A method and system for online reconfiguration of SVG reactive power capacity considering the degradation state of DC capacitors

By using the switching sampling circuit and slope difference to calculate the capacitance and equivalent series resistance in the cascaded H-bridge static var generator of SVG, and combining the shell temperature measurement to reconstruct the available reactive capacity online, the problem of difficulty in obtaining the degradation state of DC capacitors is solved, and the efficient utilization and capacity recovery of the device's output capacity are realized.

CN123121697APending Publication Date: 2026-11-17ELECTRIC POWER SCI RES INST OF STATE GRID XINJIANG ELECTRIC POWER CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611352909.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-02
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In existing technologies, it is difficult to accurately obtain the degradation state of DC capacitors online. The degradation state is disconnected from reactive power capacity control, resulting in insufficient utilization of the device's output capacity. In particular, under high modulation ratio conditions, the degradation links are difficult to bypass and reduce the load, and there is a lack of capacity recovery methods.

Method used

By using the DC capacitor terminal voltage latched by the sampling circuit at the bypass edge in the cascaded H-bridge static var generator of SVG, the capacitance and equivalent series resistance of the DC capacitor are calculated by combining the slope difference and step difference, and the degradation index is calculated by combining the case temperature measurement. This enables online reconfiguration of available reactive power capacity, and the capacity is recovered by thermal load reduction sorting and DC voltage reference reconfiguration.

Benefits of technology

It achieves accurate acquisition of DC capacitor capacitance and equivalent series resistance during normal operation, ensuring that the measurement process does not interfere with reactive power output. Slope compensation eliminates the influence of device time delay drift, resulting in high measurement accuracy. The capacitance reconstruction process converges smoothly, ensuring that the capacitor core temperature of the degraded link is within the allowable range, thus improving the device's output capability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123121697A_ABST
    Figure CN123121697A_ABST
Patent Text Reader

Abstract

This invention relates to the field of reactive power compensation technology in power systems, and more specifically, to an online reconfiguration method and system for SVG reactive power capacity considering the degradation state of DC capacitors. The method includes: Step 1, using the bypass edge when a link switches from the engaged state to the bypass state as the measurement timing, performing three-point timing sampling and slope compensation using a switching edge sampling circuit to jointly measure the capacitance and equivalent series resistance of the DC capacitors in each link online; Step 2, synthesizing the capacitor current, calculating the estimated capacitor core temperature based on the shell temperature measurement, and obtaining the capacitance degradation rate and resistance degradation ratio; Step 3, setting the allowable core temperature according to the degradation index, reconfiguring the available reactive power capacity under thermal and ripple constraints, limiting and reporting the reactive power command, and simultaneously achieving capacity recovery through thermal load reduction sorting and DC voltage reference reconstruction. This invention enables bidirectional adjustment of available reactive power capacity based on the shrinkage and recovery of DC capacitors, balancing reactive power output capability and equipment lifespan.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of reactive power compensation technology in power systems, specifically relating to an online reconfiguration method and system for SVG reactive power capacity considering the degradation state of DC capacitors. Background Technology

[0002] Static Var Generator (SVG) is a core component in renewable energy collection stations and impulsive load power supply systems, responsible for maintaining grid connection voltage and compensating for reactive power. Among these, the cascaded H-bridge SVG, directly connected to medium- and high-voltage power grids via a chain-link cascading method, is the most widely used. The DC capacitors within these links constantly withstand twice the power frequency ripple current and switching ripple current, making them among the components with the highest failure rate and shortest lifespan. A decrease in DC capacitor capacitance and an increase in equivalent series resistance directly compress the safe operating boundary of the device. Currently, obtaining the DC capacitor status mainly relies on offline detection during power outages, requiring the device to be taken out of operation, resulting in poor real-time performance. Existing online monitoring technologies are mostly based on twice-frequency ripple impedance calculations or recursive identification over long time windows, significantly affected by fluctuations in operating conditions and background harmonics. They typically only provide a single parameter, either capacitance or equivalent series resistance, and the measurement results are generally only used for over-limit alarms, disconnected from reactive power capacity control. In terms of capacity management, existing devices generally operate at constant values ​​as indicated on the factory nameplate, or dredge the entire unit based on the worst-performing link after degradation is detected. The former poses safety hazards due to over-ripple and over-temperature operation, while the latter leaves the output capacity of a large number of healthy links idle. Simultaneously, under high modulation ratio conditions, modulation freedom is nearly exhausted, making it difficult for degraded links to bypass and reduce load. Existing technologies lack corresponding capacity recovery methods. Therefore, there is an urgent need for a method that can accurately acquire the degradation state of DC capacitors online and directly convert the degradation state into reactive power capacity constraints and recovery control. Summary of the Invention

[0003] Therefore, the main objective of this invention is to provide an online reconfiguration method and system for SVG reactive power capacity that takes into account the degradation state of DC capacitors, so as to solve the problems in the prior art that the degradation state of DC capacitors is difficult to obtain accurately online, the degradation state and reactive power capacity control are disconnected, and the output capacity of the device is not fully utilized under degradation conditions.

[0004] The technical solution adopted in this invention is as follows: A method for online reconfiguration of SVG reactive power capacity considering the degradation state of DC capacitors, wherein the SVG is a cascaded H-bridge static var generator, each phase is composed of multiple cascaded links, and each link includes an H-bridge power unit and a DC capacitor, characterized by comprising the following steps: Step 1: The switching action edge when the link switches from the engaged state to the bypass state is recorded as the bypass edge, and the time when the bypass edge drive signal is issued is recorded as the trigger time. Using the switching edge sampling circuit connected between the DC capacitor terminals, the DC capacitor terminal voltage is latched at the advance time before the trigger time, at the trigger time, and at the delay blanking time after the trigger time to obtain the first latch value, the second latch value, and the third latch value. The second latch value minus the first latch value is recorded as the slope difference, and the third latch value minus the second latch value is recorded as the step difference. The slope difference, the step difference, the advance time, and the equivalent value of the capacitor current before commutation are used to calculate online the capacitance and equivalent series resistance of the DC capacitor of each link. The equivalent value of the capacitor current before commutation is determined by the link switching state value and the link current. Step 2: Calculate the estimated core temperature of each link based on the equivalent series resistance, capacitor current and shell temperature measurement values, and calculate the degradation index of each link based on the changes in capacitance and equivalent series resistance relative to the initial values. Step 3: Set the allowable core temperature for each link according to the degradation index. Calculate the upper limit of the capacitor current for each link based on the allowable core temperature, inlet air temperature measurement, and thermal resistance. Convert the upper limit of the capacitor current to the upper limit of the link current. Determine the available reactive power capacity of the SVG based on the upper limit of the link current of each phase's in-service links. When the reactive power command exceeds the available reactive power capacity, it is executed according to the available reactive power capacity limit and reported to the station control layer. In the link capacitor voltage equalization modulation, when the difference in capacitor voltage between links in the same phase exceeds the set equalization bandwidth, the activation order is determined according to the direction of capacitor voltage and link current. When it is within the equalization bandwidth, the activation order is determined according to the core temperature margin from large to small. The core temperature margin is the difference between the allowable core temperature and the estimated capacitor core temperature. Links with smaller core temperature margins receive a longer bypass time, their capacitor current decreases, and the upper limit of the link current increases, achieving bidirectional online reconfiguration of available reactive power capacity contraction and recovery.

[0005] Furthermore, the switching edge sampling circuit includes a DC blocking voltage divider network, a ringing suppression filter, a buffer amplifier, a three-channel sample-and-hold circuit, and a commutation detection comparator; the DC capacitor terminal voltage is filtered by the DC blocking voltage divider network to remove the DC component and attenuated according to the set voltage division ratio, then fed into the three-channel sample-and-hold circuit after passing through the ringing suppression filter and buffer amplifier, where the three-channel sample-and-hold circuit latches three latched values; the timing of the bypass edge is predetermined by the modulation comparison unit of the link controller; the advance time is taken... ~ Hidden time ~ It is greater than the sum of the power device turn-off commutation time and the equivalent series inductor oscillation decay time; the latched value is converted from analog to digital and restored according to the voltage division ratio, or it is obtained by continuous sampling of high-speed analog-to-digital converter at the same time; the commutation detection comparator is connected to the output of the DC blocking voltage divider network to detect the start time of the turn-off ringing, and the duration from the trigger time to the start time of the turn-off ringing is recorded as the commutation duration.

