt (t)=θ adapt (t - 1)+γ·(Δf(t)-Δf(t - 1))

[0176] Where γ is the learning rate;

[0177] S2, virtual inertia compensation module

[0178] Simulates the inertial response of a traditional generator set, provides additional inertia support during power grid frequency fluctuations, and ensures a smooth change in the power grid frequency; through a feedforward compensation algorithm, cancels the power reverse adjustment component generated by virtual inertia control, and decouples the coupling effect of primary frequency modulation and virtual inertia;

[0179] The specific method is as follows:

[0180] S21. Condition - Frequency Coupling Model

[0181] The condition - frequency coupling model is established as follows:

[0182]

[0183] Where:

[0184] H virtual (t) is the dynamic virtual inertia coefficient;

[0185]

[0186] Where:

[0187] H base is the basic virtual inertia coefficient;

[0188] k1 and k2 are the weight coefficients for adjusting the frequency change rate and frequency deviation;

[0189] R wind (t) is the relationship coefficient between the rotational speed of the wind turbine and the maximum output power, which is calculated based on multiple parameters:

[0190]

[0191] Where:

[0192] v max is the maximum wind speed;

[0193] ω max is the maximum rotational speed;

[0194] θ max is the maximum value of the pitch angle;

[0195] P max is the maximum power of the wind turbine;

[0196] α1, α2, α3, α4 are the adjustment coefficients;

[0197] S22. Dynamically Adjust the Feed - Forward Compensation Coefficient

[0198] Introduce the characteristics of the rotational speed of the wind turbine, pitch angle, and grid frequency fluctuation, and dynamically adjust the feed - forward compensation coefficient as follows:

[0199]

[0200] S23. Delay Compensation Correction

[0201] Introduce the delay function T delay Perform delay compensation correction as follows:

[0202]

[0203] S24. Coupling effect decoupling algorithm based on frequency-domain sensitivity analysis and dynamic optimization decoupling mechanism to achieve decoupling of the coupling effect between primary frequency regulation and virtual inertia

[0204] The specific method of step S24 is as follows:

[0205] S241. By analyzing the sensitivity of the wind turbine power response and the grid frequency change, obtain the interaction relationship between the primary frequency regulation power and the virtual inertia power in the frequency domain, and quantify the coupling degree between the two;

[0206] Assume that the frequency-domain response model for frequency change is G f (ω), G v (ω), then the coupling coefficient C(ω) of the system is expressed as:

[0207]

[0208] S242. Through frequency-domain analysis, establish a coupling matrix M between the primary frequency regulation power and the virtual inertia power, such that:

[0209]

[0210] S243. According to the real-time data of the grid frequency change rate, dynamically adjust the decoupling coefficient in the coupling matrix M;

[0211] S244. According to the real-time grid frequency deviation and the real-time operating state of the wind turbine, dynamically optimize the distribution ratio of the primary frequency regulation and the virtual inertia power. The optimization objective is to minimize the risk of power reverse peak shaving and frequency fluctuation, and optimize through the following objective function:

[0212]

[0213] Among them, w1 and w2 are weight coefficients;

[0214] S245. Through the real-time feedback regulation mechanism, dynamically correct the power distribution between the primary frequency regulation and the virtual inertia;

[0215] S3. Coupling effect modeling and decoupling

[0216] Establish a superposition model of the frequency regulation power, quantify the coupling coefficient using frequency-domain sensitivity analysis, and avoid power reverse peak shaving by jointly controlling the primary frequency regulation and the virtual inertia;

[0217] S4. Hierarchical coordinated control architecture

[0218] Upper decision-making layer: Dynamically allocate the frequency regulation weights of the virtual inertia and the droop control according to the grid frequency deviation and the operating state of the wind turbine;

[0219] Underlying execution layer: An adaptive amplitude limiting module is used to correct the power command to ensure smooth transition of the frequency modulation power and prevent it from exceeding the safe range.

[0220] S5. Multi-time scale coordination

[0221] Apply short-time window smoothing amplitude limiting to the virtual inertia link to avoid sudden changes in the power command. The time window for applying the short-time window is 10 - 100 ms; and smoothing control is also carried out within a longer time window, with the time window being 1 - 10 s.

