Multi-time scale coordinated network-type wind storage system low voltage ride through control method

Abstract

The present application relates to a kind of multi-time scale collaborative network type wind storage system low voltage ride through control method, belong to wind power generation control technical field.The method is first based on reactive power priority of network converter, current limiting amplitude constraint calculation maximum active power can send out ability, accurate quantification remaining unbalanced power;Again, according to the layered logic distribution power of millisecond level buffer of DC capacitor, rotor kinetic energy hundred milliseconds level absorption, super capacitor second level energy compensation, serious fault linkage DC discharge branch is consumed remaining energy;After fault clearance, collaborative control machine side power recovery, super capacitor SOC fall and DC voltage return to home.The method of the present application can effectively inhibit multi-index overrun, improve system fault ride through capability and recovery stability.

Description

Technical Field

[0001] This invention belongs to the field of wind power generation control technology, specifically relating to a low voltage ride-through control method for a multi-timescale coordinated grid-type wind-storage system. Background Technology

[0002] With the high proportion of renewable energy integrated into the power system, the shift from traditional grid-following control to grid-building control of renewable energy units has become an important technical path to improve the system's voltage support capability, inertial response capability, and transient stability level. Grid-building wind turbines can actively establish voltage amplitude and frequency through virtual synchronous machines, virtual impedance, and droop control, providing voltage support for weak grids and grids with a high proportion of renewable energy. However, during low-voltage faults, grid-building converters not only need to maintain voltage support and reactive current injection, but are also subject to the combined constraints of total current limiting, DC bus voltage safety boundary, and unit mechanical speed constraints. At this time, the active power output capability of the grid-side converter decreases significantly, while the mechanical input power of the wind turbine cannot be reduced instantaneously, resulting in a strong coupling power imbalance between turbine-side input power, grid-side output power, rotor kinetic energy, and energy storage absorption capacity.

[0003] For direct-drive wind power systems, power imbalance under low-voltage faults typically manifests first as an increase in DC bus voltage. When the grid-side voltage drops significantly, the grid-side converter is constrained by both voltage amplitude and current limiting, resulting in a sharp decrease in its active power output capacity. If the turbine-side converter continues to operate in maximum power point tracking mode, excess energy will be rapidly injected into the DC bus, causing DC bus overvoltage. If the turbine-side power is simply reduced significantly to convert the input energy of the wind turbine into rotor kinetic energy, it may cause rotor overspeed and increased mechanical stress in the transmission chain. If the system relies entirely on energy storage devices to absorb surplus energy, it will result in an oversized energy storage capacity and the risk of supercapacitor overcharging under severe faults or high SOC conditions. Therefore, the core of low-voltage ride-through control should not be limited to single-variable DC voltage suppression, but rather to multi-timescale energy coordination among the source side, DC side, energy storage side, and protection branches under grid current limiting constraints.

[0004] Existing low-voltage ride-through control methods mainly include generator-side power reduction control, DC bus voltage control, energy storage and energy absorption control, and DC discharge protection. Generator-side power reduction methods can reduce the power entering the DC bus from the source side, but their control effect is highly dependent on the rotor speed margin. When the fault is deep or lasts for a long time, relying solely on rotor kinetic energy to absorb unbalanced energy can easily cause the rotor speed to exceed the limit. Energy storage and energy absorption methods can absorb surplus power on the DC side through batteries, supercapacitors, or hybrid energy storage devices, but if the transient buffering capacity of rotor kinetic energy and DC capacitors is not considered, the energy storage device often bears excessive energy absorption pressure, resulting in a conservative capacity configuration. DC discharge branches can quickly consume surplus energy under extreme overvoltage conditions, but if used only as passive protection devices, problems such as delayed activation, frequent switching, or low energy utilization efficiency are likely to occur.

[0005] More importantly, existing methods typically equate the power imbalance during low-voltage faults simply to the difference between generator-side power and grid-side output power, or roughly estimate the degree of active power output limitation based on the voltage drop at the grid connection point. This approach ignores the priority of reactive power support and total current limiting constraints of grid-connected converters during faults. In fact, grid-connected converters need to prioritize reactive current support during low-voltage faults before any remaining current margin can be used for active power output. Therefore, under severe low-voltage or asymmetrical faults, the maximum active power output capacity of the grid side depends not only on the grid connection point voltage amplitude but also on the reactive current command, total current limiting, negative sequence voltage, and virtual impedance current limiting strategies. If this active power output capacity cannot be accurately predicted, subsequent generator-side power reduction and supercapacitor energy absorption commands may deviate.

[0006] Furthermore, existing wind-storage coordinated control methods mostly employ fixed-ratio allocation, SOC-based weight adjustment, or single-variable correction based on speed margin. While these methods can improve DC bus voltage control to some extent, they still have three shortcomings: First, they do not link the remaining fault time with the real-time available energy margin of each energy buffer unit, making it difficult to determine whether current control resources are sufficient to cover subsequent fault energy; second, they do not establish a hierarchical coordination relationship between rotor kinetic energy, DC capacitors, supercapacitors, and discharge branches, resulting in unclear control priorities under severe faults; third, the recovery process after fault clearance is usually insufficient, lacking coordination between rapid power recovery on the turbine side, SOC recovery of supercapacitors, and active power ramp-up on the grid side, which can easily induce secondary overvoltage and power backlash on the DC bus. Summary of the Invention

[0007] The purpose of this invention is to address the problems of limited active power transmission on the grid side, DC bus overvoltage, rotor overspeed, and supercapacitor overcharging under low-voltage faults. It provides a multi-timescale collaborative low-voltage ride-through control method for grid-connected wind-storage systems. First, under the constraints of reactive power priority and total current limiting of the grid-connected converter, the maximum active power transmission capacity is calculated, and a residual unbalanced power prediction model during the fault period is established using negative sequence correction. Second, an available energy margin model for rotor kinetic energy, DC capacitor, and supercapacitor is established. Unbalanced energy is predicted based on the remaining fault time, and a constrained effective energy margin risk factor considering power limiting, time scale, and safety boundary is constructed to determine whether conventional energy absorption resources are sufficient to support fault ride-through. Third, a constraint-aware power reduction control is proposed at the turbine side, dynamically adjusting the turbine-side power reference value according to fault depth, DC bus voltage, and speed margin. At the energy storage side, adaptive power allocation for supercapacitors and adaptive DC bus reference voltage rise control are proposed, enabling supercapacitors to primarily bear the residual unbalanced energy after rotor energy absorption and DC capacitor buffering. Furthermore, in cases of severe faults or insufficient energy margin, a DC discharge branch is introduced as the final protection unit. This branch is jointly triggered based on predicted energy margin, DC bus voltage, supercapacitor SOC, and rotor speed, achieving coordinated operation between the supercapacitor and the discharge branch. Finally, for the recovery phase after fault clearance, a collaborative recovery control system is proposed, encompassing rotor kinetic energy release, supercapacitor SOC recovery, and grid-side active power ramping constraints, to reduce secondary impacts during the recovery process.

[0008] To achieve the above objectives, the technical solution of the present invention is: a low-voltage ride-through control method for a multi-timescale coordinated grid-type wind and energy storage system, applied to a grid-type wind and energy storage system including a grid-type direct-drive wind turbine, a DC bus, a supercapacitor, and a DC discharge branch, comprising:

[0009] Real-time acquisition of grid connection point voltage, grid-side converter output current, DC bus voltage, rotor speed, supercapacitor SOC and terminal voltage;

[0010] When the grid connection point voltage is lower than the low voltage ride-through threshold, the maximum active power transmission capacity of the grid side is calculated based on the reactive power priority support and total current limiting constraint of the grid-side converter.

[0011] The remaining unbalanced power on the DC side is calculated based on the power injected into the DC bus on the machine side, the maximum active power output capacity on the grid side, and the transient power that the DC capacitor can bear.

[0012] Calculate the real-time available energy margins of rotor kinetic energy, DC capacitor, and supercapacitor respectively, and combine them with the total unbalanced energy in the prediction time domain to calculate the risk factor of constrained effective energy margin.

[0013] Based on the constrained effective energy margin risk factor and the real-time available energy margin of each energy unit, the target power of rotor kinetic energy, DC capacitor and supercapacitor is allocated according to the multi-time scale division of rotor kinetic energy absorption at the millisecond level, DC capacitor transient buffer at the millisecond level and supercapacitor residual compensation at the second level, and the power reference value of the machine-side converter is adjusted based on the allocation results.

[0014] Based on the target power to be carried by the supercapacitor and the DC bus voltage deviation, a reference power for the supercapacitor is generated, and the operation of the supercapacitor is controlled by a bidirectional DC / DC converter.

[0015] When the constraint effective energy margin risk factor reaches the preset input threshold, or when the DC bus voltage, supercapacitor SOC and rotor speed simultaneously reach the corresponding warning threshold, the DC discharge branch is started. The target power of the DC discharge branch is determined based on the difference between the remaining unbalanced power and the maximum energy absorption capacity of each energy unit, and its switching is controlled.

[0016] After the fault is cleared, based on the rotor speed deviation, DC bus voltage deviation, and the active power transmission capacity after grid-side recovery, the power reference value of the generator-side converter, the SOC recovery power of the supercapacitor, and the DC bus reference voltage are adjusted in a coordinated manner until the system returns to normal operation.

[0017] Furthermore, the calculation process for the maximum active power transmission capacity on the grid side includes:

[0018] Let the maximum allowable current of the grid-side converter be I. g,max The per-unit value of the positive sequence voltage at the grid connection point is U1, and the reactive power support factor is k. q During a low-voltage fault, the reactive current reference value shall be determined according to the reactive current priority principle:

[0019] i q * = sat[k q (0.9 - U1), -I g,max , I g,max ]

[0020] In the formula, sat[·] represents the amplitude limiting function;

[0021] Under the condition of satisfying the total current constraint, the maximum active current that the grid-side converter can use to transmit active power is:

[0022] i d,max =

[0023] Therefore, the maximum active power transmission capacity of the grid side under a three-phase symmetrical fault is expressed as:

[0024] P g,max = (3 / 2)U1·id,max

[0025] When asymmetrical low-voltage faults such as single-phase grounding or two-phase short circuits occur, negative sequence voltage can cause current imbalance and instantaneous power double-frequency oscillation. To avoid overestimating the grid-side active power transmission capacity, a negative sequence correction factor is introduced.