[0006] Furthermore, the equivalent value of the pre-commutation capacitor current is the product of the link switch state value before the bypass edge and the average value of the link current over a period of advance time before the trigger moment. The link switch state values ​​are +1, -1, or 0, corresponding to the two output polarities and bypass state of the engaged state. A droop compensation term is superimposed on the slope difference and step difference respectively. The droop compensation term is the average value of the two latched values ​​at the beginning and end of the corresponding sampling window multiplied by the ratio of the corresponding sampling window duration to the time constant of the DC blocking voltage divider network. The absolute value of the pre-commutation capacitor current equivalent value is multiplied by the advance time and then divided by the absolute value of the compensated slope difference to obtain the capacitance. The equivalent series resistance is obtained by multiplying the compensated slope difference by the ratio of the equivalent commutation time to the advance time, subtracting the compensated step difference, and then dividing by the pre-commutation capacitor current equivalent value. The equivalent commutation time is the sum of the commutation duration and the tailing equivalent time. The tailing equivalent time is given by a calibration table with the link current amplitude and power module temperature as query variables.

[0007] Furthermore, a bypass edge is selected where the absolute value of the equivalent value of the capacitor current before commutation is greater than 30% of the rated current of the link and the duration of the state before the bypass edge is greater than the advance time. The capacitance and equivalent series resistance corresponding to all effective bypass edges within one power frequency cycle are respectively taken as the arithmetic mean, and then the average is moved over 10 power frequency cycles, and updated once every 0.2s. During operation, the second harmonic components extracted by the switching edge sampling circuit and the capacitor voltage sampling channel are compared periodically. The voltage division ratio is corrected by the amplitude ratio, and the time constant of the DC blocking voltage divider network is corrected by the phase difference.

[0008] Further, in step 2, each control cycle multiplies the link switch state value with the link current sampling value to synthesize a capacitor current sequence. The effective value of the capacitor current is calculated according to the power frequency cycle. The link switch state value is +1, -1, or 0, corresponding to the two output polarities and bypass states of the engaged state. The equivalent series resistance is multiplied by the calibrated frequency correction coefficient to obtain the loss equivalent resistance. The frequency correction coefficient is the ratio of the loss-weighted equivalent series resistance under the typical current spectrum to the equivalent series resistance measured at the switch edge. The product of the core-shell thermal resistance, the loss equivalent resistance, and the square of the effective value of the capacitor current is filtered by a first-order low-pass filter and added to the shell temperature measurement value to obtain the estimated value of the capacitor core temperature. The time constant of the first-order low-pass filter is taken as the core-shell thermal time constant. The shell temperature measurement value is included in the calculation when it is within the set physical range and the rate of change is lower than the set threshold. In other cases, the estimated value of the capacitor core temperature of the adjacent link in the same cabinet plus the set margin is used instead, and a sensor maintenance prompt is generated.

[0009] Furthermore, the DC capacitor is a metallized film capacitor; the capacitance measured during the commissioning phase under set reference operating conditions is converted to... The equivalent series resistances are recorded as the initial capacitance and initial resistance, respectively; during operation, the equivalent series resistance is converted to the initial resistance using the estimated capacitor core temperature as a lookup variable according to the calibrated temperature correction curve. The baseline value will The ratio of the reference value to the initial resistance is recorded as the resistance degradation ratio. The difference between the initial capacitance and the current capacitance is divided by the initial capacitance and recorded as the capacitance degradation rate. The degradation indicators include the capacitance degradation rate and the resistance degradation ratio. When at least one of the two conditions is met, the DC capacitor of the corresponding link is determined to have reached the end of its life, a maintenance prompt is generated, and the corresponding link is cut off and replaced with a redundant link.

[0010] Furthermore, in step 3, allowable core temperatures are set according to capacitance degradation rates: when the capacitance degradation rate is between 0% and 1%, the allowable core temperature is set as follows: For every 1% increase in capacitance degradation rate, the allowable core temperature decreases accordingly. The lower limit of the allowable core temperature is taken The temperature difference is obtained by subtracting the measured inlet air temperature from the allowable core temperature. The thermal resistance is then multiplied by the sum of the core-shell thermal resistance and the heat dissipation thermal resistance, along with the equivalent resistance of the loss, to obtain the thermal resistance product. The upper limit of the capacitor current is obtained by dividing the temperature difference by the thermal resistance product and taking the square root. The heat dissipation thermal resistance is the thermal resistance from the capacitor casing to the inlet air, and during operation, it is calculated by taking the difference between the predicted and measured casing temperatures over a time constant. ~ The integral link is corrected online; when the estimated core temperature of the capacitor exceeds the allowable core temperature, the upper limit of the capacitor current of the corresponding link is reduced according to the ratio of the excess to the set temperature range, and the lower limit of the reduction coefficient is 0.

[0011] Furthermore, the ratio of the effective value of the capacitor current to the effective value of the link current in each power frequency cycle is recorded as the current proportion coefficient. The upper limit of the capacitor current is divided by the current proportion coefficient to obtain the upper limit of the thermally constrained link current. The peak-to-peak value of the DC voltage ripple of each link is statistically analyzed using the capacitor voltage sampling channel. The current effective value of the link current is multiplied by the ratio of the set allowable peak-to-peak value of the ripple to the current peak-to-peak value of the ripple to obtain the upper limit of the ripple-constrained link current. The upper limit of the thermally constrained link current, the upper limit of the ripple-constrained link current, and the rated current of the link are taken as the three values. The minimum value is the upper limit of the link current; the minimum value of the upper limit of the link current of each in-service link is taken as the upper limit of the reactive current of the corresponding phase, and the minimum value of the upper limit of the reactive current of the three phases is taken as the reactive current limit of the SVG symmetrical output; when the SVG is connected in a star configuration, the reactive capacity can be multiplied by 3 by the product of the reactive current limit and the effective value of the phase voltage at the grid connection point; when the SVG is connected in a delta configuration, the link current is the delta arm current, and the reactive capacity can be multiplied by 3 by the product of the reactive current limit and the effective value of the line voltage at the grid connection point; the reactive capacity can be multiplied by 3 by the product of the reactive current limit and the effective value of the line voltage at the grid connection point; the reactive capacity can be multiplied by 3 by the product of the reactive current limit and the effective value of the line voltage at the grid connection point. Update once, adjust the hysteresis band and rate limit. The hysteresis band is 2% of the rated reactive capacity, and the rate of change is within 10% of the rated reactive capacity per second. The equalization bandwidth is 2% to 5% of the rated capacitor voltage of the link. The input sequence is updated once per modulation cycle. Each link retains the minimum input duty cycle for compensating self-discharge, and the minimum input duty cycle is within 5%.

[0012] Furthermore, it also includes DC voltage reference reconfiguration: the maximum number of links that need to be put into operation in a certain phase during the power frequency cycle reaches the number of in-service links in the same phase and continues to do so. At that time, the DC voltage reference values ​​of all in-service links except the link with the smallest core temperature margin are increased by the same amount, so that the sum of the DC voltages of the links involved in the increase reaches or exceeds the peak value of the in-phase modulation voltage reference value. The increase amount is within 10% of the rated DC voltage of the link. The increase is performed according to the rate limit through the DC voltage outer loop and the capacitor voltage equalization loop. The DC voltage reference value of the link with the smallest core temperature margin is simultaneously decreased. The decreased value is above the minimum input voltage of the link power supply. The minimum input voltage is 40% to 60% of the rated DC voltage of the link. The life loss rate multiple during the increase period is the voltage acceleration exponent raised to the power of 1 and the sum of the increase amount. The voltage acceleration exponent is 7 to 10. The difference between the life loss rate multiple and 1 is accumulated with the activation time of DC voltage reference reconstruction. When the accumulated value reaches the set life loss budget, the DC voltage reference reconstruction is exited and a redundancy replacement request is generated.

[0013] An online reactive power reconfiguration system for SVG considering the degradation state of DC capacitors includes a switch edge sampling circuit, a case temperature sensor, and a link controller installed in each link; an inlet air temperature sensor installed in each power cabinet; a link current sensor for each phase and a voltage sensor at the grid connection point; and a central controller connected to each link controller via an optical fiber link. The case temperature sensor is mounted on the DC capacitor casing, the switch edge sampling circuit is connected between the DC capacitor terminals, and the output of the switch edge sampling circuit is connected to the link controller. The link controller sends up the latch value, commutation duration, trigger time marker, and link switch state value. The central controller buffers the link current sequence and reads the corresponding interval according to the trigger time marker. The central controller includes a parameter measurement module, a core temperature soft measurement module, and a capacity reconfiguration module, which respectively execute the calculation of capacitance and equivalent series resistance, the calculation of capacitor core temperature estimate and degradation index, and the calculation and allocation of available reactive power capacity. The output of the capacity reconfiguration module is connected to a reactive current command limiting stage, a link connection sequence generation stage, a DC voltage reference allocation stage, and a station communication interface. The system is used to implement any one of the above methods.