[0222] S6. Anti-peaking suppression logic

[0223] During the frequency recovery stage, introduce reverse power gradient limitation to force the frequency modulation power to exit according to the exponential decay curve, avoiding power mutation. The specific method is as follows:

[0224] S61. Dynamically adjust the reverse power gradient limitation coefficient as follows:

[0225]

[0226] Among them,

[0227] κ1 and κ2 are adjustment coefficients;

[0228] S62. Introduce an adaptive feedback mechanism to adaptively adjust the reverse power gradient limitation coefficient according to the amplitude and rate of grid frequency fluctuations, changes in wind turbine response, and grid load status as follows:

[0229]

[0230] Among them:

[0231] λ0 is the initial reverse power gradient limitation coefficient;

[0232] α1, α2, and α3 are adaptive coefficients;

[0233] S63. Introduce an improved smooth exit algorithm to ensure smooth exit of the frequency modulation power. The exit curve adopts the following form:

[0234] P exit (t) = P max · exp(-α(t) · t) · (1 + β1 · ΔP load + β2 · ΔP wind )

[0235] Among them, α(t) is the exponential decay coefficient;

[0236]

[0237] Among them:

[0238] α0 is the initial attenuation coefficient;

[0239] γ1, γ2, and γ3 are adaptive coefficients.

[0240] Figure 1 The figure shows the strategy framework structure diagram of this embodiment, demonstrating the overall architecture of the control strategy of the present invention and the collaborative work among various modules. The core objective of this figure is to ensure that the wind turbine does not generate reverse power peaking during the primary frequency regulation process and maintain the stability of the power grid frequency through effective dynamic amplitude limiting control and virtual inertia compensation.

[0241] In this framework structure, first, the control strategy obtains input signals starting from the change in the power grid frequency, mainly including the power grid frequency deviation and the frequency change rate. These signals are processed in real time by the upper decision-making layer and serve as key parameters for adjusting the output power of the wind turbine. According to the real-time condition of the power grid and the current operating state of the wind turbine (such as rotational speed, pitch angle), the upper decision-making layer calculates the target value of the frequency regulation power and determines the weight allocation between virtual inertia and primary frequency regulation control.

[0242] Next, based on the calculated frequency regulation target, the system dynamically limits the frequency regulation command through the adaptive amplitude limiting module. This module adjusts the amplitude limiting threshold according to the real-time working conditions (including power grid frequency change, wind turbine operating state, etc.) to ensure that the power output of the wind turbine is within a reasonable range and avoid the reverse peaking phenomenon caused by excessive frequency regulation power output.

[0243] Meanwhile, the virtual inertia compensation module provides additional inertia support for the power grid by simulating the inertial response of traditional generator sets, alleviating the power grid frequency fluctuation. The virtual inertia module adopts a feed-forward compensation algorithm to ensure the collaborative work between virtual inertia compensation and primary frequency regulation power. This module not only smooths the power grid frequency fluctuation but also can make timely adjustments at specific frequency change stages to avoid power reverse peaking caused by over-response.

[0244] Finally, the entire control system adopts a hierarchical coordinated control architecture. The upper decision-making layer dynamically adjusts the weight allocation between virtual inertia and frequency regulation by comprehensively considering the power grid frequency deviation and the wind turbine operating state. The lower execution layer corrects and executes the frequency regulation command through the adaptive amplitude limiting module to ensure a smooth transition of the power output and avoid power fluctuations caused by too fast frequency changes.

[0245] As Figure 2The figure shows a schematic diagram of the feed - forward compensation process of the virtual inertia compensation module, demonstrating how virtual inertia provides inertia support to stabilize the grid frequency during grid frequency fluctuations. The virtual inertia compensation module simulates the inertial response of traditional generator sets. It provides additional inertia support through a feed - forward compensation algorithm at the initial stage of frequency fluctuations, thus effectively slowing down the drastic change of frequency. When the grid frequency drops, the virtual inertia module will respond immediately, releasing energy by simulating the inertia of the generator to alleviate the speed and amplitude of the frequency drop.

[0246] The core of the feed - forward compensation process lies in real - time predicting the grid frequency change and pre - adjusting the power. At the initial stage of the grid frequency drop, the system predicts the trend of the frequency change according to the frequency change rate and the current frequency deviation, and adjusts the power output of the fan in advance. In this way, virtual inertia can provide the necessary inertia support when the grid frequency has not fluctuated significantly, avoiding too rapid a frequency drop. During the recovery stage of frequency fluctuations, the virtual inertia compensation module gradually exits the compensation process by adjusting the power output to ensure that the grid frequency returns to a stable state.