[0026] χ² = 1 - k²·U² / (U₁ + ε)

[0027] In the formula, U2 is the magnitude of the negative sequence voltage at the grid connection point, k2 is the negative sequence influence coefficient, and ε is a very small positive number to prevent the denominator from being zero; the maximum active power transmission capacity of the grid side after considering the influence of the negative sequence voltage is:

[0028] P g,max ' = χ2·P g,max .

[0029] Furthermore, the calculation process for the residual unbalanced power on the DC side includes:

[0030] The transient power that the DC capacitor can handle is defined based on the current DC voltage margin:

[0031] P dc allow = C dc ·U dc (U dc,max - U dc ) / T dc

[0032] In the formula, T dc For the DC capacitor to participate in the equivalent time constant, C dc U is the capacitance value of the DC capacitor. dc U is the real-time voltage of the DC bus. dc,max This is the maximum allowable voltage for a DC capacitor;

[0033] The remaining unbalanced power that needs to be borne by rotor kinetic energy, supercapacitors, or DC discharge branches during the fault is:

[0034] ΔP rem = [P s - P g,max ' - P dc allow ] +

[0035] In the formula, P s The power injected into the DC bus of the machine-side converter, [·] + This indicates a non-negative truncation function, meaning that the value inside the parentheses is zero when it is less than zero.

[0036] Furthermore, the calculation process for the real-time available energy margin of the rotor kinetic energy, DC capacitor, and supercapacitor includes:

[0037] The available kinetic energy margin of the rotor at the current speed is:

[0038] E rot av = 0.5J(ω r,max 2 -ω r 2 )

[0039] Where J is the moment of inertia of the wind turbine, ω r,max ω is the maximum permissible speed of the rotor. r This refers to the real-time rotor speed.

[0040] The available voltage energy margin of the DC capacitor under the current DC voltage is:

[0041] E dc av = 0.5C dc (U dc,max 2 - U dc 2 )

[0042] Where C dc U is the capacitance value of the DC capacitor. dc U is the real-time voltage of the DC bus. dc,max This is the maximum allowable voltage for a DC capacitor;

[0043] The available energy absorption margin of a supercapacitor at the current terminal voltage or SOC state is:

[0044] E sc av = 0.5C sc (U sc,max 2 - U sc 2 )

[0045] Where C sc U is the equivalent capacitance of a supercapacitor. sc,max U is the maximum allowable terminal voltage of the supercapacitor. sc This represents the real-time terminal voltage of the supercapacitor.

[0046] Furthermore, the calculation process for the constrained effective energy margin risk factor includes:

[0047] For any energy unit x, x∈{rot, dc, sc}, where rot, dc, and sc represent the rotor kinetic energy, DC capacitance, and supercapacitor, respectively, we first define the achievable absorbed energy of x after being constrained by power limiting:

[0048] E x lim (t) = min{ E x av (t), P x,max abs (t)·T_p}

[0049] In the formula, E x av (t) represents the theoretical available energy margin for the corresponding energy unit, P x,max abs (t) represents the maximum allowable absorbed power under the current state, and T_p represents the prediction time domain;

[0050] Considering the physical response speed of different energy units, a time-scale correction factor is defined:

[0051]

[0052] in, For the duration of the fault, T dc T rot and T sc The DC capacitor, rotor kinetic energy, and supercapacitor are allowed to participate in the equivalent time constant, respectively, satisfying T. dc < T rot < T sc ;

[0053] To ensure that risk factors reflect the approaching safety boundaries of rotor speed, DC bus voltage, and supercapacitor SOC in advance, a safety boundary correction factor is introduced:

[0054] θ x (t) = {sat[(z x,max - z x (t)) / (z x,max - z x,warn + ε), 0, 1]} m_x

[0055] z rot =ω r , z dc =U dc , z sc =SOC sc

[0056] In the formula, z x,warn For the warning threshold, zx,max For safety, m_x is the nonlinear penalty coefficient, sat[·] represents the amplitude limiting function, ε is a very small positive number to prevent the denominator from being zero, and ω r For real-time rotor speed, U dc This represents the real-time voltage of the DC bus, SOC. sc It is the ratio of the actual amount of charge currently stored in the supercapacitor to its rated maximum charge storage capacity;

[0057] The constraint-effective available energy margin of energy unit x is defined as:

[0058] E x eff (t)=η x ·μ x (t)·θ x (t)·E x lim (t)

[0059] In the formula, η x Energy utilization coefficient or conversion efficiency;

[0060] To avoid simply integrating the power difference and ignoring the severity of the fault and the prediction error on the side of predicting unbalanced energy, a fault severity correction coefficient is defined:

[0061] ρ(τ)=1+k u [1-U1(τ)] + +k2U2(τ) / (U1(τ)+ε)+k i |i q *(τ)| / I g,max

[0062] In the formula, U1(τ) and U2(τ) are the positive-sequence and negative-sequence voltage amplitudes at the prediction time, respectively, i q *(τ) is the reference value for reactive current, I g,max The maximum allowable current of the grid-side converter is k. u k2, k i These are the voltage sag compensation coefficient, negative sequence influence coefficient, and reactive current limiting adjustment coefficient, respectively; ρ(τ) increases with the deepening of voltage sag, the increase of negative sequence component, or the increase of reactive current occupancy ratio.

[0063] E Σ risk (t)=∫ t t+T_p ρ(τ)[P w pre (τ)-P g,max pre' (τ)] + dτ + k σ σE (t)

[0064] Among them, E Σ risk (t) represents the risk-corrected predicted imbalance energy, P w pre (τ) represents the wind turbine input mechanical power in the predicted time domain, P g,max pre' (τ) represents the maximum active power transmission capacity of the network side in the predicted time domain, σ E (t) represents the prediction error or uncertainty margin, k σ To predict the conservative coefficient, [·] + Represents a non-negative truncation function;

[0065] Meanwhile, to avoid masking the situation where the overall system margin is sufficient but individual energy units saturate prematurely, we define the energy allocation requirements and local bottleneck risk terms for each energy unit in the prediction time domain:

[0066] E x req (t)=∫ t t+T_p λ x ' (τ)[P w pre (τ)-P g,max pre' (τ)] + dτ

[0067] R b (t)=max x∈{rot,dc,sc} { E x req (t) / (E x eff (t)+ε)}

[0068] Among them, E x req (t) represents the energy required for the allocation of energy unit x in the predicted time domain, λ x ' (τ) represents the target-sharing weight of energy unit x in the prediction time domain, R b (t) represents the risk of the weakest local unit;

[0069] The real-time security boundary approach risk item is defined as:

[0070] B(t) = max{B dc (t), B rot (t), B sc (t)}

[0071] B dc=[(U dc -U dc,warn ) / (U dc,max -U dc,warn +ε)] +

[0072] B rot =[(ω r -ω warn ) / (ω r,max -ω warn +ε)] +

[0073] B sc =[(SOC sc -SOC warn ) / (SOC max -SOC warn +ε)] +

[0074] In the formula, B(t) represents the total effective margin risk of the system. dc (t), B rot (t), B sc (t) represents the boundary approximation of the DC bus voltage, the rotor speed, and the supercapacitor SOC, respectively. dc,warn ω is the DC bus voltage warning threshold. warn The rotor speed warning threshold, SOC warn The SOC (State of Charge) warning threshold for supercapacitors. max This refers to the maximum charge that a supercapacitor can store.

[0075] Ultimately, the constrained effective energy margin risk factor is defined as:

[0076] R E c (t)=αE Σ risk (t) / (E rot eff +E dc eff +E sc eff +ε)+βR b (t)+γB(t)

[0077] α+β+γ=1, α≥0, β≥0, γ≥0

[0078] In the formula, α, β, and γ are the total margin risk weighting coefficient, the local bottleneck risk weighting coefficient, and the boundary approach risk weighting coefficient, respectively.

[0079] Furthermore, the process of allocating the target's power includes:

[0080] Calculate the initial weighting λ for rotor kinetic energy, DC capacitor, and supercapacitor. x = E x av / (E rot av + E dc av +E sc av + ε), x ∈ {rot, dc, sc}, where rot, dc, and sc represent the rotor kinetic energy, DC capacitance, and supercapacitor, respectively, and E x av ε represents the real-time available energy margin of energy unit x, where ε is a minimum positive number divisible by zero;

[0081] Adjust the initial weighting by time scale: λ rot ' = , λ dc ' = , λ sc ' = 1 - λ rot ' - λ dc ' ,in For the duration of the fault, T dc T rot and T sc These are the allowable participation time constants for DC capacitor, rotor kinetic energy, and supercapacitor, respectively.

[0082] The target power to be undertaken by each energy unit is the product of the corresponding corrected weight and the remaining unbalanced power.

[0083] Furthermore, the adjustment process for the power reference value of the machine-side converter includes:

[0084] The power reference value during normal operation is the maximum power point tracking reference value P. s0 * = P MPPT = k opt ω r 3 , where k opt For the optimal torque coefficient, ω r This refers to the real-time rotor speed.

[0085] After fault switching, the reference power value on the machine side is P. s * = P s0 * - ΔPMSC * ΔP MSC * This is an adaptive power reduction command, expressed as ΔP. MSC * = K m σ u σ ω P N + K dc (U dc - U dc * ), where K m σ is the adjustment coefficient for the coupling of fault depth and rotor margin. u σ is the fault voltage drop factor. ω P is the rotor speed margin coefficient. N K is the rated power of the unit. dc U is the DC voltage feedback coefficient. dc U is the real-time voltage of the DC bus. dc * This is the DC bus reference voltage;

[0086] The power reduction command is limited between 0 and the maximum permissible power reduction value, where the maximum permissible power reduction value is ΔP. MSC,max = min[J(ω r,max 2 - ω r 2 ) / (2t f ), ΔP conv,max , P s0 * - P s,min ], where J(ω r,max 2 -ω r 2 ) / (2t f ΔP represents the maximum average power that the rotor is allowed to absorb during the remaining fault time. conv,max Due to the capacity constraint of the machine-side converter, P s0 * - P s,min For the minimum power constraint on the turbine side, J is the moment of inertia of the wind turbine, and ω r,max This is the maximum permissible speed of the rotor. For the duration of the fault, P s,min This is the minimum operating power on the machine side;

[0087] The reference value for the electromagnetic torque on the machine side is T. e * = P s * / ω rAnd the rate of change of torque is limited to a preset upper limit R. T within;

[0088] After the fault is cleared, the turbine-side power reference value is set to dP. s * / dt = [(P MPPT - P s * ) / T p ] · [1 - (U dc -U dc * ) / (U dc,max - U dc * The regularity of T] smoothly recovers to the maximum power tracking reference value. p This is the power recovery time constant on the machine side.