[0014] By adopting the above technical solution, the present invention achieves the following beneficial effects: The present invention uses the switching action of the link itself as the measurement excitation. Through timed sampling and slope compensation at the bypass edge, the capacitance and equivalent series resistance of the DC capacitor are simultaneously obtained during normal device operation. The measurement process does not interfere with the reactive power output. Slope compensation deducts the system deviation caused by charging and discharging during commutation. Online measurement of commutation time eliminates the influence of device delay drift with operating conditions, ensuring both measurement accuracy and adaptability to operating conditions. The switching edge measurement directly corresponds to the ohmic component change caused by the self-healing degradation of the metallized film capacitor. The measured quantity is consistent with the degradation mechanism, and the dual indicators of capacitance degradation rate and resistance degradation ratio cover the main failure modes. Steady-state capacitance... The limit value is based on the inlet air temperature, eliminating the potential for the limit value to rise automatically with the load. The online correction of heat dissipation thermal resistance makes the limit value follow the changes in heat dissipation conditions. The available reactive capacity automatically shrinks with the degradation state, ensuring that the capacitor core temperature of the degraded link is within the allowable range. The thermal load reduction sorting and DC voltage reference reconstruction recover capacity from the link level and the phase level, making the device output capacity significantly higher than the result of overall derating based on the worst link. The DC voltage reference reconstruction is also constrained by the lifetime loss budget, and the capacity recovery is achieved at a controllable lifetime cost. The update cycle of each control link is separated step by step and combined with hysteresis and rate limits. The reconstruction process converges smoothly. The available reactive capacity, along with the time stamp, is reported to the station control layer, which facilitates station-level coordination and scheduling. Attached Figure Description

[0015] Figure 1 The overall circuit schematic diagram of the cascaded H-bridge SVG and its monitoring and control system provided in the embodiments of the present invention; Figure 2 A circuit schematic diagram of the switching edge sampling circuit provided in an embodiment of the present invention; Figure 3 The sampling timing and voltage waveform diagram near the bypass edge provided in the embodiment of the present invention are shown in (a) and (b) respectively. (a) is the timing diagram of the link switch state and the three latch pulses, and (b) is the waveform diagram of the AC component of the DC capacitor terminal voltage. Figure 4 A waveform diagram illustrating the capacitor current synthesis and current proportion coefficient provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the equivalent thermal network for soft measurement of capacitor core temperature provided in an embodiment of the present invention. Figure 6 A characteristic curve showing the variation of the upper limit of capacitor current with capacitance degradation rate provided in an embodiment of the present invention; Figure 7 This is a timing diagram illustrating the switching states of degraded links before and after thermal load reduction sequencing, provided in an embodiment of the present invention. Figure 8 The distribution diagram of DC voltage reference values ​​of each link before and after DC voltage reference reconstruction is provided for embodiments of the present invention; Figure 9 A comparison curve of available reactive power capacity provided for embodiments of the present invention. Detailed Implementation

[0016] All features disclosed in this specification, or all steps in all disclosed methods or processes, may be combined in any way, except for mutually exclusive features and / or steps.

[0017] Any feature disclosed in this specification (including the abstract and figures) may be replaced by other equivalent or similar features, unless otherwise stated. That is, unless otherwise stated, each feature is merely one example of a series of equivalent or similar features.

[0018] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. The following embodiments use a certain... This paper focuses on the chain-type static var generator installed in a voltage-level new energy collection station. The component parameters, thresholds, and update cycles given in this paper are all verified and achievable values. Those skilled in the art can make adaptive adjustments to these values ​​within the scope of this invention based on device capacity, component selection, and heat dissipation conditions.

[0019] See Figure 1 In this embodiment, the SVG is a star-connected cascaded H-bridge static var generator with a rated reactive power capacity of [missing information]. The effective value of the line voltage at the grid connection point is Each phase consists of 20 cascaded in-service links connected to the grid point via a connecting reactor. The nominal reactance of the connecting reactor is 6%. Each phase is also equipped with two redundant links, and the ends of all phase links converge at the floating neutral point. Each link includes an H-bridge power unit and a DC capacitor. The DC capacitor is a metallized film capacitor with an initial capacitance of [missing information]. , The initial equivalent series resistance is as follows: The rated DC voltage of the chain link is The rated current of the chain link is The effective value of the link current under rated operating conditions is approximately Each link is equipped with a switching edge sampling circuit connected between the DC capacitor terminals. A case temperature sensor using a negative temperature coefficient thermistor is mounted in the middle of the capacitor casing. An inlet temperature sensor is installed at the air inlet of each power cabinet. A link current sensor is installed for each phase, and a voltage sensor is installed at the grid connection point. The link controller is based on a field-programmable logic device (FPGA) and is responsible for modulation pulse generation, latching timing control, and measurement data transmission. The central controller is connected to each link controller via a fiber optic link and is internally divided into a parameter measurement module, a core temperature soft measurement module, and a capacity reconstruction module. The output of the capacity reconstruction module is connected to a reactive current command limiting circuit, a link connection sequence generation circuit, a DC voltage reference distribution circuit, and a station communication interface.

[0020] The method in this embodiment revolves around three steps: first, online joint measurement of the capacitance and equivalent series resistance of the DC capacitors of each link; second, based on this, soft measurement of core temperature and calculation of degradation index; and finally, reconstruction of available reactive capacity under thermal and ripple constraints, and capacity recovery through thermal load reduction sorting and DC voltage reference reconstruction. Available reactive capacity can both shrink with degradation to ensure safety and recover through load reduction, forming a two-way online reconstruction of shrinkage and recovery, which differs from the one-way derating capacity management method.

[0021] Figure 2 The specific configuration of the switching edge sampling circuit is shown. The DC capacitor terminal voltage is first filtered out by a DC blocking voltage divider network and then attenuated according to a set voltage division ratio. In this embodiment, the DC blocking capacitor is [missing information - likely a specific value]. Voltage divider resistor 1 is selected Voltage divider resistor 2 is selected With a voltage divider ratio of 1 / 25, the time constant of the DC blocking voltage divider network is approximately The corresponding cutoff frequency is approximately lower than The voltage divider ratio is calculated based on the signal at the downstream stage, including the switching ringing signal. Linear range selection: The peak value of the AC component of the DC voltage ripple is approximately half of the allowable peak-to-peak ripple value. The off-ring amplitude is approximately to The sum of the two after 1 / 25 attenuation is approximately Within the linear range, the subsequent circuit operates entirely in the linear region, requiring no additional clamping protection or saturation recovery circuitry. The time constant following the voltage divider network is approximately... The ringing suppression filter is composed of Filter resistor and The filter capacitor is used to construct the off-ringing frequency in this embodiment, which is approximately... The voltage is attenuated by approximately 3.8 times by this filter, while the voltage step to be measured is... Ten time constants are established within the blanking time, so the waveform fidelity is basically unaffected. The buffer amplifier is a follower formed by high input impedance operational amplifiers. Its output is simultaneously fed into the first, second, and third sample-and-hold circuits. The three latched values ​​are converted to digital values ​​using a 16-bit analog-to-digital converter and restored according to the voltage division ratio before being used in the calculation. The input of the commutation detection comparator is connected before the output of the DC blocking voltage divider network, i.e., before the ringing suppression filter, in order to completely preserve the initial characteristics of the ringing. The comparator detects the start time of ringing cut-off, and the timer... The resolution records the duration from the trigger moment to that moment, which is used as the switching duration.