[0247] This feed - forward compensation process not only enhances the response speed of virtual inertia but also avoids the lag of traditional inertial response. By predicting and adjusting the power output in advance, the feed - forward compensation can take effect at the initial stage of frequency fluctuations, thus ensuring a smoother change in grid frequency and avoiding instability caused by too rapid a frequency change.

[0248] As Figure 3 shown is the frequency - domain sensitivity analysis diagram of the combined effect of primary frequency regulation and virtual inertia in this embodiment, demonstrating the coupling relationship between primary frequency regulation and virtual inertia and its impact on the power output of the fan during grid frequency fluctuations. Through frequency - domain sensitivity analysis, this figure quantifies the coupling coefficient between the two and further reveals how to optimize the interaction between frequency regulation and inertia compensation under different frequency changes to maximize the frequency - regulation success rate and prevent reverse power peaking.

[0249] In frequency - domain sensitivity analysis, it is first necessary to model the response characteristics of primary frequency regulation and virtual inertia and analyze their influence on grid frequency changes. On the one hand, primary frequency regulation directly responds to the deviation of the grid frequency and adjusts the power output of the fan to stabilize the frequency; on the other hand, virtual inertia provides an inertial response similar to that of traditional generator sets, alleviating frequency fluctuations, especially when the frequency changes rapidly, by increasing inertia support to smooth the frequency change. There is a strong coupling relationship between the two, and the power response of frequency regulation and the power support of virtual inertia jointly act on the change of grid frequency.

[0250] When performing frequency-domain sensitivity analysis, the response speeds and amplitudes of both at different frequencies are considered. The response of primary frequency modulation is usually fast and can adjust immediately according to the grid frequency deviation; while the response of virtual inertia is relatively slow, but it has a strong smoothing effect. Through the frequency-domain sensitivity function, the analysis quantifies the coupling degree between these two responses, thus providing a theoretical basis for jointly regulating the power outputs of both. Through frequency-domain sensitivity analysis, the optimal coupling coefficient can be determined to ensure that when the grid frequency fluctuates, the frequency modulation and inertia response of the wind turbine can act synergistically, without generating reverse power peak shaving, and ensuring the stability of the system.

[0251] This analysis diagram further emphasizes how to dynamically adjust the weight distribution of primary frequency modulation and virtual inertia through real-time monitoring of frequency deviation and rate of change, so that the wind turbine can maintain the optimal frequency modulation effect under various frequency change conditions and minimize the sudden change of power output.

[0252] As Figure 4 shown is the schematic diagram of the reverse power peak shaving suppression logic in this embodiment, which shows how the wind turbine frequency modulation system suppresses the power reverse peak shaving phenomenon caused by the coupling effect of primary frequency modulation and virtual inertia. When the grid frequency fluctuation gradually recovers and the system needs to avoid power reverse peak shaving to prevent excessive oscillation or over-adjustment of the grid frequency leading to instability.

[0253] In the frequency recovery stage (i.e., when the rate of change of frequency changes from negative to positive), when the system detects that the grid frequency has started to recover, the reverse power peak shaving suppression logic will be enabled. At this time, in order to avoid excessive power feedback or sudden change of power, the control system will introduce a reverse power gradient limit to force the frequency modulation power to exit according to a preset exponential decay curve. This control mechanism ensures that the power of the wind turbine gradually decreases and there will be no frequency oscillation caused by excessive frequency modulation.

[0254] Specifically, the reverse power peak shaving suppression logic will monitor the speed of frequency change during the grid frequency recovery process and limit the frequency modulation power of the wind turbine. The reverse power peak shaving suppression logic dynamically adjusts the exit strategy according to the power change curve, so that the frequency modulation power of the wind turbine can exit smoothly, avoiding sudden change of power command and excessive rebound of frequency.

[0255] Through this suppression logic, the power output during the frequency modulation process no longer suddenly stops or increases excessively, thus avoiding excessive frequency fluctuation and ensuring the stable recovery of the grid frequency. At the same time, the stability of the system is also enhanced, avoiding potential adverse effects on the grid caused by frequent adjustment and over-adjustment.