[0089] Furthermore, the reference power generation process of the supercapacitor includes:

[0090] The supercapacitor reference power is composed of the superposition of the feedforward component and the feedback compensation component: ,in To provide power for the supercapacitor target. The output of the DC voltage outer-loop PI controller is expressed as follows: , , These are the proportional and integral coefficients of the outer loop DC voltage, respectively. This is the DC bus reference voltage;

[0091] During a fault, the DC bus reference voltage adaptively increases with the depth of voltage drop at the grid connection point, with a target increase amount. ,in This represents the rise in the DC bus reference voltage for every 0.1 pu decrease in the grid connection point voltage. The target rise is the per-unit value of the grid connection point voltage. After being smoothed by a first-order inertial element, it is superimposed on the rated DC bus reference voltage, and the upper limit of the rise is 1.2 times the rated value.

[0092] After the fault is cleared, the DC bus reference voltage is adjusted according to... The regular pattern of smooth return to the rated value , This is the DC bus reference voltage. This is the reference voltage recovery time constant.

[0093] Furthermore, the control process of the DC discharge branch includes:

[0094] The target power of the DC discharge branch is P dump *= sat(ΔP rem - P sc,max - P rot,max - P dc allow , 0,P dump,max ), where ΔP rem For the remaining unbalanced power, P sc,max P represents the maximum allowable power absorbed by the supercapacitor. rot,max P is the maximum allowable energy absorption power of the rotor. dc allow P is the transient power that the DC capacitor can withstand. dump,max This is the maximum power output of the unloading branch.

[0095] The PWM duty cycle of the DC discharge branch is D dump = sat[P dump * R dump / U dc 2 [, 0, 1],R dump U is the equivalent resistance of the discharge resistor. dc The real-time voltage of the DC bus is given by sat[·], which represents the limiting function.

[0096] The exit condition for the DC discharge branch is that the DC bus voltage is lower than the preset exit threshold U. dump,off Furthermore, the effective energy margin risk factor is constrained to be lower than the preset exit threshold R. off And satisfy U dump,off < U dump,on R off < R on U dump,on For the DC discharge branch, the input voltage threshold, R on To constrain the input threshold of effective energy margin risk factors.

[0097] Furthermore, the collaborative adjustment process during the fault recovery phase includes:

[0098] The reference value for engine-side power is P s * = P MPPT + K ω,rec (ω r - ω r0 ) - K dc,rec (U dc - U dc * ), where K ω,rec ω is the rotor overspeed recovery coefficient. r0 K represents the rotor speed before the fault. dc,recThe DC bus voltage suppression coefficient is given, and the generator-side power change rate is limited to a preset recovery upper limit R. P,rec within;

[0099] The SOC recovery power of the supercapacitor is P sc,rec * = -sat[K soc (SOC sc - SOC ref ), 0, P sc,dis,max ], where K soc SOC adjustment coefficient, SOC ref For the target state of charge, SOC sc P is the ratio of the actual amount of charge currently stored in a supercapacitor to its rated maximum storage capacity. sc,dis,max Let P be the maximum allowable discharge power of the supercapacitor, and let sat[·] represent the limiting function. sc,rec * Satisfy P sc,rec * ≤[P g,maxrec - P s * ]+, P g,maxrec This represents the maximum active power that can be transmitted from the grid side after the fault is cleared.

[0100] When the DC bus voltage deviation, rotor speed deviation, and supercapacitor SOC deviation are all less than the corresponding preset thresholds, the system switches to normal operation mode, the generator-side converter resumes MPPT control, the supercapacitor resumes normal SOC management, and the grid-side converter resumes normal grid operation.

[0101] Compared with the prior art, the present invention has the following beneficial effects:

[0102] 1. A method for predicting the maximum active power transmission capacity of the grid side under the constraint of grid current limiting is proposed. The method incorporates the reactive power priority support, total current limiting and negative sequence voltage influence into the low voltage ride-through energy distribution calculation, which improves the accuracy of unbalanced power estimation during faults.

[0103] 2. Establish a multi-layer energy margin coordination model for rotor kinetic energy, DC capacitor, supercapacitor and DC discharge branch, and propose a constrained effective energy margin risk factor based on residual fault time prediction to achieve adaptive allocation of unbalanced energy among buffer units at different time scales.

[0104] 3. A constraint-sensing type machine-side power reduction and supercapacitor adaptive energy absorption control strategy is proposed to form a closed-loop coordination between rotor kinetic energy priority absorption, DC capacitor transient buffering and supercapacitor residual compensation, so as to avoid relying solely on machine-side power reduction or solely on energy storage absorption.

[0105] 4. For severe fault conditions such as deep low voltage and near-zero voltage, a coordinated control logic for the supercapacitor-DC discharge branch is designed to promptly activate the discharge branch when the rotor and supercapacitor energy absorption capacity are insufficient, thereby improving the system's safety boundary under extreme fault conditions.

[0106] 5. For the transient recovery process after fault clearance, a collaborative recovery control method is proposed under the constraints of rotor kinetic energy release, supercapacitor SOC recovery, and grid-side active power ramping, to suppress DC bus secondary overvoltage and power backlash during the fault recovery phase. Attached Figure Description

[0107] Figure 1 Topology diagram of a grid-type direct-drive wind and energy storage system;

[0108] Figure 2 An adaptive allocation surface plot of the supercapacitor sharing coefficient as a function of rotational speed and SOC;

[0109] Figure 3 This is a graph showing the adaptive variation of the DC bus reference voltage.

[0110] Figure 4 This is a graph showing the relationship between the minimum grid-side power and the actual output during low-voltage ride-through.

[0111] Figure 5 The simulation diagram shows the dynamic response of the system under a deep low voltage fault. Detailed Implementation

[0112] The technical solution of the present invention will now be described in detail with reference to the accompanying drawings.

[0113] This invention provides a low-voltage ride-through control method for a multi-timescale coordinated grid-type wind and energy storage system, applicable to a grid-type wind and energy storage system including a grid-type direct-drive wind turbine, a DC bus, a supercapacitor, and a DC discharge branch, comprising:

[0114] Real-time acquisition of grid connection point voltage, grid-side converter output current, DC bus voltage, rotor speed, supercapacitor SOC and terminal voltage;

[0115] When the grid connection point voltage is lower than the low voltage ride-through threshold, the maximum active power transmission capacity of the grid side is calculated based on the reactive power priority support and total current limiting constraint of the grid-side converter.

[0116] The remaining unbalanced power on the DC side is calculated based on the power injected into the DC bus on the machine side, the maximum active power output capacity on the grid side, and the transient power that the DC capacitor can bear.

[0117] Calculate the real-time available energy margins of rotor kinetic energy, DC capacitor, and supercapacitor respectively, and combine them with the total unbalanced energy in the prediction time domain to calculate the risk factor of constrained effective energy margin.

[0118] Based on the constrained effective energy margin risk factor and the real-time available energy margin of each energy unit, the target power of rotor kinetic energy, DC capacitor and supercapacitor is allocated according to the multi-time scale division of rotor kinetic energy absorption at the millisecond level, DC capacitor transient buffer at the millisecond level and supercapacitor residual compensation at the second level, and the power reference value of the machine-side converter is adjusted based on the allocation results.

[0119] Based on the target power to be carried by the supercapacitor and the DC bus voltage deviation, a reference power for the supercapacitor is generated, and the operation of the supercapacitor is controlled by a bidirectional DC / DC converter.

[0120] When the constraint effective energy margin risk factor reaches the preset input threshold, or when the DC bus voltage, supercapacitor SOC, and rotor speed simultaneously reach the corresponding warning threshold, the DC discharge branch is activated. The target power of the DC discharge branch is determined based on the difference between the remaining unbalanced power and the maximum energy absorption capacity of each energy unit, and its switching is controlled.

[0121] After the fault is cleared, based on the rotor speed deviation, DC bus voltage deviation, and the active power transmission capacity after grid-side recovery, the power reference value of the generator-side converter, the SOC recovery power of the supercapacitor, and the DC bus reference voltage are adjusted in a coordinated manner until the system returns to normal operation.

[0122] The following is a detailed implementation process of the present invention.

[0123] 1. Constraint-sensing type machine-side power reduction and rotor kinetic energy priority absorption control strategy

[0124] In grid-connected direct-drive wind-storage systems, the core of low-voltage fault ride-through is coordinating the mechanical input power of the wind turbine, the fault output capability of the grid side, the rotor kinetic energy absorption capability, and the energy storage absorption capability. Traditional turbine-side power reduction relies solely on DC voltage deviation, which is prone to insufficient or excessive power reduction. To address this, a constraint-aware turbine-side power reduction control strategy is proposed. This strategy prioritizes rotor kinetic energy as the energy absorption unit in the initial stage of a fault, incorporating speed safety margin, DC bus voltage deviation, and the degree of voltage drop at the grid connection point into the correction of the turbine-side power reference value. This allows the turbine-side converter to proactively adjust the power injected into the DC side of the wind turbine based on the real-time constraint status of the system, rather than passively reducing power solely based on DC voltage.

[0125] During normal operation, the generator-side converter is controlled according to Maximum Power Point Tracking (MPPT), with a power reference value of:

[0126] P s0 * = P MPPT = k opt ω r 3

[0127] In the formula, k opt For the optimal torque coefficient, ω r This represents the real-time rotational speed of the rotor.

[0128] When the grid connection point voltage U t When the voltage drops below the low voltage ride-through threshold, the generator side switches to fault-based power reduction mode, and the power reference value is corrected as follows:

[0129] P s * = P s0 * - ΔP MSC *

[0130] In the formula, ΔP MSC * The adaptive power reduction command on the engine side is determined by the fault depth, speed margin, and DC voltage deviation.