[0022] This embodiment selects the measurement timing as follows. The switching action edge when the link switches from the engaged state to the bypass state is denoted as the bypass edge. Before the bypass edge arrives, the DC capacitor carries a current of the same amplitude as the link current. After the commutation is turned off, the capacitor current approaches zero, and the internal voltage of the capacitor remains constant. Choosing the bypass edge as the measurement timing ensures that the latter half of the latch window, within a flat region where the voltage remains constant, results in measurement results containing only two deterministic components: the charge / discharge slope and the resistive step. These conditions cannot be obtained at the engaged edge. The timing of the bypass edge is predetermined by the modulation comparison unit of the link controller. Based on this, the link controller issues three latch pulses: the first sample-and-hold circuit latches the first latch value a time advance before the trigger time; the second sample-and-hold circuit latches the second latch value at the trigger time; and the third sample-and-hold circuit latches the third latch value a time delay after the trigger time. The trigger time is the time when the bypass edge drive signal is issued. In this embodiment, the advance time is... Hidden time The blanking time is greater than the sum of the power device turn-off commutation time and the oscillation decay time of the equivalent series inductance of the DC capacitor. There is a trade-off in the value of the advance time: if the value is too short, the slope difference amplitude is small and the signal-to-noise ratio is low; if the value is too long, the change in the power frequency current within the window is difficult to ignore. Considering both requirements, the range of the advance time is [insert range here]. to . Figure 3 (a) shows the latch timing near the bypass edge: the link switch state changes from engaged to bypass at the trigger time, and the three vertical dashed lines correspond to the first latch time, the second latch time, and the third latch time, respectively. Approximately after the trigger time... At the start of the ringing interruption, the blanking time ends on the flat segment after the ringing has fully decayed.

[0023] Let the values ​​of the first latch, second latch, and third latch, after being restored according to the voltage divider ratio, be as follows: , and ,Will reduce This is denoted as the slope difference. ,Will reduce denoted as step difference The high-pass characteristic of the DC blocking voltage divider network causes a droop in the latch difference that is proportional to the instantaneous ripple value. This droop originates from the slow discharge of the DC blocking capacitor within the latch window. The slope difference is compensated for using the following formula: ;in, The slope difference after compensation. This is the first latch value. This is the second latch value. In order to advance the time, The time constant of the DC blocking voltage divider network; That is, the average of the two latch values ​​at the beginning and end of the window, representing the average level of the signal within the window, multiplied by the window duration. The ratio of these values ​​yields the droop within the window, and this compensation is strictly accurate for linearly varying ripples within the window. The step difference is compensated using the following formula: ;in, The compensated step difference value, This is the third latch value. This is the blanking time. In this embodiment, the instantaneous ripple value is... When the slope difference is near, the droop deviation is approximately It accounts for approximately 4% of the slope difference, and the residual droop after compensation is less than 0.5% of the slope difference; the step difference window is... The compensation amount for the same instantaneous ripple value is approximately .

[0024] The capacitance and equivalent series resistance are calculated together using the following two formulas: ; ;in, For capacitance, For equivalent series resistance, This is the equivalent value of the capacitor current before commutation. This is the equivalent commutation time. The value is equal to the product of the link switching state value before the bypass edge and the average value of the link current over an interval of time equal to the lead time before the trigger moment. The link switching state values ​​are +1, -1, or 0, corresponding to the two output polarities of the engaged state and the bypass state, respectively. The capacitance is defined as the ratio of the charge increment to the voltage increment, and the charge increment corresponds to the interval average value; therefore, the interval average value is used instead of a single-point value. The high-frequency component of the switching ripple superimposed on the link current is simultaneously suppressed in the averaging process, and its systematic influence is eliminated. and The ratio is the voltage slope within the latch window, and this slope is equal to... and The ratio of these values ​​is used to obtain the capacitance; in the formula for calculating the equivalent series resistance, the slope compensation term... Multiply and The ratio is exactly equal to the voltage change caused by residual charging and discharging during the commutation shutdown period, which cancels out the corresponding component in the step difference, leaving only the purely resistive step. The system deviation caused by charging and discharging during the commutation period is thus completely deducted. As a set of actual measurement examples, a link nearing the end of its lifespan was measured... for The difference in slope after compensation is Therefore, the capacitance is obtained. The measured commutation time in this measurement was: The equivalent time for the trailing effect is obtained by looking up the table. The equivalent commutation time is The compensated step difference is After substituting, the equivalent series resistance is approximately . Figure 3 (b) in the figure shows the corresponding AC component waveform of the DC capacitor terminal voltage. The geometric meanings of the three latch values, slope difference, and step difference are marked in the figure. The amplitude of the turn-off ringing exceeds the range of the vertical axis of the figure.

[0025] The equivalent commutation time needs to be obtained independently because, at any current level, the compensated step difference value and... The ratio is always equal to the following formula: ;Right now and Combination and The coupling remains in a fixed form, and the two cannot be separated solely by latched data from the bypass edge. Increasing the number of bypass edge acquisitions under different currents also fails to separate them; this limitation stems from the measurement principle itself. Therefore, this embodiment decomposes the equivalent commutation time into the sum of the commutation duration and the tailing equivalent time: the commutation duration is measured edge-by-edge by the commutation detection comparator, covering the turn-off delay and current drop duration; this part directly follows the changes in collector current and junction temperature through edge-by-edge measurement; the tailing equivalent time corresponds to the equivalent duration of the tailing charge, and its variation with current and junction temperature is reflected in a calibration table using the link current amplitude and power module temperature as lookup variables. In this embodiment, the calibration table is based on the link current... to Module temperature to Values ​​within a range to During the commissioning phase, the calibration table is corrected based on the equivalent series resistance measured by the impedance analyzer. The threshold of the commutation detection comparator is set proportionally to the link current amplitude. This is because the ringing amplitude is proportional to the rate of change of current, and thus proportional to the current being turned off. A fixed threshold would introduce a time scale deviation that varies with the current. When the comparator output count is 0 or the link current amplitude is lower than the set threshold within the set detection window, the equivalent commutation time is taken from the calibration table value to ensure that the measurement link has a definite output under any operating condition.

[0026] Not every bypass edge is included in the calculation. Bypass edges that simultaneously meet the following two conditions are selected as valid bypass edges: Condition 1, The absolute value is greater than 30% of the rated current of the link, ensuring that the voltage step and slope difference are within the resolvable range of the amplification and quantization stages. In this embodiment, the resistive step, after being converted to voltage division, is approximately... The order of magnitude, 16-bit analog-to-digital conversion in ± The resolution within the range is approximately Sufficient quantization margin; Condition 2, the duration of the link being in the engaged state before the bypass edge is greater than the advance time, ensuring that the capacitor current remains constant throughout the entire window from the first latching moment to the second latching moment. The link controller transmits three latch values, commutation duration, trigger time marker, and link switch status values ​​via optical fiber. The central controller stores the most recent data at a fixed sampling rate. The link current sequence is read according to the trigger time mark, and the corresponding interval is read and the pairing calculation is completed. The calculation results of all effective bypass edges within one power frequency cycle are arithmetically averaged, and then averaged over 10 power frequency cycles. The capacitance and equivalent series resistance are then calculated. Updated once. During normal operation, the effective bypass edges in each power frequency cycle are usually more than 20. The averaging process suppresses random noise while preserving the trend of slow parameter changes.

[0027] The parameter measurement link is configured with two levels of calibration. During the commissioning phase, under the set baseline operating conditions (50% rated reactive power and inlet air temperature),... The above measurement procedure was run nearby, and the measured capacitance was converted to... The equivalent series resistance is recorded as the initial capacitance and initial resistance of each link. Field self-calibration ensures that the initial reference and online measurement share the same link, and inherent deviations in the link are naturally canceled out when making ratios. The initial values ​​of the time constant and voltage division ratio of the DC blocking voltage divider network are calibrated at the factory. During operation, the second harmonic component extracted by continuous sampling along the sampling circuit of the switch is periodically compared with the second harmonic component extracted by the existing capacitor voltage sampling channel. The voltage division ratio is corrected by the amplitude ratio, and the time constant of the DC blocking voltage divider network is corrected by the phase difference. The voltage division ratio is determined by the resistance and is stable over a long period, while the time constant is determined by the DC blocking capacitor and drifts slowly with capacitor aging. The drift mechanisms of the two are different, so they are corrected separately. The time constant is determined by... Drift to At that time, the amplitude of the second harmonic component changed by only about 0.2%, but the phase changed from approximately Become approximately Phase difference is a sensitive and identifiable observation. The power pulsation of the link under pure reactive operation is located at the second harmonic, and this component naturally exists simultaneously in both channels, requiring no additional excitation for verification.