[0256] The foregoing has shown and described the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and what is described in the above embodiments and the specification is only to illustrate the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for preventing power reverse peaking by coupling virtual inertia in the primary frequency regulation of a network-forming fan, characterized in that, It includes the following steps: S1. Adaptive dynamic amplitude limiting control; Based on the frequency deviation, the rate of change of frequency, the current operating state of the wind turbine, and the wind speed prediction data, the dynamic amplitude limiting threshold is calculated in real time through fuzzy logic or dynamic weight algorithm, and the amplitude limiting threshold and compensation coefficient are corrected online by using a fuzzy logic controller; S2. Virtual inertia compensation module; Simulate the inertial response of traditional generator sets, provide additional inertia support when the grid frequency fluctuates, and ensure the smooth change of the grid frequency; through the feed-forward compensation algorithm, cancel the power reverse regulation component generated by the virtual inertia control, and decouple the coupling effect of primary frequency modulation and virtual inertia; S3. Coupling effect modeling and decoupling; Establish a superposition model of frequency modulation power, quantify the coupling coefficient by using frequency domain sensitivity analysis, and avoid power reverse regulation peaks by jointly controlling primary frequency modulation and virtual inertia; S4. Hierarchical coordinated control architecture Upper decision-making layer: Dynamically allocate the frequency modulation weights of virtual inertia and droop control according to the grid frequency deviation and the operating state of the wind turbine; Lower execution layer: Use an adaptive amplitude limiting module to correct the power command to ensure the smooth transition of the frequency modulation power and not exceed the safe range; S5. Multi-time scale coordination; Apply a short-time window smoothing amplitude limit to the virtual inertia link to avoid sudden changes in the power command; S6. Anti-reverse regulation peak suppression logic; In the frequency recovery stage, introduce a reverse power gradient limit to force the frequency modulation power to exit according to the exponential decay curve, and avoid sudden changes in power.

2. The method for preventing power reverse peaking by coupling virtual inertia in the primary frequency regulation of a network-forming fan is characterized in that, The dynamic amplitude limiting threshold is calculated as follows: Where: P limit (t) is the dynamic amplitude limiting threshold; P wind (t) is the output power of the fan; is the rate of change of the power grid frequency; Δf(t) is the grid frequency deviation; ΔP load (t) is the change in the grid load; ΔP wind (t) is the fluctuation amount of the fan power; γ1(t) is the adaptive adjustment coefficient; Where: γ 1,0 is the initial adaptive adjustment coefficient; α1 is the coefficient adjustment factor; The input variables of the fuzzy logic controller are: frequency deviation, frequency change rate, fan speed deviation, and wind speed deviation. The fuzzy logic controller outputs a limit amplitude correction factor ΔP limit (t), and this correction factor is combined with the limit amplitude threshold: P th,adjusted = P limit (t) + ΔP limit (t) Through real-time calculation by fuzzy logic, ΔP limit (t) is automatically corrected according to the change of working conditions; The compensation coefficient is corrected as follows: θ adapt (t) = θ adapt (t - 1) + γ·(Δf(t) - Δf(t - 1)) Where γ is the learning rate.

3. The method for preventing power reverse peak regulation by coupling virtual inertia in the primary frequency regulation of a network-forming fan according to claim 1, wherein, The specific method of step S2 is as follows: S21. Operating condition-frequency coupling model The operating condition-frequency coupling model is modeled as follows: Where: H virtual (t) is the dynamic virtual inertia coefficient; Where: H base Base virtual inertia coefficient; k1 and k2 are the weight coefficients for adjusting the rate of change of frequency and the frequency deviation; R wind (t) is the relationship coefficient between the rotational speed of the fan and the maximum output power, and is calculated based on multiple parameters: Where: v max is the maximum wind speed; ω max is the maximum rotational speed; θ max is the maximum value of the pitch angle; P max is the maximum power of the fan; α1, α2, α3, and α4 are the adjustment coefficients; S22. Dynamically adjust the feed-forward compensation coefficient Introduce the characteristics of the wind turbine speed, pitch angle, and grid frequency fluctuation, and dynamically adjust the feed-forward compensation coefficient as follows: S23. Delay compensation correction Introduce the delay function T delay Perform delay compensation correction as follows: S24. A coupling effect decoupling algorithm based on frequency domain sensitivity analysis and dynamic optimization decoupling mechanism is used to decouple the coupling effect of primary frequency modulation and virtual inertia.