[0131] ΔP MSC * = K m σ u σ ω P N + K dc (U dc - U dc * )

[0132] Where, σ u = (0.9 - U t ) / (0.9 - U min ) is the fault voltage sag factor (U min This factor (the lower limit of voltage under severe faults) characterizes the impact of fault depth on the grid-side active power output capability; the deeper the voltage drop, the greater the voltage drop. u The larger the value, the higher the degree of power imbalance on the DC side; σ ω = (ω r,max - ω r ) / (ω r,max - ω r0 ) is the rotor speed margin coefficient (ω) r,max ω is the maximum permissible speed of the rotor. r0 (where K is the rotor speed before the fault). The larger this coefficient is, the farther the rotor is from its maximum speed limit, and the larger the kinetic energy space available for absorbing fault energy; conversely, it indicates that the rotor's ability to continue absorbing energy has decreased. m K is the coupling adjustment coefficient. dc P is the DC voltage feedback coefficient. NThis represents the unit's rated power. The above formula demonstrates that the power reduction amplitude on the turbine side is no longer determined by a single voltage deviation, but rather by a combination of "fault depth, speed margin, and DC voltage deviation." When the fault is deep and the speed margin is sufficient, the turbine side actively reduces electromagnetic power, allowing more energy to be transferred to the rotor's kinetic energy; when the rotor approaches its maximum speed limit, σ... ω The power reduction on the machine side is automatically reduced to avoid rotor overspeed.

[0133] To ensure that the power reduction process on the turbine side meets mechanical and electrical safety constraints, the power reduction command must satisfy the following:

[0134] 0 ≤ ΔP MSC * ≤ ΔP MSC,max

[0135] ΔP MSC,max = min[J(ω r,max 2 - ω r 2 ) / (2t f ), ΔP conv,max , P s0 * - P s,min ]

[0136] In the formula, J(ω) r,max 2 - ω r 2 ) / (2t f ΔP represents the maximum average energy absorption power during the rotor's remaining fault time. conv,max For the capacity constraint of the machine-side converter, P s0 * - P s,min For the minimum power constraint on the machine side, J is the moment of inertia, and t is the moment of inertia. f P represents the remaining downtime. s,min This is the minimum operating power on the machine side.

[0137] The reference value for the electromagnetic torque on the machine side is: T e * = P s * / ω r

[0138] To avoid mechanical shock to the transmission chain caused by sudden changes in electromagnetic torque during a fault, a rate of change constraint is added to the torque command:

[0139] |dT e * / dt| ≤ R T

[0140] RT This represents the upper limit of the torque change rate.

[0141] After the fault is cleared, the power on the turbine side does not directly recover to the MPPT reference value, but rather recovers smoothly based on the recovery status of the DC bus voltage and the rotor speed deviation.

[0142] dP s * / dt = [(P MPPT - P s * ) / T p ] · [1 - (U dc - U dc * ) / (U dc,max - U dc * )]

[0143] In the formula, Tp is the generator-side power recovery time constant. This recovery method can avoid the rapid recovery of generator-side power after fault clearance, which could lead to secondary overvoltage on the DC bus, while allowing the rotor speed to gradually return to the normal operating range.

[0144] 2. Multi-timescale energy coordination mechanism of wind turbine-DC bus-supercapacitor

[0145] During low-voltage fault ride-through, the power imbalance in the grid-connected wind-storage system is not instantaneously borne by a single unit, but rather shared by the DC capacitor, rotor kinetic energy, and supercapacitor at different time scales. Traditional wind-storage coordinated control often treats the energy storage system as an additional power compensation unit on the DC side, lacking a characterization of the division of labor among rotor kinetic energy, transient buffering of the DC capacitor, and rapid energy absorption by the supercapacitor. This results in a conservative energy storage capacity design and control strategies that are difficult to adapt to different fault depths and operating conditions.

[0146] To address this, this invention proposes a multi-timescale energy coordination mechanism. This mechanism divides the unbalanced energy during fault ride-through into three levels: the first level is a DC capacitor providing millisecond-level transient buffering, primarily suppressing voltage spikes at the initial moment of the fault; the second level is rotor kinetic energy absorbing energy at the hundred-millisecond level, mainly handling source-side power reduction from the initial to the middle stages of the fault; and the third level is a supercapacitor providing rapid energy storage at the second level, primarily handling the remaining unbalanced energy after rotor margin depletion. This layered coordination approach avoids relying solely on supercapacitor energy absorption or rotor speed-up, thereby improving fault ride-through capability and the economic efficiency of energy storage configuration.

[0147] Using the DC bus as the power balancing node:

[0148] C dc U dc · dU dc / dt = P s - P g - P sc

[0149] In the formula, C dc For DC bus capacitor, U dc P is the DC bus voltage. s Injecting power to the machine side, P g P is the grid-side output power. sc This is for the supercapacitor to absorb power. The equation shows that when a low-voltage fault causes P... g When P decreases, if s If the voltage cannot be reduced synchronously, the excess power will first enter the DC capacitor, causing the DC bus voltage to rise.

[0150] From the perspective of overall system energy, the following conditions are met during a fault:

[0151] P w - P g = P rot + P dc + P sc + P loss

[0152] P rot = P w - P s

[0153] P dc = C dc U dc · dU dc / dt

[0154] Among them, P w P is the mechanical power input to the wind turbine. rot P is the power absorbed by the rotor kinetic energy. dc P is the power absorbed by the DC capacitor. loss This refers to system losses.

[0155] This relationship indicates that the power imbalance between the wind turbine input power and the grid-side output power during a fault should not be entirely borne by the supercapacitor, but should be allocated preferentially using the system's existing energy buffer resources. Based on this, the total unbalanced energy is decomposed as follows:

[0156] ΔP = ΔP dc + ΔP rot + ΔP sc

[0157] Corresponding energy relationship:

[0158] = ΔEdc + ΔE rot + ΔE sc + E loss

[0159] Where, ΔE dc = 0.5C dc (U dc 2 - U dc0 2 ), ΔE rot = 0.5J(ω r 2 - ω r0 2 ), These represent the energy absorbed by the DC capacitor, rotor kinetic energy, and supercapacitor, respectively, E. loss U represents the total energy loss of the system. dc0 This is the initial voltage before the DC bus fault. This represents the initial moment of a low-voltage fault. Define the energy absorption ratio η of the rotor kinetic energy, the supercapacitor, and the DC capacitor. rot = ΔE rot / E total η sc = ΔE sc / E total η dc = ΔE dc / E total (E) total = (Total unbalanced energy), used to quantify synergistic effects.

[0160] 3. Fault ride-through control strategy on the supercapacitor side

[0161] 3.1 Supercapacitor Adaptive Power Allocation Strategy

[0162] Define rotor storage ratio coefficient ( Equivalent average unbalanced power during the fault period , This represents the total mismatch energy within the system during low-voltage ride-through. During the period of failure, The rated rotor speed, (Rated power), characterizing the rotor's ability to share unbalanced power. The initial power sharing factor between the rotor and the supercapacitor is:

[0163]

[0164] In the formula, and These are the initial sharing coefficients for rotor energy storage and supercapacitor energy storage, respectively. The above formula ensures the conserved distribution of unbalanced power between the two types of energy storage units.

[0165] The maximum power absorbed by a supercapacitor is constrained by the maximum charging current, the terminal voltage difference, and the remaining capacity, which can be expressed as:

[0166]

[0167] In the formula, U sc I is the terminal voltage of the supercapacitor. ch,max For the maximum charging current, R sc E is the equivalent series resistance. max E sc These represent the maximum storage energy and the current storage energy, respectively, T. c The sampling period.

[0168] To uniformly characterize the real-time energy absorption capacity of the rotor and supercapacitor, the normalized energy absorption margin coefficient of the rotor is defined as:

[0169]

[0170] In the formula, For actual adjustable power, Characterizing the rotor's power carrying capacity relative to a reference. Available energy absorption margin.

[0171] The normalized energy absorption margin of a supercapacitor is:

[0172]

[0173] In the formula, Characterizing the rotor's power carrying capacity relative to a reference. Available energy absorption margin; and These are the lower and upper limits of the supercapacitor's state of charge, respectively. For SOC sensitivity factor, when At this time, the energy absorption weight of supercapacitors decays faster in the high SOC region, which can reduce the risk of overcharging. .

[0174] Furthermore, the adaptive power allocation function is constructed as follows:

[0175]

[0176] Introducing deviation correction amount ( (As a correction factor, taken as 0.2~0.4), the final supercapacitor sharing factor is:

[0177]

[0178] The supercapacitor absorbs power of .

[0179] Figure 2 for Adaptive assignment surface plot as rotational speed and state of charge (SOC) change.

[0180] 3.2 Adaptive Control of DC Bus Reference Voltage

[0181] Introducing DC voltage outer loop compensation, the additional regulation of the supercapacitor is:

[0182]

[0183] In the formula, , These are the actual voltage and reference voltage of the DC bus, respectively. , These are the proportional and integral coefficients of the DC voltage outer loop, respectively, satisfying... , , This is the rated voltage of the DC bus. This represents the upper limit allowed for DC buses.

[0184] During a fault, the reference voltage rises as the grid connection point voltage drops:

[0185]

[0186]

[0187]

[0188] In the formula, This represents the rise in the DC bus reference voltage for every 0.1 pu drop in the grid connection point voltage. (Most severe voltage drop) At that time, there should be . The target rise is determined by the voltage drop depth at the grid connection point. This represents the smooth rise in the actual DC bus reference voltage. Reference voltage rise time constant. The duration of the low-voltage ride-through is primarily determined by the duration of the low-voltage ride-through. According to the engineering steady-state criterion for a first-order inertial element, to ensure that the reference voltage essentially completes its rise during the low-voltage ride-through, the following conditions must be met: According to low penetration time Select The reference voltage completes its main rise within approximately 1.8 seconds.

[0189] After the fault is cleared, to avoid a rapid drop in the DC bus reference voltage causing a power recovery surge, It recovers to its rated value according to the first-order inertial law.

[0190]

[0191] In the formula, The reference voltage recovery time constant, =0.65s is the recovery time constant.

[0192] During low voltage ride-through, when Depend on Towards When falling, along Figure 3 The reference voltage adaptive rise trajectory shown by the solid line is... Smoothly increase to no more than After the fault is cleared, when When it returns to the normal range, along Figure 3 The reference voltage smooth recovery trajectory, shown by the dashed line, gradually returns to its rated value. This strategy improves the DC-side energy buffer margin during low-level breakdown and reduces the power surge during fault recovery.

[0193] The final supercapacitor reference power is .

[0194] 3.3 Supercapacitor Capacitance Design

[0195] To balance fault ride-through capability with the economics of energy storage configuration, supercapacitors only handle the remaining unbalanced energy. In the initial stage of a fault, limitations imposed by grid connection voltage dips and current limiting... The difference between the two is the surplus power that the supercapacitor needs to absorb. As the voltage recovers, Gradually increase, and in Timely satisfaction ,like Figure 4 As shown.