[0028] Step 2, the core temperature soft measurement, begins with capacitor current synthesis. The central controller performs this measurement in each control cycle (in this embodiment...). Multiply the switch state value of the chain link by the sampled current value of the chain link to synthesize the capacitor current sequence, and calculate the effective value of the capacitor current according to the power frequency period. . Figure 4 The synthesis process under rated operating conditions is given: the peak value is shown above. Valid value The link current is shown in the middle, representing the link switching state. The link voltage reference is phased with the link current. The duty cycle of this link is maximum near the zero-crossing of the link current and minimum near the peak value of the link current. Below is the capacitor current obtained by multiplying the two values ​​point by point; its effective value is approximately... The ratio of the current to the effective value of the link current is approximately 0.65; this ratio is the current proportion factor. Both the switch status and the link current are existing information from the control system; the acquisition of the capacitor current does not depend on any additional sensors.

[0029] Loss calculations use the equivalent resistance of the loss: ;in, The equivalent resistance for losses. The frequency correction factor is set to 1.08 in this embodiment. The equivalent series resistance measured at the switching edge reflects the ohmic component at high frequencies, while the capacitor heating is generated by the ripple current of each frequency band on the equivalent series resistance at the corresponding frequency point. There is a fixed proportional relationship between the two, and the frequency correction factor is used to complete the conversion between the two. Defined as the ratio of the loss-weighted equivalent series resistance under typical current spectrum to the equivalent series resistance measured at the switching edge, it is obtained from the factory impedance spectrum calibration. In this embodiment... The equivalent series resistance at is High-frequency ohmic components are After spectral weighting, the value is 1.08. The self-healing degradation of metallized film capacitors manifests as an increase in the resistance of the metallization layer and the gold plating layer, i.e., an increase in the ohmic component. The switching edge measurement directly corresponds to this component, and the measured value is consistent with the degradation mechanism. As degradation progresses, the proportion of the ohmic component in the total loss increases. The value gradually changes from 1.08 to approximately 1.04. The value of 1.08 is used throughout the process, which is slightly conservative in the later stages of degradation. Based on this, the estimated value of capacitor core temperature and the upper limit of capacitor current are both on the safe side.

[0030] The estimated capacitor core temperature is calculated using the following formula: ;in, This is an estimated value for the capacitor core temperature. This is the shell temperature measurement value. The thermal resistance from the capacitor core to the casing is, in this embodiment, During calculation, , and The product of the squares of the three terms is then subjected to a first-order low-pass filter before being combined with... The filtering time constant is taken as the core-shell thermal time constant, which is used in this embodiment. This is to reflect the thermal inertial response of core temperature to changes in loss. The core-shell thermal time constant is calibrated using a calibration sample of the embedded thermocouple: a capacitor is passed through the sample... The effective value of the current, the loss is approximately The core-shell temperature difference under thermal steady state is approximately Thus we obtain The core-shell thermal time constant is fitted by the rise process of the core-shell temperature difference. Figure 5 The corresponding equivalent thermal network is given. The heat loss flows through the core-shell thermal resistance to the shell temperature node, and then through the heat dissipation thermal resistance to the inlet air temperature node. The parallel thermal capacitance at the core temperature node corresponds to the core-shell thermal time constant. The shell temperature measurement value is taken out at the shell temperature node. The shell temperature measurement value is verified for rationality before being used in the calculation: the value is within... to Within the physical range and at a rate of change lower than The time-based determination is reliable; in other cases, the estimated capacitor core temperature of adjacent links in the same cabinet is added. The system provides a safety margin for replacement and generates sensor maintenance prompts, allowing the system to switch to safe-side operation when a single thermistor fails, rather than directly causing false protection. The effective value of the capacitor current and the estimated value of the capacitor core temperature are... Updated once.

[0031] The degradation index is calculated as follows. Using the estimated capacitor core temperature as the lookup variable based on the calibrated temperature correction curve, the equivalent series resistance is converted to... The reference value is that the temperature coefficient of the ohmic component of a metallized thin-film capacitor is approximately per liter of [value missing]. Increased by 0.35%, The conversion factor for the vicinity is approximately 1.16; The ratio of the baseline value to the initial resistance is denoted as the resistance degradation ratio, and the difference between the initial capacitance and the current capacitance divided by the initial capacitance is denoted as the capacitance degradation rate. When at least one of the following two conditions is met—a capacitance degradation rate of 5% and a resistance degradation ratio of 2.0—the DC capacitor of the link is determined to have reached the end of its lifespan. A maintenance prompt is generated, and the link is disconnected and replaced with a redundant link. The reason for using dual criteria is that the self-healing failure of metallized film capacitors is mainly manifested by capacitance loss, while the degradation of the gold plating layer is led by an increase in resistance. The two indicators corroborate each other, covering the main failure paths. In the aforementioned measured example, the estimated core temperature of the link is approximately... , measured The temperature correction curve above was used to calculate the result. benchmark value approximately , with initial resistance Compared to a resistance degradation ratio of 2.0, the capacitance degradation rate is 5%. Both criteria are met simultaneously, indicating that the device has reached the end of its lifespan.

[0032] Step 3: First, set the allowable core temperature for each link according to the capacitance degradation rate: when the capacitance degradation rate is between 0% and 1%, the allowable core temperature is... For every 1% increase, the corresponding decrease The lower limit is taken The grading relationship is calculated by inversely from the lifespan temperature characteristics of the capacitor: the capacitor in this embodiment... The expected lifespan at core temperature is approximately 100,000 hours. For every decrease in core temperature... With the lifespan doubled, and assuming a remaining time of 80,000 hours before the target capacitor replacement period, the remaining lifespan is calculated based on the capacitance degradation rate. Then, the upper limit of the core temperature required to guarantee the target lifespan is determined step-by-step and rounded down. The integer increments. The more severely degraded the link, the lower the core temperature it is constrained to operate at, and its subsequent degradation rate slows down accordingly, thus breaking the positive feedback loop where heating and aging accelerate each other.

[0033] The upper limit of capacitor current under thermal constraint is calculated using the following formula: ;in, This is the upper limit of the capacitor current. For the allowable core temperature, This is the measured value of the inlet air temperature. In this embodiment, the thermal resistance from the capacitor casing to the air intake is considered. initial value Typical intake air temperature in summer is In the formula, the equivalent resistance of the loss is taken as the online measured value. When the chain link operates continuously at this limit, the capacitor core temperature approaches the allowable core temperature, and the equivalent resistance of the loss simultaneously approaches the value at the allowable core temperature. The limit and the core temperature are related. The solution converges to a self-consistent solution during the update cycle. The steady-state limit is based on the inlet air temperature rather than the case temperature because the case temperature itself increases with the load. If the case temperature is used as the reference, the actual core temperature will exceed the allowable core temperature after loading according to the limit. However, when the inlet air temperature is used as the reference, the steady-state core temperature after continuous loading according to the limit is exactly within the allowable core temperature. Online correction during operation: Calculate the predicted shell temperature using the following formula. ;in, The predicted shell temperature is calculated as the difference between the predicted and measured shell temperature over a time constant. Correction of the integral process Correction when power loss is greater than The threshold is set when the input is within ±30% of the initial value and the correction amount is within this range. Under light load, the residual information is low, and setting a threshold can prevent correction drift. This mechanism allows the capacity limit to follow slow changes in heat dissipation conditions such as dust accumulation in the air duct and filter blockage. Because... and Proportional to the square root, the impact of resistance measurement deviation on capacitance limits is halved; for example, a 10% resistance deviation only causes about a 5% limit deviation, thus enhancing the fault tolerance of the measurement link. In the metallized thin-film capacitor of this embodiment, the capacitance loss of the metallization layer and the increase in resistance of the gold-plated layer proceed synchronously. For every 1% increase in capacitance degradation rate, converted to... The equivalent series resistance increases by about 20% relative to the initial resistance. When the capacitance degradation rate reaches 5%, the resistance degradation ratio is exactly 2.0, which is consistent with the two life end criteria mentioned above. This correspondence was obtained by fitting the accelerated aging test data of capacitors of the same model. Figure 6 Based on this relationship, the characteristic curve of the upper limit of capacitor current as a function of capacitance degradation rate is given: the initial state is approximately As degradation progresses, the curve gradually shifts downwards, with the allowable core temperature corresponding to each segment marked above the curve. When the capacitance degradation rate is 4%, it is approximately... When the value reaches 5%, the link is cut off, and the curve terminates. When the estimated capacitor core temperature exceeds the allowable core temperature, the excess amount is adjusted accordingly. The ratio of the temperature bands reduces the upper limit of the capacitance current of the link, and the lower limit of the reduction coefficient is 0; by using proportional reduction instead of direct blocking, the oscillation of the limit near the critical point is avoided.