4. The method for preventing power reverse peak regulation by coupling virtual inertia in the primary frequency regulation of a network-forming fan according to claim 3, characterized in that The specific method of step S24 is as follows: S241. By analyzing the sensitivity of the wind turbine power response and the grid frequency change, obtain the interaction relationship between the primary frequency modulation power and the virtual inertia power in the frequency domain, and quantify the coupling degree between the two; Assume that the frequency-domain response models with frequency variations are G f (ω) and G v (ω), then the coupling coefficient C(ω) of the system is expressed as: S242. Through frequency domain analysis, establish a coupling matrix M between the primary frequency modulation power and the virtual inertia power, such that: S243. According to the real-time data of the grid frequency change rate, dynamically adjust the decoupling coefficient in the coupling matrix M; S244. According to the real-time grid frequency deviation and the real-time operating state of the wind turbine, dynamically optimize the distribution ratio of the primary frequency modulation and virtual inertia power. The optimization goal is to minimize the risk of power reverse regulation peak and frequency fluctuation, and the optimization is carried out through the following objective function: Where w1 and w2 are the weight coefficients; S245. Dynamically correct the power distribution between primary frequency regulation and virtual inertia through a real-time feedback adjustment mechanism.

5. The method for preventing power reverse peak regulation by coupling virtual inertia in the primary frequency regulation of a network-forming fan according to claim 1, characterized in that With a multi-time scale coordinated control mechanism, the frequency modulation response of the wind turbine can be divided into short-term and long-term responses: the short-term response is that virtual inertia control and droop control give priority to handling the rapid changes in grid frequency; the long-term response is that primary frequency modulation control handles the long-term fluctuations in grid frequency. The time window with a short time window applied in step S5 is 10 - 100 ms; and smooth control is also carried out within a longer time window, with the time window being 1 - 10 s. By optimizing the allocation of time scales, the frequency modulation accuracy of the system can be improved, and the negative impact caused by frequent adjustment of frequency modulation power in a short time can be avoided.

6. The method for preventing power reverse peaking by coupling virtual inertia in the primary frequency regulation of a network-forming fan according to claim 1, wherein The implementation of the decoupling algorithm can be divided into the following steps: (1) Real-time collect grid frequency data: Obtain grid frequency deviation and frequency change rate data. (2) Wind turbine status monitoring: Real-time monitor the rotational speed, pitch angle, and power output of the wind turbine. (3) Calculate the coupling matrix: Calculate the coupling matrix M based on frequency domain sensitivity analysis and dynamically adjust it according to the grid frequency change and wind turbine status. (4) Optimize power distribution: Calculate the optimal power distribution strategies P f1 (t) and P v1 (t); (5) Output the decoupled frequency modulation command: Generate a frequency modulation power command optimized after decoupling to ensure the stability and frequency modulation accuracy of the system. Among them, according to the real-time data of the grid frequency change rate, dynamically adjust the decoupling coefficient in the coupling matrix M; when the frequency changes violently, the decoupling coefficient M will be automatically adjusted so that the responses of primary frequency regulation and virtual inertia do not interfere with each other, avoiding power reverse peaking caused by too fast response; according to the real-time grid frequency deviation and the real-time operating status of the wind turbine, dynamically optimize the power distribution between primary frequency regulation and virtual inertia; by introducing a modular machine learning algorithm, online adjust the decoupling coefficient of the system to optimize the allocation ratio between primary frequency regulation and virtual inertia. Assuming the optimization goal is to minimize the risk of power reverse peaking and frequency fluctuations, it can be optimized through the following objective function: where: w1 and w2 are weight coefficients, representing the weights of frequency deviation and power reverse peaking respectively.

7. The method for preventing power reverse peak shaving by coupling virtual inertia in the primary frequency regulation of a network-forming fan according to claim 1, characterized in that The specific method of step S6 is as follows: S61. Dynamically adjust the reverse power gradient limit coefficient as follows: where, κ1 and κ2 are adjustment coefficients. S62. Introduce an adaptive feedback mechanism to adaptively adjust the reverse power gradient limit coefficient according to the amplitude and rate of grid frequency fluctuations, the changes in wind turbine response, and the grid load status as follows: where: λ0 is the initial reverse power gradient limit coefficient. α1, α2, and α3 are adaptive coefficients. S63. Introduce an improved smooth exit algorithm to ensure the smooth exit of frequency modulation power, and the exit curve adopts the following form: P exit P(t) = P max · exp(-α(t)·t)·(1 + β1·ΔP load + β2·ΔP wind ) where, α(t) is the exponential decay coefficient. where: α0 is the initial decay coefficient. γ1, γ2, and γ3 are adaptive coefficients.

Images