[0196] Therefore, the capacity of a supercapacitor is determined by the remaining unbalanced energy after rotor energy storage. The area enclosed by the two inner curves represents the energy required for the supercapacitor to absorb:

[0197]

[0198] This is the minimum injection power on the machine side;

[0199] Considering the capacitance decay at the end of the supercapacitor's lifespan, and verifying it at 80% of the rated capacity, the rated capacitance should meet the following requirements:

[0200]

[0201] In the formula, k0 is the initial capacitance coefficient (taken as 0.7~0.8), u CON For the rated voltage, u sc,0 u sc,t These are the voltages before and after the fault, respectively.

[0202] 4. Multi-energy margin collaborative control strategy under network current limiting constraints

[0203] 4.1 Prediction of grid-side active power output capacity under grid current limiting constraints

[0204] Let the maximum allowable current of the grid-side converter be I. g,max The per-unit value of the positive sequence voltage at the grid connection point is U1, and the reactive power support factor is k. q During a low-voltage fault, the reactive current reference value shall be determined according to the reactive current priority principle:

[0205] i q * = sat[k q (0.9 - U1), -I g,max , I g,max ]

[0206] In the formula, sat[·] represents the amplitude limiting function;

[0207] Under the condition of satisfying the total current constraint, the maximum active current that the grid-side converter can use to transmit active power is:

[0208] i d,max =

[0209] Therefore, the maximum active power transmission capacity of the grid side under a three-phase symmetrical fault is expressed as:

[0210] P g,max = (3 / 2)U1·i d,max

[0211] When asymmetrical low-voltage faults such as single-phase grounding or two-phase short circuits occur, negative sequence voltage can cause current imbalance and instantaneous power double-frequency oscillation. To avoid overestimating the grid-side active power transmission capacity, a negative sequence correction factor is introduced.

[0212] χ² = 1 - k²·U² / (U₁ + ε)

[0213] In the formula, U2 is the magnitude of the negative sequence voltage at the grid connection point, k2 is the negative sequence influence coefficient, and ε is a very small positive number to prevent the denominator from being zero; the maximum active power transmission capacity of the grid side after considering the influence of the negative sequence voltage is:

[0214] P g,max ' = χ2·P g,max

[0215] P g,max ' This is the corrected maximum active power transmission capacity of the grid side.

[0216] The transient power that the DC capacitor can handle is defined based on the current DC voltage margin:

[0217] P dc allow = C dc ·U dc (U dc,max - U dc ) / T dc

[0218] In the formula, T dc For the DC capacitor to participate in the equivalent time constant, C dc U is the capacitance value of the DC capacitor. dc U is the real-time voltage of the DC bus. dc,max This is the maximum allowable voltage for a DC capacitor;

[0219] The remaining unbalanced power that needs to be borne by rotor kinetic energy, supercapacitors, or DC discharge branches during the fault is:

[0220] ΔP rem = [P s - P g,max ' - P dc allow ] +

[0221] In the formula, P s The power injected into the DC bus of the machine-side converter, [·] + This represents a non-negative cutoff function, meaning it takes the value of zero when the value within the parentheses is less than zero. This formula directly couples reactive power priority, current limiting, active power limitation, and DC-side energy distribution in grid control, enabling it to identify the remaining unbalanced power during a fault more accurately than simply estimating the power deficit based on the voltage drop depth.

[0222] 4.2 Coordinated Allocation of Multiple Energy Margins Based on Remaining Failure Time Prediction

[0223] During low-voltage ride-through, the control objective is not to enable a single energy storage unit to independently absorb all the unbalanced power, but rather to perform constrained energy distribution among rotor kinetic energy, DC capacitors, and supercapacitors. Since fault duration, fault depth, initial rotational speed, and supercapacitor state of charge (SOC) all affect the available energy absorption space, adjusting the power sharing ratio solely based on instantaneous voltage deviation or instantaneous SOC is insufficient to guarantee safety under sustained fault conditions.

[0224] Therefore, this invention further introduces residual time-of-failure prediction and multi-energy margin assessment. Let the current time be t, the prediction time domain be T_p, and the imbalance energy within the prediction period be:

[0225] E Σ pre (t) = ∫ t t+T_p [P w pre (τ) - P g,max pre' (τ)] + dτ

[0226] In the formula, P w pre (τ) represents the wind turbine input mechanical power in the predicted time domain, P g,max pre' (τ) represents the maximum active power transmission capacity of the network side in the predicted time domain.

[0227] The available kinetic energy margin of the rotor at the current speed is:

[0228] E rot av = 0.5J(ω r,max 2 -ω r 2 )

[0229] The available voltage energy margin of the DC capacitor under the current DC voltage is:

[0230] E dc av = 0.5C dc (U dc,max 2 - U dc 2 )

[0231] The available energy absorption margin of a supercapacitor at the current terminal voltage or SOC state is:

[0232] E sc av = 0.5C sc (U sc,max 2 - U sc 2 )

[0233] Based on this, the system energy margin risk factor is defined as follows:

[0234] R E = E Σ pre (t) / (Erot av + E dc av + E sc av + ε)

[0235] The above formula can be used as a basic energy risk criterion at the total energy level: when R E When R < 1, it indicates that the theoretically available energy absorption margin of the rotor, DC capacitor, and supercapacitor can cover the predicted fault energy; when R E A value ≥ 1 indicates that existing buffer resources may be insufficient. Considering that this basic indicator does not differentiate between the power absorption limit, response time scale, and safety boundary approach of each energy unit, a risk factor for constrained effective energy margin is further defined.

[0236] Improvement of the constraint effectiveness of energy margin risk factors

[0237] To avoid the risk criterion being merely a simple ratio between the predicted imbalance energy and the theoretical total energy margin, this paper further modifies the theoretical available energy of the rotor kinetic energy, DC capacitor, and supercapacitor into "constrained effective available energy." This effective margin simultaneously considers energy capacity, maximum absorbed power, physical response time, and margin from the safety boundary, thus characterizing the actual energy absorption capacity that can be mobilized within the predicted fault time.

[0238] For any energy unit x, x∈{rot, dc, sc}, where rot, dc, and sc represent the rotor kinetic energy, DC capacitance, and supercapacitor, respectively, we first define the achievable absorbed energy of x after being constrained by power limiting:

[0239] E x lim (t) = min{ E x av (t), P x,max abs (t)·T_p}

[0240] In the formula, E x av (t) represents the theoretical available energy margin for the corresponding energy unit, P x,max abs (t) represents the maximum allowable absorption power under the current state, and T_p represents the prediction time domain. The above equation shows that even if a certain unit still has a large theoretical energy space, if its absorption power in the prediction time domain T_p is insufficient, the entire theoretical energy margin cannot be regarded as a callable margin.

[0241] Considering the physical response speed of different energy units, a time-scale correction factor is defined:

[0242]

[0243] in, For the duration of the fault, T dc T rot and T sc The DC capacitor, rotor kinetic energy, and supercapacitor are allowed to participate in the equivalent time constant, respectively, satisfying T. dc < T rot < T sc The above formula is used to illustrate that DC capacitors are suitable for rapid buffering in the early stage of a fault, rotor kinetic energy is suitable for absorption on a medium to short time scale, and supercapacitors are suitable for continuous absorption in the middle and later stages of a fault.

[0244] To ensure that risk factors reflect the approaching safety boundaries of rotor speed, DC bus voltage, and supercapacitor SOC in advance, a safety boundary correction factor is introduced:

[0245] θ x (t) = {sat[(z x,max - z x (t)) / (z x,max - z x,warn + ε), 0, 1]} m_x

[0246] z rot =ω r , z dc =U dc , z sc =SOC sc

[0247] In the formula, z x,warn z is the warning threshold. x,max To provide a safety upper limit, m_x is a nonlinear penalty coefficient. m_x can be a positive number greater than 1, used to ensure that the effective margin decreases faster as the energy unit approaches the safety boundary, thereby avoiding control lag; ω r For real-time rotor speed, U dc This represents the real-time voltage of the DC bus, SOC. sc This is the ratio of the actual amount of charge currently stored in a supercapacitor to its rated maximum charge storage capacity.

[0248] The constraint-effective available energy margin of energy unit x is defined as:

[0249] E x eff (t)=η x μ x (t)·θ x (t)·E x lim (t)

[0250] In the formula, η x Energy efficiency or conversion efficiency; compared with direct use of E x av In comparison, E x eff It can simultaneously reflect energy capacity constraints, power absorption constraints, time scale constraints, and safety boundary constraints.

[0251] To avoid simply integrating the power difference and ignoring the severity of the fault and the prediction error, a fault severity correction coefficient is defined on the side of predicting the unbalanced energy:

[0252] ρ(τ)=1+k u [1-U1(τ)] + +k2U2(τ) / (U1(τ)+ε)+k i |i q *(τ)| / I g,max

[0253] In the formula, U1(τ) and U2(τ) are the positive-sequence and negative-sequence voltage amplitudes at the predicted time, respectively, i q *(τ) is the reference value for reactive current, I g,max The maximum allowable current of the grid-side converter is k. u k2, k i These are the voltage sag compensation coefficient, negative sequence influence coefficient, and reactive current limiting adjustment coefficient, respectively; ρ(τ) increases with the deepening of voltage sag, the increase of negative sequence component, or the increase of reactive current occupancy ratio.