[0034] The upper limit of capacitor current needs to be converted into a current constraint at the link level. The ratio of the effective value of capacitor current to the effective value of link current in each power frequency cycle is denoted as the current proportion coefficient. Dividing the upper limit of capacitor current by the current proportion coefficient yields the upper limit of thermally constrained link current, initially set as follows: Divide by 0.65 approximately Exceeding the rated current of the chain link This indicates that thermal constraint does not constitute a bottleneck under healthy conditions. Ripple constraint parallel calculation: Using the existing capacitor voltage sampling channel, the peak-to-peak value of the DC voltage ripple for each link is statistically analyzed according to the power frequency cycle. The current effective value of the link current is multiplied by the ratio of the allowable peak-to-peak value of the ripple to the current peak-to-peak value of the ripple, yielding the upper limit of the ripple constraint link current. The allowable peak-to-peak value of the ripple is taken as 16% of the rated DC voltage of the link. The measured peak-to-peak ripple value under rated operating conditions is approximately The corresponding upper limit of the ripple constraint link current is approximately The basis for the proportional extrapolation is that the peak-to-peak value of the ripple is approximately proportional to the reactive current. When the effective value of the link current is less than 10% of the rated value, the current peak-to-peak value of the ripple is too small and the ratio is distorted. This item uses the result of the most recent calculation. Finally, the minimum value of the upper limit of the thermal constraint link current, the upper limit of the ripple constraint link current, and the rated current of the link is taken as the upper limit of the link current of the link.

[0035] The available reactive capacity can be synthesized and executed as follows: Take the minimum value among the upper limits of the link current of all in-service links in each phase as the upper limit of the reactive current for that phase. Then, take the minimum value among the upper limits of the reactive current of the three phases as the reactive current limit for the symmetrical output of the SVG. The available reactive capacity can then be obtained using the following formula: ;in, Available reactive power capacity This is the effective value of the phase voltage at the grid connection point. As a reactive current limit, the effective value of the phase voltage in this embodiment is approximately [value missing]. The reactive current limit is subject to the rated reactive current. Given the constraints, the smaller of the two values ​​is taken; the available reactive power capacity under healthy conditions is the rated value. Available reactive power capacity per Updated once, adjusting the settings to include hysteresis band and rate limit. The hysteresis band is set to 2% of the rated reactive power capacity. The rate of change is 10% of the rated reactive power capacity per second. Within / s; when the reactive power command exceeds the available reactive power capacity, it is executed according to the available reactive power capacity limit. The available reactive power capacity, along with the time stamp, is reported to the station control layer via the station communication interface for use by the station-level reactive power and voltage control system. The activation sequence update cycle is... The available reactive power capacity update cycle is The thermal resistance correction time constant is: The three parameters differ by more than one order of magnitude. The fast loop considers the slow loop parameters to be constant, and the slow loop considers the fast loop to have converged. With the hysteresis band and rate limit, the capacity reconstruction process converges monotonically and smoothly. No capacity oscillation was observed in the actual test.

[0036] The first step in capacity recovery is thermal load reduction sequencing. During link capacitor voltage equalization modulation, when the voltage difference between in-phase links exceeds the set equalization bandwidth, the activation order is determined based on the capacitor voltage level and the direction of the link current, bringing the capacitor voltage back within the equalization bandwidth. The equalization bandwidth is set to 3% of the rated capacitor voltage of the link. When the capacitor voltage difference is within the equalization bandwidth, the activation order is determined by the core temperature margin from largest to smallest. The core temperature margin is the difference between the allowable core temperature and the estimated capacitor core temperature. Links with larger core temperature margins are activated first, and links with smaller margins are bypassed first. This order is maintained for each modulation cycle. Execute once. The priority order of voltage balancing and thermal load reduction ensures that load reduction does not come at the expense of voltage balancing. Figure 7 The sorting results are presented as follows: Under full-capacity output and a peak requirement of 20 links, before thermal load reduction sorting, links are sorted by capacitor voltage, with equal opportunities for activation, and the duty cycle of degraded links is approximately 63%. After thermal load reduction sorting, their activation needs are taken over by healthy links, which are only forced to activate during periods when the number of required links exceeds the number of other active links, reducing the duty cycle to approximately 8%. The peak period for required links occurs near the peak modulation voltage, at which point the link current is close to 0. Degraded links bear very little current during forced activation periods, and their effective capacitor current decreases from approximately... dropped to about The current ratio coefficient decreased from 0.65 to 0.11, and the upper limit of the thermal constraint link current correspondingly increased; the effective value of the capacitive current of the 19 healthy links that were replaced increased to approximately The current ratio factor rises to 0.66, shifting the capacity bottleneck from degraded links to healthy links. This leads to a rebound in the upper limit of reactive current and available reactive capacity for that phase, resulting in available reactive capacity exceeding the overall derating result based on the worst-performing link. The minimum duty cycle is maintained because the link controller's power supply and voltage equalization resistor continuously draw power from the DC capacitor; in this embodiment, the power draw is approximately... A complete bypass will cause the DC voltage to drop slowly and conflict with voltage balancing control; the DC voltage drop caused by this power draw within one power frequency cycle is approximately When the capacitor voltage drops to the lower edge of the equalization bandwidth, the voltage equalization loop can be briefly engaged during the charging period of the link current to compensate for the voltage drop. The equivalent minimum engagement duty cycle is 1%.

[0037] The second execution stage of capacity recovery is DC voltage reference reconstruction, used in deep operating conditions where sequencing degrees of freedom are exhausted. This occurs when the maximum number of links that need to be put into service in a phase within the power frequency cycle reaches the number of in-service links in the same phase and remains so. At this point, it is determined that sorting alone is insufficient to provide bypass time for degenerate links. In this embodiment, this condition occurs when the device is operating at full capacity: the effective value of the phase voltage at the grid connection point is approximately... The effective value of the voltage connected to the reactor is approximately The sum of the two yields a calculated peak phase modulation voltage requirement of approximately The sum of the DC voltages of the 20 in-service links is This requires the voltages of all 20 links to participate in the synthesis. Therefore, the DC voltage reference values ​​of the 19 in-service links (excluding the link with the smallest core temperature margin) are each increased by 5.3% to... The minimum upsampling required to synthesize this peak value from 19 links is 3.5%. Taking into account a voltage margin of approximately 2%, we take 5.3%. The sum of the voltages of the 19 links reaches... If the modulation peak demand is exceeded, the link with the smallest core temperature margin is thus granted bypass conditions; the DC voltage adjustment via the outer loop and capacitor voltage equalization loop is executed slowly at a limited rate, with the adjustment range within 10% of the link's rated DC voltage. The DC voltage reference value of the link with the smallest core temperature margin is simultaneously reduced to... To reduce voltage stress, the adjusted value is above the minimum input voltage of the power supply for the chain link. In this embodiment, the input range of the power supply is... to The minimum input voltage is 41% of the rated DC voltage, which is within the range of 40% to 60%. Figure 8 The distribution of DC voltage reference values ​​for each link before and after reconfiguration is shown. The 20 links are unified. Adjusted to 19 links With 1 link .

[0038] DC voltage reference reconstruction trades capacity for a controllable lifetime cost, constrained by a lifetime loss budget. The lifetime of metallized film capacitors decreases power-lawfully with increasing voltage. The lifetime loss rate during the upward adjustment period is multiplied by the voltage acceleration exponent raised to the power of 1 and the sum of the upward adjustment magnitude. The voltage acceleration exponent is taken as 7 to 10, and 8 in this embodiment. The cumulative lifetime loss is calculated by the following formula: ;in, To accumulate lifespan loss, To increase the range, The voltage acceleration index, For the first The activation duration is summed and iterated through each activation. When the increase is 5.3%, the lifespan loss rate multiplier is approximately 1.51, meaning that every hour of activation consumes an additional 0.51 hours of lifespan. In this embodiment, the annual lifespan loss budget is taken as 5% of the rated annual duration, i.e., 438 hours, corresponding to a total annual activation duration of approximately 860 hours. When the increase is 8%, the lifespan loss rate multiplier is approximately 1.85, meaning that every hour of activation consumes an additional 0.85 hours of lifespan. The total annual activation duration corresponding to the same budget is shortened to approximately 515 hours. The activation duration decreases power-law with the increase, thus directly binding the duration of capacity recovery to voltage stress in the lifespan loss budget. When the cumulative value reaches the budget, the DC voltage reference reconstruction is exited and a redundant replacement request is generated, and the subsequent handling is handed over to the maintenance system. Figure 9 Two capacity management methods were compared. The horizontal axis of the graph represents the capacitance degradation rate of the worst-performing link in the same phase, and the capacitance degradation rate of the remaining links in the same phase is taken as 0.48 times this value, to reflect the dispersion of capacitance degradation within the same batch of capacitors. When the worst-performing link is derated as a whole, under the condition that the capacitance degradation rate of the worst-performing link is 4%, the available reactive power capacity is reduced to approximately [missing value]. After adopting the thermal load reduction sequencing and DC voltage reference reconstruction of this embodiment, the available reactive power capacity under the same operating conditions is maintained at approximately [percentage missing]. The device output capacity is increased by approximately 30%; when the worst-case link capacitance degradation rate is 2% and 3%, the available reactive power capacity based on the overall derating of the worst-case link is approximately... and After adopting this embodiment, they are respectively maintained at approximately The output capacity is increased by about 9% and about 27% respectively, and the capacity advantage is more obvious as the degradation is more severe. Under this deep operating condition, the bypass time of the degraded link is further increased from about 92% after thermal load reduction to about 99%. That is, after the duty cycle is reduced to the minimum duty cycle and the DC voltage reference value is reduced to 50% of the rated value, the life loss rate calculated according to the same power law is less than 1% of the rated voltage, and the degradation process basically stops.

[0039] As an optional implementation, the SVG can also use a delta connection, in which case the link current is the delta arm current, the link current sensor is set in each delta arm, and the available reactive capacity is obtained by multiplying the product of the effective value of the grid connection point line voltage and the reactive current limit of the arm by 3, and the rest of the steps of the method remain the same.

[0040] As an optional implementation, the three latch values ​​can also be obtained by continuously sampling the output of the buffer amplifier using a high-speed analog-to-digital converter and selecting them from the sampling sequence at the same time, with the sampling rate being [value missing]. The quantization error at the selected time point is within ± Within this range, the impact on the capacitance and equivalent series resistance measurement results is less than 0.4% and 2% respectively, and the selection accuracy meets the requirements. At this time, the first sample-and-hold circuit, the second sample-and-hold circuit, and the third sample-and-hold circuit are equivalently implemented by digital point selection, and the commutation detection can also be completed by the slope discrimination of the high-speed sampling sequence, further simplifying the circuit.

[0041] As an optional implementation, the DC capacitor can also be an aluminum electrolytic capacitor. In this case, the life end criterion is adjusted to at least one of the following: the capacitance degradation rate reaches 20% and the resistance degradation ratio reaches 2.0. The frequency correction coefficient is increased to 2 to 4 due to the strong frequency dependence of the equivalent series resistance of the electrolytic capacitor. It needs to be checked regularly during the maintenance window in conjunction with offline measurement. The remaining steps of the method remain the same.

[0042] As an optional implementation, each setting value can be adjusted within the following range: advance time to Hidden time to The effective bypass current threshold is taken as 30% to 50% of the rated current of the link, the equalization bandwidth is taken as 2% to 5% of the rated capacitor voltage of the link, the minimum duty cycle is within 5%, the hysteresis band is taken as 1% to 3% of the rated reactive capacity, the rate of change is taken as 5% to 20% of the rated reactive capacity per second, and the inlet air temperature is substituted according to the actual measured value of the season. These adjustments do not change the implementation method and the nature of the effect.

[0043] In summary, this embodiment completes the online joint measurement of DC capacitor capacitance and equivalent series resistance during normal device operation, transforms the degradation state step by step into allowable core temperature, upper limit of capacitor current, upper limit of link current and available reactive capacity, and realizes capacity recovery through thermal load reduction sorting and DC voltage reference reconstruction. Measurement, constraint and execution are connected step by step to form a closed loop of online reactive capacity reconstruction considering the degradation state of DC capacitor.

[0044] While specific embodiments of the present invention have been described above, those skilled in the art should understand that these specific embodiments are merely illustrative. Those skilled in the art can omit, substitute, and modify the details of the above methods and systems in various ways without departing from the principles and essence of the present invention. For example, combining the above method steps to perform substantially the same function and achieve substantially the same result in substantially the same manner still falls within the scope of the present invention. Therefore, all equivalent transformations and modifications made within the principles and spirit of the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for online reconfiguration of SVG reactive power capacity considering the degradation state of DC capacitors, wherein, SVG is a cascaded H-bridge static var generator, where each phase consists of multiple cascaded links. Each link includes an H-bridge power unit and a DC capacitor. Its key feature is the inclusion of the following steps: Step 1: The switching action edge when the link switches from the engaged state to the bypass state is recorded as the bypass edge, and the time when the bypass edge drive signal is issued is recorded as the trigger time. Using the switching edge sampling circuit connected between the DC capacitor terminals, the DC capacitor terminal voltage is latched at the advance time before the trigger time, at the trigger time, and at the delay blanking time after the trigger time to obtain the first latch value, the second latch value, and the third latch value. The second latch value minus the first latch value is recorded as the slope difference, and the third latch value minus the second latch value is recorded as the step difference. The slope difference, the step difference, the advance time, and the equivalent value of the capacitor current before commutation are used to calculate online the capacitance and equivalent series resistance of the DC capacitor of each link. The equivalent value of the capacitor current before commutation is determined by the link switching state value and the link current. Step 2: Calculate the estimated core temperature of each link based on the equivalent series resistance, capacitor current and shell temperature measurement values, and calculate the degradation index of each link based on the changes in capacitance and equivalent series resistance relative to the initial values. Step 3: Set the allowable core temperature for each link according to the degradation index. Calculate the upper limit of the capacitor current for each link based on the allowable core temperature, inlet air temperature measurement, and thermal resistance. Convert the upper limit of the capacitor current to the upper limit of the link current. Determine the available reactive power capacity of the SVG based on the upper limit of the link current of each phase's in-service links. When the reactive power command exceeds the available reactive power capacity, it is executed according to the available reactive power capacity limit and reported to the station control layer. In the link capacitor voltage equalization modulation, when the difference in capacitor voltage between links in the same phase exceeds the set equalization bandwidth, the activation order is determined according to the direction of capacitor voltage and link current. When it is within the equalization bandwidth, the activation order is determined according to the core temperature margin from large to small. The core temperature margin is the difference between the allowable core temperature and the estimated capacitor core temperature. Links with smaller core temperature margins receive a longer bypass time, their capacitor current decreases, and the upper limit of the link current increases, achieving bidirectional online reconfiguration of available reactive power capacity contraction and recovery.

2. The method according to claim 1, characterized in that, The switching edge sampling circuit includes a DC blocking voltage divider network, a ringing suppression filter, a buffer amplifier, a three-channel sample-and-hold circuit, and a commutation detection comparator. The DC capacitor terminal voltage is filtered by the DC blocking voltage divider network to remove the DC component and attenuated according to the set voltage division ratio. It is then fed into the three-channel sample-and-hold circuit after passing through the ringing suppression filter and the buffer amplifier. The three-channel sample-and-hold circuit latches three latched values. The timing of the bypass edge is predetermined by the modulation comparator unit of the link controller. The advance time is 10μs~50μs, and the blanking time is 2μs~10μs, which is greater than the sum of the power device turn-off commutation time and the equivalent series inductor oscillation decay time. The latched value is converted from analog to digital and restored according to the voltage division ratio, or it is obtained by continuous sampling by high-speed analog-to-digital converter at the same time. The commutation detection comparator is connected to the output of the DC blocking voltage divider network, detects the start time of the turn-off ringing, and records the duration from the trigger time to the start time of the turn-off ringing as the commutation duration.