[0254] E Σ risk (t)=∫ t t+T_p ρ(τ)[P w pre (τ)-P g,max pre' (τ)] + dτ + k σ σ E (t)

[0255] Among them, E Σ risk (t) represents the risk-corrected predicted imbalance energy, P w pre (τ) represents the wind turbine input mechanical power in the predicted time domain, P g,max pre' (τ) represents the maximum active power transmission capacity of the network side in the predicted time domain, σ E (t) represents the prediction error or uncertainty margin, k σ To predict the conservative coefficient, [·] +Represents a non-negative truncation function;

[0256] Meanwhile, to avoid masking the situation where the overall system margin is sufficient but individual energy units saturate prematurely, we define the energy allocation requirements and local bottleneck risk terms for each energy unit in the prediction time domain:

[0257] E x req (t)=∫ t t+T_p λ x ' (τ)[P w pre (τ)-P g,max pre' (τ)] + dτ

[0258] R b (t)=max x∈{rot,dc,sc} { E x req (t) / (E x eff (t)+ε)}

[0259] Among them, E x req (t) represents the energy required for the allocation of energy unit x in the predicted time domain, λ x ' (τ) represents the target-sharing weight of energy unit x in the prediction time domain, R b (t) represents the risk of the weakest local unit;

[0260] The real-time security boundary approach risk item is defined as:

[0261] B(t) = max{B dc (t), B rot (t), B sc (t)}

[0262] B dc =[(U dc -U dc,warn ) / (U dc,max -U dc,warn +ε)] +

[0263] B rot =[(ω r -ω warn ) / (ω r,max -ω warn +ε)] +

[0264] B sc =[(SOCsc -SOC warn ) / (SOC max -SOC warn +ε)] +

[0265] In the formula, B(t) represents the total effective margin risk of the system. dc (t), B rot (t), B sc (t) represents the boundary approximation of the DC bus voltage, the rotor speed, and the supercapacitor SOC, respectively. dc,warn ω is the DC bus voltage warning threshold. warn The rotor speed warning threshold, SOC warn The SOC (State of Charge) warning threshold for supercapacitors. max This refers to the maximum charge that a supercapacitor can store.

[0266] Ultimately, the constrained effective energy margin risk factor is defined as:

[0267] R E c (t)=αE Σ risk (t) / (E rot eff +E dc eff +E sc eff +ε)+βR b (t)+γB(t)

[0268] α+β+γ=1, α≥0, β≥0, γ≥0

[0269] In the formula, α, β, and γ are the total margin risk weighting coefficient, the local bottleneck risk weighting coefficient, and the boundary approximation risk weighting coefficient, respectively. In engineering implementation, when R... E c (t) <R on At this time, the system maintains the conventional coordinated energy absorption of the rotor, DC capacitor, and supercapacitor; when R E c (t)≥R on At that time, the power reduction amplitude on the machine side is increased in advance, the charging power of the supercapacitor is limited, and the DC discharge branch is put into preparation state; when R E c (t) <R off Furthermore, when the DC bus voltage is lower than the exit threshold, the unloading preparation or unloading input state is exited.

[0270] Calculate the initial weighting λ for rotor kinetic energy, DC capacitor, and supercapacitor. x = Ex av / (E rot av + E dc av +E sc av + ε), x ∈ {rot, dc, sc}, x∈{rot,dc,sc}, where rot, dc, and sc represent the rotor kinetic energy, DC capacitance, and supercapacitor, respectively, and E x av ε represents the real-time available energy margin of energy unit x, where ε is a minimum positive number divisible by zero;

[0271] Adjust the initial weighting by time scale: λ rot ' = , λ dc ' = , λ sc ' = 1 - λ rot ' - λ dc ' ,in For the duration of the fault, T dc T rot and T sc These are the allowable participation time constants for DC capacitor, rotor kinetic energy, and supercapacitor, respectively.

[0272] The target power to be undertaken by each energy unit is the product of the corresponding corrected weight and the remaining unbalanced power.

[0273] 4.3 Coordinated Control of Supercapacitor-Discharge Branch under Severe Fault Conditions

[0274] Under conditions of deep low voltage, near-zero voltage, or persistent faults, the grid-side active power transmission capacity drops sharply. If the wind turbine speed is already high before the fault occurs, or the supercapacitor is in a high-charge state (SOC), relying solely on rotor kinetic energy and supercapacitor energy absorption may not simultaneously meet the constraints of DC bus overvoltage, rotor speed overspeed, and supercapacitor overcharging. Therefore, this invention includes a DC-side surplus energy discharge branch as a final protection unit under extreme fault conditions.

[0275] The offloading branch does not participate in the normal power distribution during low-voltage faults; it is only activated when the constrained effective predicted energy margin is insufficient or multiple safety boundaries are approaching simultaneously. Its first type of triggering condition is when the constrained effective energy margin risk factor reaches the activation threshold.

[0276] R E c≥ R on

[0277] The second type of triggering condition is when the DC bus, supercapacitor, and rotor simultaneously approach the safety boundary:

[0278] U dc ≥ U dump,on SOC sc ≥ SOC high ω r ≥ ω safe

[0279] In the formula, U dump,on To set the voltage threshold for the discharge branch, SOC high ω is the high-charge state threshold of a supercapacitor. safe The rotor safety warning speed.

[0280] The target power of the unloading branch is determined based on the remaining unbalanced power, the maximum absorption capacity of the supercapacitor, the maximum energy absorption capacity of the rotor, and the allowable buffering capacity of the DC capacitor.

[0281] P dump * = sat(ΔP rem - P sc,max - P rot,max - P dc allow , 0, P dump,max )

[0282] Where ΔP rem For the remaining unbalanced power, P sc,max P represents the maximum allowable power absorbed by the supercapacitor. rot,max P is the maximum allowable energy absorption power of the rotor. dc allow P is the transient power that the DC capacitor can withstand. dump,max This is the maximum power output of the discharge branch. When the rotor and supercapacitor still have sufficient margin, P... dump * The value is zero; the unloading branch is only put into operation as a shortfall when neither of the two is sufficient to cover the remaining unbalanced power.

[0283] When a resistor discharge branch is used, the duty cycle of the discharge branch can be expressed as:

[0284] D dump = sat[P dump * R dump / U dc 2 , 0, 1]

[0285] Rdump U is the equivalent resistance of the discharge resistor. dc This represents the real-time voltage of the DC bus; to avoid frequent switching of the discharge branch, a hysteresis exit condition is set:

[0286] U dc ≤ U dump,off R E c < R off

[0287] U dump,off < U dump,on R off < R on The discharge branch will only be deactivated once the DC bus voltage drops and the risk of the effective energy of the constraint returns to a safe range.

[0288] The capacity of the discharge resistor should be checked based on the maximum discharge power under the most severe fault. To ensure that the maximum discharge power can still be released when the DC bus voltage reaches the upper limit of the allowable range, the discharge resistor can be selected according to the following formula:

[0289] R dump ≤ U dc,max 2 / P dump,max

[0290] Under severe fault conditions, the control priority settings of this invention are as follows: First, the grid-side converter performs grid voltage support and reactive current priority; Second, the generator side reduces power and uses rotor kinetic energy to absorb some of the surplus energy; Third, the supercapacitor absorbs the remaining unbalanced power that the rotor and DC capacitor cannot bear; Fourth, when the predicted energy margin is insufficient or the real-time safety boundary is approaching simultaneously, the DC side discharge branch is activated.

[0291] 4.4 Collaborative Recovery Control After Fault Clearance

[0292] After a low-voltage fault is cleared, if the generator-side power immediately recovers to the maximum power tracking reference value while the supercapacitor remains in a high state of charge (SOC), problems such as secondary overvoltage on the DC bus, grid-side active power recovery impact, and excessively rapid rotor speed drop can easily occur. Therefore, this invention introduces coordinated control of rotor kinetic energy release, DC bus reference value drop, and supercapacitor SOC recovery during the fault recovery phase.

[0293] After the fault is cleared, the turbine-side power reference value does not directly return to P. MPPT Instead, it is corrected based on the rotor speed deviation and the DC bus voltage deviation:

[0294] P s * = P MPPT + K ω,rec (ωr - ω r0 ) - K dc,rec (U dc - U dc * )

[0295] In the formula, K ω,rec K is the rotor overspeed recovery coefficient. dc,rec This is the DC bus voltage suppression coefficient. When the rotor speed is higher than the speed before the fault, the electromagnetic power on the machine side is appropriately increased to release the rotor kinetic energy; when the DC bus voltage is too high, the rapid rise of the power on the machine side is suppressed to prevent secondary overvoltage of the DC bus.

[0296] To avoid drivetrain shock and DC-side power backlash caused by sudden changes in machine-side power, a constraint is set on the rate of change of machine-side power recovery:

[0297] |dP s * / dt| ≤ R P,rec

[0298] After the fault is cleared, the supercapacitor enters SOC recovery mode. When the SOC is higher than the target value, the supercapacitor releases energy to the DC side according to the limited power; when the SOC is close to the target value, the recovery power gradually decreases.

[0299] P sc,rec * = -sat[K soc (SOC sc - SOC ref ), 0, P sc,dis,max ]

[0300] In the formula, the negative sign indicates that the supercapacitor is transitioning from an energy absorption state to a discharge recovery state, and SOC (State of Charge) is... ref For the target state of charge, P sc,dis,max This represents the maximum allowable discharge power of the supercapacitor. To prevent the supercapacitor's recovery process from blocking the active power recovery channel after a grid-side fault, the supercapacitor's recovery power should also meet the following requirements:

[0301] |P sc,rec * | ≤ [P g,max rec - P s * ] +

[0302] In the formula, P g,max recThis represents the maximum active power that can be delivered from the grid side after the fault is cleared. If the active power channel on the grid side has not been fully restored immediately after the fault is cleared, the recovery of the supercapacitor's SOC will be delayed, prioritizing the stability of the DC bus and the safe recovery of the rotor.

[0303] The DC bus reference voltage may be adaptively raised during a fault. After the fault is cleared, to avoid a sudden drop in the reference value causing an impact on the outer voltage loop, the DC bus reference value will fall back to its rated value according to a first-order inertial law.

[0304] dU dc * / dt = (U dc,N - U dc * ) / T U,rec

[0305] In the formula, U dc,N T is the rated reference voltage of the DC bus. U,rec The reference voltage recovery time constant. The end condition for the recovery phase can be set as follows:

[0306] |U dc - U dc,N | ≤ δ U ,|ω r - ω r0 | ≤ δ ω |SOC sc - SOC ref | ≤ δ SOC

[0307] When the above conditions are met simultaneously, the system exits the fault recovery mode, the generator-side converter resumes normal MPPT control, the supercapacitor resumes standby or normal SOC management, and the grid-side converter resumes normal grid operation.

[0308] This recovery strategy ensures a clear sequence for energy release after fault clearance: first, stabilizing the DC bus; second, smoothly releasing the rotor kinetic energy; and third, restoring the supercapacitor's state of charge (SOC) based on the remaining capacity of the grid-side active power channels. This avoids secondary overvoltage and active power surges caused by multiple control loops acting rapidly simultaneously during fault clearance.

[0309] Simulation verification

[0310] To verify the effectiveness of the proposed control method, based on Figure 1 A simulation model of the grid-connected direct-drive wind and energy storage integrated system is established. The grid-side converter uses VSG control to achieve grid-connected operation, and the supercapacitor is connected to the DC bus via a bidirectional DC / DC converter for DC-side power regulation during faults. The simulation system uses a 3MW grid-connected direct-drive wind turbine as the research object. The rated capacitance of the supercapacitor is determined according to the capacity design method in Chapter 3, and is taken as... The upper limit of the DC bus reference voltage is taken as... The lift and recovery time constants are respectively taken as , To verify the effectiveness of the proposed rotor-supercapacitor coordinated energy balance strategy under low-voltage fault conditions, a low-voltage fault simulation was set up. A deep low-voltage fault (voltage drops to 0.2 pu) was simulated, with the fault occurring at 0.5 s, lasting for 0.625 s, and clearing at 1.125 s, for a simulation duration of 2 s. The effects of traditional grid-connected control, machine-side power reduction control alone, and the coordinated control of this invention were compared. Under the low-voltage fault condition, the grid-connected voltage drops to 0.2 pu, a drop depth of 80%. The main system variable responses under this condition are as follows: Figure 5 As shown. By Figure 5 As shown in (b), under power reduction control only on the machine side, P s The P value is reduced by approximately 0.35 pu under the method of this invention. s The power consumption was reduced by approximately 0.22 pu, a decrease of 37.1%. Compared to minor faults, deep faults significantly reduce grid-side output capability and significantly increase DC-side unbalanced power, thus requiring supercapacitors to deeply participate in transient energy absorption. Figure 5 (c) Figure 5 As can be seen from (d), the method of the present invention shares the unbalanced power through the coordinated adjustment of rotor kinetic energy and the energy absorption of supercapacitors, thus avoiding the limitations of relying solely on speed control.

[0311] Table 1 Performance Comparison of Different Control Methods

[0312] Control strategy <![CDATA[P S / pu]]> <![CDATA[u dc / pu]]> <![CDATA[ω rmax / pu]]> <![CDATA[T rec / s]]> Traditional network control 0 1.21 1.04 1.23 Machine-side power reduction control 0.35 1.12 1.22 1.42 Collaborative Control 0.22 1.00 1.13 1.34

[0313] As shown in Table 1, under traditional network control, u dc The peak value was 1.21 pu, exceeding the 1.2 pu limit; under the machine-side power reduction control alone, the peak rotor speed was 1.22 pu, indicating a speed exceeding the limit. In contrast, the method of this invention controls the UDC peak value within the normal range, the peak speed is controlled at 1.13 pu, and the speed recovery time is 1.34 s. This demonstrates that the method proposed in this invention can simultaneously suppress DC overvoltage and rotor overspeed, improve the deep undervoltage fault ride-through capability, and effectively solve the problem of multiple indicators exceeding the limit.

[0314] The above are preferred embodiments of the present invention. Any changes made to the technical solution of the present invention that do not exceed the scope of the technical solution of the present invention shall fall within the protection scope of the present invention.

Claims

1. A low-voltage ride-through control method for a multi-timescale coordinated grid-type wind-storage system, applied to a grid-type wind-storage system including a grid-type direct-drive wind turbine, a DC bus, a supercapacitor, and a DC discharge branch, characterized in that... include: Real-time acquisition of grid connection point voltage, grid-side converter output current, DC bus voltage, rotor speed, supercapacitor SOC and terminal voltage; When the grid connection point voltage is lower than the low voltage ride-through threshold, the maximum active power transmission capacity of the grid side is calculated based on the reactive power priority support and total current limiting constraint of the grid-side converter. The remaining unbalanced power on the DC side is calculated based on the power injected into the DC bus on the machine side, the maximum active power output capacity on the grid side, and the transient power that the DC capacitor can bear. Calculate the real-time available energy margins of rotor kinetic energy, DC capacitor, and supercapacitor respectively, and combine them with the total unbalanced energy in the prediction time domain to calculate the risk factor of constrained effective energy margin. Based on the constrained effective energy margin risk factor and the real-time available energy margin of each energy unit, the target power of rotor kinetic energy, DC capacitor and supercapacitor is allocated according to the multi-time scale division of rotor kinetic energy absorption at the millisecond level, DC capacitor transient buffer at the millisecond level and supercapacitor residual compensation at the second level, and the power reference value of the machine-side converter is adjusted based on the allocation results. Based on the target power to be carried by the supercapacitor and the DC bus voltage deviation, a reference power for the supercapacitor is generated, and the operation of the supercapacitor is controlled by a bidirectional DC / DC converter. When the constraint effective energy margin risk factor reaches the preset input threshold, or when the DC bus voltage, supercapacitor SOC and rotor speed simultaneously reach the corresponding warning threshold, the DC discharge branch is started. The target power of the DC discharge branch is determined based on the difference between the remaining unbalanced power and the maximum energy absorption capacity of each energy unit, and its switching is controlled. After the fault is cleared, based on the rotor speed deviation, DC bus voltage deviation, and the active power transmission capacity after grid-side recovery, the power reference value of the generator-side converter, the SOC recovery power of the supercapacitor, and the DC bus reference voltage are adjusted in a coordinated manner until the system returns to normal operation.

2. The low voltage ride-through control method for a multi-timescale coordinated grid-type wind-storage system according to claim 1, characterized in that, The calculation process for the maximum active power transmission capacity on the grid side includes: Let the maximum allowable current of the grid-side converter be I. g,max The per-unit value of the positive sequence voltage at the grid connection point is U1, and the reactive power support factor is k. q During a low-voltage fault, the reactive current reference value shall be determined according to the principle of reactive current priority. i q * = sat[k q (0.9 - U1), -I g,max , I g,max ] In the formula, sat[·] represents the amplitude limiting function; Under the condition of satisfying the total current constraint, the maximum active current that the grid-side converter can use to transmit active power is: i d,max = Therefore, the maximum active power transmission capacity of the grid side under a three-phase symmetrical fault is expressed as: P g,max = (3 / 2)U1·i d,max When asymmetrical low-voltage faults such as single-phase grounding or two-phase short circuits occur, negative sequence voltage can cause current imbalance and instantaneous power double-frequency oscillation. To avoid overestimating the grid-side active power transmission capacity, a negative sequence correction factor is introduced. χ² = 1 - k²·U² / (U₁ + ε) In the formula, U2 is the magnitude of the negative sequence voltage at the grid connection point, k2 is the negative sequence influence coefficient, and ε is a very small positive number to prevent the denominator from being zero; the maximum active power transmission capacity of the grid side after considering the influence of the negative sequence voltage is: P g,max ' = χ2·P g,max 。 3. The low voltage ride-through control method for a multi-timescale coordinated grid-type wind-storage system according to claim 2, characterized in that, The calculation process for the residual unbalanced power on the DC side includes: The transient power that the DC capacitor can handle is defined based on the current DC voltage margin: P dc allow = C dc ·U dc (U dc,max - U dc ) / T dc In the formula, T dc For the DC capacitor to participate in the equivalent time constant, C dc U is the capacitance value of the DC capacitor. dc U is the real-time voltage of the DC bus. dc,max This is the maximum allowable voltage for a DC capacitor; The remaining unbalanced power that needs to be borne by rotor kinetic energy, supercapacitors, or DC discharge branches during the fault is: ΔP rem = [P s - P g,max ' - P dc allow ] + In the formula, P s The power injected into the DC bus of the machine-side converter, [·] + This indicates a non-negative truncation function, meaning that the value inside the parentheses is zero when it is less than zero.

4. The low voltage ride-through control method for a multi-timescale coordinated grid-type wind-storage system according to claim 1, characterized in that, The calculation process for the real-time available energy margin of the rotor kinetic energy, DC capacitor, and supercapacitor includes: The available kinetic energy margin of the rotor at the current speed is: E rot av = 0.5J(ω r,max 2 -oh r 2 ) Where J is the moment of inertia of the wind turbine, ω r,max ω is the maximum permissible speed of the rotor. r This refers to the real-time rotor speed. The available voltage energy margin of the DC capacitor under the current DC voltage is: And dc av = 0.5C dc (The dc,max 2 - The dc 2 ) Where C dc U is the capacitance value of the DC capacitor. dc U is the real-time voltage of the DC bus. dc,max This is the maximum allowable voltage for a DC capacitor; The available energy absorption margin of a supercapacitor at the current terminal voltage or SOC state is: And sc av = 0.5C sc (The sc,max 2 - The sc 2 ) Where C sc U is the equivalent capacitance of a supercapacitor. sc,max U is the maximum allowable terminal voltage of the supercapacitor. sc This represents the real-time terminal voltage of the supercapacitor.