3. The method according to claim 2, characterized in that, The equivalent value of the capacitor current before commutation is the product of the link switch state value before the bypass edge and the average value of the link current over a period of time equal to the advance time before the trigger moment. The link switch state values ​​are +1, -1, or 0, corresponding to the two output polarities and bypass state of the engaged state. A droop compensation term is added to the slope difference and step difference respectively. The droop compensation term is the average value of the two latched values ​​at the beginning and end of the corresponding sampling window multiplied by the ratio of the corresponding sampling window duration to the time constant of the DC blocking voltage divider network. The absolute value of the equivalent value of the capacitor current before commutation is multiplied by the advance time and then divided by the absolute value of the compensated slope difference to obtain the capacitance. The equivalent series resistance is obtained by multiplying the compensated slope difference by the ratio of the equivalent commutation time to the advance time, subtracting the compensated step difference, and then dividing by the equivalent value of the capacitor current before commutation. The equivalent commutation time is the sum of the commutation duration and the tailing equivalent time. The tailing equivalent time is given by a calibration table with the link current amplitude and power module temperature as query variables.

4. The method according to claim 3, characterized in that, The effective bypass edge is selected when the absolute value of the equivalent value of the capacitor current before commutation is greater than 30% of the rated current of the link and the duration of the state before the bypass edge is greater than the advance time. The capacitance and equivalent series resistance corresponding to all effective bypass edges in one power frequency cycle are respectively taken as arithmetic average, and then the average is moved over 10 power frequency cycles, and updated once every 0.2s. During operation, the second harmonic components extracted from the switching sampling circuit and the capacitor voltage sampling channel are periodically compared. The voltage division ratio is corrected by the amplitude ratio, and the time constant of the DC blocking voltage divider network is corrected by the phase difference.

5. The method according to claim 1, characterized in that, In step 2, each control cycle multiplies the link switch state value with the link current sampling value to synthesize the capacitor current sequence. The effective value of the capacitor current is calculated according to the power frequency cycle. The link switch state value is +1, -1, or 0, corresponding to the two output polarities and bypass states of the engaged state. The equivalent series resistance is multiplied by the calibrated frequency correction factor to obtain the loss equivalent resistance. The frequency correction factor is the ratio of the loss-weighted equivalent series resistance under the typical current spectrum to the equivalent series resistance measured at the switch edge. The product of the core-shell thermal resistance, the loss equivalent resistance, and the square of the effective value of the capacitor current is filtered by a first-order low-pass filter and added to the shell temperature measurement value to obtain the estimated value of the capacitor core temperature. The time constant of the first-order low-pass filter is taken as the core-shell thermal time constant. The shell temperature measurement value is included in the calculation when it is within the set physical range and the rate of change is lower than the set threshold. In other cases, the estimated value of the capacitor core temperature of the adjacent link in the same cabinet is replaced by a set margin, and a sensor maintenance prompt is generated.

6. The method according to claim 5, characterized in that, The DC capacitor is a metallized film capacitor. During the commissioning phase, the capacitance measured under the set reference operating conditions and the equivalent series resistance converted to 25°C are recorded as the initial capacitance and initial resistance, respectively. During operation, the equivalent series resistance is converted to the 25°C reference value using the estimated capacitor core temperature as the lookup variable according to the calibrated temperature correction curve. The ratio of the 25°C reference value to the initial resistance is recorded as the resistance degradation ratio. The difference between the initial capacitance and the current capacitance is divided by the initial capacitance and recorded as the capacitance degradation rate. The degradation indicators include the capacitance degradation rate and the resistance degradation ratio. When at least one of the two conditions, the capacitance degradation rate reaches 5% and the resistance degradation ratio reaches 2.0, is met, the DC capacitor of the corresponding link is determined to have reached the end of its life, a maintenance prompt is generated, and the corresponding link is cut off and replaced with a redundant link.

7. The method according to claim 6, characterized in that, In step 3, allowable core temperatures are set according to capacitance degradation rates: when the capacitance degradation rate is between 0% and 1%, the allowable core temperature is 85℃; for every 1% increase in capacitance degradation rate, the allowable core temperature is reduced by 5℃, and the lower limit of the allowable core temperature is 70℃; the allowable core temperature is subtracted from the measured air intake temperature to obtain the temperature difference; the sum of the core shell thermal resistance and the heat dissipation thermal resistance is multiplied by the loss equivalent resistance to obtain the thermal resistance product; the temperature difference is divided by the thermal resistance product and the square root is taken to obtain the upper limit of the capacitor current; The thermal resistance is the thermal resistance from the capacitor casing to the air intake. During operation, the difference between the predicted casing temperature and the measured casing temperature is corrected online through an integral process with a time constant of 1 min to 10 min. When the estimated capacitor core temperature exceeds the allowable core temperature, the upper limit of the capacitor current of the corresponding link is reduced by the ratio of the excess to the set temperature range, and the lower limit of the reduction coefficient is 0.

8. The method according to claim 7, characterized in that, The ratio of the effective value of the capacitor current to the effective value of the link current in each power frequency cycle is recorded as the current ratio coefficient. The upper limit of the capacitor current is divided by the current ratio coefficient to obtain the upper limit of the thermally constrained link current. The peak-to-peak value of DC voltage ripple for each link is statistically analyzed using the capacitor voltage sampling channel. The effective value of the current link current is multiplied by the ratio of the set allowable peak-to-peak value of ripple to the current peak-to-peak value of ripple to obtain the upper limit of the ripple-constrained link current. The maximum link current is the minimum of the three: the maximum current limit of the thermal constraint link, the maximum current limit of the ripple constraint link, and the rated current of the link. The minimum value of the upper limit of the link current of each in-service link is taken as the upper limit of the reactive current of the corresponding phase. The minimum value of the upper limit of the reactive current of the three phases is then taken as the reactive current limit of the SVG symmetrical output. When the SVG is connected in a star configuration, the available reactive capacity is the product of the reactive current limit and the effective value of the phase voltage at the grid connection point multiplied by 3. When the SVG is connected in a delta configuration, the link current is the delta arm current, and the available reactive capacity is the product of the reactive current limit and the effective value of the line voltage at the grid connection point multiplied by 3. The available reactive capacity is updated once every 0.2s. The hysteresis band and rate limit are adjusted. The hysteresis band is 2% of the rated reactive capacity, and the rate of change is within 10% of the rated reactive capacity per second. The equalization bandwidth is 2% to 5% of the rated capacitor voltage of the link. The input sequence is updated once every modulation cycle. Each link retains the minimum input duty cycle for compensating self-discharge, and the minimum input duty cycle is within 5%.

9. The method according to claim 1, characterized in that, It also includes DC voltage reference reconfiguration: When the maximum number of links to be put into operation in a certain phase reaches the number of in-service links in the same phase and lasts for 1 second, the DC voltage reference value of each in-service link in the same phase, except for the link with the smallest core temperature margin, is increased by the same amount, so that the sum of the DC voltages of the links involved in the increase reaches or exceeds the peak value of the in-phase modulation voltage reference value. The increase amount is within 10% of the rated DC voltage of the link. The increase is executed according to the rate limit through the DC voltage outer loop and the capacitor voltage equalization loop. The DC voltage reference value of the link with the smallest core temperature margin is simultaneously decreased. The decreased value is above the minimum input voltage of the link power supply. The minimum input voltage is 40% to 60% of the rated DC voltage of the link. The life loss rate multiple during the increase period is the voltage acceleration exponent raised to the power of 1 and the sum of the increase amount. The voltage acceleration exponent is 7 to 10. The difference between the life loss rate multiple and 1 is accumulated with the activation time of DC voltage reference reconfiguration. When the accumulated value reaches the set life loss budget, DC voltage reference reconfiguration is exited and a redundancy replacement request is generated.

10. An SVG reactive power capacity online reconfiguration system considering the degradation state of DC capacitors, characterized in that, The system includes a switching edge sampling circuit, a case temperature sensor, and a link controller installed in each link; an inlet air temperature sensor installed in each power cabinet; a link current sensor for each phase and a voltage sensor at the grid connection point; and a central controller connected to each link controller via an optical fiber link. The case temperature sensor is mounted on the DC capacitor casing, the switching edge sampling circuit is connected between the DC capacitor terminals, and the output of the switching edge sampling circuit is connected to the link controller. The link controller sends up the latch value, commutation duration, trigger time marker, and link switch status value. The central controller buffers the link current sequence and reads the corresponding interval according to the trigger time marker. The central controller includes a parameter measurement module, a core temperature soft measurement module, and a capacity reconstruction module, which respectively execute the calculation of capacitance and equivalent series resistance, the calculation of capacitor core temperature estimate and degradation index, and the calculation and allocation of available reactive power capacity in the method. The output of the capacity reconstruction module is connected to the reactive current command limiting stage, the link connection sequence generation stage, the DC voltage reference allocation stage, and the station communication interface. The system is used to implement the method of any one of claims 1 to 9.