5. The low voltage ride-through control method for a multi-timescale coordinated grid-type wind-storage system according to claim 4, characterized in that, The calculation process for the constrained effective energy margin risk factor includes: For any energy unit x, x∈{rot, dc, sc}, where rot, dc, and sc represent the rotor kinetic energy, DC capacitance, and supercapacitor, respectively, we first define the achievable absorbed energy of x after being constrained by power limiting: E x lim (t) = min{ E x av (t), P x,max abs (t)·T_p} In the formula, E x av (t) represents the theoretical available energy margin for the corresponding energy unit, P x,max abs (t) represents the maximum allowable absorbed power under the current state, and T_p represents the prediction time domain; Considering the physical response speed of different energy units, a time-scale correction factor is defined: in, For the duration of the fault, T dc T rot and T sc The DC capacitor, rotor kinetic energy, and supercapacitor are allowed to participate in the equivalent time constant, respectively, satisfying T. dc < T rot < T sc ; To ensure that risk factors reflect the approaching safety boundaries of rotor speed, DC bus voltage, and supercapacitor SOC in advance, a safety boundary correction factor is introduced: θ x (t) = {sat[(z x,max - z x (t)) / (z x,max - z x,warn + ε), 0, 1]} m_x With rot =ω r ,With dc =U dc ,With sc =SOC sc In the formula, z x,warn For the warning threshold, z x,max The upper limit is defined by m_x, the nonlinear penalty coefficient is defined by sat[·], the amplitude limiting function is defined by ε, and ω is a very small positive number to prevent the denominator from being zero. r For real-time rotor speed, U dc This represents the real-time voltage of the DC bus, SOC. sc It is the ratio of the actual amount of charge currently stored in the supercapacitor to its rated maximum charge storage capacity; The constraint-effective available energy margin of energy unit x is defined as: E x eff (t)=η x ·m x (t)·θ x (t)·E x lim (t) In the formula, η x Energy utilization coefficient or conversion efficiency; To avoid simply integrating the power difference and ignoring the severity of the fault and the prediction error on the side of predicting unbalanced energy, a fault severity correction coefficient is defined: p(τ)=1+k u [1-U1(t)] + +k2U2(τ) / (U1(τ)+ε)+k i |i q *(t)| / I g,max In the formula, U1(τ) and U2(τ) are the positive-sequence and negative-sequence voltage amplitudes at the prediction time, respectively, i q *(τ) is the reference value for reactive current, I g,max The maximum allowable current of the grid-side converter is k. u k2, k i These are the voltage sag compensation coefficient, negative sequence influence coefficient, and reactive current limiting adjustment coefficient, respectively; ρ(τ) increases with the deepening of voltage sag, the increase of negative sequence component, or the increase of reactive current occupancy ratio. E Σ risk (t)=∫ t t+T_p p(t)[P w pre (t)-P g,max pre' (t)] + dτ + k σ s E (t) Among them, E Σ risk (t) represents the risk-corrected predicted imbalance energy, P w pre (τ) represents the wind turbine input mechanical power in the predicted time domain, P g,max pre' (τ) represents the maximum active power transmission capacity of the network side in the predicted time domain, σ E (t) represents the prediction error or uncertainty margin, k σ To predict the conservative coefficient, [·] + Represents a non-negative truncation function; Meanwhile, to avoid masking the situation where the overall system margin is sufficient but individual energy units saturate prematurely, we define the energy allocation requirements and local bottleneck risk terms for each energy unit in the prediction time domain: E x req (t)=∫ t t+T_p l x ' (t)[P w pre (t)-P g,max pre' (t)] + dt R b (t)=max x∈{rot,dc,sc} { E x req (t) / (E x eff (t)+ε)} Among them, E x req (t) represents the energy required for the allocation of energy unit x in the predicted time domain, λ x ' (τ) represents the target-sharing weight of energy unit x in the prediction time domain, R b (t) represents the risk of the weakest local unit; The real-time security boundary approach risk item is defined as: B(t)=max{B dc (t), B rot (t), B sc (t)} B dc =[(U dc -IN dc,warn ) / (IN dc,max -IN dc,warn +ε)] + B rot =[(ω r -oh warn ) / (ω r,max -oh warn +e)] + B sc =[(SOC sc -SOC warn ) / (SOC max -SOC warn +ε)] + In the formula, B(t) represents the total effective margin risk of the system. dc (t), B rot (t), B sc (t) represents the boundary approximation of the DC bus voltage, the rotor speed, and the supercapacitor SOC, respectively. dc,warn ω is the DC bus voltage warning threshold. warn The rotor speed warning threshold, SOC warn The SOC (State of Charge) warning threshold for supercapacitors. max This refers to the maximum charge that a supercapacitor can store. Ultimately, the constrained effective energy margin risk factor is defined as: R E c (t)=αE Σ risk (t) / (E rot eff +E dc eff +E sc eff +e)+βR b (t)+γB(t) α+β+γ=1, α≥0, β≥0, γ≥0 In the formula, α, β, and γ are the total margin risk weighting coefficient, the local bottleneck risk weighting coefficient, and the boundary approach risk weighting coefficient, respectively.

6. The low voltage ride-through control method for a multi-timescale coordinated grid-type wind-storage system according to claim 1, characterized in that, The process of allocating the target's power includes: Calculate the initial weighting λ for rotor kinetic energy, DC capacitor, and supercapacitor. x = E x av / (E rot av + E dc av + E sc av +ε), x ∈ {rot, dc, sc}, where rot, dc, and sc represent the rotor kinetic energy, DC capacitance, and supercapacitor, respectively, and E x av ε represents the real-time available energy margin of energy unit x, where ε is a minimum positive number divisible by zero; Adjust the initial weighting by time scale: λ rot ' = , λ dc ' = , λ sc ' = 1- λ rot ' - λ dc ' ,in For the duration of the fault, T dc T rot and T sc These are the allowable participation time constants for DC capacitor, rotor kinetic energy, and supercapacitor, respectively. The target power to be undertaken by each energy unit is the product of the corresponding corrected weight and the remaining unbalanced power.

7. The low voltage ride-through control method for a multi-timescale coordinated grid-type wind-storage system according to claim 1, characterized in that, The process of adjusting the power reference value of the machine-side converter includes: The power reference value during normal operation is the maximum power point tracking reference value P. s0 * = P MPPT = k opt ω r 3 , where k opt For the optimal torque coefficient, ω r This refers to the real-time rotor speed. After fault switching, the reference power value on the machine side is P. s * = P s0 * - ΔP MSC * ΔP MSC * The adaptive power reduction command is expressed as ΔP. MSC * = K m σ u σ ω P N + K dc (U dc - U dc * ), where K m σ is the adjustment coefficient for the coupling of fault depth and rotor margin. u σ is the fault voltage drop factor. ω P is the rotor speed margin coefficient. N K is the rated power of the unit. dc U is the DC voltage feedback coefficient. dc U is the real-time voltage of the DC bus. dc * This is the DC bus reference voltage; The power reduction command is limited between 0 and the maximum permissible power reduction value, where the maximum permissible power reduction value is ΔP. MSC,max = min[J(ω r,max 2 - ω r 2 ) / (2t f ), ΔP conv,max , P s0 * - P s,min ], where J(ω r,max 2 - ω r 2 ) / (2t f ΔP represents the maximum average power that the rotor is allowed to absorb during the remaining fault time. conv,max Due to the capacity constraint of the machine-side converter, P s0 * - P s,min For the minimum power constraint on the turbine side, J is the moment of inertia of the wind turbine, and ω r,max This is the maximum permissible speed of the rotor. For the duration of the fault, P s,min This is the minimum operating power on the machine side; The reference value for the electromagnetic torque on the machine side is T. e * = P s * / ω r And the rate of change of torque is limited to a preset upper limit R. T within; After the fault is cleared, the turbine-side power reference value is set to dP. s * / dt = [(P MPPT - P s * ) / T p ] · [1 - (U dc - U dc * ) / (U dc,max - U dc * The regularity of T] smoothly recovers to the maximum power tracking reference value. p U is the power recovery time constant on the machine side. dc,max This is the maximum allowable voltage for a DC capacitor.

8. The low voltage ride-through control method for a multi-timescale coordinated grid-type wind-storage system according to claim 1, characterized in that, The reference power generation process of the supercapacitor includes: The supercapacitor reference power is composed of the superposition of the feedforward component and the feedback compensation component: ,in To provide power for the supercapacitor target. The output of the DC voltage outer-loop PI controller is expressed as follows: , , These are the proportional and integral coefficients of the DC voltage outer loop, U. dc This is the real-time voltage of the DC bus. This is the DC bus reference voltage; During a fault, the DC bus reference voltage adaptively increases with the depth of voltage drop at the grid connection point, with a target increase amount. ,in This represents the rise in the DC bus reference voltage for every 0.1 pu decrease in the grid connection point voltage. The target rise is the per-unit value of the grid connection point voltage. After being smoothed by a first-order inertial element, it is superimposed on the rated DC bus reference voltage, and the upper limit of the rise is 1.2 times the rated value. After the fault is cleared, the DC bus reference voltage is adjusted according to... The regular pattern of smooth return to the rated value , This is the DC bus reference voltage. This is the reference voltage recovery time constant.

9. The low voltage ride-through control method for a multi-timescale coordinated grid-type wind-storage system according to claim 1, characterized in that, The control process of the DC discharge branch includes: The target power of the DC discharge branch is P dump * = sat[ΔP rem - P sc,max - P rot,max - P dc allow , 0,P dump,max ], where ΔP rem For the remaining unbalanced power, P sc,max P represents the maximum allowable power absorbed by the supercapacitor. rot,max P is the maximum allowable energy absorption power of the rotor. dc allow P is the transient power that the DC capacitor can withstand. dump,max This is the maximum power output of the unloading branch. The PWM duty cycle of the DC discharge branch is D dump = sat[P dump * R dump / U dc 2 [, 0, 1],R dump U is the equivalent resistance of the discharge resistor. dc The real-time voltage of the DC bus is given by sat[·], which represents the limiting function. The exit condition for the DC discharge branch is that the DC bus voltage is lower than the preset exit threshold U. dump,off Furthermore, the effective energy margin risk factor is constrained to be lower than the preset exit threshold R. off And satisfy U dump,off < U dump,on R off < R on U dump,on R is the threshold voltage for the DC discharge branch. on To constrain the input threshold of effective energy margin risk factors.

10. The low voltage ride-through control method for a multi-timescale coordinated grid-type wind-storage system according to claim 1, characterized in that, After the fault is cleared, based on the rotor speed deviation, DC bus voltage deviation, and the restored active power output capacity of the grid side, the power reference value of the generator-side converter, the restored SOC power of the supercapacitor, and the DC bus reference voltage are adjusted in a coordinated manner until the system returns to normal operation. Specifically, this includes: The reference value for engine-side power is P s * = P MPPT + K ω,rec (ω r - ω r0 ) - K dc,rec (U dc - U dc * ), where P MPPT K is the maximum power point tracking reference value. ω,rec ω is the rotor overspeed recovery coefficient. r ω represents the real-time rotor speed. r0 U represents the rotor speed before the fault. dc U is the real-time voltage of the DC bus. dc * K is the DC bus reference voltage. dc,rec The DC bus voltage suppression coefficient is given, and the generator-side power change rate is limited to a preset recovery upper limit R. P,rec within; The SOC recovery power of the supercapacitor is P sc,rec * = -sat[K soc (SOC sc - SOC ref ), 0, P sc,dis,max ], where K soc SOC adjustment coefficient, SOC ref For the target state of charge, SOC sc P is the ratio of the actual amount of charge currently stored in a supercapacitor to its rated maximum storage capacity. sc,dis,max Let P be the maximum allowable discharge power of the supercapacitor, and let sat[·] represent the limiting function. sc,rec * Satisfy P sc,rec * ≤[P g,maxrec - P s * ]+, P g,maxrec This represents the maximum active power that can be transmitted from the grid side after the fault is cleared. When the DC bus voltage deviation, rotor speed deviation, and supercapacitor SOC deviation are all less than the corresponding preset thresholds, the system switches to normal operation mode, the generator-side converter resumes MPPT control, the supercapacitor resumes normal SOC management, and the grid-side converter resumes normal grid operation.

Images